CMP v4 PDF

‫هب انم خدا‬
‫‪www.iran-mavad.com‬‬
‫ﻣﺮﺟﻊ ﻣﻬﻨﺪﺳﯽ ﻣﻮاد و ﻣﺘﺎﻟﻮرژي‬
COMPREHENSIVE MATERIALS
PROCESSING
EDITOR-IN-CHIEF
Saleem Hashmi
Dublin City University, Dublin, Ireland
ASSOCIATE EDITORS-IN-CHIEF
Gilmar Ferreira Batalha

University of São Paulo, São Paulo, Brazil
Chester J. Van Tyne
Colorado School of Mines, Golden, CO, USA
Bekir Yilbas
King Fahd University of Petroleum and Minerals, Dhahran, Saudi Arabia
VOLUME 4
FILMS AND COATINGS: TECHNOLOGY AND RECENT DEVELOPMENT

VOLUME EDITOR
David Cameron
Lappeenranta University of Technology, Finland
AMSTERDAM • BOSTON • HEIDELBERG • LONDON • NEW YORK • OXFORD

PARIS • SAN DIEGO • SAN FRANCISCO • SINGAPORE • SYDNEY • TOKYO
www.iran-mavad.com
Elsevier
Radarweg 29, PO Box 211, 1000 AE Amsterdam, Netherlands
The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, UK
225 Wyman Street, Waltham, MA 02451, USA
Copyright 2014 Elsevier, Ltd. unless otherwise stated. All rights reserved.
No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means electronic,
mechanical, photocopying, recording or otherwise without the prior written permission of the publisher.
Permissions may be sought from Elsevier’s Science & Technology Rights department in Oxford, UK: phone (+44) (0) 1865 843830;
fax (+44) (0) 1865 853333; email: permissions@elsevier.com. Alternatively you can submit your request online by visiting the Elsevier
website at http://elsevier.com/locate/permissions and selecting Obtaining permission to use Elsevier material.
Website links cited in the Encyclopedia were accepted as accurate when first submitted to the publisher, however, the Publisher is unable
to ensure that they will continue to remain so.
Notice
No responsibility is assumed by the publisher for any injury and/or damage to persons or property as a matter of products liability,
negligence or otherwise, or from any use or operation of any methods, products, instructions or ideas contained in the material herein,
Because of rapid advances in the medical sciences, in particular, independent verification of diagnoses and drug dosages should be made
British Library Cataloguing in Publication Data

A catalogue record for this book is available from the British Library
Library of Congress Cataloging-in-Publication Data

A catalog record for this book is available from the Library of Congress
ISBN: 978-0-08-096532-1
For information on all Elsevier publications

visit our website at store.elsevier.com
Printed and bound in Italy

14 15 16 17 18 19 10 9 8 7 6 5 4 3 2 1
Project Manager: Gemma Tomalin

Associate Project Manager: Joanne Williams
www.iran-mavad.com
EDITOR-IN-CHIEF
Saleem Hashmi graduated from East Pakistan University of Engineering and Technology in 1967,
and in 1972 he completed his master’s and PhD degrees from the University of Manchester. He
was appointed as a postdoctoral fellow with UMIST, Manchester, in January 1973, and in October
1973, he joined Sheffield Hallam University as a lecturer. In 1976, he was promoted to senior
lecturer, then to principal lecturer in 1977, and to reader in 1984.
In January 1987, Professor Hashmi founded the School of Mechanical & Manufacturing Engi-
neering at Dublin City University. While in the process of establishing this school, he attracted
substantial external research funding to set up the Material Processing Research Centre (MPRC),
a research center of excellence in the field of material processing. To this day, he continues his
affiliation with Dublin City University as an emeritus professor in the school he founded 27 years ago.
In 1990, Professor Hashmi established Advances in Materials and Processing Technologies
(AMPT), a much needed international conference in materials processing. He continues to be the
Chairperson of the Steering Committee for this series of conferences. In 1998, he was appointed
as editor-in-chief of Elsevier Journal of Material Processing Technology and continued in this role
until 2008.
In 1995, Professor Hashmi was awarded the higher doctorate degree of DSc, by the University of
Manchester, the highest academic degree any university in the United Kingdom can award.
Professor Hashmi has supervised or co-supervised 110 PhD and 55 MEng research students to successful completion. Over the years,
he has also acted as an external examiner and expert assessor for PhD candidates and engineering departments with universities in Ireland,
the United Kingdom, India, Pakistan, Bangladesh, Hong Kong, Canada, Australia, and Malaysia. He is a Chartered Engineer (CEng), a Fellow
of the Institution of Mechanical Engineers, UK (FIMechE) and also a Fellow of the Institution of Engineers of Ireland (FIEI). He has
published over 430 papers and 12 books and is still publishing.
v
www.iran-mavad.com
This page intentionally left blank
www.iran-mavad.com
EDITORIAL BOARD
Chester J. Van Tyne is the FIERF (Forging Industry Education and Research Foundation) professor
at Colorado School of Mines (CSM) in Golden, Colorado. He received his undergraduate and
graduate degrees from Lehigh University and worked for Bethlehem Steel as a research engineer for
a short period before joining Lafayette College in the Metallurgical Engineering Department as
a professor in 1980. In 1988, he joined the Metallurgical and Materials Engineering Department at
CSM. He is actively engaged in the research operations of the Advanced Steel Processing and
Products Research Center at CSM. His research interests are in metal deformation processes
especially forging, sheet stamping, and friction during metalworking operations.
Gilmar F. Batalha graduated as a mechanical engineer from the University of Brasilia in 1982. He
completed his MSc in Mechanical Engineering at the Federal University of Santa Catarina (1987),
his PhD in Mechanical Engineering (1995) and habilitation/livre-docencia in manufacturing
processes (2010) at the Department of Mechatronics and Mechanical Systems, Polytechnic School
of Engineering, University of Sao Paulo, where he has been a tenured professor since 1989. He has
had postdoctoral research experience at the Friedrich Alexander University Erlangen–Nuremberg,
Germany (1997–98). He was a visiting professor at Ecole Centrale de Lille, France (2004–06), and
at the Faculty of Mechanical Engineering of the Silesian University of Technology, Gliwice, Poland
(2012–13). He served as an adjunct professor at the Faculty of Engineering, UDESC, Joinville, Brazil
(1984–89). He has been an advisor on several PhD theses and master’s dissertations, coordinator
of the Laboratory of Manufacturing Engineering of the EPUSP, and is responsible for several
international and national cooperation agreements for exchange of students and researchers as well
as for consultancy with the industry. He is the volume editor for the Materials Modeling and
Characterization volume in Elsevier’s Comprehensive Materials Processing series.
Sergio T. Button received his PhD in Mechanical Engineering from School of Mechanical Engi-
neering, University of Campinas, Brazil, in 1990. He is currently full professor of Manufacturing
Processes at the University of Campinas. His main research fields include metal forming and
numerical simulation. Currently, his research is focused on cross wedge rolling and hot stamping.
He is the coauthor of 2 books and more than 100 papers in international journals and conferences.
He is the editor and reviewer of many international journals, and participates in the organization of
three national and international scientific conferences.
vii
www.iran-mavad.com
viii Editorial Board
David Cameron is currently scientific advisor to Miktech Oy, Finland, and adjunct professor in
Dublin City University, Ireland. From 2003 to 2013, he was professor of Material Technology in
Lappeenranta University of Technology, Finland, where he established the Advanced Surface
Technology Research Laboratory (ASTRaL). He obtained his doctorate from the University of
Glasgow in 1980 and has spent his research career working on thin-film technology. He was
a senior research fellow at the Royal Signals and Radar Establishment in the United Kingdom until
1982 when he joined the School of Electronic Engineering at Dublin City University, becoming
associate professor. His research focus is on atomic layer deposition and its industrialization.
Joseph A. McGeough is an honorary professorial fellow in the School of Engineering at the

University of Edinburgh. He is a graduate from the universities of Glasgow (BSc and PhD) and
Aberdeen (DSc). He held research appointments at the universities of Leicester, Queensland, and
Strathclyde. As a student he gained numerous county, university, and national championship
athletic awards. His industrial training began as an undergraduate vacation apprentice for a firm of
electrical contractors based at ICI ltd Nobel Division, and he subsequently undertook postgraduate
experience at International Research and Development, Newcastle upon Tyne, before returning to
university service. He was successively a lecturer, senior lecturer, and reader in Engineering at
Aberdeen University, and then moved to Edinburgh University to take up its Regius Chair of
Engineering. His research work has dealt mainly with manufacturing processes for which he has
been awarded Institution of Mechanical Engineers and the Society for Underwater Technology
prizes. He has been the recipient of a Royal Society/SERC (Science and Engineering (now Engi-
neering and Physical Sciences) Research Council) Industrial Fellowship in order to transfer the
results of his research to industry. His publications include Principles of Electrochemical Machining,
Advanced Methods of Machining, and Micromachining of Engineering Materials (editor). His latest book
is The Engineering of Human Joint Replacements (2013), which contains details of the materials and
manufacturing processes used in this field. He is a fellow of the Institution of Mechanical Engi-
neers, International Academy for Production Engineering, Royal Society of Edinburgh, and Royal
Academy of Engineering.
Emin Bayraktar is a full university professor and research director at Supmeca/Paris, School of
Mechanical and Manufacturing Engineering, Paris, France. He was awarded a PhD degree in
Mechanical Engineering by IST/OSU and a second PhD degree in Materials Science and Engi-
neering in Paris. Subsequently, Professor Bayraktar was awarded a higher doctorate degree of
habilitation (DSc) in Mechanics of Materials in recognition of his research contributions at
Paris 6, University Marie-Curie. His teaching and research expertise lies on the processing and
manufacturing of the composite materials and damage analysis. Recently he has been appointed
as an adjunct professor in UNICAMP-Campinas, Brazil, on the new design of composites for the
application in aeronautical engineering. He is on the editorial board of a number of scientific
journals. He is also a member of the Steering Committee for the International Conference Series
on Advances in Materials and Processing Technologies (AMPT).
www.iran-mavad.com
Editorial Board ix
Bekir S. Yilbas received his PhD degree in Mechanical Engineering from Birmingham University.
He was awarded the Doctor of Engineering degree in 2005 by Birmingham University due to his
significant contribution to his field of study. He has published 375 journal papers in reputable
international journals and presented over 60 papers at international conferences. He has served
as an editorial board member of international journals, including International Journal of Machine
Tool and Manufacture Design, Research and Application, International Journal of Subsurface Sensing
Technologies and Applications, Journal of Materials Processing Technology, Journal of Achievements in
Materials and Manufacturing Engineering, International Journal of Nanomanufacturing, and Archives of
Materials Science and Engineering. He has received numerous awards in recognition of his research
work.
Nabil Bassim was born and raised in Cairo, Egypt. Dr. Bassim received a BSc in Chemical Engi-
neering from Cairo University in 1965. He went to the University of Virginia, United States, where
he received his MSc in Chemical Engineering (Nuclear Engineering option), and a PhD in Materials
Science in 1973. He moved to Canada where he is now professor of Materials Science and Engi-
neering at the University of Manitoba. He served for years as associate dean of research at the
Faculty of Engineering and was director of International Programs for the faculty. His areas of
research interest include mechanisms and mechanics of plastic deformation, dislocation theory,
fracture and fatigue, and nondestructive evaluations. He has patents on the use of acoustic emission
for monitoring of structures. He also has an interest in studying the behavior of materials at high
strain rates and large strains as they affect processes such as metal forming and machining. He has
published over 120 papers and supervised numerous MSc and PhD theses.
Syed H. Masood is professor of Mechanical and Manufacturing Engineering at Swinburne

University of Technology in Melbourne, Australia, where he is also the program coordinator of the
master’s program in advanced manufacturing technology. He obtained his PhD from the University
of Queensland, Australia; MEng from the University of New Brunswick, Canada; PG Dip from IIT
Delhi; and BSc Eng Hons from Aligarh Muslim University, India. Professor Masood has worked
previously at Central Queensland University, Rockhampton, and PNG University of Technology,
Papua New Guinea. He has published over 275 technical articles in international journals, books,
and conferences in different areas of advanced manufacturing. His current research interests include
additive manufacturing technologies, especially fused deposition modeling and direct metal
deposition, plastics manufacturing processes, material development, and material characterization.
www.iran-mavad.com
x Editorial Board
Mustafizur Rahman is currently a professor at the Department of Mechanical Engineering at

National University of Singapore (NUS), and has been with NUS since 1982. He obtained his BSc
in Engineering from Bangladesh University of Engineering and Technology (BUET), MEng and PhD
from Tokyo Institute of Technology, Japan. His research interests include micro-/nanomachining,
design and development of miniature machine tools and devices for micro-/nanomachining, and
high-speed machining of difficult-to-machine materials. He has received Outstanding Achievement
Awards from the Singapore Institute of Engineers and the Japan Society of Mechanical Engineers in
2003 and 2005, respectively.
George Krauss is currently university emeritus professor at the Colorado School of Mines and
a metallurgical consultant specializing in steel microstructural systems. Dr. Krauss received the BSc
in Metallurgical Engineering from Lehigh University in 1955 and the MSc and ScD degrees in
Metallurgy from the Massachusetts Institute of Technology in 1958 and 1961, respectively, after
working at the Superior Tube Company as a development engineer in 1956. In 1962–63, he was an
NSF postdoctoral fellow at the Max Planck Institut für Eisenforschung in Düsseldorf, Germany. He
served at Lehigh University as assistant professor, associate professor, and professor of Metallurgy
and Materials Science from 1963 to 1975, and in 1975 joined the faculty of the Colorado School of
Mines as the AMAX Foundation professor in Physical Metallurgy. He was the John Henry Moore
professor of Metallurgical and Materials Engineering at the time of his retirement from the Colo-
rado School of Mines in 1997.
In 1984, Dr. Krauss was a principal in the establishment of the Advanced Steel Processing and
Products Research Center, an NSF Industry-University cooperative research center at the Colorado
School of Mines, and served as its first director until 1993. He has authored the book Steels: Pro-
cessing, Structure, and Performance, (ASM International, 2005), coauthored the book Tool Steels, fifth
edition, (ASM International, 1998), and edited or coedited conference volumes on tempering of
steel, carburizing, zinc-based coatings on steel, and microalloyed forging steels. He has published
over 330 papers and lectured widely in technical conferences, universities, corporations, and ASM
International chapters, presenting a number of keynote, invited, and honorary lectures. He presented the Edward DeMille Campbell
Memorial Lecture of ASM International in 2000, the Howe Memorial Lecture of ISS in 2003, and the Adolf Martens Memorial Steel Lecture of
AIST in 2010.
Dr. Krauss has served as the president of the International Federation of Heat Treatment and Surface Engineering (IFHTSE), 1989–91,
and as President of ASM International, 1996–97. He is a fellow of ASM International, TMS, and IFHTSE. He has been awarded the Adolf
Martens Medal of the German Society for Heat Treatment and Materials, the Charles S. Barrett Silver Medal of the Rocky Mountain Chapter
of ASM, the George Brown Gold Medal of the Colorado School of Mines, and several other professional and teaching awards, including the
ASM Albert Easton White Distinguished Teacher Award in 1999. He is an honorary member of the Iron and Steel Institute of Japan,
a distinguished member of the Iron and Steel Society, an honorary member of ASM International, and an honorary member of the Japan
Institute of Metals.
www.iran-mavad.com
CONTRIBUTORS TO VOLUME 4
E. Arab-Tehrany L. Hultman
Université de Lorraine, France Linköping University, Linköping, Sweden
R. Bandorf J. Koskinen
Fraunhofer-Institut für Schicht- und Oberflächentechnik Aalto University, Espoo, Finland
IST, Braunschweig, Germany
M. Leskelä
P. Baroch University of Helsinki, Finland
University of West Bohemia, Plzen, Czech Republic
X. Liu
G. Bräuer Shanghai Institute of Ceramics, Shanghai, PR China
Fraunhofer-Institut für Schicht- und Oberflächentechnik
A. Lotz
IST, Braunschweig, Germany
Max Planck Institute for Polymer Research, Mainz,
E. Broitman Germany
Linköping University, Linköping, Sweden
S. Matsui
D. Cameron University of Hyogo, Hyogo, Japan
Lappeenranta University of Technology, Lappeenranta,
P.H. Mayrhofer
Finland
Vienna University of Technology, Wien, Austria
P.F. Carcia
J. Musil
DuPont Central Research and Development,
University of West Bohemia, Plzen, Czech Republic
Wilmington, DE, USA
J. Niinistö
S. Desobry
University of Helsinki, Finland
Université de Lorraine, France
P. Pichat
D.P. Dowling
CNRS/Ecole Centrale de Lyon (STMS), Ecully CEDEX,
University College Dublin, Dublin, Ireland
France
L. Duque
Y. Qiao
Shanghai Institute of Ceramics, Shanghai, PR China
Germany
R. Rachbauer
R. Förch
Montanuniversität Leoben, Leoben, Austria
Germany and Institut für Mikrotechnik Mainz GmbH, M. Ritala
Mainz, Germany University of Helsinki, Finland
L. Guan F. Rovere
Hebei University, Baoding, China Montanuniversität Leoben, Leoben, Austria and RWTH
Aachen University, Aachen, Germany
D. Hegemann
Empa, St. Gallen, Switzerland J.M. Schneider
RWTH Aachen University, Aachen, Germany
D. Holec
Montanuniversität Leoben, Leoben, Austria Y. Setsuhara
Osaka University, Osaka, Japan
xi
www.iran-mavad.com
xii Contributors to Volume 4
V. Sittinger A. Yli-Pentti
Fraunhofer-Institut für Schicht- und Oberflächentechnik Metropolia University of Applied Sciences, Vantaa,
IST, Braunschweig, Germany Finland
R. Tietema P. Zeman
Hauzer Techno Coating BV, Venlo, The Netherlands University of West Bohemia, Plzen, Czech Republic
N. Toyoda D. Zhang
University of Hyogo, Hyogo, Japan Huazhong University of Science and Technology,
Wuhan, China
P. Vuoristo
Tampere University of Technology, Tampere, Finland
www.iran-mavad.com
CONTENTS OF VOLUME 4
Editor-in-Chief v
Editorial Board vii
Contributors to Volume 4 xi
Preface xv
4.01 Introduction to Films and Coatings: Technology and Recent Development 1

D Cameron
4.02 Cathodic-Arc and Thermal-Evaporation Deposition 3
J Koskinen
4.03 Magnetron Sputtering 57
G Bräuer
4.04 High Power Impulse Magnetron Sputtering e HIPIMS 75
R Bandorf, V Sittinger and G Bräuer
4.05 Atomic Layer Deposition 101
M Leskelä, J Niinistö and M Ritala
4.06 Laser Ablation 125
D Zhang and L Guan
4.07 Surface Processing Using Cold Atmospheric Pressure Plasmas 171
D P Dowling
4.08 Ion Beam Deposition: Recent Developments 187
N Toyoda and S Matsui
4.09 Plasma Polymer Deposition and Coatings on Polymers 201
D Hegemann
4.10 Thermal Spray Coating Processes 229
P Vuoristo
4.11 Electroplating and Electroless Plating 277
A Yli-Pentti
4.12 Plasma Sources in Thin Film Deposition 307
Y Setsuhara
4.13 Hard Nanocomposite Coatings 325
J Musil, P Zeman and P Baroch
4.14 Protective Transition Metal Nitride Coatings 355
P H Mayrhofer, R Rachbauer, D Holec, F Rovere and J M Schneider
xiii
www.iran-mavad.com
xiv Contents of Volume 4
4.15 Advanced Carbon-Based Coatings 389

E Broitman and L Hultman
4.16 Photocatalytic Coatings 413
P Pichat
4.17 Biocompatible Coating 425
Y Qiao and X Liu
4.18 Antimicrobial Bioactive Polymer Coatings 449
R Förch, L Duque and A Lotz
4.19 Thin-Film Diffusion Barriers for Electronic Applications 463
P F Carcia
4.20 Diffusion Barrier Layers for Edible Food Packaging 499
S Desobry and E Arab-Tehrany
4.21 Large-Scale Industrial Coating Applications and Systems 519
R Tietema
www.iran-mavad.com
PREFACE
In recent years considerable advances have taken place in materials and their processing technologies, the pace
of which is progressing at an accelerated rate. It is therefore vitally important that researchers, academics, design
engineers, and manufacturing technologists be up to date with these new developments. Such developments
open up excellent opportunities to improve effective utilization of material resources, improve efficiency,
reliability, durability, and cost effectiveness of the products. This comprehensive series will serve researchers all
over the world in keeping them fully abreast of such new developments.
Comprehensive Materials Processing is a one-stop resource consolidating and enhancing the literature of the
materials processing and manufacturing universe. The work addresses recent progress in materials technologies
and science as well as covering recent developments in specific manufacturing processes involved with materials
for applications in engineering, biomedical, environmental protection, health and safety, and sensor materials
and technologies. The authoritative analysis of these processes will assist scientists and engineers in the selec-
tion, design, and usage of materials, whether required in the lab or in action across industry.
As the editor-in-chief of this comprehensive series, a truly collaborative work, I am greatly indebted to the
volume editors, internationally renowned experts in their fields, for guiding and selecting the topics for their
respective volumes, commissioning authors, and reviewing the content so meticulously. Their true dedication to
the scientific community is reflected in the time and energy they have given to this project. My sincere thanks are
also due to all of the authorseresearchers and materials practitioners who have contributed extensive coverage
of literature review as well as recent works of research to this huge project. Their insight and specialist knowledge
in their respective fields is reflected in the high quality of this work.
Myself, the three associate editors-in-chief, and all of the volume editors are greatly appreciative of all the
hard work undertaken to turn this idea into a publishable work. Our special thanks go to Gemma Tomalin, the
project manager, along with Joanne Williams, and the rest of the team at Elsevier who served successively to
keep the project on track through friendly nudges in order to ensure timely completion. We are also hugely
grateful to Mohamed Imrankhan for his dutiful coordination of the proofs.
The extensive academic discussion of core theories and applications, supplemented by applied case studies
and advanced multimedia features has drawn together numerous areas of research and I sincerely hope that this
work will prove to be of great help to both the young and experienced members of the international research
community, academics, and industrial practitioners for many years to come.
Saleem Hashmi
Editor-in-Chief
xv
www.iran-mavad.com
www.iran-mavad.com
4.01 Introduction to Films and Coatings: Technology and Recent Development
D Cameron, Lappeenranta University of Technology, Lappeenranta, Finland
2014 Elsevier Ltd. All rights reserved.
The topic of films and coatings is an extensive one covering everything from paints to nanolayered surface modifications. It has
a long history stretching for at least 40 millennia in the form of cave paintings and is also at the leading edge of current research with
carbon nanostructured layers being an example here. There are many applications of high-technology coatings, from friction
reduction in motor engines to decorative coatings on jewelry and microelectronics to barrier layers for protection of food.
The development of functional coatings has led to enormous increases in performance in machinery operation, efficiency in
manufacture, and reduction in the production of waste materials.
It is extremely difficult to do justice to such an enormous field of work in one volume and, to an extent, it is an eclectic mixture of
topics. I have based the contents of this volume, loosely, on thin films. I have, however, stretched the definition of ‘thin’ films,
typically thought of as in the micrometer range and lower, to include thicker layers such as those produced by the techniques of
thermal spraying and electroplating where thicknesses in the hundreds of micrometers are typical. Again, the choice of topics is
arbitrary to an extent but covers both the coating processes themselves and some diverse applications. Some of these processes have
been in use for many decades; for example, electroplating and physical vapor deposition. However, the striking feature is that they
are all, even the most ancient processes, under a continual process of development and improvement, and the level of research
continues to increase. A search for research articles on the topic of ‘thin films’ shows that over the last 20 years, approximately
195 000 research articles have been published, and the current publication rate is approximately 16 000 per annum.
The chapter deals with the deposition techniques and covers low-pressure physical vapor deposition techniques such as sput-
tering, arc, ion beam, and laser deposition. It also highlights the most recent developments, for example, high-power impulse
magnetron sputtering, which has led to new developments in magnetron and power supply design and new ways of influencing the
deposition process. The topic of atomic layer deposition is covered, as well. This technique was initially proposed almost 50 years
ago, but there has been an explosion of interest in the last 5 years for a range of applications. Atmospheric pressure processes are
widely used for thicker films, particularly in thermal spraying and its variants such as plasma spraying. Atmospheric pressure
chemical vapor deposition processes for thin films are increasingly considered because they avoid the complexities and costs of
vacuum processing.
In many of these processes, the existence of plasma is a crucial part of the operation, and an account of plasma sources used in
thin-film deposition is included. Wet processes with a long industrial history such as electroplating are of enormous economic
importance, and these are also covered.
The chapters on the applications of coatings include mechanical engineering applications such as hard, wear resistant, and
protective coatings with an emphasis on their large-scale industrial usage. The very wide range of other uses is illustrated in such
examples as in food packaging, photocatalysis, and biocompatibility. In many cases, coatings are applied to relatively soft polymer
materials for both decorative and strengthening purposes, so this is covered in one chapter. The broad range of techniques used to
provide decorative enhancement to objects is detailed. The use of coatings to enhance appearance – perhaps one of the oldest uses –
has a chapter unto itself. New types of materials are continuously being developed, and the chapter on carbon-based coatings gives
an account of recent work in this field.
Overall, this volume provides a robust picture of the current capabilities and state-of-the-art broad range of coatings and their
applications.
Comprehensive Materials Processing, Volume 4 http://dx.doi.org/10.1016/B978-0-08-096532-1.00400-3 1

www.iran-mavad.com
www.iran-mavad.com
4.02 Cathodic-Arc and Thermal-Evaporation Deposition
J Koskinen, Aalto University, Espoo, Finland
4.02.1 Introduction 3
4.02.2 Thermal Evaporation 4
4.02.2.1 Basics 4
4.02.2.2 Evaporation Geometry 6
4.02.2.3 Evaporation Sources 6
4.02.2.4 Web Coating 7
4.02.2.5 Corrosion Coating 7
4.02.2.6 Solar Cells 8
4.02.2.7 Polymeric Film Deposition 10
4.02.2.8 Spotless Vacuum Arc Process 10
4.02.3 Cathodic-Arc Deposition Process 10
4.02.3.1 Cathodic Arcs from the First Experiments to Modern Industrial Deposition 10
4.02.3.2 Unique Features of Cathodic-Arc Deposition 12
4.02.3.3 Ignition and the Cathode Spot 14
4.02.3.3.1 Rapid Sequence of Microexplosions 15
4.02.3.4 Arc Plasma 17
4.02.3.5 DC Vacuum Arc 19
4.02.3.6 Pulsed Vacuum Arc 20
4.02.3.7 Macroparticles 23
4.02.3.7.1 Particle Effects 23
4.02.3.7.2 Control of Particle Evolution 25
4.02.3.7.3 Particle Filtering 27
4.02.3.8 Typical Coatings by Cathodic-Arc Deposition 30
4.02.3.8.1 ta-C 30
4.02.3.8.2 Nitrides 34
4.02.3.8.3 Oxides 36
4.02.3.9 Ultrathin Films and Nanolaminates 39
4.02.3.9.1 Multilayers, Superlattices, and Nanolaminates 40
4.02.3.9.2 MAX Phase 41
4.02.3.10 Applications of Cathodic-Arc Deposited Films 41
4.02.3.10.1 Tribology 41
4.02.3.10.2 Optical Cathodic-Arc Coatings 47
4.02.3.10.3 Decorative Coatings 49
4.02.3.10.4 Biological Applications 51
References 53
4.02.1 Introduction
Cathodic-arc deposition and thermal-evaporation processes have almost a century-long history as technical solutions to produce
thin films. The two technologies have developed to the specialized directions owing to the particular physical characteristics of the
processes. They both share common features: solid material as a source, melting of source material, limited heat load to the
substrate, line-of-sight process, and upscalability at moderate cost. However, they also have big differences such as the ionization
and kinetic energy of the growing species. The biggest problem related to cathodic-arc deposition is the inherent evolution of
macroparticles. This has possibly slowed its development in comparison to the recent new deposition technologies such as high-
power impulse magnetron sputtering (HIPIMS) (see Chapter 4.04, HIPIMS, in this book). Thermal evaporation allows limited
control of the growing film microstructure, which has led to the development of various plasma-assisted deposition processes.
In this chapter, an overview of cathodic-arc deposition and thermal evaporation is presented. As regards the cathodic-arc process,
the recent and excellent book by Dr Andre Anders from the Lawrence Berkeley Laboratory in the United States, Cathodic Arcs: From
Fractal Spots to Energetic Condensations, has been a major source of material for this chapter. The main effort of this chapter has been
to collect the new research results published in the short period from 2010 until mid-2012. A strong weight on research on
tetrahedrally coordinated diamondlike carbon (DLC, specifically ta-C) films in this chapter is a result of the personal background in
this field by the author. A long-lasting interest in cathodic-arc processes has led me to write on that topic much more extensively
than on thermal evaporation, a choice that is not necessarily justified by the technical relevance of this coating process.
Comprehensive Materials Processing, Volume 4 http://dx.doi.org/10.1016/B978-0-08-096532-1.00409-X 3

www.iran-mavad.com
4 Cathodic-Arc and Thermal-Evaporation Deposition
4.02.2 Thermal Evaporation
Thermal evaporation is one of the most traditional vacuum physical vapor deposition (PVD) methods. Owing to the possibility to
obtain a high deposition rate, the process is relatively easily upscalable. It is applied to many different substrate materials, including
polymers, metals, and glasses. At the moment, thermal evaporation is used to produce thin-film coatings at economic prices in large
volumes. In research into high-performance thin films, other PVD methods that apply plasma and energetic ions have replaced thermal
evaporation. However, there is a growing interest in using this method for the web coating of polymers and processing of nanomaterials.
4.02.2.1 Basics
The basic setup of thermal evaporation includes a vacuum chamber, a source of evaporated material, and a substrate holder, often
including substrate movement, as shown in Figure 1.
In Table 1, some differences of thermal evaporation and magnetron sputtering are listed:
The vacuum level for thermal evaporation is a practical question of the impurity requirements of the coating and substrate. When
the film has to be free of impurities, a UHV vacuum system capable of pressure down to 109 mbar is needed and the deposition
conditions are similar to those of molecular beam epitaxy (MBE) systems. If the coating functionality is not sensitive to water vapor
contaminants, a moderate 104 mbar vacuum level is adequate. An example is the deposition of barrier films for packaging
applications. The molecule mean free path is inversely proportional to the pressure and affects, for example, the kinetic energy of the
molecule hitting the growing surface.
Figure 1 Schematic presentation of a thermal-evaporation system used for silicon wafers or optical lenses.
www.iran-mavad.com
Cathodic-Arc and Thermal-Evaporation Deposition 5
Table 1 Comparison of thermal evaporation with magnetron sputtering (partially obtained from textbooks)
Evaporation Sputtering
1
Deposition rate Up to mm s <1 nm s1
Choice of materials Limited Almost unlimited
Purity Excellent (at ultrahigh vacuum (UHV)) Impurities from deposition system
Substrate heating Very low Could be high
In situ cleaning Not available (without addition of ion sources) Easily with sputter etching
Stoichiometry Little control (needs multiple sources with good control) Easier to control
Change of source material Easy Expensive
Radiation damage X-ray with e-beam Energetic ions
Decomposition of material High Low
Scaling up Resistive heating easier, e-beam difficult Good
Uniformity Difficult Easy over large surface
Capital costs Low High
In situ thickness control Easy Difficult
Number of depositions One per charge Many per target
Material consumption Material left in crucibles Only part of target material can be used
Adhesion Poor Excellent
Shadowing effect Large Small
Film properties Difficult to control, porous Controllable with bias, pressure, substrate
temperature, dense
Lift-off masking Easy Difficult
Step coverage Poor Good, could be controlled
System layout flexibility Low High
The thermal-evaporation rate depends on the vapor pressure of the material. The deposition flux jvap is:
jvap ¼ aav ðpeq pi Þ$ð2pmkTÞ1=2 [1]
where peq is the equilibrium vapor pressure of the material at the vapor source; pi is vapor pressure at the growing surface, m is the
mass of the molecule; T is temperature; k is the Boltzmann constant; and aav is the vaporization coefficient, which is the proportion
of those surface-impinging molecules that condense on the growing surface. peq increases as temperature increases, and a practical
deposition speed is achieved when peq is about 102 mbar. As a general rule, this vapor pressure is achieved close to the melting
temperature of the material. In Figure 2, the practical evaporation temperature and the melting point of some materials are shown
(1).
Figure 2 A practical evaporation temperature for several metallic elements when the vapor pressure is about 102 mbar. After Mahan, J. E. Physical
Vapor Deposition of Thin Films; John Wiley & Sons, 2000.
www.iran-mavad.com
4.02.2.2 Evaporation Geometry

Often, the evaporation source is a point source, as shown in Figure 3.
For a point source, the dependence of the deposition flux to the geometry is:
jvap ¼ aav ðpeq pi Þ$ð2pmkTÞ1=2 cosðQÞcosðfÞ=pR2 [2]
Thus, the deposition rate is dependent on the distance, and on the angle from the point source Q and angle f of the specimen
normal to the point source. If the specimen is located in a radial geometry, as indicated in Figure 3(b), then cos(Q) ¼ cos(f) ¼ R/2r.
The jvap reduces to:
jvap ¼ aav ðpeq pi Þ$ð2pmkTÞ1=2 =4pr 2 [3]
Then the growth rate of the film is independent of the angles and the distance R.
4.02.2.3 Evaporation Sources

The different evaporation sources could be listed as follows:
l Resistive heating
B boats of conductive material (refractory metal or carbon)
B wire baskets
B wire-heated crucibles
B plated rods
B box sources
l Inductive heating
l Electron beam heating
Resistive heating is suitable for materials with a low melting temperature. For the wire sources, the evaporant has to wet the
filament, and the melt has to stay at the filament by surface tension. A typical amount is maximum 1 g. For the boat, the amount is up
to a few grams. The box source is a resistively heated crucible covered with a plate allowing the vapor to exit through small orifices. The
plate is used in order to limit droplets from the splashing melt, which is typical for melts like silicon monoxide or zinc sulfide.
The resistive-heating boats may be equipped with wire feeding to replenish the evaporated materials. Mechanical feeding to the
melt from a wire spool inside the deposition chamber can be used to upscale the evaporation, for example, of copper and aluminum
with several g min1 and tens of kg amounts with one source in one charge.
A high current of focused electrons may be steered by magnetic lenses to a crucible in order to evaporate materials. The electron
beam power can be adjusted from one kW to hundreds of kW. The crucible is commonly cooled by water, which minimizes the
contaminants in the deposition. If the evaporated material has high thermal conductivity, a crucible of thermally insulated
refractory material may be used. A typical electron beam source is schematically shown in Figure 4 (2).
Figure 3 The geometry of thermal evaporation with an evaporation source and specimen. (a) Specimen at arbitrary position and (b) specimen position at
radial symmetry.
www.iran-mavad.com
Figure 4 Electron beam evaporation. After Bringham Young University, Provo, UT, USA. http://www.cleanroom.byu.edu/metal.phtml.
The selection of materials that may be deposited by thermal evaporation is quite extensive, including most of the solid elemental
materials, oxides, carbides, and sulfides. In reactive evaporation, typically oxygen, nitrogen, or carbon-containing gases are added to
control the stoichiometry of oxides, nitrides, and carbides.
Tables and selection guidance for the selection of evaporation sources for the different materials are available from equipment
manufacturers and in textbooks. A comprehensive table is provided online by Brigham Young University (3).
4.02.2.4 Web Coating

The upscalability of the deposition rate and the low thermal load to the substrate have enabled the development of web-coating
processes with high productivity.
The coated polymers include PE (polyethylene), PET (polyethylene terephthalate), PP (polypropylene), CPP (cast poly-
propylene), and BOPP (biaxially oriented polypropylene). The metal-coated polymer webs have similar electrical, optical, and gas
barrier properties to metal foils. The applications are, for instance (4):
l electric capacitors,
l gas, light, and ultraviolet (UV) barriers for flexible food packaging,
l thermal radiation screens for building envelopes,
l reflective coatings
l decorative wrappings.
The technology for coating polymeric webs with metals for the packaging industry has been available for decades. Yet, today the
demand for better, low-cost barriers against oxygen and water is high. The evaporated aluminum and alumina coatings provide an
oxygen transmission rate (OTR) better than the substrate PET polymer by a factor of up to 100 (Figure 5,5).
The transparent aluminum oxide stacked with thin aluminum layers provides a coating with interference colors. The color
depends on the angle of inspection and provides a strong decorative appearance to the polymer surface. It is also used as an
anticounterfeiting application to protect production brand names.
The emerging technology of printed electronics is another field of production with high expectations to produce electronic
devices integrated to low-cost mass products such as lab-on-ship, smart packaging, and electronic paper. The standard process for
printed electronics is gravure printing of functional inks, hot embossing, and lamination. The demand for barriers (e.g., for organic
light-emitting diode (OLED) films) is high but not yet fulfilled by applying ink-based transparent layers. The construction of hybrid
web systems, including vacuum deposition and in particular the high deposition rate of thermal evaporation, has growing potential.
In Figure 6, a schematic illustration of a web coater with thermal evaporation is shown (6).
The very high production rate of a web up to 2500 mm wide that is coated with a speed up to 17 m s1 requires an array of high-
capacity evaporation sources. In Figure 7, an array of aluminum sources used in a web coater to produce layers of aluminum and
alumina is shown.
4.02.2.5 Corrosion Coating

Yet another large-scale coating process of evaporated aluminum is the deposition of corrosion-resistant coatings for aircraft
components. The standard solution, which is used in aircrafts still in action today, was designed originally in the 1960s and relies on
www.iran-mavad.com
Figure 5 OTR as a function of layer thickness of the thermally evaporated film. After Ludwig, R.; Josephson, L. Clear Barriers and High Volume Productivity,
Applied Materials Ltd. Polymers Laminations, Adhesives Coatings Extrusions PLACE 2006 Conference, September 17-21, 2006, Cincinnati, Ohio.
cadmium anticorrosion coatings. Serious environmental problems related to cadmium have forced the industry to search for
alternative solutions. Electrolytic plating of aluminum requires toxic electrolytes due to the high affinity of aluminum for oxygen.
A process called directed vapor deposition (DVD) has been developed by the University of Virginia (8). In this process, thermally
evaporated aluminum interacts with a supersonic gas jet with a pressure in the range of 1–50 Pa. The interaction with the gas jet
enables steering of the evaporated aluminum out of the line-of-sight path and depositing aluminum on 3D components, as well as
into holes and cavities.
4.02.2.6 Solar Cells

Photovoltaic cells are an application with very high growth potential. According to solar energy predictions, large surface areas of
low-cost high-efficiency solar cells will be assembled in places with intense sunshine such as deserts. The high surface area and low
cost require scalable thin-film processes where thermal evaporation has been shown to be a viable manufacturing technology.
Figure 6 Schematic illustration of a web-coating device with thermal evaporation. After Ludwig, R.; Kukla, R.; & Josephson, E. Vacuum Web Coating –
State of the Art and Potential for Electronics. Proc. IEEE 2005, 93 (8), 1483–1490.
www.iran-mavad.com
Figure 7 Array of aluminum evaporation. With permission of Leybold Optics GmbH.
Thermal evaporation has been demonstrated in coatings of copper–indium–gallium–selenide (CIGS) cells. The growth of columnar
crystals at low temperature with large crystal size enables good transport properties in the semiconducting crystals. The solar cells are
deposited in large volumes and require an affordable substrate, such as soda–lime glass. The use of this substrate limits the possible
high substrate temperatures. Coevaporation of Cu, In, Ga, and Se has provided promising results at a moderate substrate
temperature of 450 C (9). The limited global resources of indium and gallium have motivated researchers to find alternative
photon-absorbing materials. One potential candidate is Cu2ZnSn(S,Se)4 (CZTS), which has a kesterite structure, where every two In
or Ga atoms in the chalcopyrite structure are replaced by a Zn and Sn atom. Coevaporation of Cu, Zn, Sn, and Se has enabled very
high efficiency of the CZTS cell. The cross section of the deposited film is shown in Figure 8 (10).
The solar cell large-area evaporation industry also has a demand to deposit uniform coating thickness at low deposition
temperatures. New linear evaporation sources have been developed for the growing market of solar cell production. One example is
Figure 8 Scanning electron microscopy: (a) cross section and (b) plan view of coevaporated CZTS film. Reproduced from Repins, I., et al.
Co-Evaporated Cu2ZnSnSe4 Films and Devices. Sol. Energy Mater. Sol. Cells 2012, 101(C), 154–159.
www.iran-mavad.com
Pressure
compartment
and slit
Vapor-shaping channel
Vapor
Crucible
Substrate
Figure 9 Schematic of the VPT linear vapor source. After Vacuum Process Technology, LLC, http://www.vptec.com/products_deposition.htm.
a VPT (Vacuum Process Technology LLC) linear vapor source where the vapor flow is directed by vapor-shaping guides. The guides
are metal sheet plates directing vapor from higher vapor pressure regions to the growing surface, as shown in Figure 9 (11).
4.02.2.7 Polymeric Film Deposition

Thermal evaporation is also used to produce polymeric coatings. Thermal evaporation by resistive heating of polymers is limited by
the low thermal conductivity of polymers. On the other hand, electron beam evaporation of organic materials is limited by the
decomposition of the molecules. However, there are some applications where polymeric films have been grown successfully by
thermal evaporation, such as PTFE and nylon in metal–polymer nanocomposite films (12). The main problem is the decrease of
molecular weight at thermal heating. Small polymers up to several thousands of g mol1, equivalent to tens of monomeric units,
have been successfully deposited by vacuum thermal evaporation (13). Polythiophenes (PTh) are important molecules for
producing polymeric semiconducting materials. PTh molecules have successfully been coated by thermal evaporation at 300 C.
A deposition rate of 1 nm min1 has been achieved. The films were highly crystalline without additional annealing. Polymer solar
cell devices have been demonstrated by using these evaporated PTh coatings (14).
4.02.2.8 Spotless Vacuum Arc Process

Thermal evaporation may be assisted by plasma in order to induce ion bombardment to the substrate. In the ion-plating process,
a glow discharge is induced, typically in an argon-containing ambient. The cathode is normally the substrate, and the other electrode
is the vacuum chamber. When arc-discharge plasma is induced between the heated cathode and anode, a deposition process often
called a spotless vacuum arc results (Figure 10) (15). This process has been applied to various metals and nitrides when nitrogen
ambient is used. When an ambient gas is used, the cathode surface needs to supply a sufficient number of thermionic electrons to
allow an arc discharge. The ionization rate of the metal species has been reported to be up to 50%. In order to avoid harmful coating
accumulation on the cooled anode, a heated anode was added that takes part in the evolution of metal plasma in the process
(Figure 11) (16). As the arc does not produce cathodic spots as in the case of a cathode arc on an unheated surface (to be discussed
in detail in this chapter), no droplets are produced.
4.02.3 Cathodic-Arc Deposition Process

4.02.3.1 Cathodic Arcs from the First Experiments to Modern Industrial Deposition
The history of the cathode arc has been recently documented in great detail by A. Anders (17), and the following short history relies
on that source.
The history of cathodic-arc deposition begins along with the history of electricity and the ability to generate electric current.
Experimentation on pulsed arcs or sparks parallels the development of capacitors. Observations of continuous current arcs date to
the times of the development of the battery. The basic phenomena of the cathodic arc were already known but not properly
understood in the early 1900s. Industrial cathodic-arc coating emerged in the last quarter of the twentieth century.
Going to the early history of the cathodic arc, sparks obtained from short-circuiting capacitors were recorded when the Leyden jar
was developed. It is a capacitor constructed from a glass bottle with tin foil inside and outside. Notes of sparks were recorded in the
1750s. Priestly (1734–1804) noted arc spots, metal particle evolution, and a dark-layer deposition, most probably oxidized metal
films. Transparent coatings with Newton rings that were hard to wipe off by the fingertip were also observed.
The development of a battery composed of stacked piles of metal plates of Zn and Ag or Cu followed the inventions of Ales-
sandro Volta (1745–1827). Originally, electricity was associated with animals, which were capable of both producing electricity
www.iran-mavad.com
Figure 10 Schematic illustration of a spotless arc-activated deposition.
(e.g., eels) and reacting to electricity (e.g., frogs). The electric effect of the first stacks of metal batteries was evaluated by recording the
sensation of a human to the shock (much like the quality of wine is evaluated by human sensing today, but perhaps not using as
poetic descriptions). The first battery, of about 60 pairs of metal plates, was used to demonstrate electrochemistry and electro-
plating. Humphrey Davy (1778–1829) constructed a battery pile of about 150 pairs of Cu/Zn in a wooden box. He observed pulsed
arcs using different metal electrodes in air. Vasilii Petrov (1761–1834) in St. Petersburg built an extremely large Volta’s pile of 4200
pairs of Cu/Zn capable of producing continuous current arcs.
Low-pressure experiments were possible after the invention of the vacuum pump. Davy demonstrated the carbon arc lamp at low
pressure for the Royal Society using a Volta’s pile of 2000 pairs of plates.
The invention of induction, known from reports of M. Faraday (1791–1867), enabled researchers to produce high voltages for
capacitor charging. A German, H. D. Rühmkorff (1803–1877), developed a practical high-voltage pulsed current source by applying
induction and an interrupted primary current into a spark coil.
Figure 11 Spotless arc-assisted deposition with an electron beam–heated titanium cathode and anode. Reproduced from Scheffel, B.; Modes, T.;
Metzner, C. Spotless Arc Activated High-Rate Deposition Using Novel Dual Crucible Technology for Titanium Dioxide Coatings. In Extended 20 Abstract
14th International Conference on Plasma Surface Engineering, September 10–14, 2012, in Garmisch-Partenkirchen, Germany.With permission of
Fraunhofer FEP.
www.iran-mavad.com
The evolution of vacuum came into public awareness with a famous show arranged by O. Guericke in Magdeburg in 1654. He
demonstrated the bonding of two hollow hemispheres pumped to low pressure. The piston pump of that time was limited by a leak
in the piston to probably about 1 Pa. J. Geissler (1814–1879) developed a pump with mercury as a sealant in the piston, reducing
the minimum pressure to the vapor pressure of Hg. By using the new power supply and the mercury pump, J. Plücker reported in
1858 the cathodic-arc deposition of metallic films on glass with a mirror appearance. M. Faraday studied systematically the
deposition of different metallic films on a similar device as Geissler by using metals such as Cu, Sn, Fe, Pb, and Zn.
The arc discharge was used by G. Stokes (1831–1879) to investigate the spectral lines in the optical emission of metal plasmas.
Until late 1800s, cathodic-arc development was motivated mainly as a technical solution to illumination.
The early 1900s was the time of the evolution of modern physics. A number of publications on vacuum arcs appeared in relation
to the discovery of the electron and ionization. Experiments demonstrated the evolution of the cathodic spot, steering of the spot,
arc voltage, and probes for plasma measurement. The distinction between cathodic and thermionic arcs was established, and the
effect of oxygen on the cathode spot was recorded. Until the 1920s, the technical applications of mercury arc rectifiers became
particularly important in electric power systems.
The cathodic arc had a role in the massive development of uranium enrichment in the United States during the Second World
War. In the latter half of the 1900s, knowledge of cathodic arcs increased as a result of the advent of ultrafast cameras. In the 1960s,
the current density at the cathode spot was determined to be at least 1011 A m2, and eventually in the 1970s, values of up to
1012 A m2 were reported. Ion velocity was determined to be at a supersonic value of 104 m s1. An explanation for ion acceleration
was available only in the 1990s.
The evolution of industrial cathodic-arc coating started in the 1960s and 1970s. Thermal evaporation was the standard method
of producing thin films at the time. The main application was decorative coating. The deposition of refractory coatings as super-
conducting films (Ta, Nb, and V) was being reported by the 1950s.
In the Soviet Union, strong development of arc processing started in the 1960s. The main application of arc deposition was for
vacuum getter pumps. They were developed in the National Science Center at the Kharkov Institute of Physics and Technology
(NSC KIPT). Research on various hard coatings processed from refractory nitrides and carbides and experiments on DLC were
carried out. The intense development of the still-ongoing thermonuclear fusion reactor included the development of magnetic
confinement of plasma. The KIPT group of V. M. Khoroshikh, V. E. Strel’nitsij, I. I. Aksenov, and V. A. Belois applied the quarter
torus to demonstrate the classic 90 plasma filter for the cathodic-arc source. The industrial cathodic-arc process was based on
several models of a plasma source called Bulat, starting in 1974. During the 1960s–80s, a number of Soviet patents on cathodic-
arc processes were made.
In the United States and Japan, the first cathodic-arc reports were given at the International Symposium of Discharges and
Electrical Insulation (ISDEV) in 1960. In 1979, the Bulat and related plasma sources were applied by a US company, Nobel Field
(now Multi Arc Vacuum Systems Inc.). Cathodic arc has been the core technology of a number of vacuum-coating developers and
job coaters in the Western world since the 1980s. The strength of the cathodic-arc coating industry is its ability to produce high-
quality thin films at competitive costs. The major markets have been decorative and wear-resistant coatings. The largest drawback
is the evolution of microparticles and defects in the coatings that these particles cause. A balance of quality and price has driven
the development and optimization of filtering and steering of the arc in order to apply cathodic arcs to industrial coating
processes.
The needs for twenty-first century thin films often do not allow macroparticles or defects in the films, which is driving the
evolution of highly effective filtering. Industrial deposition devices with sophisticated filtering have been demonstrated: S-shaped,
large-area double-bent, and other filter designs are described more in Section 4.02.3.7. The benefit of high ionization has motivated
the development of cathodic-arc processes for research into and application of nanostructured thin films, ultrathin films, multi-
component films, self-lubricating coatings, and hybrid deposition systems combining arcs with magnetron sputtering and other ion
sources.
4.02.3.2 Unique Features of Cathodic-Arc Deposition

When cathodic-arc plasma erodes, a solid electrode and the ionized species condense on a solid surface at a low ambient pressure,
and a PVD process called vacuum cathodic-arc deposition is obtained. The cathodic-arc discharge is localized in small spots on the
cathode. The high-discharge current density at the cathode spots is the reason for the benefits and problems of the cathodic-arc
deposition process when compared to other PVD processes.
Cathodic-arc deposition is a unique thin-film process due to the fact that (1) the film grows from practically fully ionized species,
and (2) the ions have a kinetic energy Ekin of tens of eV. The cathodic arc also has a unique disadvantage with the evolution of
macroparticles of the cathode material.
Energetic ions are the key to all nonthermal PVD processes. The growth of the film is associated with the impact of an ion with
kinetic energy well above the thermal energy of vapor. A collision cascade and penetration into the subsurface atomic layers result.
Surface bombardment sputters loose species from the surface, which allows dense material to grow, aids mobility at the surface, and
compensates for the substrate temperature often required for thermal vapor processes. The effects of ion energy and substrate
temperature are described in a phenomenological model by Thorton. A. Anders has modified the diagram and relates the originally
used parameters of PVD pressure and substrate temperature to more physical and universal parameters (18). The modified structure
zone model is shown in Figure 12.
www.iran-mavad.com
Figure 12 Structure zone model of thin films grown in energetic ion bombardment as a function of temperature modified from Anders, A. A Structure
Zone Diagram Including Plasma-Based Deposition and Ion Etching. Thin Solid Films 2010, 518 (15), 4087–4090.
There are three parameters in the T*, E*, and t*. T* is the generalized temperature:
T
¼ Ts=T þ Tpot ; [4]
m
where Tpot ¼ Epot/(kNmoved); Ts is the substrate temperature; and Tm is the melting temperature. Epot is the potential energy related to
the energy of sublimation and ionization gained when an ion enters the surface. The energy is divided by the number of atoms
Nmoved rearranged at the heated location of the incoming ion.
E* is the normalized energy related to the displacements and heating caused by the incoming ion. In the simplest case of a film
growing from single species of energetic ions,
E
¼ Ekin=E ; [5]
c
where Ec is the cohesive energy of the material.

The vertical axis of the graph t* is the net thickness of the film.
In the model, the different zones 1–T–2–3 with particular types of crystalline structure are visualized. The zones are dependent
on the ion energy and substrate temperature. Zone 1 has very low adatom mobility, resulting in a fine-grained textured structure
with lots of imperfections. Zone 2 is a uniform columnar structure, and Zone 3 a recrystallized dense film with large crystal size.
Zone T is a transition zone with homogeneous structure of V-shaped grains throughout the whole film. Zone T is reached by
applying a deposition process with a sufficiently high ion energy and ionization rate such as HIPIMS and, in particular, cathodic-arc
processes. In the structural zone model, regions that are not accessible are indicated. The high E* contributes to elevated T* due to
the Tpot term, and thus T* cannot be arbitrarily low at high E* values. When E* is even higher, the self-sputtering coefficient becomes
over unity and net etching of the substrate results.
The high ion energy and self-sputtering are important means to obtain optimal adhesion. Accordingly, cathodic-arc coatings
generally have good adhesion compared to coatings processed with other PVD methods.
The cathodic-arc process does not require an ambient gas since the species of the plasma originate from the solid cathode. The
low ambient pressure leads to a line-of-sight process. As a result, uniform deposition requires sample manipulation or a set of
cathodic-arc sources around the substrate.
The cathodic-arc process can be pulsed or direct current (DC). In the pulsed mode, the plasma density can be very high and the
deposition rate enormous during the active phase of the plasma pulse. A momentary deposition rate of over 2 mm s1 is achievable.
In DC plasma, this would be impossible due to the enormous accumulating heat load to the substrate. However, in pulsed mode,
the typical active period is only 1/1000 and heat dissipation is easily arranged. Thus, benefits to the film properties due to pulsed
plasma are achieved and will be discussed in more detail in this chapter. However, the growth of reactive coatings (e.g., aluminum)
in the pulsed mode leaves the thin-film surface prone to the oxidizing species that adsorb and react on the surface between the
pulses. This leads to difficulties in contamination control, especially in films with high purity requirements.
www.iran-mavad.com
4.02.3.3 Ignition and the Cathode Spot

The cathode spot and its ignition are the critical phenomena in cathodic-arc deposition. The cathode spot is at the interface of
plasma and the solid conducting material (cathode). A special feature of the cathodic-arc plasma is that the plasma current is
concentrated into isolated bright spots that move on the cathode surface. When the different kinds of plasmas are observed as
a function of the discharge current, a transition from the glow discharge to the arc discharge takes place. The arc plasma is char-
acterized by significant lowering of the discharge voltage.
Cathodic-arc evolution is an unwanted breakdown event occurring in high-voltage accelerators. The electric field strength
needs to be maximized in order to maximize the acceleration of ions or electrons in vacuum. Experiments and numerical
simulations have been extensively done to investigate the breakdown using metal electrodes with an electric gap of about
20 mm and a voltage of 4–6 kV. In the case of copper electrodes, the breakdown can occur at about a 200–300 MV m1
electric field (19).
The cathodic-arc spot has a very dynamic behavior. The following phenomena are characteristic to it: a sudden evolution or
explosive emission of electrons, followed by the burst of cathode atoms and ions; the movement of the cathode spot; and eventually
the decay of the spot. The physics of the cathode processes is elaborated in great detail by A. Anders (17).
The spot evolution could be divided into four stages: (1) a pre-explosion stage, (2) an explosive stage, (3) a cooling period with
high electron emission and evaporation, and (4) a final cooldown. The different evolution steps of the cathode spot are illustrated in
Figure 13.
The ignition and other mechanisms that keep the arc burning are complex and the subject of ongoing research. There are
several phenomena that are important, such as electron work function w, electric field strength, surface geometry, plasma
parameters, ion bombardment, surface impurities, and dielectric layers on the cathode (e.g., oxides). These issues are briefly
discussed here.
The ignition at the pre-explosion stage requires a high local electron emission. This is proportional to the local electric field
strength and inversely proportional to the electronic work function of the cathode material. The work function is the amount of
(a) (b)
Electron emission
Electrons
Atoms
Electrons -
-
+
+
- -
- High plasma
- Low impedance
- current density
+
+ -
Hotspot -
Ions
Macroparticles
High local electric field strength
(c) (d)
Electrons
High impedance Cooling down

-
Hot zone with

molten cathode Atoms
material
Macroparticles
Figure 13 The evolution of the cathode spot in four stages. (a) Pre-explosion; (b) explosive stage; (c) cooling with molten cathode material; and (d) final
cooling.
www.iran-mavad.com
energy required for an electron to escape from the conductive material from the Fermi level to the vacuum. The work function is
a material-specific parameter. It depends on the electronic structure of the material (Fermi energy) but also on the crystal structure,
surface reconstruction, and impurities on the surface. A typical value of the work function of metals is about 4 eV. The electron
emission as a function of the electric field follows the Fowler–Nordheim formula, according to which the electron emission depends
roughly on w as 1/w$exp(w3/2). Thus, a lowering of work function rapidly increases the electron emission.
The electric field on the cathode surface is related to the potential gradient at the plasma sheath. The plasma sheath is the
quasineutral boundary in the vicinity of the cathode. It self-adjusts the space charge distributions in order to sustain the balance of
the flow of negative and positive charges. The thickness of the sheath is proportional to the voltage drop, plasma density, and
electron temperature.
The ignition of the cathode spot has been observed to be related to local protruding tips. The local electric field is enhanced by
geometry so that a protrusion or a tip enhances the field as a function of the inverse of the local radius of curvature. In order to ignite
the arc plasma, an avalanche of electron emission is required. In numerical models, local protrusion tips of tens of nm in height and
radius with a field enhancement factor from 10 to 1000 have been suggested. In the case of cathode spots, which result in rapidly
solidified melt pools, such tips have indeed been observed. However, it is unlikely that in general, tips of such dimensions are
present during the pre-explosion stage. On the other hand, breakdown does not happen if no field enhancement is present.
Energetic ions bombard the cathode with the kinetic energy owing to the potential difference of the plasma and the cathode.
Depending on the ion energy, secondary electrons are emitted. The bombardment by the ions is a major reason for local high
emission of electrons.
The arc plasma is quite turbulent and unstable. The turbulence may cause local thickness variations of the cathode sheath. This
implies local ‘holes’ with reduced thickness in the sheath, resulting in high field strengths. Such holes in the plasma sheath can create
conditions favorable for arc ignition.
Adsorbed atoms on the cathode surface can lower the work function of metals provided that the adsorbed atom has a lower work
function than the cathode. Thus, impurities that in practice always decorate surfaces (except in UHV conditions) are important in
cathodic-arc ignition. The impurity effect can also be dynamic. For example, a slowly moving ion approaching the cathode surface
enhances the electron emission. An important effect similar to that of impurity atoms is the role of a dielectric or semiconducting
surface on the cathode. The nonmetallic surface affects the value of the work function, but, more importantly, a field enhancement is
associated with the formation of a thin dielectric film on the cathode. Typically, such a film is metal oxide. A metal with high affinity
to oxygen may chemisorb oxygen molecules on the surface originating from the impurities of the ambient vacuum (e.g., chamber
walls), or in the case of reactive deposition from the gas flown to the chamber. A positive charge is built on the thin oxide surface,
resulting in a high electric field leading to current breakthrough. This mechanism has been shown to be a very efficient enhancement
of arc initiation.
In summary, a field enhancement of the order of 100 is required for the ignition of arc. The origin of the field enhancement could
be a combination of spatial and temporal phenomena, including a field-enhancing tip, a lowering of work function, an increased
electric field by impurities, and a dielectric or semiconducting layer.
4.02.3.3.1 Rapid Sequence of Microexplosions

At the explosive stage, a high-density electron burst occurs. About 1011 electrons are emitted in 10 ns. Due to resistive heating, an
emission of cathode atoms follows. A large number of atoms are ionized by the bombardment of the ejected electrons. A proportion
of the ions are accelerated by the electron flux away from the cathode. A proportion of ions are accelerated back to the cathode
surface, resulting in collision cascades (depending on the maximum kinetic energy). The process is also described as a thermal run-
away where the electron current exceeds the arc current. Then, single or multiple charged ions are emitted from the cathode spot,
compensating for the excess of electron flow. In the case of copper, up to 4 1010 A cm2 ions have been observed. The heat flux
due to incoming ions could be dramatic in the case of a sufficiently high ion flux bombarding the cathode. H. Timco et al. (19) has
numerically simulated the overlap of collision cascades in the case of 10 keV Cuþ ions.
The time evolution of the arc discharge has been experimentally characterized and modeled by using numerical methods. In the
case of metal ions, an evolution of multiple charged ions is observed. In the case of copper, the average ionization stage is about 2,
and with tungsten it is 3. A. Anders has comprehensively tabulated the ion charge stage distributions, covering a large part of the
periodic table of elements (Appendix B in Ref. (17)). Typically, the plasma of the vacuum arc is composed of ions and electrons with
a very low number of atomic neutrals.
A simulation of the evolution of the arc as a function of time is shown in Figure 14.
The lifetime of cathode spots can vary between a few to hundreds of ns. There are three main reasons for spot decaying: (1)
resistive heating causing increased electrical resistance (metals), which reduces the electron emission. However, in the case of
semiconducting materials, the conductivity increases due to elevated temperature. (2) At the surface solid metal becomes liquid and
vapor. The conductivity of metal vapor is much less than the bulk metal conductivity. The electron emission decreases and the
plasma density decreases. (3) There is competition among emission sites due to increase of resistance of the spot, with a high
probability of igniting a new cathode spot.
Two types of cathode spots may be identified. Type 1 cathode spots are generally related to a contaminated surface. The
contamination of, for example, an oxide film enhances the ignition of the spot. The spots are isolated and small. The Type 2 spots are
larger and a result of higher current, and close overlapping of each other forms arc spot traces. This type normally operates once the
surface has become less contaminated as a result of, for example, Type 1 cathode spots.
www.iran-mavad.com
Figure 14 Time evolution of the plasma initiation process in four time steps obtained by numerical modeling. Reproduced from Timko, H. Modelling
Vacuum Arcs: From Plasma Initiation to Surface Interactions. Report Series in Physics HU-P-D188, Theses, 2011.
In the case of metals, the arc current is transmitted at cathode spots, which have a limited lifetime. An apparent movement of the
location of the plasma results from a sequence of ignition and distinction of individual burning cathode spots.
The nature of the arc discharge is largely stochastic and random. In the case of titanium, the arc current and the size of the
individual spot vary on a large scale between 1 and 50 mm in diameter (20). A self-similarity of the cathode spot size has been
observed. A. Anders has analyzed the chaotic nature of the arc discharge with random movement of the location of cathode spots
and with a large variation of arc current and voltage. A fractal nature of cathode spots both spatially and temporally was observed.
The movement of the spots is a stochastic Brownian movement. Sometimes, an individual arc is extinguished for a period of about
10 ns and then ignited, followed by an almost uninterrupted sequence of arc spots. In this case, the individual arc spot is formed,
separated by several mm distance from a track of overlapping arc spots.
The movement is rapid in particular for Type 1 spots. For Type 2 spots, the movement is slower and the new cathode spot is often
ignited in the vicinity of the old spot.
In the presence of a magnetic field, the motion of spots opposite to the direction of the Ampere force has been observed. As
the Ampere force bends the plasma column to the direction of j 3 B, the spot moves counterintuitively in the opposite direction.
The mechanism of the retrograde direction is not well understood, but it seems to be related to the microplasma jet bursts parallel to
the cathode surface with a preferential direction relative to the auxiliary magnetic field. For a magnetic field of 15 mT, the speed of
movement is 10–15 m s1 (21). Magnetic fields may be used to steer the arc spot movement, as described in the following sections.
Empirically, it has been observed that the arc-burning voltage is proportional to the cohesive energy of the cathode material.
This is known as the ‘cohesive energy rule’. The typical current of ions generated by the arc-burning current is in the range of 10%,
known as the ‘10% rule’. The actual value varies between a few percent to 20%. The input energy to the cathode spot from the
plasma scales with the arc voltage. The higher the voltage, the higher is the available kinetic energy of ions at the cathode surface.
www.iran-mavad.com
However, only some portion of the plasma energy goes for the vaporation of the cathode material. A. Anders has observed that
with high-cohesive-energy materials, the discharge current fluctuates at higher frequency than with low-cohesive-energy mate-
rials. This reflects the difficulty of igniting the emission sites of high-cohesive-energy materials. A similar rule is known for the
sputter yield. There could be a connection between these rules since the energy of the incoming ions has a role in the ignition of
the cathode spot.
The most challenging feature of the vacuum arc deposition process is the evolution of microparticles. In terms of material
erosion, the major portion of material removal from the cathode is in the form of particles. Accordingly, the particles comprise
between 10 and 90% of the mass of the growing film.
The particle evolution is most commonly understood as a result of a molten metal pool ejecting molten droplets. In some cases,
the particle hits the growing film while not solidified, producing a flattened droplet on the substrate. On the other hand, metals with
high melting temperature tend to solidify before hitting the substrate. In that case, the particles might bounce off the surface or stick
to the growing surface with a lower probability. That could explain the observation that the fraction of particles of the total mass of
the deposited film was about 3.5% in the case of tungsten and 90% in the case of lead.
There is a broad distribution of the size of particles. When investigating a 50 nm thick copper film, the diameter of the particles
varies between <0.1 mm and >10 mm. If plotted to a log–log presentation, there is a linear correlation with the size and amount of
particles, as shown in Figure 15 (22). This linear distribution is also interpreted as proof of the fractal nature of the cathode spot.
In the case of high-melting-point materials such as carbon, the formation of a molten pool is less probable. The accumulated
thermal stresses caused by the high localized current densities could be the main reason for particle evolution. Also, the micro-
structure and porosity of the graphite cathode have a large effect on the number of microparticles.
4.02.3.4 Arc Plasma

The arc plasma originates from individual cathode spots, with one spot active at a time. Thus, the plasma expands from a point
source.
With no external magnetic field, a plasma pressure gradient and ion–electron coupling dominate. In the case of a large cylindrical
anode symmetric to the cathode, the plasma expands symmetrically with a distribution as a function of angle q from the normal to
the cathode surface. The intensity of plasma is cosaq. The exponent a is related to the pinching of the plasma. If a < 1, there is
spherical distribution, and when a > 1, the plasma is pinched and forms a focused jet.
Figure 15 Macroparticle size distribution for copper and platinum. After Anders, S.; et al. On the Macroparticle Flux from Vacuum Arc Cathode Spots.
IEEE Trans. Plasma Sci. 1993, 21 (5), 440–446.
www.iran-mavad.com
The acceleration of ions takes place near the cathode spot. After this, the plasma expands with a constant velocity. Thus, the
density of plasma reduces as 1/r2 as a function of distance r from the cathode.
The acceleration process that takes place in the immediate vicinity of the cathode spot is assumed to be a result of the following
phenomena: (1) pressure gradient of ions, (2) pressure gradient of electrons, (3) collective acceleration of ions by electron–ion
coupling, and (4) acceleration by an electric field (potential hump). According to Hantzsche (23), the high flux of electrons collides
or interacts via coulomb forces with the metal atoms flowing from the cathode spot. Since the electrons move with much higher
velocity and in much higher quantity than the atoms, kinetic energy is transferred to the atoms. Electrons are stripped off from the
atoms, causing ionization. The very dense plasma has such an electron and ion pressure gradient that the ions are accelerated to their
final kinetic energy in a distance less than 10 mm from the cathode. Within this distance, a dense nonequilibrium plasma with
multiple ionized ions and electrons is formed. Beyond this distance, called the ‘freezing zone’ the charge state distribution is
stabilized. The ionization rate of ions varies and is characteristic for each element. The typical ion energy values vary from a few tens
of eV to over a 100 eV.
An external magnetic field is used to steer the spot motion on the cathode and to guide the plasma. Thus, magnetic fields have
a big effect on the plasma. In the vicinity of the cathode spot, the plasma density is high, even exceeding that of atmospheric
pressure. As the plasma expands, the density reduces to a value where electrons become magnetized due to the external magnetic
field. This means that the electron motion is governed by the gyration motion in the magnetic field. This distance from cathode is
typically a few mm. The electron motion is not affected when the motion is parallel to the magnetic field lines, whereas the motion
perpendicular to the field lines is possible only as a result of collisions or collective effects such as turbulences. Thus, the plasma
flows along the magnetic field lines.
The charge state distribution stays constant as the plasma expands, provided that the chamber pressure is sufficiently low and the
mean free path of the ions is comparable to the vacuum chamber dimensions.
An external magnetic field has been shown to have an effect on the charge state distribution. When the magnetic field is increased
to a value of about 1 T, ionization is increased as shown in Figure 16. The magnetization of the electrons is assumed to be the reason
for the increased ionization. The discharge current creates an internal magnetic field. Increased ionization is observed when the arc
current is increased above 300 A.
The ion energy originates from processes in the close proximity of the cathode spot, and remains constant in the case of vacuum
with no external magnetic field. However, the magnetic field can affect the ion energy significantly.
Ion energy distributions and electron temperature have been measured extensively for several cathode materials (24). By using
a time-of-flight measurement to determine a pulsed plasma velocity, it was found that there is one sharp peak at the velocity
spectrum of ions with a value between 1000 and 3000 m s1. Taking into account that there is a broad distribution of different ion
charge states, the acceleration has to be due to the hydrodynamic effect near the cathode spot independent of the charge state of the
ion. If the acceleration would be caused by an electric field, one would expect to see separate maxima for each charge state.
Figure 16 Increase of average ionization as a function of magnetic field for various metals. After Anders, A. Cathodic Arcs: From Fractal Spots to
Energetic Condensations. In Atomic, Optical, and Plasma Physics; Springer: New York, 2010; Vol. 50.
www.iran-mavad.com
160
ECE (×10)
140 Ekin
Te (×10)
120
100
energy (eV)
80
60
40
20
0
0 20 40 60 80 100
atomic number, Z
Figure 17 The average ion kinetic energy (Ekin) obtained both by modeling and experimentally for different atom numbers. Comparison with electron
temperature (Te) and cohesive energy (EEC). After Anders, A. Cathodic Arcs: From Fractal Spots to Energetic Condensations. In Atomic, Optical, and
Plasma Physics; Springer: New York, 2010; Vol. 50.
In Figure 17, the mean ion velocity for various elements is shown. There is a connection with the cohesive energy of the cathode
material to the arc spot phenomena as described in Section 4.02.3.3. Accordingly, there is also a correlation of ion energy and
cohesive energy of elements (see Figure 17).
In the presence of a magnetic field, the movement of electrons is impeded perpendicular to the magnetic field lines. Due to this,
the plasma impedance increases, resulting in higher arc voltage. This results in an increase of electron temperature, higher pressure
gradient at the cathode spot, higher ion energy, and an increase of the average charge state.
A. Anders recently reviewed vacuum cathodic-arc plasma comprehensively (25).
The gradient of the magnetic field is used to confine plasmas finding practical application, such as in thermo-nuclear fusion
vessels. The electron is either accelerated or decelerated when moving in the magnetic field gradient. When the magnetic field
increases, the electron velocity decreases. Since the ions and electrons are coupled, ion energy decreases or increases accordingly. The
magnetic field applied in particle filters could have a significant effect on ion energy, thus having an effect on the deposited film
structure.
The ambient gas pressure has an effect on ion energy as a result of collisions with neutral atoms. This is particularly the case in
reactive cathodic-arc deposition with several Pa ambient pressure. Then, the mean free path of ions becomes less than the
dimension of the chamber.
4.02.3.5 DC Vacuum Arc

There are two main technical approaches for vacuum arc deposition devices: continuous current DC and pulsed deposition. The DC
method, due to its simplicity and upscalability, is the most common industrial method. Normally, the arc current is in the range of
40–150 A. Increase of the deposition rate and upscaling are performed in a straightforward manner by increasing the DC current to
higher values. The limiting issue is cooling. On the other hand, pulsed deposition provides some advantages such as higher ion
charge stage and lower particle emission. The pulsed arc is quite suitable for the deposition of carbon coatings and the construction
of simple filtered arc sources. In the following, DC vacuum arc sources are separately described in more detail.
The DC vacuum arc system is composed of a cooled cathode and an annular anode. A separate triggering device is also included.
The basic setup is illustrated in Figure 18 (26). Typically, the system is cylindrically symmetric.
A mechanical trigger is commonly used in DC vacuum arc systems. A spring-loaded electrode rod is pushed to contact with the
cathode, causing a trigger spark that initiates an arc. Often, the DC system is optimized to be sufficiently stable with less frequent
need for triggering. Thus, a mechanical trigger could have sufficiently long service life. Welding of the trigger electrode to the cathode
and contaminants from the trigger electrode are two possible problems related to the mechanical trigger. Often, refractory metal
trigger rods are used.
The arc is ignited either by a high-voltage pulse using surface conductivity between the cathode and an ignition electrode or by
using a trigger electrode brought mechanically to a contact with the cathode. A number of different geometries have been developed
to construct the cathode with cooling and to insulate the cathode from the anode and the trigger electrode. Often, the constructions
have been described in patents starting from year 1971 by Snaper.
A magnetic field coil has often been included in the construction that provides several effects, including cathode spot steering
and enhancing the plasma ionization, and ion energy by a magnetic field gradient. Without an external magnetic field, the process is
called ‘random arc’ where the location of the arc is dependent on geometry and local spatial and temporal conditions of the cathode
www.iran-mavad.com
Figure 18 Basic setup of a DC cathodic-arc plasma source.
and the entire system. The system is easy to construct, but the cathode can wear quite unevenly. Also, the spot could move slowly,
resulting in local heating of the cathode to such an extent that it evaporates neutral cathode atoms at the expense of ionization. Such
a process is behind the concept of vacuum arc evaporation.
By controlling the movement of the cathode spot, one can wear the cathode more evenly, and by having the cathode spots move
faster, local heating of the cathode is restricted. The movement may be controlled by the use of an external magnetic field, and the
process is called ‘steered arc’.
The movement of the cathode spot is only virtual; instead, it is the location of the ignition of the consecutive individual arc spots
that is moving. Steering of the arc can be done by using magnetic or electric fields. Steering by a magnetic field is governed by two
different rules: the (1) retrograde motion rule, and (2) acute-angle rule.
The retrograde movement means that the cathode spot moves in the opposite direction of the Ampere force –B 3 j, which was
introduced in Section 4.02.3.3. The acute-angle rule means that the spot moves to the direction of the inclination of the magnetic
field line relative to the normal of the cathode surface. The applications of these rules are utilized in guiding the arc spot to the
wanted location of the cathode and also in increasing the speed of spot motion.
A common means of steering the arc is to have a magnetic coil around or behind the cathode. When using a cylindrical or
blunted conical-shaped cathode, the magnetic field lines drive the cathode spot to the end of the cathode according to the acute-
angle rule (see Figure 19). This helps drive the spot from the location of the ignition electrode to a location more advantageous to
the deposition process. Often, such cathode geometry is applied in filtered arcs. Another arrangement is to provide an arched field
line where the field lines are parallel to the cathode surface as in magnetron-sputtering targets. The acute-angle rule keeps the
cathode spots in a race track trajectory commonly seen in magnetron sputtering.
Permanent magnets can be used to steer the arc as well. There are examples where a permanent magnet is placed behind the
cathode and rotated to ensure a more uniform wear of the cathode, as shown in Figure 20 (27).
The geometry of a cylindrical rotating cathode has been developed particularly to the upscale deposition process. A permanent
magnet can be mechanically moved inside the cooling channel of a hollow cathode rod.
In the case of long cylindrical cathode rods, an interesting configuration for arc steering is accomplished by the direction of the
DC current. The spot tends to ignite at the location of the lowest potential. Thus, the arc moves to the direction of the more negative
side of the electrode. If the spot motion is detected by a sensor, the polarity of the steering voltage can be changed and the
movement of the spot is reversed (see Figure 21) (28).
Some cathode materials require special conditions and design. Semiconducting materials such as boron may be used only if the
conductivity is sufficiently high. This can be accomplished by heating the cathode.
4.02.3.6 Pulsed Vacuum Arc

Pulsed vacuum arc offers some unique benefits compared to DC. It is possible to miniaturize sources; they can have lower average
power and thus an easier thermal management. Also, the thermal load to the substrate may be kept low, allowing deposition on
polymeric substrates, for example.
www.iran-mavad.com
Figure 19 (a) Magnetic field lines through a conical cathode and (b) arched magnetic field lines. After Anders, A. Cathodic Arcs: From Fractal Spots to
Energetic Condensations. In Atomic, Optical, and Plasma Physics; Springer: New York, 2010; Vol. 50.
A schematic illustration of a pulsed cathodic-arc device is shown in Figure 22. The cathode is connected to a capacitor or a pulse-
forming network. The plasma is ignited by a spark on the surface of an insulating ceramic material.
Miniaturized pulsed sources are possible when the pulse length is short enough for the spot to stay on the cathode surface. The
DC random arc or steered arc generally requires a cathode with a minimum diameter of 5 cm. Miniature-pulsed cathode sources
with dimensions as low as 3 mm in diameter have been developed (17). Naturally miniaturized sources have a limited reservoir for
deposited material, and the sources could be applied for very-thin-film deposition such as catalytic layers. However, a feed of a thin,
long cathode rod is possible. Miniature arc sources have also been used as ion sources capable of producing highly ionized metal
beams and as microthrusters for space applications.
The duty cycle can be arbitrarily low, which may be achieved by means of shorter pulses and longer intermediate time between
the pulses. The duty cycle can be used to control the heat load to the growing film and substrate.
Pulsed sources can have very high deposition rates, even higher than when using DC. Since the arc voltage in DC is constant
(about 40 V) in all vacuum arc processes, the deposition rate is proportional to the average arc power. In pulsed mode, a very high
arc current may be obtained (in the kA range). Then, several cathode spots burn simultaneously. Two arc spots repel each other by
the retrograde spot movement rule due to the magnetic fields created by the arc plasma filaments. In order to control the wear of the
cathode, the arc could be ignited at the center of the cylindrical cathode. The arc spots drift on average radially to the edge of the
cathode, as seen in Figure 23.
A high deposition rate source can be coupled with a solenoid plasma filter to obtain a high deposition rate of highly ionized
plasma with a reduced amount of particles.
www.iran-mavad.com
Figure 20 Arc spot steering by using a rotating permanent magnet placed behind the cathode. Reproduced from Ramalingam et al. Control of Vacuum
Arc Material Operation, Method and Apparatus. US Patent, 4,673,477, 1987. For details of the figure please see the source line reference.
Miniaturized plasma sources allow a design of multiple cathodes feeding a single plasma filter. A system with two arc sources
coupled with a 90 filter is shown in Figure 24 (29).
Some problems such as contamination of the film arise with the pulsed deposition. For example, the hydrogen level can increase
up to 0.09 at.% in Al2O3 coating when using pulsed cathodic arc. The reason is that during the low-duty cycle, ambient gas
molecules stick and are chemisorbed on the growing film. In the case of DC, the adsorbed molecules are more likely to be knocked
off from the surface before chemisorption due to the continuous flux of the incoming ions bombarding the growing film (30).
The vacuum arc may be triggered by several different ways: by a mechanical contacting electrode, by a trigger electrode causing
a surface discharge on a contaminated insulator, or by applying a laser pulse.
The ignition arc can easily be obtained when a short (some ms) high-voltage pulse is fed to an electrode that is separated from the
cathode or from another trigger electrode by an insulator surface. The insulator surface obtains a conducting ‘contamination’ film
and produces a spark plasma due to the high-power density in the thin film. The conducting coating is automatically replenished
from the cathode material. The trigger electrode could be located at the center of the cathode or at the outer rim of the cathode, or be
placed cylindrically around the cathode in the area between the cathode and the anode. The different trigger concepts are shown in
Figure 25. A simple variation of the surface spark triggering is to separate the cathode and anode with an insulating ceramic. The arc
power source is first feeding a high-voltage pulse for triggering, followed by the actual discharge pulse.
Laser beams have been used for triggering the arc plasma developed by FhG-IWS in Dresden (see refs within (31)). A pulsed laser
is focused through a vacuum chamber window to the location of the cathode surface where the arc is to be ignited. The laser
triggering offers a unique benefit in the use of multimaterial cathodes. The laser spot is directed optically to the selected segment of
www.iran-mavad.com
Figure 21 Arc deposition with cathode rods (14) equipped with a switchable potential (60) difference between the cathode rod ends steering the
movement of the arc, and with arc spot sensors (28,29). Vergasol, G.E. Electric Arc Vapor Deposition Device. US patent 5,037,522 ,1991. For details of the
figure please see the source line reference.
the multimaterial cathode. The use of a cylindrical rotating cathode with sectors of different materials (e.g., graphite and titanium)
allows the deposition of multilayers (shown schematically in Figure 50).
4.02.3.7 Macroparticles
4.02.3.7.1 Particle Effects
The arc discharge–deposited films are affected by macroparticles. The hitting of a particle with the growing film has several effects.
The particle may stick to the surface. Depending on the size of the particle, it either is embedded into the film or causes shadowing
and nodular structures to evolve. Metals of low melting temperature could decorate the growing film with circular, rapidly solid-
ifying splats. Various types of macroparticles in the cathodic-arc deposited coatings are shown in Figure 26. A clear reduction of
macroparticles is observed by using magnetic filtering. Since the filtering efficiency is not 100%, a fraction of macroparticles hit the
growing surface causing defects in the growth process of the film due to, for example, shadowing effects. Defects, pinholes, and
cavities in the films are formed. The hitting and embedding of macroparticles could cause defects on the growing film, resulting in
nodular defects even in the case of amorphous films such as carbon.
In applications such as tool coatings, the cathodic-arc coatings have excellent tribological properties. Commonly, they provide
better abrasion resistance compared to other PVD coatings. The overall quality of the film around the macroparticles often
compensates the problems caused by the defects induced by the macroparticles. In applications where an absence of defects and
pinholes is critical (e.g., in electronics and barrier coatings), effective filtering of macroparticles is necessary.
The effect of macroparticles on tribology is discussed in more detail in Section 4.02.3.10.1.
www.iran-mavad.com
Figure 22 A simple pulsed cathodic-arc device.
Figure 23 Arc spot traces of a pulsed arc on a Cu cathode. Photo courtesy and with permission of Dr Peter Siemroth.
www.iran-mavad.com
Figure 24 Dual-cathode filtered source equipped with a pulsed bias source and computer control for designing composite thin films. After Anders, A.;
Pasaja, N.; Sansongsiri, S. Filtered Cathodic Arc Deposition with Ion-Species-Selective Bias. Rev. Sci. Instrum. 2007, 78 (6), 063901.
Figure 25 Different trigger constructions: (a) trigger at the center of the cathode; (b) at the outer rim of the cathode; or (c) placed cylindrically around the
cathode in the area between the cathode and anode.
4.02.3.7.2 Control of Particle Evolution

In pulsed arc, the discharge current is large compared to in DC arc. The arc evolves from rapidly created new cathode
spots. This reduces the thermal stresses at the cathode spot. This has been shown to reduce both the size and number of
particles (17).
In reactive arc discharge deposition, poisoning of the metallic cathode by a nonmetallic layer may occur. This layer
promotes the evolution of cathode spots, thus also reducing the lifetime of individual spots. Thus, a reduction of particle
evolution results.
www.iran-mavad.com
Figure 26 Macroparticles in cathodic-arc TiN deposited for 2 h by using a (a) filtered arc; and (b) direct arc. Reproduced from Gorokhovsky, V. I.;
Bhattacharya, R.; Bhat, D. G. Characterization of Large Area Filtered Arc Deposition Technology: Part I – Plasma Processing Parameters. Surf. Coat.
Technol. 2001, 140 (2), 82–92.(c) Nodular defects in a ta-C coating without filtering. Reproduced from Drescher, D., et al. A Model for Particle Growth in
Arc Deposited Amorphous Carbon Films. Diam. Relat. Mater. 1998, 7 (9), 1375–1380.
Both evaporation and net weight gain of macroparticles in plasma can occur. In the case of material with high vapor pressure,
evaporation has been reported to increase the portion of neutral atoms in the plasma (34). Attempts to reduce the number of
macroparticles by maximizing their evaporation has not been successful; on the contrary, in some cases a weight gain of the
macroparticles due to collision with ions and neutrals is possible (17).
H. Takikawa recently reviewed the methods of reduction of macroparticles in cathodic-arc deposition (35). Four methods are
mentioned: (1) Steered arc, in which the rapid movement of arc spot location lowers the local temperature, causing less thermal
stress and particle evolution. (2) A limited reduction of particle evolution could be obtained by a ‘current-controlled arc’. If pulsed
DC is used, the arc current is suppressed before the cathode spot becomes hot. The limited effect is understandable since the cathode
spots are composed of a series of spots of varying size and lifetime in a fractal manner. (3) ‘Distributed arc’ or thermionic arc occurs
www.iran-mavad.com
when the cathode surface is at a sufficiently high temperature in order to provide high electron thermal emission, which is
comparable to the discharge current. The explosive cathode spot processes are not present, and high deposition speed with little or
no macroparticles is possible. However, the ionization rate and the ion energy are also low, and thus lack the essential plasma
characteristics typical for cathodic-arc processes. (4) ‘Shunting arc’ is a process where a pulsed current heats a rod, resulting in the
evaporation of the surface. Then, by further increasing the discharge current, an arc discharge occurs in the vapor. In principle, no
macroparticles evolve, but again the ionization and the ion energy that would evolve from the cathode spot process are not present.
In summary, the reduction of macroparticle evolution is possible with several methods. However, a number of macroparticles
evolve due to the cathode spot process, and the only option to clean the arc plasma from particles is by filtering.
4.02.3.7.3 Particle Filtering

The evolution of particles is inherently connected to the cathodic-arc process. There are various means to reduce and control the
evolution of particles, as described in the previous section. The most efficient means to reduce the particles is filtering of the plasma.
The main idea is to guide the plasma by an electro-magnetic field so that the massive and mostly neutral particles are not affected.
The first magnetic filters for arc deposition were introduced by Aksenov et al. (36). In the following, the basic physical principles of
the filters are introduced, followed by a description of the filter design and filter applications as described in Ref. (17).
4.02.3.7.3.1 Physical principle of filters

The particle filter most often used is a curved magnetic field produced by electric coils causing the arc plasma to curve. The particles
move in a straight line and are collected separately. As described in Section 4.02.2.4, the physical modeling of arc plasma is
challenging. The bending of plasma by a magnetic field can be described only qualitatively in this context. The Lorentz force of the
magnetic field on the charged particle moving with a certain velocity forces the particle in a curvilinear trajectory. The gyration radius
of electrons is small and should be less than the dimension of the plasma filter (i.e., have a smaller diameter). The ions, being orders
of magnitude more massive than electrons, have a gyration radius typically larger than the filter dimensions. The magnetic field lines
cause electron charge density to have a gradient, and roughly speaking, the magnetic field lines are equipotent lines in the plasma
that are responsible for the guidance of ions. Thus, the magnetic field lines affect the electrons in the plasma, which then pull the
ions to the curvilinear path by electrostatic force.
The collisions between ions, neutrals, and electrons interfere with the magnetization of charged particles in the plasma. The
collision frequency is proportional to the density of species in the plasma. A typical value of the magnetic field (or induction) is
10 mT. The electrons with kinetic energy of 2–4 eV are able to make several gyrations up to an electron density of 1021 m3 before
a collision. Thus, electrons are magnetized generally in most practical plasma and gas densities.
By electric probe measurements, it has been observed that the magnetic field induces electrostatic potentials of the order of
500–1000 V m1 inside the filter. This is in most cases sufficient to bend the energetic ions in the plasma. Experimentally, it has been
observed that a large portion of ions are lost in the filter. This may be a result of several effects, and is not easy to quantify with simple
plasma models. The ions are oscillating radially while transmitted through the filter. The oscillation amplitude is dependent on the
magnetic field strength. The ions with sufficient radial oscillation amplitude will be lost to the walls of the filter. By applying magneto-
hydrodynamic models, the ability of a filter to transmit ions is exponentially proportional to the magnetic field strength, achieving the
highest efficiency at high field strength. Typically, a 90 bent filter has an ability to transmit 10–30% of the ions entering the filter.
4.02.3.7.3.2 Different filter designs

The design of filters is a question of optimizing the filtering efficiency, growth rate of the coating, size, and cost. In the case of low-
melting-temperature metals (e.g., zinc), the droplets are at least partly molten and have a high probability of sticking to the wall
where they hit. On the other hand, most particles of high-melting-temperature materials (such as refractory metals and carbon)
bounce elastically from the walls and after a finite number of collisions could land on the substrate. To reduce the probability of
particles bouncing to the growing film, particles are trapped by fins and other geometrical obstructions. Magnetic filters can reduce
the number of particles by several orders of magnitude, but typically not to zero. Recently, Takikawa et al. (35) has presented the
various magnetic filter types shown schematically in Figure 27 (37).
The straight filter (see Figure 27(a)) mainly focuses on the plasma and reduces the number of particles at the substrate. This
approach is commonly used in industrial applications. Typically, the side wall is equipped with plate fins to trap particles. Another
method is to block the line of sight to the substrate by inserting an additional magnetic coil or simply a plate on the axis of the
straight filter (see Figure 27(b)).
The straight filter can be bent off-axis, eliminating the line of sight (see Figure 27(c–e, g)). By applying bent-focused arc plasma,
slightly different trajectories of different ion masses have been observed when using composite cathodes. This allows composition
modification of the substrate by adjusting the magnetic filter parameters during deposition (38).
The dome-shaped filter applies the cathode–anode geometry to block the line of sight to the substrate (Figure 27(h)).
There are several variations of the bent filters. The plasma duct could be bent into a knee shape, L shape, or quarter torus. The
magnets could be individual coils, segments, or even continuous coils winding along the duct. Generally, the ion transport is higher
and particle filtering less efficient if the filter duct is shorter and less bent. Double-bent filters, as shown in Figure 27(e), represent the
most commonly used filters for highest efficiency. Even 120 J-shaped filters have been used.
A filter with high efficiency is a bent filter with linear geometry. This geometry has been applied by the Tel Aviv group for a large-
area deposition of optical coatings (39) (Figure 28).
www.iran-mavad.com
Figure 27 Different types of filter systems. (a) Rectilinear; (b) rectilinear with a magnetic island; (c) knee; (d) 90 duct; (e) off-plane double bent;
(f) Venetian blind; (g) S-filter; (h) dome; and (i) stroboscopic filter. After Takikawa, H. & Tanoue, H. Review of Cathodic Arc Deposition for Preparing
Droplet-Free Thin Films. IEEE Trans. Plasma Sci. 2007, 35 (4), 992–999 and Anders, A. Approaches to Rid Cathodic Arc Plasmas of Macro- and
Nanoparticles: A Review. Surf. Coat. Technol. 1999, 120, 319–330.
Figure 28 Schematic illustration of a cathodic-arc deposition system with a linear bent filter. Reproduced from Zhitomirsky, V. N., et al. Properties of
SnO2 Films Fabricated Using a Rectangular Filtered Vacuum Arc Plasma Source. Thin Solid Films 2008, 516 (15), 5079–5086.
www.iran-mavad.com
Figure 29 Schematic image of a T-shaped magnetic filter with two plasma sources, including several magnetic coils for plasma optics. After Aksyonov,
D.; et al. 2010. Deposition of Ti-Al-N Coatings Using Two-Channel T-Shaped Magnetic Filter. In 24th International Symposium on Discharges and
Electrical Insulation in Vacuum (ISDEIV), pp 494–496.
The 90 bent-filter concept has been extended to have two arc sources facing each other in a T shape and sharing some of the
magnetic coils. Then, there is a source of two plasmas from different cathodes from one direction to the substrate (Figure 29) (40).
A commercial version of this principle is called large-area filtered arc deposition (LAFAD), where one or more T-shaped filtered
sources are mounted to a deposition chamber (32), including also magnetron-sputtering sources.
To aim at a large deposition area, Venetian blind filters have been demonstrated (Figure 27(f)). An array of blades blocks the line
of sight to the substrate. When a high current is driven through the blades, a magnetic field is induced, enhancing the filter efficiency
to transport plasma through the blocked line-of-sight path. Filters of 30 30 cm have been demonstrated. The filter has been
optimized to reduce the number of particles to an acceptable level (e.g., in metal nitride deposition (41)).
For mainly research purposes, open-coil bent filters have been used (42,43). There, the particles are allowed to fly through the
filter coil (see Figure 30). In the case of pulsed arc, the arc current is driven through the bent coil, resulting in a very simple device.
Such filtered sources have been miniaturized (44,45). Small filtered plasma sources offer the flexibility to construct setups of
multiple sources in a single chamber for research. A. Anders has demonstrated a miniaturized filtered source with two individual
cathodes within one filter coil. Composite or nanolaminate coatings are possible by controlled firing of different cathode materials
(see Figure 24).
Figure 30 Open curved filter for carbon plasma. The hot macroparticles exit the filter. Reproduced from Anders, A., et al. Ultrathin Diamond-Like Carbon
Films Deposited by Filtered Carbon Vacuum Arcs. IEEE Trans. Plasma Sci. 2001, 29 (5), 768–775.
www.iran-mavad.com
Filters with no magnetic field have also been investigated. One approach is an electrostatic filter, as demonstrated by the Dresden
group (see Figure 50). One can also simply align the flat substrate so that the angle of incidence of the plasma is 90 . The ion
trajectories are bent to the surface due to sheath potential at the substrate. Also, a negative bias potential may be used. The ions are
bent near the substrate surface, thus reducing the number of particles on the substrate. Finally, a stroboscopic method to filter
particles hitting the substrate has been investigated (Figure 27(i)). A rotating disk with a hole is synchronized with the pulse
ignition.
The magnetic filtering effect could also be enhanced by applying an external radial electric field. A. Anders showed that there is an
optimum bias voltage inside the plasma filter in order to enhance the filtering of a 90 filter. The optimum value was 10–25 V. If the
bias voltage is increased, more electrons are attracted from the plasma, thus interfering with the magnetic filtering effect.
4.02.3.8 Typical Coatings by Cathodic-Arc Deposition

4.02.3.8.1 ta-C
Carbon as an elemental material has an astonishing number of extreme properties. Amorphous thin films of DLC are covered in
Chapter 4.15 in this book, but the DLC coatings obtained in particular with cathodic-arc deposition are elucidated in more detail in
this chapter. ta-C coatings may be deposited by using energetic ion beams such as cathodic arc, mass-separated ion beams, pulsed
laser deposition, and lately also HIPIMS (46). Traditionally, DLC also contains hydrogen, mainly when the source material of
carbon is hydrocarbon gas. ta-C coatings can also be alloyed with hydrogen by using reactive vacuum arc deposition.
The elastic modulus and hardness values of ta-C films correlate with the proportion of the sp3 bonds formed. The bond structure
of the coating depends strongly on the ion energy of the carbon ions (47). An optimum ion energy window for high sp3 content
reported by several groups is between 40 and 200 eV. The highest sp3 contents by using cathodic-arc deposition are over 80%. Such
ion kinetic energy is possible, particularly in pulsed-arc deposition with sufficiently high arc current. A value of 40 eV has been
measured when using a 3 kA arc current (48). A straightforward way to increase the kinetic energy is by using substrate biasing. The
energetic ion hits the surface and is slowed and stopped at a certain depth depending on the ion energy. A sequence of effects with
different time frames results. The duration of the collision effects is about 0.1 ps, the thermal phase about 1 ps, and the diffusion
type relaxation about 100 ps. The process of carbon film growth has been modeled by using a subplantation model.
In the subplantation model, the energetic ion penetrates the substrate and causes a local stress field and local energy dissipation,
resulting in the formation of the dense, fourfold, coordinated metastable sp3 structure. As the ion has been stopped, the absorbed
kinetic energy is distributed to a small volume called a thermal spike. For a 40 eV ion, the diameter of the thermal spike is about 5 Å,
thus affecting only the nearest neighbors of the carbon atom. If the ion energy is lower, the incoming ion range is too small to cause
such a local stress field. Then, the resulting bond structure is mainly sp2. In cases in which the ion energy is too large, the dissipated
energy is sufficient to allow the relaxation of the local compressive stress, leading again to an sp2-bonded structure. The relaxation
process can be significant. A detailed comparison of the subplantation model of densification has been reported by M. Bilek et al. (49).
The high sp3 content is obtained only at low temperature (typically less than 100 C). Higher deposition temperatures cause
relaxation of carbon to the sp2 bonding. The sp3 content is observed to be quite independent of the deposition temperature until
a transient temperature. After the transient temperature, the film is growing with an sp2 structure, as shown in Figure 31, where
a pulsed vacuum arc was used (48).
It may be assumed that the energetic ion growth causes an sp3-bonded structure. This transition effect could be fitted to a simple
model that assumes relaxation by zero-order reaction kinetics. Then the sp2 bond content increases due to relaxation of the sp3
bonds as exp(Ea/kT), where Ea is the activation energy, k the Boltzmann constant, and T the temperature.
Figure 31 The effect of deposition temperature on sp3 content by using pulsed vacuum arc.
www.iran-mavad.com
Table 2 The effect of deposition growth rate on the transition temperature above
which the cathodic-arc films are sp2 bonded
Deposition rate (nm s1) Transition temperature ( C)
2380 400
620 350
370 300
1.5 180
The transition temperature has been reported to vary depending on the deposition conditions. The effect of the deposition rate
has been observed to affect the transition temperature. Interestingly, the transition temperature increases as the deposition rate
increases. In Table 2, the effect of deposition rate on the sp3 content is shown (48).
The density of the film is closely related to the sp3/sp2 bonding structure of the film. There is a direct correlation of the density
with the sp3 content. The highest sp3-structured films have a density up to 3.3 g cm3, which is close to that of crystalline diamond
(3.45 g cm3) (50).
The ta-C films are very smooth when deposited by filtered arc. Values of atomically smooth 0.1 nm Ra values are reported. The
energetic ion deposition allows surface mobility of atoms, thus enabling the growth of an even smoother surface than the original
substrate roughness (51). The relaxation to form sp2 structures is due to the diffusion of atoms to the surface. This causes increased
roughness of the surface. The evolution of surface roughness is pronounced when the sp3 content of the film is lowest.
The uniqueness of ta-C is its high value of hardness and elastic modulus. The values of various types of ta-C and doped ta-C
collected from literature are shown in Figures 32 and 33 (50). The high elastic modulus is a manifestation of the fourfold bonding
of sp3. The E values for ta-C vary typically between 300 and 800 GPa. The highest elastic modulus is that of crystalline diamond
1000 GPa. The amorphous ta-C coating has been shown to have a very high elastic modulus, even as an ultrathin film of less than
10 nm. Analysis of the bond structure of ta-C coatings by using electron energy loss spectroscopy (EELS) in transmission electron
microscopy has shown that 1–2 nm of the top surface is mainly sp2 bonded. Beyond this, a constant level of sp3 content is reached
throughout the bulk of the film.
The very high value of the elastic modulus is applied in acoustic wave filters and for microelectromechanical systems.
The high hardness value makes ta-C films unique. H values of above 50 GPa are generally called ultrahard. Hardness and elastic
modulus values are usually measured by using nanoindentation. There is large scatter, particularly in the maximum reported values
for hardness. Several reasons cause this uncertainty. These include the large elasticity of the ta-C coatings observed in the
Figure 32 The correlation of hardness and density of various DLC coatings deposited by different methods.
www.iran-mavad.com
Figure 33 The correlation of elastic modulus and density of various DLC coatings deposited by different methods.
indentation; the hardness of the ta-C is close to the hardness of the indenter diamond tip, and the hardest ta-C films are relatively
thin due to the high intrinsic stress.
The hardness and elastic modus are observed to correlate roughly as H/E ¼ 0.1.
High intrinsic compressive stress is inherently related to the high content of sp3 bonds. Values above 10 GPa are reported. The
high compressive stress is detrimental for the adhesion of the coating. Thus, control of stress has been investigated. Thermal
annealing after the deposition of the coating has been shown to reduce the intrinsic stress to almost zero after 600 C. At this
temperature, roughly 2% of the sp3 bonds convert to sp2 (52). In the process, these sp2 structures align so that the p bond is
perpendicular to the surface and the film volume is decreased, lowering the compressive intrinsic stress. Temperatures up to 1000 C
are needed to convert ta-C to sp2 a-C in vacuum or in an inert atmosphere. The stress has also been lowered by high-energy ion
implantation. Doping of the film, particularly with boron, has also been observed to lower the compressive stress by almost 90%
while maintaining the very high E and H values (53).
ta-C coatings have excellent mechanical properties, but the critical boundary condition in all applications is the adhesion. The
interface of the substrate and ta-C film is the critical place for a detrimental crack to grow. The driving forces are high due to internal
compressive stress of the coating, sharp change of the elastic modulus, and often lack of chemical bonding. Adhesion of ta-C is
managed by applying the tricks common to all thin films: cleaning the interface, applying bond layers, inducing interface mixing,
and controlling the internal compressive stresses. Adhesion promotion is an art for all coaters but is reported rarely, and generally
not described in detail or even mentioned at all in the literature.
Prior to coating, the substrate surface is often sputter etched by using a separate ion source or glow discharge with argon ions.
The adhesion-promoting interface layers are often refractory metals that have high affinity to carbon. They form metallic alloys
with the substrate material (particularly in the case of metals) and have sufficient mechanical properties (e.g., shear stress).
Commonly applied thin films are Cr and Ti with a thickness of some tens of nm. The adhesion of the interlayer is critical.
Particularly with cathodic arc, the use of metal plasma for cleaning has been very efficient. The self-sputtering rate of metal ions is
over 1 when the ion energy is typically higher than a few hundred eV. In the case of aluminum, the critical ion energy is about
500 eV, and titanium 1200 eV. The self-sputter coefficient of some metals is shown in Figure 34.
The use of particularly high ion energy to enhance adhesion of ta-C is a common solution (54,55). The high ion energy sputters
off adsorbed molecules from the surface and allows ions to penetrate through the often contaminated and defected surface layers.
High ion energy causes ion-beam mixing and in most cases results in a diffuse interface at the boundary of the substrate and
coating. The high ion energy is often produced by substrate biasing in a pulsed form. This reduces the heat load to the substrate
and prevents arcing of the substrate, which otherwise could become an arc cathode. Negative bias voltages from some hundred eV
to one kV or more have been used. The pulse duty cycles can be varied. When the pulsed bias is in kV or tens of kV values, the
process is often called ‘plasma-based ion implantation and deposition’ (PBII&D) (56). PBII was originally developed to implant
www.iran-mavad.com
Figure 34 Calculated self-sputter rate of various metals. After Anders, A. Cathodic Arcs: From Fractal Spots to Energetic Condensations. In Atomic,
Optical, and Plasma Physics; Springer: New York, 2010; Vol. 50.
metals (or semiconductors) with gaseous ions, particularly with nitrogen by surrounding the substrate in a nitrogen glow
discharge plasma. Then, the substrate was pulsed negatively, forming a wide plasma sheath accelerating the ions to the substrate.
PBII has been combined with vacuum arc, which is a natural source of highly ionized metal plasma, to reduce the costs in metal
ion implantation.
The ability to deposit highly stressed ta-C films with thickness of over 1 mm has been demonstrated. High adhesion has been also
demonstrated by depositing such films on (100) silicon wafers. Instead of delamination of the stressed film by interfacial cracking,
the bulk of the silicon wafer is cracked, as shown in Figure 35. One means to relax the high stress caused by the ta-C film is to apply
a soft substrate or a soft interlayer.
Graphitic particles cause defects in the growing coating. In films with thickness less than 1 mm, pinholes are found. They are
observed by optical microscopy. The large variation of particle size causes a variety of defects in the coating. Small particles are
embedded in the coating. It has been observed that the number of pinholes reduces as the thickness of the coating increases.
Pinholes can be detected by using acid, which dissolves the substrate material. The growing film seems to repair the defects, causing
embedding of the particles. In the cross section of the coating, nodular growth is observed for coatings of over 1 mm thickness (57).
Shadowing and low mobility of carbon atoms result in conical growth nodules, which are easily detached from the coating. The
growth direction of nodules is observed to follow the angle of incidence of the ions (33). The behavior is similar to the tilted
columns observed in magnetron sputtering.
The number of particles hitting the growing film is controlled by using filtering methods, as discussed in detail in Section
4.02.3.7.3.
The particles are most harmful in applications such as integrated circuits, where one defect is likely to destroy the operation of the
entire circuit. A high deposition rate by using filtered cathodic arc to coat with ta-C films is difficult to achieve. However, there are
numerous applications that are not sensitive to defects on the coating, such as protective coatings in tools and mechanical
components. On some occasions, the roughness caused by particles is even beneficial to the application, as in the case of lubricated
tribological contacts. There, the voids on the surface act as lubrication pockets that transport lubrication molecules to the critical
contact zone, to be discussed in more detail in Section 4.02.3.10.1.
The quality of the cathode material has an effect on the particle emission. Graphite is produced by high-temperature processing.
The material is polycrystalline, and a high density is difficult to obtain. Grades of graphite vary in density between 1.6 and
1.9 g cm3. The yield of macroparticles is affected by the density of the graphite cathode. With dense graphite cathodes, fewer
particles are produced.
The cathode erosion forms complicated shapes and structures on the surface of graphite. The cathode spot seems to burn in
valleys of protrusions with less erosion. Eventually, the yield of carbon plasma reduces, as there is a limited line of sight from the
cathode spot to the substrate. These obstructions have been observed to also reduce the number of particles acting as miniature-
sized mechanical particle filters.
Doping of ta-C has been investigated extensively. Different doping and alloying elements such as hydrogen, nitrogen, boron,
metals, fluorine, and polymers have been reported.
Hydrogen is a natural doping element of carbon films produced by PVD or plasma-enhanced chemical vapor deposition
(PECVD) from hydrocarbon gas. Hydrogen bonds with carbon and terminates the dangling bond. It prevents the sp2 double carbon
www.iran-mavad.com
Figure 35 High adhesion and high stress of ta-C film causing the decohesion of a silicon wafer surface. Reproduced from Alakoski, E., et al. Adhesion
and Quality Test for Tetrahedral Amorphous Carbon Coating Process. Diam. Relat. Mater. 2003, 12 (12), 2115–2118.
bond formation and lowers the conductivity of the amorphous carbon structure. The low content of hydrogen in sp3-bonded
carbon yields very similar structure and properties compared to hydrogen-free material. Thus, the material is called hydrogenated
tetrahedral amorphous carbon (ta-C:H). Hydrogen may be added to ta-C simply by feeding hydrogen or hydrocarbon gas into the
ambient during the deposition. Hydrogen contents between 0 and 30% have been observed.
4.02.3.8.2 Nitrides
Nitrides are an important PVD thin-film type. The process needs a metal target and typically nitrogen and argon ambient. The
cathodic arc provides a high ionization rate, which is ideal for nitride thin-film processes. Typical nitride films produced by
industrial cathodic-arc processes are listed in Section 4.02.3.10. In this section, some important nitride cathodic-arc processes are
discussed in more detail.
4.02.3.8.2.1 TiN
Titanium nitride coatings have been the success story of PVD coating development. TiN has excellent wear properties for tool
coatings and a gold color that is perfect for decorative coatings. Cathodic arc provides a highly ionized plasma allowing the
deposition of dense, fine-grained TiN coatings. Titanium is often used as an adhesion-promoting layer for other coatings in PVD
processes. Thus, the TiN-protective coating process may be relatively easily optimized to provide sufficient adhesion to most of the
substrate materials. The main limitation in the use of cathodic arc to deposit TiN is the particle emission. However, by applying
steered arc, tribological TiN coatings are produced successfully.
The high level of ionization has allowed one to optimize the TiN hardness. The development of titanium-based nitrides has led
to ternary nitride nanocomposites and multilayer coatings, which are described in more detail further in this chapter.
Owing to the huge ionization of the metal plasma, the TiN coatings have a very high internal compressive stress. Cheng et al.
(58) used a large-area filtered arc deposition (LAFAD) device to produce multilayers of TiN/Ti. By applying 40 eV bias voltage, the
TiN coating had a 2.2 GPa compressive stress. When a 3 mm coating with alternating layers of TiN and Ti 240 and 20 nm of thickness
www.iran-mavad.com
was deposited, the internal stress decreased to 1.6 GPa. This was explained to be a result of strain release of compressed TiN layers by
the deformation of the softer Ti layer. The expense of soft Ti layers was the reduction of microhardness from 32 to 20 GPa.
4.02.3.8.2.2 CrN
CrN is a widely used PVD coating, particularly in tooling (e.g., in polymer molds). CrN is less brittle than titanium-based nitrides.
Ion implantation of metals to modify the surfaces of metals and coatings has been well known since the 1980s. Weng et al. have
used a hybrid device that had a filtered arc and an MEVVA ion source (59). About 3 mm thick CrN films were implanted by using
a 45 kV 2 mA ion beam. The ions were Cþ and V2þ with an average ionization of 1þ and 2þ, respectively. The range of the implanted
ions was about 150 nm for vanadium and about 200 nm for carbon. Also, dual implantation of Cþ and V2þ was done. Formation of
V16N1.5 and Cr23C6 was observed by X-ray diffraction (XRD) as a result of V2þ and Cþ implantations. The intensive ion implantation
was observed to smooth the cathodic-arc-deposited coating surface, which was decorated by the chromium particles and nodular
structures. The submicron diameter particles were reduced in size due to sputtering, and weakly bound particles over mm in size were
pushed out from the surface, leaving cavities behind. The internal stress of the CrN coating was about 1.4 GPa. The compressive
stress increased to 4 GPa due to carbon implantation and to 2.4 GPa due to V2þ implantation. The microhardness was observed
to increase from approximately 3000 HV to approximately 4000 HV as a result of the implantations. In dry sliding against a WC
ball, an increase of wear resistance of the V2þ- and Cþ-implanted CrN up to 10 and 1.5, respectively, was recorded. Also, the
corrosion resistance was significantly better as determined by polarization measurements.
4.02.3.8.2.3 TiAlN
The addition of aluminum to TiN has been widely applied in order to enhance the coating’s thermal stability in machining tools.
Aluminum forms a thin protective oxide on the surface of the coating subjected to high local maximum temperatures of over
1000 C.
The increase of hardness in Ti–Al–N compounds is explained as a result of the formation of a NaCl-type supersaturated nitride
where Al substitutes Ti (when x < 0.7 in (Ti1–xAlx)N). When the material is annealed (e.g., during the high-temperature application),
fine-scale TiN and AlN precipitates are formed, increasing the hardness. With further annealing, a thermodynamically stable cubic
NaCl-type TiN and ZnS–Wurtszite-type AlN are formed (60).
Aksyonov et al. (36) used the T-shaped source to deposit TiAlN coatings. By applying Ti and Al cathodes, it was possible to
deposit TiAlN coating with 40 wt.% of aluminum. The spatial variation of the aluminum content was less than 3 wt.%. On the other
hand, the aluminum content could be adjusted to values between 8 and 58 wt.%. Thus, a flexible filtered cathodic-arc source to
optimize the Ti/Al composition in TiAlN films was demonstrated.
Liu et al. (61) investigated the thermal stability and oxidation of TiAlN coatings, which were cathodic arc deposited from Ti–Al
alloy targets in a nitrogen–argon ambient. The (Ti–50Al)N coating remained nitride even up to 900 C annealing in air, whereas the
(Ti–30Al)N coating was completely oxidized. The coating remained electrically conductive and prevented diffusion of Cr from steel
alloys. These properties are important for solid-oxide fuel cell (SOFC) applications to stabilize the steel interconnects.
4.02.3.8.2.4 Other ternary nitrides

Xie et al. (62) reported on the processing of 1.3 mm thick multilayered (bilayer or trilayer) TiN/Ti and TiN/Ti–Si–N and their
mechanical properties. The structure is composed of 300 nm TiN layers and about 100 nm layers of Ti or Ti–Si–N. The TiN was
a columnar structure. The predominant mechanism of deformation was intergranular under indentation on a ductile steel substrate.
This was due to the interlayers between the TiN. The crack propagation in the columnar TiN layer is terminated in the interlayer (see
Figure 36). Both types of interlayers, the ductile Ti layer and hard and brittle nanocrystalline Ti–Si–N, acted as barriers against
intergranular shear sliding. The deformation resistance is increased by two mechanisms: (1) intercolumnar shear area increase as the
columnar shear of the TiN is interrupted by the interlayers, and (2) energy consumed in the deformation of the interlayer.
The increase of the number of interlayers in the structure also increased the deformation energy. The film with the nanocrystalline
Ti–Si–N interlayer needed much higher deformation energy than that with the Ti interlayer.
In the case of depositing Ti–Si–N film by DC cathodic arc from a composite target Ti0.8Si0.2 in N2/Ar ambient, filtering of
macroparticles increased the corrosion resistance of the coating. Corrosion potential was further increased for the depositions,
where substrate bias voltage getting an optimum at 120 V was used. This bias voltage also resulted in the lowest roughness value
(ra ¼ 10 nm). The film composition was analyzed to have 15 nm size TiN nanocrystals embedded in an amorphous silicon nitride
matrix (63).
Addition of magnesium to TiN has been shown to enhance the thermal stability of the coating compared to pure TiN (64). The
atomic ratio of Mg–Ti in the TiMgN alloy could be varied from 0 to 0.55. The deposition was done by using a hybrid pulsed arc for Ti
and magnetron sputtering of Mg in a nitrogen/argon ambient. The Mg alloying did not affect the hardness of the coating (value
measured 34–38 GPa). But the oxidation kinetics of TiMgN was clearly different from that of TiN. The TiO2 rutile was detected in the
coatings by XRD after annealing of TiN in air at 450 C and of Ti-0.41MgN at 550 C. TiMgN coatings have color varying from
golden-copper to violet-gray as the atomic ratio of Mg/Ti is varied from 0 to 0.55.
An industrial cathodic-arc device has been applied to coating with TiBN films (65). Sintered-arc Ti/B alloy targets of two different
compositions were used: 5/3 and 5/1. By varying the N2/Ar gas ratio between 0 and 1, the obtained film compositions are shown in
a simplified Ti–B–N ternary in Figure 37. The boron content in the film was in all cases roughly 50% of that of the target
composition. The loss of boron was assumed to be due to gas scattering. At pure Ar, ambient TiB and Ti phases were observed in the
www.iran-mavad.com
Figure 36 The deformation of a multilayer TiN with Ti or Ti–Si–N nanocrystalline interlayers. Reproduced from Xie, Z. H., et al. Deformation Mechanisms
of TiN Multilayer Coatings Alternated by Ductile or Stiff Interlayers. Acta Mater. 2008, 56 (4), 852–861.
film by using XRD. When nitrogen was added, a nanocomposite structure with Ti(B)N crystals with a size of 8–30 nm was observed.
With a low N2/Ar ratio (0–0.1), B–O and B–Ti and pure B signals were detectable by X-ray photoelectron spectroscopy (XPS) on the
surface. With low pressure of nitrogen, the 8–20 nm size Ti(B)N nanocrystals were surrounded by an amorphous TiBx matrix. At
high pressure of nitrogen, Ti(B)N nanocrystals were embedded in an amorphous BN matrix. With N2/Ar over 0.25, B–N bonds
dominated, as shown by XPS. By analyzing the mechanical properties of the films deposited from a Ti/B 5/3 target, an optimum of
hardness and wear resistance was obtained with N2/Ar ration 0.1. The hardness was 40 GPa, and the corresponding wear resistance
was 10 higher than that of other compositions measured in dry sliding against an Al2O3 ball. Otherwise, the hardness and wear
resistance of the Ti(B)N were about the same as for TiN. The optimal hardening mechanism was attributed in the case of high boron
concentration to the mechanically strong TiBx amorphous matrix, and in the case of lower B concentration to the effective solid
solution hardening of Ti(B)N nanocrystals compared to TiN nanocrystals.
4.02.3.8.3 Oxides
Oxide coatings deposited by cathodic-arc methods are often applied as optical coatings. In this section, some important recently
investigated oxide coatings are introduced in more detail.
4.02.3.8.3.1 ZnO
ZnO is a transparent n-type semiconducting film (66). It has a wide band gap of 3.37 eV and a high exciton binding energy (about
60 meV). ZnO is applied in photodetectors, light-emitting diodes (LEDs), electroluminescence (PL), and photovoltaic and gas
sensors (67,68). Compared to other thin-film deposition methods, vacuum arc has the advantage of a high ionization rate
contributing to the film’s structure, which prevents the growth of columnar crystals and enhances adhesion to the substrate.
Typically, the deposition is carried out by using a filtered arc, either pulsed or DC. The DC current varies normally between 70
and 150 A, and the substrate temperature is elevated to about 200 C depending on the substrate. The deposition rate also varies
www.iran-mavad.com
Figure 37 Different compositions of Ti–B–N coatings obtained from cathodes with Ti/B ratio of 5/1 and 5/3, shown in a simplified ternary phase
diagram. Reproduced from Neidhardt, J., et al. Wear-Resistant Ti–B–N Nanocomposite Coatings Synthesized by Reactive Cathodic Arc Evaporation. Int. J.
Refract. Met. Hard Mater. 2010, 28 (1), 23–31.
from 2 to over 200 nm min1. The oxygen pressure is about 0.1 Pa. The films are normally nanocrystalline, with the crystal size
varying from about 10 to 20 nm depending on the process.
The oxygen flow rate and the oxygen partial pressure have been varied in order to find optimal structures and properties of the
ZnO films. A filtered 60 DC arc was used to deposit ZnO films with varying oxygen flow rates. With the optimal oxygen flow rate of
5 sccm, the films had the lowest tensile intrinsic stress and the highest conductivity of 5 103 (U cm)1 (69). The intensities of the
PL spectra depended on the oxygen flow, and the corresponding transition could be identified when compared to the numerical
simulations of different intrinsic defect levels in the ZnO.
ZnO thin films have been deposited by DC-filtered arc on UV-fused silica. The arc current was 150 A, the oxygen pressure was
0.67 Pa, and the deposition rate was 10 nm s1. The films were post annealed in atmosphere at temperatures between 400 and
600 C to enhance the optical properties. The annealing resulted in stoichiometric ZnO, and the crystal size increased from 17 to
21 nm. At the visual range, the optical transmittance increased from 85 to 90% as a result of annealing. The conductivity decreased
dramatically from 102 (U cm)1 to 104 (U cm)1 (70).
Pulsed filtered arc with a 90 filter has been used to deposit ZnO films. Deposited at a substrate temperature of 25 C, the average
crystal size was 17 nm. The optical transmittance at the visible region (400–700 nm) was over 90%. The photoconductivity was
better with thinner films (170 nm) than with thicker ones (740 nm). This was attributed to the surface state sensitization due to the
difference in oxygen bonding on the surface of the films compared to the bulk of the films (71).
4.02.3.8.3.2 ZnO:Al
Aluminum-doped ZnO (AZO) is a good candidate to replace the standard indium tin oxide (ITO) as a transparent conductive film.
One of the essential issues is the sufficiently high deposition speed. A high deposition rate was obtained by using DC and a 90
helical filter with a 50 mT magnetic field. The distance between the cathode and substrate was about 300 mm. The arc current was
25–70 A, the oxygen flow rate 20–50 sccm, and the total pressure between 1 and 5 mTorr. The substrate was heated to 200–400 C.
Owing to the short distance between the cathode and the substrate, the ions have fewer collisions before hitting the substrate.
Together with the high deposition rate, a high surface temperature exceeding the substrate temperature for a crystalline AZO film
resulted. This helped to restrict unnecessary substrate heating and permitted depositions on lower-cost glass substrates. The
resistivity of the film was 4.8 104 U cm, and the mobility 50 3 cm2 V1 s1. The growth rate was 250 nm min1 (72).
Post annealing of the AZO has been observed to enhance the transmittance while lowering the conductivity of the film. AZO film
deposited by pulsed filtered arc at 0.06 Pa oxygen pressure was rapidly thermal annealed for 10 s. The optimum temperature was
600 C, which increased the transmittance from about 20 to 58% at 2500 nm wavelength, yet maintained a resistivity of only
3.6 103 U cm. The effect was attributed to the increased crystallinity and to the decrease of charge carrier concentration as a result
of post annealing (73).
The mechanical properties of AZO coatings were optimized by using a filtered one-eighth torus DC arc deposition. The devise
contained two cathodes 50 mm from the central axis of the plasma duct. Each cathode was equipped with an individual power
www.iran-mavad.com
Figure 38 Schematic diagram of a laser-assisted filtered cathodic vacuum arc deposition system. After Yuen, C.; et al. Room Temperature Deposition of
p-Type Arsenic Doped ZnO Polycrystalline Films by Laser-Assist Filtered Cathodic Vacuum Arc Technique. J. Appl. Phys. 2007, 101 (9), 094905.
supply. The hardness and wear properties were optimal at 0.5 Pa oxygen, with the Zn/Al concentration of 1.2 at a deposition
temperature of 500 C (74).
4.02.3.8.3.3 ZnO:As
ZnO films are n-type semiconductors. The need of p-type ZnO has previously been fulfilled, for example, by using molecular beam
epitaxy (MBE) at a high substrate temperature of 1000 C (75). C. Yuen et al. used a hybrid filtered DC arc and pulsed laser
deposition to deposit ZnO doped with arsenic. An off-plane double-bent filter (90 and 45 with respect to the axis of the cathodic-
arc source) and Zn3As2 target for pulsed laser deposition were used (Figure 38).
p-type ZnO:As was successfully deposited at room temperature, even on polymeric substrates at a 10 nm min1 deposition rate.
The high ionization rate was observed to produce similar optical properties and crystal structure as when using high substrate
temperature. It was assumed that high ionization corresponded to a surface temperature of about 300 C. A functioning p-n ZnO-
based homojunction was deposited on polymeric substrate (76).
4.02.3.8.3.4 TiO2
Titanium oxide is the most investigated metal oxide in surface science. TiO2 has two important phases (rutile and anatase) and
a number of applications, including catalyst, photocatalyst, photovoltaic, solar thermal, gas sensors, white pigments, corrosion
protection, and optical coating in ceramics and in electronic devices. Huge effort has been put into the development of photo-
catalytic surfaces as thin films. Owing to the high refractive index (2.9 rutile, 2.5 anatase), TiO2 is extensively used for thin-film
optical-interference coating (77).
Bias voltage has been shown to affect the mechanical and optical properties of the DC-filtered, vacuum-arc-deposited TiO2 films
at room temperature. When coated on insulating substrates, the DC bias voltage, however, resulted in charge buildup and limited
the growth of the crystal structure in the film, and the film structure was amorphous (78). The use of pulsed bias at 5 and 50 kHz
resulted in an anatase structure when focusing of the plasma was used, resulting in a higher current density. The critical current
density was 0.5 mA cm2. With lower value, the structure remained amorphous (79). Possibly the high current density and pulsed
bias voltage elevated the flash temperature on the surface, assisting crystal growth. When DC vacuum arc deposition was carried out
at elevated temperature with no additional bias voltage, the coatings were amorphous until, at 300 C deposition temperature,
crystalline anatase started to form (80).
Oxide coatings are also developed for corrosion protection and for barrier coatings. In the following, some examples of recent
results of such cathodic-arc-deposited oxide films are discussed.
4.02.3.8.3.5 CrOx and TaOx

Filtered vacuum arc has been used to deposit 10–50 nm thick CrOx and TaOx coatings at room temperature on low-alloyed carbon
steel (AISI52100), and the corrosion properties were investigated (81). Prior to deposition, the steel surface was etched by 350 eV
www.iran-mavad.com
argon ions at 0.5 mA cm2 for 30 min or longer. By using time-of-flight secondary ions mass spectroscopy (ToF-SIMS), tantalum
carbide formation was observed at the substrate-coating interface. Carbon was assumed to be due to surface contaminations. The
stability of the interlayer was attributed to be the reason for the better sealing property against corrosion current for the Ta2O5 film as
compared to the Cr2O3 film. The increase of the film thickness from 10 to 50 nm did not increase the sealing property of the
coatings. The best corrosion resistance was after a long 90 min argon ion etching.
4.02.3.8.3.6 ErOx
Erbium oxide has high stability in a strongly reducing atmosphere and has the capability of suppressing hydrogen isotope
permeation. Filtered-arc-deposited Er2O3 films have been investigated to be applied to reduce tritium permeation and accumu-
lation into fusion reactor first wall materials (82,83).
4.02.3.8.3.7 Al–Cr–O
An industrial vacuum arc deposition process utilizing pulse modification of DC arc sources and reactive gas control has been
developed by OC Oerlikon Balzers AG (84). (Al1xCrx)2O3 coatings with x varying from 1 to 0.5 were produced at 550 C
temperature and in pure oxygen ambient of about 0.3 Pa. The microstructure of the coatings was a solid solution of a-alumina and
eskolite chromia. With a lower content of chromium (x < 0.15), the coatings were amorphous. The (Al1xCrx)2O3 coating deposited
on top of a TiCN coating was demonstrated to prevent the oxidation and degradation of the TiCN when annealed at 900 C in air.
Multilayers of AlCrN/AlCrO were deposited, demonstrating the potential to produce various multilayer structures by using multiple
plasma sources and switching of reactive gas.
4.02.3.9 Ultrathin Films and Nanolaminates

Films with a thickness not more than 10 nm are an interesting challenge both in film deposition aspects and from an applications
viewpoint. Such films are called ultrathin. The cathodic-arc process due to its high ionization of the plasma and high ion energy at
the deposition has distinct benefits in the growth of ultrathin films.
Metal ultrathin films are conductive and transparent in the visual part of the spectrum and thus applicable to optical coatings.
Byon et al. have grown thin Ag films with both filtered cathodic arc and magnetron sputtering on ZnO-coated glass (85). The
coalescence of the island structure was monitored by in situ conductivity measurements. The onset of conductivity occurred much
earlier with cathodic-arc coating compared to magnetron-sputtered coating (Figure 39).
The reason for this was in the difference of the ion energy, which was about 70 eV for cathodic arc and 1–3 eV for magnetron
sputtering, respectively. With low ion energy, Ag atoms remained on the substrate surface and minimized the energy in the
formation of islands. The more energetic Ag ions were planted to a depth of 1.5 nm, which was several atomic layers beneath the
surface. Furthermore, the islands had a reduced height/width aspect ratio, leading to an earlier contact between islands. A spon-
taneous increase of conductivity after the interruption of the deposition was observed. The spontaneous coalescence of islands was
faster with cathodic-arc deposition than with magnetron sputtering. The finding indicates the potential for optimization by using
energetic metal ions for ultrathin metal layers for optical coatings.
Figure 39 Sheet resistance of silver coating on ZnO/glass substrate as a function of coating thickness, produced by filtered cathodic arc and mag-
netron sputtering. After Byon, E., Oates, T. W. H.; Anders, A. Coalescence of Nanometer Silver Islands on Oxides Grown by Filtered Cathodic Arc
Deposition. Appl. Phys. Lett. 2003, 82 (10), 1634.
www.iran-mavad.com
Figure 40 The sp2 content of ta-C film as a function of coating thickness measured with XPS and near-edge X-ray absorption fine structure (NEXAFS).
Reproduced from Soin, N., et al. Thickness Dependent Electronic Structure of Ultra-Thin Tetrahedral Amorphous Carbon (Ta-C) Films. Thin Solid
Films 2012, 520 (7), 2909–2915.
ta-C films were described in Section 4.02.3.8.1 in more detail. However, the ultrathin ta-C layers are of particular interest in
several applications and in nanomaterials research. N. Soin et al. have deposited ta-C films, varying the thickness between 1 and
10 nm by using filtered cathodic arc (86). Several surface-sensitive spectroscopic methods were used to measure the film-bonding
structure as a function of film thickness. In Figure 40, the sp2 content as a function of coating thickness is shown. A 1 nm thick film
shows mainly contaminated hydrogen-terminated carbon sp3 bonds. In the intermediate thickness of 2–5 nm, close to 40% of sp2
bonds are observed and films thicker than 5 nm have about 32% of sp2 and thus higher sp3 content. Thus, the films of already 2 nm
thickness have a high sp3 content, and films of 6 nm or more in thickness are typical sp3 ta-C coatings. Surface acoustic waves have
been used to measure ultrathin ta-C films deposited by pulsed cathodic arc on silicon (87). The elastic modulus of a 5 nm film was
430 GPa, typical of ta-C film with a high content of sp3. Thus, the ta-C films are applicable as a protective ultrathin layer for optical
and magnetic storage devices.
The mechanical and tribological properties of ultrathin ta-C are naturally affected by the substrate properties. G. Wang et al. have
demonstrated an ultrathin version of the duplex process. A classical duplex coating is, for example, a few mm of TiN coated on top of
a plasma-nitrided steel surface, which has a hardened surface zone of about 100 mm in thickness. In the ultrathin version of duplex
coating, a silicon substrate with a Fe–Ni (tunneling magnetoresistant) alloy coating was magnetron sputtered with 3 nm of silicon.
The layer was plasma nitride, forming a 3 nm gradient Si–N layer on the magnetic metal films. Finally, deposition of 5 nm of ta-C by
using a filtered cathodic arc was done. The mechanical and corrosion properties were found significantly better compared to the
nonnitrided substrate (88).
4.02.3.9.1 Multilayers, Superlattices, and Nanolaminates

The multilayer deposition results in interesting phenomena when the thickness of individual layers becomes sufficiently thin (in the
range of 7 nm). The hardness of the layer and even the elastic modulus of the coating are higher than the values of individual
sublayers. This is called the superlattice effect (60). It was first observed for single-crystal TiN/VN layers epitaxially grown on MgO
substrate by using MBE. In a typical multilayer of material, A and B each has a constant layer thickness, and the structure has
a thickness modulation period l. The value of l has an optimum in the range of 2–10 nm, and then the structure is called
a superlattice. Since the first observations, numerous material combinations with the superlattice effect have been observed,
including transition metal nitrides, carbides, and carbonitrides. Such films have also been deposited by cathodic arc or by using
hybrid methods combining cathodic arc and sputtering.
Several superlattice mechanisms have been identified. One of the mechanisms is the stress caused by lattice mismatch at the
interfaces. The hardness increase is due to the blocking of the dislocation movement. The reason for this is the difference in shear
moduli of the individual layers and the coherency strain, resulting in periodic stress–strain fields caused by the lattice mismatch. In the
TiN/AlN system with a l value of less than 3 nm, a metastable cubic crystalline structure for the AlN layer is formed. For superlattice
films with AlN layers thicker than 2 nm, a stable hexagonal structure of AlN is formed. The maximum hardness was obtained with
a l value of 2.5 nm. When the modulation period is reduced to a sufficiently small value, a TiAlN solid solution is formed.
LAFAD has been used to deposit multilayer films of TiN/Ti and TiCrN/TiCr (89). Nanolaminates WC–CrAlN with a bilayer
period of 2–10 nm have been deposited by Lee et al. (90). In both cases, the microhardness was close to that of the monolithic film,
but the internal stress was decreased and the adhesion and tribological properties were significantly enhanced.
The inherent challenge of ta-C films is the high intrinsic compressive stress. Wang et al. investigated the use of filtered cathodic
arc to produce layered structures by varying the bias voltage in different sequences from 0 to 100 V. The 50 nm thick graded layered
www.iran-mavad.com
films were processed by applying higher bias voltage at the beginning of the deposition and lower values at the top layers of the
coating. The high nanohardness of 59 GPa was reached, while the compressive stress was reduced to 3.3 GPa from the value of
5.2 GPa of the single-layer ta-C film (91).
4.02.3.9.2 MAX Phase

Multilayers or nanolaminates form a single-phase material (e.g., in the case of Ti3SiC2). This material was first discovered in 1990 at
Drexel University and is called the MAX phase. The MAX phase is produced as both bulk and thin films, and a recent review has been
published by P. Eklund (92). The MAX phase combines metallic and ceramic properties. The composition is Mnþ1AXn, where M is
a transition metal, A is from the A group of the periodic table (usually IIIA or IVA), and X is carbon or nitrogen. n may have values of
1, 2, or 3. The structure is inherently nanolaminate where Mnþ1Xn layers are interleaved with a single layer of A. The laminate
crystalline structure provides, for instance, high toughness, thermal stability, and good tribological properties. Pulsed cathodic arc
with Ti, Al, and C cathodes was used to deposit Ti2AlC MAX-phase films at 900 C (93). There are indications that the high
ionization rate of the cathodic arc allows the MAX phase to be deposited at lower substrate temperatures than when using other PVD
methods.
4.02.3.10 Applications of Cathodic-Arc Deposited Films

The technology of thin films is directly connected to the industrial applications and needs during its history. There are benefits and
limitations regarding cathodic-arc deposition technology. The best features of the technology are demonstrated in the following
sections analyzing the tribological, optical, decorative, and biological applications.
The commercial use of cathodic-arc coatings has been summarized in Table 3, following A. Anders (17,94):
4.02.3.10.1 Tribology
4.02.3.10.1.1 Potential of tribological coatings
One of the main application fields of PVD thin films is to protect material against wear and to provide a low coefficient of friction
(95). The needs for surfaces and coatings in tribological contact are very much dependent on the conditions and requirements of the
mechanical contact. The coefficient of friction is due to a complicated process where typically engineering lubricants are also
involved. The reduction or control of wear is the most common function of tribological coatings. Wear is a system parameter
depending on numerous things such as the substrate material, countersurface, load, speed of sliding, wear debris, and lubrication.
Both friction and wear are also nonlinear processes where a stable condition is reached after the tribological system has reached
a balance. Thus, the selection and optimization of surface coatings in tribological applications comprise a challenging task.
The reliability and lifetime of mechanical devices and production systems have such high demands that the challenge would be
impossible to overcome without the use of wear-resistant coatings and surface treatments. The economic impact of wear and friction
exceeds US$100 billion per annum. One example of the impact and potential of advanced coatings for tribology is the recent
publication by Holmberg et al. (96). The fuel consumption of an average passenger car was carefully analyzed. It was estimated that
about 28% of the energy is used to overcome direct frictional losses without taking into account the braking. This is equivalent to 7.3
million TJ per year. A special evaluation was carried out on the published data of low-friction coatings emerging on the market and
of coatings that are currently in the development phase. Based on this information, it was estimated that by applying surface
coatings, a reduction of frictional losses by 18% in the short term (5–10 years) and 61% in the long term (15–25 years) could
Table 3 Commercial uses of various typical cathodic-arc coatings
Coating category Application Typical coating material
Hard coatings Cutting tools TiN, TiAlN, TiAlYN, multilayers, ta-C

Forming tools CrN, multilayers
Plastic molds TiN, TiAlN, CrN, CrTiN, DLC
Hard coatings with solid lubricants Cutting tools TiAlN/W–C:H, TiAlN/WC/a-C, TiAlN/MoS2
Forming tools CrN/W–C:H
Decorative coatings Faucets, shower heads ZrN, CrN, TiAlN
Door hardware ZrN
Automotive Cr on polymeric interlayer
Functional coatings Engine parts, fuel injection, automotive interior parts CrN, metal–DLC, AlSn, CrN, DLC
Textile machinery TiCN, metal–DLC
Electromagnetic interference (EMI) or radiofrequency Al, Cu, Ni–Cr
interference (RFI) shielding
Reflective coatings Al, Cr, stainless steel, Ni–Cr
Biocompatible coatings for medical tools and implants Ti, DLC, TiN, TiCN, ZrN, CrN, Ag, Au, Cu
Metallization and diffusion barrier for semiconductors Ta, TaN, Cu
Protection of magnetic disks and read-write heads Ultrathin DLC
After Anders, A. Cathodic Arcs: From Fractal Spots to Energetic Condensations. In Atomic, Optical, and Plasma Physics; Springer: New York, 2010; Vol. 50 and http://www.osti.
gov/bridge/purl.cover.jsp?purl¼/810482-HCSxLp/810482.pdf.
www.iran-mavad.com
Figure 41 The trends of coefficient of friction in four different categories of passenger cars: (a) typical 10-year-old car in use today: Car2000; (b) an
advanced car that uses the best coatings commercially available today: Car2010; (c) a car that uses the best coatings demonstrated in laboratories
in research programs today: Lab2010; and (d) a car that is possible to develop based on the best guess of experts on tribology: Car2020. Reproduced from
Holmberg, K.; Andersson, P.; Erdemir, A. Global Energy Consumption due to Friction in Passenger Cars. Tribiol. Int. 2012, 47(C), 221–234.
be obtained. In Figure 41, the development trend of lowering friction by the use and development of low-friction coatings in cars
is shown.
4.02.3.10.1.2 Nitride coatings

The quick development of tribological coatings is closely related to the discovery of TiN coatings. The coating is hard and chemically
stable, adheres well to metals, and bears a distinct golden color. In the 1980s, TiN coating became a standard for high-performance
cutting tools made of high-speed steel and hard metal. Cathodic-arc deposition was one of the coating methods to produce high-
quality TiN coatings on tools.
The cathodic-arc process provides highly ionized plasma, which is ideal for growing hard layers with high density and small grain
size. The growth rate is high and limited mainly by the heat management of the deposition process. The most important limitation
in the vacuum arc process is the evolution of droplets resulting in a coating with pinholes and other defects. In many tribological
applications, these defects are detrimental, but in some cases they are even beneficial. As discussed in Section 4.02.3.7.3, the
nonfiltered vacuum arc has the highest coating speed and is easiest to upscale to a commercial process. The surface is decorated with
defects caused by macroparticle bombardment. However, the high quality of the film ‘around’ the defects has excellent properties
that are difficult to obtain by other means. Due to the high ionization rate, the deposition temperature may be kept lower compared
to, for example, magnetron sputtering in order to provide good adhesion and dense microstructure. This is an advantage in the
selection of substrate materials. For example, the use of hardened steel for components is possible since the substrate temperature
can be kept below 100 C.
In recent years, the expectations for coatings in applications have become more demanding and thin-film defects are often not as
acceptable as before. Filtering has also been adapted more for large-scale deposition systems.
The effect of filtering has been investigated in numerous publications. For example, Hsu et al. have studied the tribological
properties of filtered and nonfiltered TiAlN (97). The abrasion resistance of both coatings was almost identical.
The difference in the surface morphology of filtering compared to the nonfiltered deposition of CrN is illustrated in Figure 42
(98). The nanohardness (21 GPa) and elastic modulus (490 GPa) of both coatings were quite similar. In dry reciprocal sliding
against a Si3N4 ball, the defect-decorated surface of the nonfiltered coating had low wear resistance due to wear debris removal
caused by macroparticles and micropits. The filtered CrN showed mild and oxidative wear.
4.02.3.10.1.3 Multilayers
The layered structures of thin films and nanolaminates have been one of the major development strategies to create coatings for high
loading conditions (95). The interfaces of indifferent materials and laminates of thin intermediate layers have a strong effect on the
mechanical properties. The hard materials deform by crack propagation, and the discontinuities in the film structure have several
possible mechanisms that toughen the film: (1) the crack may create a void at the tip that reduces the stress driving the crack; (2)
a change in stress at the interface may bend the propagation of the crack, preventing it from penetrating to the substrate; and (3) in
the case of tougher laminate, plastic deformation lowers the stress sufficiently at the crack tip to stop the crack growth (60).
In tribological applications, the increase in hardness is beneficial against abrasive wear. The wear phenomenon against harder-
wearing counterparticles is related to the ability of the particles or countersurface asperities to penetrate into the surface and plow
material. Then, the wear rate is roughly inversely proportional to the hardness of the surface. The increase in hardness is associated
with brittleness, the ability for cracks to grow in the hard material. The toughness of the material is a measure of load required to
grow a crack in the brittle material. Often, the toughness of the material is simply evaluated by measuring the hardness and elastic
modulus. The effective modulus is defined E* ¼ E/(1n2), where n is the Poisson’s ratio of material. H, E*, and also the ratio H3/E*2
www.iran-mavad.com
Figure 42 The difference in surface morphology of filtering (a) compared to the nonfiltered (b) deposition of CrN. Reproduced from Mo, J. L.; Zhu, M. H.
Tribological Characterization of Chromium Nitride Coating Deposited by Filtered Cathodic Vacuum Arc. Appl. Surf. Sci. 2009, 255 (17), 7627–7634.
could be used to analyze the toughness of the film. The higher values of H3/E*2 are more tough. Both H and E are measured by using
nanoindentation, which could be used to predict the toughness of the film (99).
A hybrid of a filtered arc deposition and an unbalanced magnetron sputtering (FAD–UBM) device has been used to deposit
layered TiCrN and TiBC coatings (100). A complicated nanolaminate and gradient coating architecture was produced, as sche-
matically shown in Figures 43 and 44. The load supporting the ductile layer is a multilayer of TiCr/TiCrN with a gradient of carbon.
The carbon gradient is obtained by gradual increase of CH4 in the ambient during film growth. Gradually, magnetron sputtering is
applied to obtain B4C. A harder top layer is processed. The wear tests simulating the unwanted interruption of oil lubrication
demonstrated the coating against an M50 steel ball in reciprocating sliding (3 GPa contact load, 0.36 m s1 sliding). By measuring
the time to increase the sliding friction from 0.02 to the cut-off valued of 0.15, the survival time up to 450 s was observed with the
coating. The corresponding value for an uncoated surface was less than 100 s.
www.iran-mavad.com
Figure 43 The nanolaminate and gradient-coating architecture deposited on an aerospace bearing material. Reproduced from Gorokovsky, V. I., et al.
Deposition and Characterization of Hybrid Filtered Arc/Magnetron Multilayer Nanocomposite Cermet Coatings for Advanced Tribological Applications.
Wear 2008, 265 (5–6), 741–755.
Figure 44 (a) High-resolution transmission electron microscopy image of the top segment shown in Figure 42 and (b) a higher magnification showing
the nanolaminates of the top end of the top segment. Reproduced from Gorokovsky, V. I., et al. Deposition and Characterization of Hybrid Filtered
Arc/Magnetron Multilayer Nanocomposite Cermet Coatings for Advanced Tribological Applications. Wear 2008, 265 (5–6), 741–755.
Z. Ottmar et al. (101) have investigated multilayer cathodic-arc deposition to grow thick films up to 100 mm in thickness. In the
cathodic-arc process, the growing film thickness bears the nodular growth of defects, resulting in a very rough surface. In CrN/TiN
multilayer film with about 100 nm single-layer thickness, healing of defects was observed. The deposition was carried out at 0.8 Pa
pressure of nitrogen from Cr and Ti targets with an arc current of 100 A and a substrate temperature of 400 C. As seen in Figure 45,
the embedded particles are covered with a multilayer structure that smoothes the surface, showing a completely opposite effect to
the nodular growth. The smoothing and self-healing of defects are not well understood. It was assumed to be related to the high
surface diffusion of species due to high ion energy and the suppression of crystal growth in multilayer films, particularly when the
single layer is smaller (in the range of 5–10 nm). The high deposition rate of 5 mm h1 and the ability to grow thick films with a low
number of defects open up the potential to produce tribological coatings for tools and components.
4.02.3.10.1.4 The ta-C coatings

The tribology of DLC is often analyzed by dividing the DLC coatings into two categories: hydrogenated and hydrogen-free films.
Commonly, the former are fabricated by PECVD-type processes, and the latter by vacuum-arc deposition.
ta-C coatings are among the hardest thin films deposited at low temperature. ta-C:H is a tetrahedrally coordinated hard DLC film
with hydrogen added up to 30 at.% content. The ranges of hardness and coefficient of friction are shown in Figure 46 (102). One
can see that the range of hardness from hard to superhard coatings is possible with several types of coatings, but the DLC coatings
have distinctly low coefficients of friction.
www.iran-mavad.com
Figure 45 A cross section of a CrN/TiN multilayer on a 1.3343 steel substrate showing healing of defects. Reproduced from Zimmer, O.; Kaulfub, F.
PO4046-Hard Coatings with High Film Thickness Prepared by PVD. Plasma Process. Polym. 2009, 6 (S1), S152–S156.
The cause of the low coefficient of friction for DLC films could be due to both intrinsic and extrinsic factors. The intrinsic factors
include the sp2/sp3 bonding. The extrinsic factors include the parameters of the sliding contact, the ambient atmosphere, and the
temperature. Thus, the coefficient of friction (COF) is dependent on a large number of parameters and is both a material parameter
and an engineering system parameter. The surface roughness is one of the most important material parameters. For DLC and
diamond, there is almost a linear correlation with the smoothness and the COF and the amount of wear. The smoother films have
a lower COF and wear rate. For arc-deposited films, the surface roughness varies in a large range depending on the deposition
source, film thickness, and filtering. Often, a post polishing of the ta-C film is required before application for sliding contact.
The external factors have a profound significance on COF. ta-C films from a clean system (e.g., UHV vacuum and high
temperature) occupy the dangling sigma bonds on the surface, which form chemical bonding over the contact surface as
demonstrated by numerous molecular dynamic simulations. In the case that the ambient contains hydrogen or water, or when the
DLC film contains hydrogen, the sigma bonds are terminated by hydrogen or OH groups. Such a surface is inert, very low interaction
results between the surfaces, and the COF of the films is low. When the surfaces are clean and not terminated, COF values of up to
0.7 have been reported. In contrast, if DLC has up to 50 at.% hydrogen and the surface is smooth, COF values as low as 0.001 in
sliding tests in dry nitrogen atmosphere have been achieved. Such low COF is difficult to obtain, even in lubricated conditions, and
the phenomenon is called superlubricity.
Often in applications in tooling and mechanical engineering, the DLC coating has a countersurface of metal, ceramic, or
polymer. When two hard ta-C surfaces slide against each other, there is a risk of high local stresses being induced, causing fracture of
the coatings. When DLC has a counterbody of, for example, metal, the COF value depends on the formation of a transfer layer
on the countersurface. In the case of carbide-forming metal (e.g., Ti, W, Si, or Fe), a covalently bonded transfer layer is often formed
on the countersurface. This transfer layer results in low adhesion between the surface and low COF. In the case of a non carbide-
forming countersurface such as Cu, no transfer layer forms and the COF is higher. The formation of the transfer layer is dependent
Figure 46 Schematic presentation of the coefficient of friction and hardness of various DLC and other PVD coatings. After Erdemir, A.; Donnet, C.
Tribology of Diamond-Like Carbon Films: Recent Progress and Future Prospects. J. Phys. D Appl. Phys. 2006, 39 (18), R311–R327.
www.iran-mavad.com
Type of DLC a-C:H a-CH or a-C ta-C or a-C

hydrogen free
Friction range < 0.02 0.1 – 0.2 > 0.5
Nature of interaction Van der Waals Hydrogen or
Energy (eV/bond) 0.08 0.2 0.4 – 0.8
Interaction between
two DLC surfaces C C
H H H0 OH C C
H H O O C C
C C C C
Environment Inert of UHV Humid UHV
Figure 47 Friction mechanisms of various types of DLC at the molecular level. After Erdemir, A.; Donnet, C. Tribology of Diamond-Like Carbon Films:
Recent Progress and Future Prospects. J. Phys. D Appl. Phys. 2006, 39 (18), R311–R327.
not only on the chemical properties of the countermaterial but also on the sliding parameters such as speed and load. Obviously,
the transfer film formation occurs in contacts on limited countersurface areas such as in the case of bearings, but transfer layer
formation is not possible (e.g., in cutting tool applications).
The coefficient of friction is different for different types of DLC and in different environments, and is illustrated in Figure 47.
Traditionally, the main application field of ta-C coatings has been tooling. Currently, the coatings of components (e.g., in the
automotive field) are growing in market volume.
One of the most remarkable achievements in the engineering of ta-C coatings for tribology is the application of friction modifiers
added to lubrication. Addition of glycerol mono-oleate (GMO) to a poly-alpha-olefin (PAO) oil resulted in a COF of about 0.05 in
sliding contact with two ta-C surfaces. When pure glycerol was used, the friction coefficient decreased close to 0.005 (103). Minami
et al. (104) investigated the surface phenomena in the lubricated contact experimentally by using isotope-modified GMO lubri-
cants. Martin et al. (105) carried out experimental surface analysis and numerical modeling. The conclusion was that the ta-C surface
is terminated with OH groups originating from the decomposition of glycerol at the tribological contact. A portion of glycerol
molecules decomposed to water molecules, providing superlubricity.
In lubricated contacts, it is well known that the thickness of the lubricating oil film is also dependent on the topography and
geometry of the sliding contact. By control of the topography of the ta-C coating, a benefit for oil lubrication was obtained. In
tribological tests simulating metal forming (e.g., punching), the lubricating oil was retained longer on a surface that has cavities
acting as lubrication pockets. The surface texturing could be done by allowing more macroparticles to the film, followed by pol-
ishing (cavities of 10–100 nm diameter); masking the substrate with deliberately dusted particles of a few mm in diameter before the
ta-C coating (cavities of 1–10 mm diameter); or using laser texturing (cavities of 10–30 mm diameter). The COF was measured in
a reciprocating test where an AISI316 ball of 10 mm in diameter was sliding at 4 mm s1 with a 10 N load on the textured ta-C on
M2 steel. The contact was relubricated by neat metal-forming lubricant after each 10 sliding cycles. The measured RVK value of the
surfaces was used as a measure of the number of cavities. A 15% lowering of COF was the result in laboratory tests with coatings of
the largest RVK (Figure 48). 15–100% enhancement of the tool life was obtained in real production tests in different metal-forming
processes (106).
Cathodic-arc information can be obtained from commercial companies’ online distributed data. In Table 4, such data about
commercial tribological coatings have been collected, including some properties and suggested applications. In commercial
information from various ta-C coating developers, some trends may be noticed, as discussed here.
By applying an effective filter (120 ), the lowered deposition rate is still sufficient for very thin layers typical for electronic
industry applications. Metal layers used for magnetic hard disks and read-write heads have been processed (Figure 49). The
www.iran-mavad.com
Re-lubricated
0.14
0.12
Coefficient of friction
0.1
0.08
0.06
0.04
0.02
0
0 0.05 0.1 0.15 0.2 0.25 0.3
Rvk
Figure 48 COF of textured ta-C film as a function of the RVK value, which is a measure of the volume of the cavities in the surface.
reduction of macroparticles with some different cathodic-arc processes is shown in Table 5. Given the high ionization rate, almost
conformal filling of trenches has been demonstrated with this method.
Laser ignition and compound cylindrical cathodes have been developed at Fraunhofer Institute (FhG-IWS). A multilayer coating
may be produced by focusing the ignition laser pulse on the required sector of the rotating cathode. The laser-ignited plasma source
may be equipped with an electro-static filter to reduce the number of macroparticles (Figure 50) (108).
For coating of ternary oxide films, a multisource-pulsed cathodic-arc system has been developed by Balzers Oerlikon AG; it is
schematically shown in Figure 51 (84).
4.02.3.10.2 Optical Cathodic-Arc Coatings

Cathodic-arc deposition is a natural method to produce optical transparent oxide films. When using sputtering methods, target
poisoning is a major difficulty in depositing oxide films with the desired stoichiometry. In cathodic arcs, poisoning is generally not
a problem, but instead it helps to stabilize the cathode spot ignition as described in Section 4.02.3.3. P. J. Martin et al. have made
pioneering work in the development of optical oxide coatings ((109) and references within). The high ionization rate and high ion
energy prevent the columnar growth of the oxide films, and dense films without substrate heating are possible. The cathodic-arc
optical coatings have a small extinction coefficient, and the refractive index may be selected by the choice of the coating material.
The refractive indices of different dielectric films deposited by cathodic arc are shown in Figure 52.
Optical films are used often as antireflective (AR) coatings. The sophisticated AR coatings are stacks of oxides with different
refractive indices. Such coatings on lenses of, for example, cameras need to be scratch resistant. Cathodic-arc films have an excellent
adhesion and wear resistance compared to films made using any other method.
One important type of AR coating is composed of stacks of oxides and ultrathin metallic layers. Typically, Ni, Au, or Al with
a thickness of less than 10 nm sandwiched between about 100 nm oxide layers is applied. N. B. Abally et al. have applied a filtered
cathodic arc to investigate Al2O3/Al–Al2O3 trilayer AR coatings. Based on numerical simulations, the Al layer optimum is 3 nm thick
(Figure 53) (110).
By using XPS and ellipsometry, the stoichiometry of the alumina was optimized. The deposition of the thin aluminum layer was
critical. For instance, the negative bias voltage played an important role. With a low value (<200 V), a 3 nm thick Al film was
islandlike. In the case where the negative bias voltage was too high (>1000 V), the film was smooth, but pits due to electric
discharges on the dielectric film were observed. The optimum bias value was 200 V. It was also observed that a nominal thickness
of 6 nm was required in order to obtain a 3 nm Al film thickness. Partial oxidation of the initial deposited 6 nm Al films was
assumed to be the explanation for this result (111).
The need for transparent conductive films is high due to several rapidly growing applications such as flat panel displays, solar
cells, and touchscreen devises. The standard solution is to produce transparent conductive oxides (TCOs) by depositing ITO. ITO
has a transparency of over 90% in the visible spectrum and low resistivity to 0.5 mUm for the best films. Currently, about 108 m2
of coated glass per year is produced. The quickly increasing demand of indium is affecting other important application fields of
the metal such as coating for aircraft parts, cryogenic and vacuum applications, opto-electrical devices, and fiber-optic
communication.
Other TCO solutions by using cathodic-arc deposition have been investigated with promising results. A filtered cathodic arc
equipped with a rectangular filter has been used to deposit SnO2 films. 100–200 nm thick films had an optical transmittance of
75–90% in the visible spectrum, and the conductivity was 50–90 mUm (39). The films were deposited on microscopic glass slides at
room temperature, and the film properties did not change significantly during an 8-month storage in ambient air.
Aluminum-doped ZnO (AZO) is one of the most likely candidates to replace ITO. The cathodic-arc deposition results of AZO
were discussed in Section 4.02.3.8.3.
www.iran-mavad.com
48
Table 4 Hard tribological coatings produced by commercial vacuum arc processes, obtained from commercial information distributed by the companies
Coating architecture
M ¼ monolayer,
ML ¼ multilayer, Coefficient of friction
NC ¼ nanocomposite, in dry sliding countermaterial, Maximum service
Composition G ¼ gradient Color Hardness GPa a ¼ Al2O3, s ¼ steel temperature C Application field
TiN M Gold-yellow 23–25 0.4s 0.7a 600 Tools: cutting, punching, forming, plastic molds,
drilling; components in machines; plungers
of diesel injection pumps
TiAlN M, G, NC Gray 25–35 0.35s 0.4a 700–900 Tools: drilling, milling, reaming, turning,
dry machining, ML-interrupted cutting
AlCrN M Blue-gray 30–32 0.35 900–1100 Tools, end mill
TiCN M Gray 30 0.4 600 Milling, tapping, thread forming, punching,
Cathodic-Arc and Thermal-Evaporation Deposition
forming of ferritic and austenitic steels,

plastic injection molding
CrN M Metal-silver 15–25 0.5 700 Molds, dies, forming, components: valve
gear components, piston rings, hydraulic pistons,
textile machines, food-processing machines
ZrN M White-gold 20 0.4 550 Machining aluminum and titanium alloys
www.iran-mavad.com
TiSiN NC 35 0.4s 1500 Tools

CrTiN ML Metal-silver Up to 30 0.4 600 Tools: milling, molds, dies, machine parts,
similar to CrN but lower cost
TiN/TiAlN ML 25–35 0.4s
ta-C M Black * 40–70 0.05–0.15 500 Tools: cutting, forming, molds, cutting blades,
cutting of aluminum alloys; engine components:
hydraulic components, bearings, guides, sealing
surfaces, valve lifters, pistons and rings, valves,
gears, fuel injection
TiAlCN ML, G Violet 33 0.3 500 Tools: milling, hobbing, tapping, stamping, punching
AlTiSiN NC Violet-blue 40–45 0.45 1100–1200 Tools: milling, plunging, sawing, hobbing,
drilling, stamping, tapping, heat resistant for
high-performance machining
Figure 49 Design of the new filtered-arc source for industrial usage. With permission of Arc Precision Inc. http://www.arcprecision.com/5503/4602.
html.
4.02.3.10.3 Decorative Coatings

Cathodic-arc deposition is an efficient method to produce nitride and carbon nitride films as described in Section 4.02.3.8. The
mechanical, adhesion, and corrosion properties of nitrides are excellent. When these properties are added to the ability to vary the
color of the coating, a strong method of producing decorative coatings is available.
The optical properties of surfaces are the result of visible light interacting with the surface. The appearance is determined by the
reflected light from the surface. Three important aspects are dominant: (1) the surface roughness, (2) the interference of reflected
waves in the case of transparent or slightly absorbing coatings, and (3) the absorbing properties of coatings having free charge
carriers. The surface roughness determines the shining or matte appearance of the surface requiring variation of topography often
achieved in combination of the substrate and coating roughness.
In the case of transparent films, color results from the constructive interference of light reflecting from the surface and the
coating–substrate interface. The color is dependent on the coating thickness, refractive index, and direction of the light. In the case of
a transparent coating of uniform thickness, an appearance of spectral colors of the coating depending of the view angle is obtained.
If the coating is nonuniform in thickness, Newton’s rings appear. Oxide and DLC coatings with control of coating thickness provide
interference colors (e.g., on a metal surface). When the coating is thicker, the absorption of light in the coating results eventually in
Table 5 Efficiency of droplet elimination in comparison to a commercial

state-of-the-art filter source
Number of particles > 0.1 mm

Arc source per nm film thickness, per m2
DC 1.2 $ 1010
Pulsed 7.5 $ 107
90 filtered 2.0 $ 105
120 filtered 1.5 $ 102
With permission of Arc Precision Inc. http://www.arcprecision.com/5503/4602.html.
www.iran-mavad.com
Figure 50 Laser-ignited pulsed plasma system with an option for plasma filtering.
a black appearance. DLC films with mm range thickness are applied as decorative black coating on various products such as
wristwatches, jewelry, and mobile electronics.
The absorption of light is a physical phenomenon due to free charge carrier absorption, interband absorption, and absorption
due to impurities. In the case of metals, photons in a large spectrum range may easily be absorbed by electrons at the Fermi level.
The propagating electric field of the light is exponentially damped depending on the density of the free carriers that can absorb in the
visible spectrum. In the case of nitrides, free charge carrier concentration is lower than in most metals, and the surface shows
a colored appearance. The absorption of photons determines the color of the reflected light.
Figure 51 A schematic illustration of an arc deposition device with each arc source operated with an individual pulsed power supply. Reproduced from
Ramm, J., et al. Pulse Enhanced Electron Emission (P3e™) Arc Evaporation and the Synthesis of Wear Resistant Al–Cr–O Coatings in Corundum
Structure. Surf. Coat. Technol. 2007, 202 (4–7), 876–883. For details of the figure please see the source line reference.
www.iran-mavad.com
Figure 52 Refractive index of various optical thin films deposited by cathodic arc at XCSIRO. Reproduced from Martin, P.; Bendavid, A. Review of the
Filtered Vacuum Arc Process and Materials Deposition. Thin Solid Films 2001, 394 (1), 1–14.
Figure 53 (a) Diagram of the optimized AR-coating design and (b) theoretical reflectance in the optical range with different Al-coating thickness values.
After Proc. SPIE 7045, 704508-7.
Hodroj et al. investigated the (Ti,Mg)N system by using a hybrid cathodic arc (Ti) and magnetron sputtering for an Mg target. By
varying the Ti/Mg composition, the color of the coating could be varied (64). The effect of the composition on conductivity r and
color are shown in Table 6. The color coordinates L*a*b* are as follows: L* is lightness (0 ¼ black, 100 ¼ white), a* is one
chromatic coordinate (100 ¼ green, þ100 ¼ red), and b* is the other chromatic coordinate (100 ¼ blue, þ100 ¼ yellow).
From Table 6, one can see that the color of the (Ti,Mg)N coatings changes depending on the Mg/Ti atomic ratio. When increasing
the ratio, colors from golden to copper, violet, and metallic gray are obtained.
In the Zr–C–N system of cathodic-arc deposited coatings, a change of color starting from low-carbon ZrN to a low-nitrogen-
composition ZrC is as follows: pale brass–brass–gold–bronze–copper–gray (112).
The descriptions of colors provided by some commercial coating companies are listed in Table 4 in Section 4.02.3.10.1.
The cathodic arc’s ability to deposit coatings at relatively low deposition temperatures compared to other PVD processes and
with better adhesion is an advantage in coating on polymeric and plastic surfaces. Thus, the bright metallic appearance on plastic
appliances is produced by cathodic arc either by applying electroplated intermediate coatings or by coating directly on polymers
Figure 54 (113).
4.02.3.10.4 Biological Applications

Carbon thin films have a market and vast potential in biological applications. Technology related to human health and life sciences
is a rapidly growing field allowing an important opportunity for materials science. R. Hauert has made a review on the application of
modified DLC coatings for biological applications (114). Biomaterial implantation into the human body reestablishes biological
and mechanical functions aiming to increase the quality of life. The main applications of DLC are related to blood-contacting
implants with proper bio-reactions and the use of DLC to reduce wear in load-bearing contacts.
Filtered cathodic-arc deposition has been investigated by the Tiainen and Lappalainen groups for artificial hip joints in order to
reduce wear and lower the coefficient of friction (115,116). By using a hip simulator with a ta-C-coated joint ball against ultrahigh
molecular weight polyethylene (UHMWPE), a reduction of the UHMWPE wear of a factor of 10–100 was measured compared to
www.iran-mavad.com
Table 6 L*a*b* color coordinates, and conductivity of (Ti,Mg)N coating as a function of Mg/Ti compositiona
Color coordinate values

IMg Mg/Ti A r
(A) Atomic ratio (nm) L* a* b* (mUcm)
0 0 0.4306 78.6 1.0 35.2 98

0.1 0.03 0.4294 78.4 0.0 23.3 106
0.3 0.09 0.4301 75.0 2.6 26.3 110
0.2 0.22 0.4314 65.5 11.8 25.9 175
0.6 0.24 0.4320 64.7 12.3 25.5 183
0.5 0.41 0.4327 55.7 11.3 3.9 338
0.7 0.48 0.4331 55.4 7.5 0.4 406
0.7 0.53 0.4337 56.7 1.8 1.5 457
a
Martin, P.; Bendavid, A. Review of the Filtered Vacuum Arc Process and Materials Deposition. Thin Solid Films 2001, 394 (1), 1–14.
Figure 54 Examples of decorative coatings produced by commercial coating companies. (a) Blue dial coated by Diarc Technology Oy (with permission,
copyright Sarpaneva Watches Oy). (b) Blue component inside a watch coated by Diarc Technology Oy (with permission, copyright Diarc Technology
Oy. http://www.diarc.fi/decorative.html). (c) Faucet components (photo courtesy and with permission of Hauzer Ltd).
www.iran-mavad.com
Figure 55 Difference between Ti and ta-C in their ability to resist bacterial adhesion. Circular areas of (a) titanium (b) and ta-C are patterned on silicon. The
microbial growth is observed under optical microscope. Reproduced from Myllymaa, K. Novel Carbon Coatings and Surface Texturing for Improving
Biological Response of Orthopedic Implant Materials. Dissertation, Department of Physics and Mathematics, University of Eastern Finland, Finland, 2010.
a noncoated steel ball. In the case of a metal/metal contact, the reduction of wear was by a factor of 105 when compared to the
noncoated surfaces. An increase in corrosion resistance of the metals was observed as well. The use of ta-C coatings for hip joints has
recently been reviewed by Alakoski et al. (117), who described the best potential of thick ta-C coatings against each other. Then, the
release of harmful wear particles of the polymer cup is avoided. However, there has been a very long delay in the implant-
manufacturing industry to adapt the results for clinical tests.
The biocompatibility of ta-C coatings has been tested experimentally by several groups. The hemo compatibility of filtered
cathodic-arc-coated ta-C was observed to exhibit significantly less platelet aggregation compared to the reference carbon material
(low-temperature isotropic pyrolytic carbon, or LTI). This was attributed to the differences in charge transfer properties between the
two materials (118).
The doping of DLC in general is the means to modify the biocompatibility of the films. Kwong et al. have doped ta-C with silver
by applying a dual cathode of C and Ag of a cylindrical symmetry in a filtered pulsed cathodic-arc device (119). The Ag concen-
tration on the surface was varied between 70 and 90 at.%. But silver was in the form of metallic inclusions in a sp3 carbon matrix.
A high platelet adhesion was observed while having antibacterial properties due to silver. The results suggest benefits for cardio-
vascular devices and implant applications.
Doping of ta-C with polymers has been developed by Anttila group. The pulsed cathodic arc is filtered by a curved solenoid. The
carbon plasma is transported through a polymeric tube, allowing elements from the polymer to dope the plasma and eventually the
growing film (120). The doping by polytetrafluoroethylene (PTFE) and polydimethylsiloxane (PDMS) was obtained. The surface
energy was significantly lowered, yielding a water contact angle of 112 K. Myllymaa has compared the biofilm formations by
incubation of Staphylococcus epidermidis bacterial suspensions on the surface. The ta-C with PTFE was shown to be resistant to biofilm
formation (121).
About two-thirds of the infections related to implants are caused by the Staphylococcus aureus or S. epidermidis bacteria. The
S. aureus adherence to silicon substrates coated with magnetron-sputtered films of tantalum, titanium, and chromium and pulsed
filtered arc deposited ta-C were compared. The three metals are typical in metal implants. Determined by the amount of adhered
bacteria to the different films, the ta-C was the most microbe-repellent of the materials tested. The results indicate a benefit of the
development of ta-C in preventing device-related infections. The amount of microbial coverage of the materials was roughly
as follows: Ti 23%, Ta 14%, Cr 1.4%, and ta-C 0.4%. A microscopic image of the difference in microbial growth is shown in
Figure 55 (121).
References
1. Mahan, J. E. Physical Vapor Deposition of Thin Films; John Wiley & Sons: New York, 2000.
2. http://www.cleanroom.byu.edu/metal.phtml.
3. http://www.cleanroom.byu.edu/TFE_materials.phtml.
4. Kukla, R.; Ludwig, R.; Meinel, J. Overview on Modern Vacuum Web Coating Technology. Surf. Coat. Technol. 1996, 86–87, 753–761.
5. Ludwig, R.; Josephson, L. Clear Barriers and High Volume Productivity, Applied Materials Ltd. Polymers Laminations, Adhesives Coatings Extrusions PLACE 2006 Conference,
September 17–21, 2006, Cincinnati, Ohio http://www.tappi.org/content/enewsletters/eplace/2006/06PLA32.pdf.
6. Ludwig, R.; Kukla, R.; Josephson, E. Vacuum Web Coating – State of the Art and Potential for Electronics. Proc. IEEE 2005, 93 (8), 1483–1490.
7. http://www.leyboldoptics.com/optics/packaging/machines/pro-m/pro-m.html.
8. NAVAIR Public Release 2012-119 Distribution Statement A.
9. Zhang, L., et al. Structural, Optical and Electrical Properties of Low-Temperature Deposition Cu(InxGa1x)Se2 Thin Films. Sol. Energy Mater. Sol. Cells 2012, 99(C), 356–361.
10. Repins, I., et al. Co-Evaporated Cu2ZnSnSe4 Films and Devices. Sol. Energy Mater. Sol. Cells 2012, 101(C), 154–159.
www.iran-mavad.com
11. Vacuum Process Technology, LLC, http://www.vptec.com/products_deposition.htm.

12. Grytsenko, K. P.; Schrader, S. Nanoclusters in Polymer Matrices Prepared by Co-Deposition from a Gas Phase. Adv. Colloid Interface Sci. 2005, 116 (1–3), 263–276.
13. Gritsenko, K. P.; Krasovsky, A. M. Chem. Rev. 2003, 103 (9), 3607.
14. Kovacik, P., et al. Effect of Side GROUPS on the Vacuum Thermal Evaporation of Polythiophenes for Organic Electronics. Org. Electron. 2012, 13 (4), 687–696.
15. Goedicke, K.; Scheffel, B.; Schiller, S. Plasma-Activated High Rate Electron Beam Evaporation Using a Spotless Cathodic Arc. Surf. Coat. Technol. 1994, 68, 799–803.
16. Scheffel, B.; Modes, T.; Metzner, C. Spotless Arc Activated High-Rate Deposition Using Novel Dual Crucible Technology for Titanium Dioxide Coatings. In Extended Abstract
14th International Conference on Plasma Surface Engineering, September 10–14, 2012, in Garmisch-Partenkirchen, Germany.
17. Anders, A. Cathodic Arcs: From Fractal Spots to Energetic Condensations In Atomic, Optical, and Plasma Physics; Springer: New York, 2010. Vol. 50.
18. Anders, A. A Structure Zone Diagram Including Plasma-Based Deposition and Ion Etching. Thin Solid Films 2010, 518 (15), 4087–4090.
19. Timko, H. Modelling Vacuum Arcs: From Plasma Initiation to Surface Interactions. Report Series in Physics HU-P-D188, Theses, 2011.
20. Schülke, T.; Anders, A.; Siemroth, P. Macroparticle Filtering of High-Current Vacuum Arc Plasmas. IEEE Trans. Plasma Sci. 1997, 25 (4), 660–664.
21. Ehiasarian, A. P., et al. Influence of Steering Magnetic Field on the Time-Resolved Plasma Chemistry in Cathodic Arc Discharges. J. Phys. D Appl. Phys. 2004, 37 (15),
2101–2106.
22. Anders, S., et al. On the Macroparticle Flux from Vacuum Arc Cathode Spots. IEEE Trans. Plasma Sci. 1993, 21 (5), 440–446.
23. Hantzsche, E. Two-Dimensional Models of Expanding Vacuum Arc Plasmas. IEEE Trans. Plasma Sci. 1995, 23 (6), 893–898.
24. Byon, E.; Anders, A. Ion Energy Distribution Functions of Vacuum Arc Plasmas. J. Appl. Phys. 2003, 93 (4), 1899.
25. Anders, A. The Evolution of Ion Charge States in Cathodic Vacuum Arc Plasmas: A Review. Plasma Sources Sci. Technol. 2012, 21 (3), 035014.
26. Snaper, A. A. Arc Deposition Process and Apparatus. US Patent 3,625,848, 1971.
27. Ramalingam et al. Control of Vacuum Arc Material Operation, Method and Apparatus. US Patent, 4,673,477, 1987.
28. Vergasol, G. E. Electric Arc Vapor Deposition Device. US Patent, 5,037,522, 1991.
29. Anders, A.; Pasaja, N.; Sansongsiri, S. Filtered Cathodic Arc Deposition with Ion-Species-Selective Bias. Rev. Sci. Instrum. 2007, 78 (6), 063901.
30. Schneider, J. M., et al. Hydrogen Uptake in Alumina Thin Films Synthesized from an Aluminum Plasma Stream in an Oxygen Ambient. Appl. Phys. Lett. 1999, 74 (2), 200.
31. Ratayski, U., et al. Effect of the Carbon Ion Energy on the Microstructure of Ta-C/Cr Multilayers. Surf. Coat. Technol. 2011, 206 (7), 1753–1758.
32. Gorokhovsky, V. I.; Bhattacharya, R.; Bhat, D. G. Characterization of Large Area Filtered Arc Deposition Technology: Part I – Plasma Processing Parameters. Surf. Coat.
Technol. 2001, 140 (2), 82–92.
33. Drescher, D., et al. A Model for Particle Growth in Arc Deposited Amorphous Carbon Films. Diam. Relat. Mater. 1998, 7 (9), 1375–1380.
34. Hayess, E., et al. Measurements of the Behaviour of Neutral Atom Density in a Diffuse Vacuum Arc by Laser-Induced Fluorescence (LIF). IEEE Trans. Plasma Sci. 1989, 17 (5),
666–671.
35. Takikawa, H.; Tanoue, H. Review of Cathodic Arc Deposition for Preparing Droplet-Free Thin Films. IEEE Trans. Plasma Sci. 2007, 35 (4), 992–999.
36. Aksenov, I. I.; Bolous, V. A.; Padalka, V. G.; Khoroshikh, V. M. Instrum. Exp. Tech 1978, 21, 1416–1418.
37. Anders, A. Approaches to Rid Cathodic Arc Plasmas of Macro- and Nanoparticles: A Review. Surf. Coat. Technol. 1999, 120, 319–330.
38. Paperny, V. L., et al. Mass Separation of a Multi-Component Plasma Flow Travelling through a Magnetic Transport System. Plasma Sources Sci. Technol. 2011, 20 (3),
035005.
39. Zhitomirsky, V. N., et al. Properties of SnO2 Films Fabricated Using a Rectangular Filtered Vacuum Arc Plasma Source. Thin Solid Films 2008, 516 (15), 5079–5086.
40. Aksyonov, D.; et al. 2010. Deposition of Ti-Al-N Coatings Using Two-Channel T-Shaped Magnetic Filter. In 24th International Symposium on Discharges and Electrical
Insulation in Vacuum (ISDEIV), pp 494–496.
41. Zavaleyev, V.; Walkowicz, J. Application of the Taguchi Approach of the Design of Experiments for Determination Constructional and Working Parameters of the Linear
Venetian Blind Microdroplet Filter. Vacuum 2012, 86 (9), 1248–1254.
42. Anttila, A.; Hirvonen, J. -P.; Koskinen, J. Procedure and Apparatus for the Coating of Materials by Means of a Pulsating Plasma Beam. US5078848 (A), 1992-01-07.
43. Anders, A., et al. Ultrathin Diamond-Like Carbon Films Deposited by Filtered Carbon Vacuum Arcs. IEEE Trans. Plasma Sci. 2001, 29 (5), 768–775.
44. Anders, A.; Kulkarni, A. V. Materials Research Society Synthesis of Ultrathin Ta-C Films by Twist-Filtered Cathodic Arc Carbon Plasmas; Materials Reseach Society:
Warrendale, PA, 2001; p 675.
45. Chekh, Y.; Zhirkov, I. S.; Delplancke-Ogletree, M. P. Simple Filtered Repetitively Pulsed Vacuum Arc Plasma Source. Rev. Sci. Instrum. 2010, 81 (2), 023506.
46. Aijaz, A., et al. A Strategy for Increased Carbon Ionization in Magnetron Sputtering Discharges. Diam. Relat. Mater. 2012, 23(C), 1–4.
47. Fallon, P. J.; Veerasamy, V. S.; Davis, C. A.; Robertson, J.; Amaratunga, G. A. J.; Milne, W. I.; Koskinen, J. Properties of Filtered-ion-beam-deposited Diamondlike Carbon as
a Function of Ion Energy. Phys. Rev. B 1993, 48, 4777.
48. Koskinen, J.; Hirvonen, J.-P.; Keränen, J. Effect of Deposition Temperature and Growth Rate on the Bond Structure of Hydrogen Free Carbon Film. J. Appl. Phys. 1998, 84 (1),
648–650.
49. Bilek, M. M. M.; McKenzie, D. R. A Comprehensive Model of Stress Generation and Relief Processes in Thin Films Deposited with Energetic Ions. Surf. Coat. Technol. 2006,
200 (14–15), 4345–4354.
50. Charitidis, C. A. Nanomechanical and Nanotribological Properties of Carbon-based Thin Films: A Review. Int. J. Refract. Met. Hard Mater. 2010, 28, 51–70.
51. Casiraghi, C.; Robertson, J.; Ferrari, A. C. Diamond-Like Carbon for Data and Beer Storage. Mater. Today 2007, 10 (1–2), 44–53.
52. McCulloch, D., et al. Mechanisms for the Behavior of Carbon Films during Annealing. Phys. Rev. B 2004, 70 (8), 085406.
53. Chhowalla, M., et al. Boronated Tetrahedral Amorphous Carbon (Ta-C: B). Diam. Relat. Mater. 1997, 6 (2), 207–211.
54. Alakoski, E., et al. Adhesion and Quality Test for Tetrahedral Amorphous Carbon Coating Process. Diam. Relat. Mater. 2003, 12 (12), 2115–2118.
55. Chhowalla, M. Thick, Well-Adhered, Highly Stressed Tetrahedral Amorphous Carbon. Diam. Relat. Mater. 2001, 10 (3), 1011–1016.
56. Anders, A. Metal Plasma Immersion Ion Implantation and Deposition: A Review. Surf. Coat. Technol. 1997, 93 (2), 158–167.
57. Hakovirta, M., et al. Graphite Particles in the Diamond-Like a-C Films Prepared with the Pulsed Arc-Discharge Method. Diam. Relat. Mater. 1995, 4 (12), 1335–1339.
58. Cheng, Y. H., et al. Internal Stresses in TiN/Ti Multilayer Coatings Deposited by Large Area Filtered Arc Deposition. J. Appl. Phys. 2008, 104 (9), 093502.
59. Weng, K.-W.; Lin, T.-N.; Wang, D.-Y. Tribological Property Enhancement of CRN Films by Metal Vapor Vacuum Arc Implantation of Vanadium and Carbon Ions. Thin Solid
Films 2008, 516 (6), 1012–1019.
60. Stueber, M., et al. Concepts for the Design of Advanced Nanoscale PVD Multilayer Protective Thin Films. J. Alloy. Compd 2009, 483 (1–2), 321–333.
61. LIU, X., et al. Developing TiAlN Coatings for Intermediate Temperature Solid Oxide Fuel Cell Interconnect Applications. Int. J. Hydrogen Energ. 2008, 33 (1), 189–196.
62. Xie, Z. H., et al. Deformation Mechanisms of TiN Multilayer Coatings Alternated by Ductile or Stiff Interlayers. Acta Mater. 2008, 56 (4), 852–861.
63. Cheng, Y. H., et al. Mechanical and Tribological Properties of Nanocomposite TiSiN Coatings. Surf. Coat. Technol. 2010, 204 (14), 2123–2129.
64. Hodroj, A., et al. Oxidation Resistance of Decorative (Ti, Mg)N Coatings Deposited by Hybrid Cathodic Arc Evaporation-Magnetron Sputtering Process. Surf. Coat. Technol.
2011, 205 (19), 4547–4553.
65. Neidhardt, J., et al. Wear-Resistant Ti–B–N Nanocomposite Coatings Synthesized by Reactive Cathodic Arc Evaporation. Int. J. Refract. Met. Hard Mater. 2010, 28 (1), 23–31.
66. Look, D. C.; Hemsky, J. W.; Sizelove, J. R. Residual Native Shallow Donor in ZnO. Phys. Rev. Lett. 1999, 82, 2552.
67. Matsumoto, Y.; Hirata, G. A.; Takakura, H.; Okamoto, H.; Hamakawa, Y. A New Type of High Efficiency with a Low-cost Solar Cell Having the Structure of a mc-SiC/
polycrystalline Silicon Heterojunction. J. Appl. Phys. 1990, 67, 6538.
68. Lampe, U.; Muller, J. Thin-Film Oxygen Sensors Made of Reactively Sputtered ZnO. Sens. Actuators 1989, 18, 269.
www.iran-mavad.com
69. Li, C., et al. Research on the Properties of ZnO Thin Films Deposited by Using Filtered Cathodic Arc Plasma Technique on Glass Substrate under Different Flow Rate of O2.
Appl. Surf. Sci. 2007, 253 (8), 4000–4005.
70. Çetinörgü, E.; Goldsmith, S.; Boxman, R. L. The Effect of Annealing on Filtered Vacuum Arc Deposited ZnO Thin Films. Surf. Coat. Technol. 2007, 201 (16–17), 7266–7272.
71. Kavak, H., et al. Optical and Photoconductivity Properties of ZnO Thin Films Grown by Pulsed Filtered Cathodic Vacuum Arc Deposition. Vacuum 2008, 83 (3), 540–543.
72. Mendelsberg, R. J., et al. Achieving High Mobility ZnO:Al at Very High Growth Rates by dc Filtered Cathodic Arc Deposition. J. Phys. D Appl. Phys. 2011, 44 (23),
232003.
73. Gao, F., et al. Preparation of High Transmittance ZnO: Al Film by Pulsed Filtered Cathodic Arc Technology and Rapid Thermal Annealing. Appl. Surf. Sci. 2011, 257 (15),
7019–7022.
74. Zukerman, I., et al. Vacuum Arc Deposition of Al2O3–ZrO2 Coatings: Arc Behavior and Coating Characteristics. J. Mater. Sci. 2010, 45 (23), 6379–6388.
75. Tsukazaki, A.; Ohtomo, A.; Onuma, T.; Ohtani, M.; Makino, T.; Sumiya, M.; Ohtani, K.; Chichibu, S. F.; Fuke, S.; Segawa, Y.; Ohno, H.; Koinuma, H.; Kawasaki, M. Nat.
Mater. 2005, 4, 42.
76. Yuen, C., et al. Room Temperature Deposition of p-Type Arsenic Doped ZnO Polycrystalline Films by Laser-Assist Filtered Cathodic Vacuum Arc Technique. J. Appl. Phys.
2007, 101 (9), 094905.
77. Diebold, U. The Surface Science of Titanium Dioxide. Surf. Sci. Rep. 2003, 48, 53.
78. Bendavid, A.; Martin, P. J.; Takikawa, H. Deposition and Modification of Titanium Dioxide Thin Films by Filtered Arc Deposition. Thin Solid Films 2000, 360, 241.
79. Bendavid, A.; Martin, P. J.; Preston, E. W. The Effect of Pulsed Direct Current Substrate Bias on the Properties of Titanium Dioxide Thin Films Deposited by Filtered
Cathodic Vacuum Arc Deposition. Thin Solid Films 2008, 517 (2), 494–499.
80. Kleiman, A.; Márquez, A.; Lamas, D. G. Anatase TiO2 Films Obtained by Cathodic Arc Deposition. Surf. Coat. Technol. 2007, 201 (14), 6358–6362.
81. Díaz, B., et al. Chromium and Tantalum Oxide Nanocoatings Prepared by Filtered Cathodic Arc Deposition for Corrosion Protection of Carbon Steel. Surf. Coat. Technol.
2012, 206, 1–32.
82. Levchuk, D., et al. Erbium Oxide as a New Promising Tritium Permeation Barrier. J. Nuclear Mater. 2007, 367–370, 1033–1037.
83. Chikada, T., et al. Deuterium Permeation Behavior of Erbium Oxide Coating on Austenitic, Ferritic, and Ferritic/Martensitic Steels. Fusion Eng. Des. 2009, 84 (2–6),
590–592.
84. Ramm, J., et al. Pulse Enhanced Electron Emission (P3e™) Arc Evaporation and the Synthesis of Wear Resistant Al–Cr–O Coatings in Corundum Structure. Surf. Coat.
Technol. 2007, 202 (4–7), 876–883.
85. Byon, E.; Oates, T. W. H.; Anders, A. Coalescence of Nanometer Silver Islands on Oxides Grown by Filtered Cathodic Arc Deposition. Appl. Phys. Lett. 2003,
82 (10), 1634.
86. Soin, N., et al. Thickness Dependent Electronic Structure of Ultra-Thin Tetrahedral Amorphous Carbon (Ta-C) Films. Thin Solid Films 2012, 520 (7), 2909–2915.
87. Schneider, D., et al. Testing Ultra-Thin Films by Laser-Acoustics. Surf. Coat. Technol. 2000, 126 (2), 136–141.
88. Wang, G.-G., et al. Silicon Nitride Gradient Film as the Underlayer of Ultra-Thin Tetrahedral Amorphous Carbon Overcoat for Magnetic Recording Slider. Mater. Chem.
Phys. 2011, 131 (1–2), 127–131.
89. Gorokhovsky, V. I., et al. Characterization of Large Area Filtered Arc Deposition Technology: Part II – Coating Properties and Applications. Surf. Coat. Technol. 2001,
140 (3), 215–224.
90. Lee, H. Y., et al. Characterization of WC–CrAlN Heterostructures Obtained Using a Cathodic Arc Ion Plating Process. Surf. Coat. Technol. 2003, 174–175, 303–309.
91. Wang, G.-G., et al. The Preparation and Evaluation of Graded Multilayer Ta-C films Deposited by FCVA Method. Appl. Surf. Sci. 2011, 257 (11), 5064–5069.
92. Eklund, P.; Beckers, M.; Jansson, U.; Högberg, H.; Hultman, L. The Mnþ1AXn Phases: Materials Science and Thin-Film Processing. Thin Solid Films 2010, 518 (8), 1851–1878.
93. Guenette, M. C., et al. Cathodic Arc Co-Deposition of Highly Oriented Hexagonal Ti and Ti2AlC MAX Phase Thin Films. Thin Solid Films 2010, 519 (2), 766–769.
94. http://www.osti.gov/bridge/purl.cover.jsp?purl¼/810482-HCSxLp/810482.pdf.
95. Holmberg, K.; Matthews, A. Coatings Tribology; In Tribology and Interface Engineering Series, 2nd ed. Elsevier: Amsterdam, 2009; Vol. 6.
96. Holmberg, K.; Andersson, P.; Erdemir, A. Global Energy Consumption due to Friction in Passenger Cars. Tribiol. Int. 2012, 47(C), 221–234.
97. Hsu, C.-H.; Lee, C.-C.; Ho, W.-Y. Filter Effects on the Wear and Corrosion Behaviors of Arc Deposited (Ti, Al)N Coatings for Application on Cold-Work Tool Steel. Thin Solid Films
2008, 516 (15), 4826–4832.
98. Mo, J. L.; Zhu, M. H. Tribological Characterization of Chromium Nitride Coating Deposited by Filtered Cathodic Vacuum Arc. Appl. Surf. Sci. 2009, 255 (17), 7627–7634.
99. Musil, J.; Jirout, M. Toughness of Hard Nanostructured Ceramic Thin Films. Surf. Coat. Technol. 2007, 201 (9–11), 5148–5152.
100. Gorokovsky, V. I., et al. Deposition and Characterization of Hybrid Filtered Arc/Magnetron Multilayer Nanocomposite Cermet Coatings for Advanced Tribological Applications. Wear
2008, 265 (5–6), 741–755.
101. Zimmer, O.; Kaulfuß, F. PO4046-Hard Coatings with High Film Thickness Prepared by PVD. Plasma Process. Polym. 2009, 6 (S1), S152–S156.
102. Erdemir, A.; Donnet, C. Tribology of Diamond-Like Carbon Films: Recent Progress and Future Prospects. J. Phys. D Appl. Phys. 2006, 39 (18), R311–R327.
103. Kano, M. Super Low Friction of DLC Applied to Engine Cam Follower Lubricated with Ester-Containing Oil. Tribol. Int. 2006, 39 (12), 1682–1685.
104. Minami, I., et al.., STLE Memb. Investigation of Tribo-Chemistry by Means of Stable Isotopic Tracers, Part 2: Lubrication Mechanism of Friction Modifiers on Diamond-Like
Carbon. Tribol. Trans. 2007, 50 (4), 477–487.
105. Martin, J.-M., et al. Gas-Phase Lubrication of Ta-C by Glycerol and Hydrogen Peroxide. Experimental and Computer Modeling. J. Phys. Chem. C 2010, 114 (11), 5003–5011.
106. Koskinen, J.; Tapper, U.; Andersson, P.; Varjus, S.; Kolehmainen, J.; Tervakangas, S.; Buss, W. Friction Reduction by Texturing of DLC Coatings Sliding against Steel under Oil
Lubrication. Surf. Coat. Technol. 2010, 204, 3794–3797.
107. http://www.arcprecision.com/5503/4602.html.
108. http://www.iws.fraunhofer.de/en/business_fields/pvd_nanotechnology/carbon_coatings/technologies/filter-laser-arc.html.
109. Martin, P.; Bendavid, A. Review of the Filtered Vacuum Arc Process and Materials Deposition. Thin Solid Films 2001, 394 (1), 1–14.
110. Proc. SPIE 7045, 704508-7.
111. Bilus Abaffy, N.; Partridge, J. G.; McCulloch, D. G. Proceedings of SPIE. In Photovoltaic Cell and Module Technologies II; SPIE, pp 704508–704508–10.
112. Eerden, M.; van Ijzendoorn, W.;Tietema, R.; van der Kolk, G. J. A Systematic Study of the Properties of the Zr–C–N Ternary System, Deposited by Reactive Arc Evaporation. 48th
Annual Technical Conference, Society of Vacuum Coaters, San Francisco, CA, 2003; pp 56–60.
113. http://www.diarc.fi/decorative.html.
114. Hauert, R. A Review of Modified DLC Coatings for Biological Applications. Diam. Relat. Mater. 2003, 12 (3–7), 583–589.
115. Tiainen, V. M. Amorphous Carbon as a Bio-Mechanical Coating – Mechanical Properties and Biological Applications. Diam. Relat. Mater. 2001, 10 (2), 153–160.
116. Lappalainen, R., et al. Some Relevant Issues Related to the Use of Amorphous Diamond Coatings for Medical Applications. Diam. Relat. Mater. 1998, 7 (2), 482–485.
117. Alakoski, E., et al. Load-Bearing Biomedical Applications of Diamond-Like Carbon Coatings: Current Status. Open Orthop. J. 2008, 2, 43.
118. Yu, L., et al. Haemocompatibility of Tetrahedral Amorphous Carbon Films. Surf. Coat. Technol. 2000, 128, 484–488.
119. Kwok, S. C. H., et al. Hemocompatibility and Anti-Bacterial Properties of Silver Doped Diamond-Like Carbon Prepared by Pulsed Filtered Cathodic Vacuum Arc Deposition. Diam.
Relat. Mater. 2007, 16 (4–7), 1353–1360.
120. Anttila, A., et al. Preparation of Diamond-Like Carbon Polymer Hybrid Films Using Filtered Pulsed Arc Discharge Method. Surf. Eng. December 2003, 19 (6), 425–428.
121. Myllymaa, K. Novel Carbon Coatings and Surface Texturing for Improving Biological Response of Orthopedic Implant Materials. Dissertation, Department of Physics and
Mathematics, University of Eastern Finland, Finland, 2010.
www.iran-mavad.com
www.iran-mavad.com
4.03 Magnetron Sputtering
G Bräuer, Fraunhofer-Institut für Schicht- und Oberflächentechnik IST, Braunschweig, Germany
4.03.1 The Basics of Sputtering 57

4.03.2 Sputtering from Planar Diodes 59
4.03.3 Radio-Frequency Sputtering 60
4.03.4 The Magnetron Cathode 61
4.03.5 High Target Utilization 65
4.03.6 Reactive Sputtering 66
4.03.7 Pulse Magnetron Sputtering 69
4.03.8 Special Developments 71
4.03.8.1 The Unbalanced Magnetron 71
4.03.8.2 Sputter Yield Amplification and Serial Cosputtering 72
4.03.8.3 Hot Target Sputtering 72
4.03.9 Concluding Remarks and Outlook 73
References 73
4.03.1 The Basics of Sputtering
If an energetic ion hits the surface of a solid (target), the following effects may occur:
l Sputtering
l Emission of electrons (secondary electron emission)
l Ion implantation
l Ion reflection
l Lattice vibrations (heat generation)
In a thin-film deposition process, sputtering and secondary electron emission are the essential processes (Figure 1). Ejected
atoms form the thin film, and electrons are needed to create new ions by collisions in the negative glow of the plasma. The most
important quantity describing the sputter process is the sputter yield Y given by
Y ¼ Average number of emitted particles=Number of incident ions [1]
Figure 1 The principle of sputtering.

www.iran-mavad.com
58 Magnetron Sputtering
The first model for the sputtering process was published by Sigmund in 1969 (1).
According to Figure 2, an ion with a mass m1 hits a surface of an amorphous solid (mass m2) with an energy E0 at an incident
angle Q. The energy loss within the solid may be described by
ðdE0 =dxÞtot ¼ ðdE0 =dxÞn þ ðdE0 =dxÞe [2]
where (dE0/dx)n is the part that results in sputtered atoms, while (dE0/dx)e is the part that results in secondary electrons.
The energy transfer to the target atoms will generate more high-energy collision products and a collision cascade is formed in
a certain target volume close to the surface. A typical collision cascade is several tens of nanometers in length and contains around
1000 atoms. Emitted particles have gained enough energy to overcome surface binding forces. Approximately 99% of them leave the
target as neutrals.
According to Sigmund, the sputter yield in a wide range of ion energies is given by
YðE0 ; m1 ; m2 ; QÞ ¼ aðm2 =m1 Þ$Sn ðE0 Þ$U01 $ðcos QÞf ðm2 =m1 Þ [3]
where Sn ðE0 Þ ¼ ðdE0 =dxÞn =N [4]
N is the number of target atoms per volume, Sn(E0) is called ‘nuclear stopping power,’ and U0 is the surface binding energy (heat of
sublimation).
Figure 3 shows the principal behavior of Y as a function of ion energy. A certain threshold energy in the range of 50 eV is
necessary to generate a collision cascade. Up to 5–10 keV, Y increases linearly with E0. For E0 > 10 keV, a decrease is
observed due to ion implantation. For thin-film deposition by sputtering, the discharge voltages range from 200 V (low-
voltage magnetron) to a few thousand volts (diode), meaning that one is always working in a regime governed by a linear
increase of Y.
Figure 2 An ion hits a surface.
Figure 3 Qualitative dependence of sputter yield Y on ion energy.
www.iran-mavad.com
Magnetron Sputtering 59
Figure 4 Energy distributions of sputtered nickel atoms. Reproduced from Hippler, R.; Kersten, H.; Schmidt, M.; Schoenbach, K.-H. Low Temperature
Plasmas; Wiley-VCH, 2008. ISBN: 978-3-572-40673-9.
From eqn [3], it is quite clear that Y w 1/U0. Values for the heat of sublimation are in the range of several electronvolts. Materials
with high values are C (7.36 eV) and W (8.66 eV), thus these are difficult to sputter. Zn has a very low U0 (1.34 eV).
The bombarding ions are in general formed from an inert gas. Since the sputter yield strongly increases with the mass number of
the ion, Ne is not very effective. Ar is a good compromise in particular for industrial coating systems, since Kr and Xe are too
expensive. The sputter yield increases with the angle of incidence Q up to an angle of approximately 60 . A rough estimation in the
range from perpendicular incidence (Y0) and 60 may be done by
YðQÞ ¼ Y0 =cos Q [5]
However, the exact shape depends on parameters such as target mass m1 and ion energy E0. For glazing incidence, a steep
decrease of Y is observed due to reflection of the ions at the target surface.
The angular distribution of sputtered particles is given by
NðFÞ ¼ N0 $cosn F [6]
where F is the emission angle.
Depending on the various parameters, n varies between 1 and 2. The distribution function indicates that independent of the
angle of ion incidence Q most of the ejected atoms follow the target normal.
Of particular interest with respect to the film growth mechanisms is the energy of sputtered particles. The first significant results
were published by Thompson (2) who found that the energy distribution dY/dE of sputtered particles is independent of the energy
E0 of the incident ion. It shows the characteristic behavior of a Maxwell distribution with a maximum at U0/2 (see also Figure 4 (3)).
We can conclude that sputtered atoms in general have an energy of some electronvolts, which is approximately 10 times higher than
in case of evaporated particles. This is the reason why sputtered films exhibit a better quality than evaporated films (if the evap-
oration is done without additional plasma support).
It should be pointed out that the sputter yield is independent of target temperature as far as the latter is kept well below the melting
point (for sputtering with hot targets see Section 4.03.8.3). It should further be noted that 85% of the incident ion energy is converted to
heat, thus sputter targets have to be effectively cooled. Only approximately 2% of the incident energy is used to create sputtered particles.
4.03.2 Sputtering from Planar Diodes
Figure 5 outlines a setup for diode sputtering. The substrates to be coated are placed on an anode (if they are conductive, they may
act as anodes themselves). Often the anode is represented by the complete vacuum chamber. The cathode is surrounded by
a grounded dark space shield in order to avoid additional plasma discharges on the rear surface of the cathode. A sputter diode is
operated at a typical pressure of 5 $ 102 mbar to 101 mbar, the discharge voltage is in the range of 1000–3000 V, and the discharge
current is of the order of magnitude of 1 mA cm2. It is unnecessary to say that only conductive materials can be sputtered when
applying DC voltage (for sputtering of insulating targets see Section 4.03.3). Circular (100–500 mm) as well as rectangular target
shapes may be used. The target-to-substrate distance (TSD) is 50–150 mm.
The efficiency of a deposition process is described by the deposition rate
S ¼ d=t [7]
where d is the film thickness and t is the deposition time.
www.iran-mavad.com
Figure 5 Principal setup of a diode sputter system.
The unit of S is nm s1 or mm h1. Typical deposition rates for diode sputtering of metals are 0.5–1 nm s1.
It is obvious that S w Y and S w P ¼ U$I (P is the power delivered to the target, U is the discharge voltage, and I is the discharge
current). S also depends on the total pressure and the TSD (scattering of sputtered particles).
Diode sputtering suffers from some severe limitations. Due to the poor deposition rates the productivity is very low. Because of
bombardment with energetic electrons, there is a high thermal load to the substrate making the coating of plastic materials such as
polycarbonate nearly impossible. At a pressure of 101 mbar, the mean free path of sputtered particles is around 0.5 mm. On their
way to the substrate, they will lose substantial energy due to many collisions.
Due to these drawbacks nowadays diode sputtering is applied only in some special cases.
4.03.3 Radio-Frequency Sputtering
If a sputter system with an insulating target is powered by DC voltage, it will act like a capacitor. According to Figure 6, the charge of
the positive ions impinging on the nonconductive surface cannot be compensated by electrons from the power supply. The surface
will charge up, and the bombardment by further ions is suppressed. The sputter process will stop.
Figure 6 The application of DC voltage to an insulating target.
www.iran-mavad.com
Figure 7 The principle of RF sputtering.
Figure 8 The formation of target voltage in an RF sputter process.
It is well known from electrodynamics that alternating current can be driven through a capacitor if the frequency is high enough.
Figure 7 shows the principle of sputtering using radio frequency (RF) at 13.56 MHz. At such frequencies, the heavy ions are not able
to follow the fast changes of polarity. Due to reasons explained below, the surface areas of target electrode and substrate electrode
are different (often one is talking about an ‘asymmetric RF discharge’). Since the mobility of the electrons is excellent, they can
follow the field changes without any delay. Consequently, a DC bias will develop in particularly on the target electrode if the target
material is insulating (for conductive targets, the coupling to the RF power supply has to be accomplished by a blocking capacitor).
It should be mentioned that a DC bias may also develop on an insulating substrate electrode. However, based on the equation
UT =US ¼ ðAS =AT Þ4 [8]

where
UT is the DC bias at target electrode
US is the DC bias at substrate electrode
AT is the surface area of target electrode
AS is the surface area of substrate electrode
The DC bias at the substrate electrode will be very small if AS >> AT (which is the case for real sputter systems).
Figure 8 shows the superposition of RF voltage and DC bias. Based on the DC voltage, the ions will be attracted to the surface of
the insulator making sputtering on nonconductive compounds such as ceramics possible.
4.03.4 The Magnetron Cathode
The invention of the planar magnetron cathode by Chapin in 1974 (patent issued in 1979) marked a new era in vacuum
coating technology. Magnetron sputtering has become the most important technology for the deposition of thin films when
www.iran-mavad.com
excellent performance is required. Today, it is one of the key processes for manufacturing of innovative products such as all
kinds of disks for data storage and entertainment, flat displays, smart windows, or thin-film solar cells. The magnetron cathode
combines the advantages of economic deposition even on large areas and the ability to coat very temperature-sensitive plastic
substrates.
The planar magnetron consists of a diode equipped with an additional inhomogeneous magnetic field as shown in Figure 9. The
field is generated by permanent magnets (e.g., NdFeB) that are mounted on a ferromagnetic yoke. Electrons are trapped by the
Lorentz force K ¼ e (v B), which reaches a maximum in the area where Bx (the component parallel to the target surface) is large. It
decreases toward the center and the outer areas of the target surface (Bjjv). The trajectories of electrons are obtained by solving the
differential equation
me $dn=dt ¼ e$E þ eðn BÞ [9]
According to Figure 10, the electrons form a ring-shaped current on the target surface called the ‘race track.’ Due to the trapping
of the electrons in the race track area, the formation of ions is substantially enhanced, resulting in a remarkable drop of target voltage
(lower plasma impedance) and a jump in deposition rate by a factor of 10–20. The process pressure can be reduced by at least one
order of magnitude thus leading to an increase of the mean free path. Figure 11 shows I–U characteristics for Al sputtered from
a diode and a magnetron. Deposition rates for some target materials as function of power density are given in Figure 12. Planar
magnetrons can be realized with circular or rectangular shapes. An inherent drawback is the nonuniform erosion (causing
a nonuniform layer thickness distribution), which is caused by the nonuniform ion current density. In particular, this holds if the
substrate is resting below the magnetron during deposition. Large plates (such as glass panes) or foils can be arranged to pass
rectangular magnetrons at constant speed, in these cases, uniformity problems can be easily eliminated.
Figures 13–15 show magnetron cathodes used in industrial coaters.
In conclusion, the main advantages of magnetron sputtering are as follows:
l Low plasma impedance and thus high discharge currents from 1 to 100 A (depending on cathode length) at typical voltages
around 500 V
l Deposition rates in the range from 1 to 10 nm s1
l Low thermal load to the substrate
l Coating uniformity in the range of a few percent even for several meters long cathodes
Figure 9 Principle of a planar magnetron.
Figure 10 Magnetic field and electron trajectories (schematic).
www.iran-mavad.com
Figure 11 U –I characteristic for diode and magnetron.
Figure 12 Deposition rates for magnetron sputtering of different metals as function of power.
l Easy to scale up
l Dense and well-adherent coatings
l Large variety of film materials available (nearly all metals and compounds)
l Broadly tunable film properties
www.iran-mavad.com
Figure 13 Scheme of a planar magnetron cathode. Source: LEYBOLD.
Figure 14 3.75-m-long magnetron for architectural glass coating. Source: LEYBOLD.
www.iran-mavad.com
Figure 15 Magnetron cathodes in a machine for hard coatings. Source: Fraunhofer IST.
Some essential problems accompanying magnetron sputtering have been solved through substantial R&D efforts during the past
30 years. Such efforts mainly focused on the following goals:
l Improvement of target material utilization
l Stabilization of the reactive process in the transition regime
l Long-term stable high-rate deposition of dielectric films
l Improved film properties
l Higher deposition rates
4.03.5 High Target Utilization
Due to the strong inhomogeneous ion current density, the target material utilization of a planar magnetron is limited to 25–30%
(this holds for a standard magnet set with fixed magnets).
Around 1985, the development of circular magnetrons was driven by the need for fast and cost-effective metallization of a new
entertainment medium, the compact disc (CD).
One of the first attempts to improve the target utilization was the so-called interpoles target-hollow magnetron (IPT-HM) (4)
outlined in Figure 16. In such a configuration, a race track-shaped target is located between the poles of the magnet system (please
note that for a standard system, the magnets are behind the target). A very flat magnetic field is realized by soft magnetic pole pieces
positioned on top of permanent magnets. The erosion zone is broadened, and the material utilization increases up to 60%. Such
Figure 16 The interpoles target-hollow magnetron.
www.iran-mavad.com
Figure 17 The rotatable magnetron.
cathodes were mainly used in the early single disc-coating machines for the metallization of CDs. They marked the beginning of an
era of high-rate disc metallization processes with deposition rates around 50 nm s1. Self-sputtering was observed for Cu at high
power densities (5). Self-sputtering (sputtering without inert gas) occurs if the power density is high enough to ionize sufficient
sputtered material so that the metal ions can sustain the sputter process themselves. This might be seen as an early approach to high
power impulse magnetron sputtering (HIPIMS).
Using the IPT-HM together with a superimposed variable magnetic field created by an electromagnetic coil allows more degrees
of freedom: the area of maximum erosion can be shifted across the target surface.
Further approaches to improve the life of planar sputtering targets were oscillating permanent magnet arrays (6) or the use of
ferromagnetic sheets (7), where the modified field lines create two race tracks. Such configurations resulted in a material utilization
of 50–60%.
In the early 1980s, a tubular magnetron was developed (8,9) and was transferred to industrial coaters as ‘C-MAG’ (cylindrical
magnetron) a few years later. According to Figure 17, cathode body and target are tubes rotating around a fixed magnet array with
a frequency of about 1 Hz. New target material is continuously ‘turned’ into the plasma zone. With a material utilization up to 90%
and other advantages, such rotatable magnetrons are today state of the art in many industrial coaters.
4.03.6 Reactive Sputtering
Magnetron sputtering of metals is not a big challenge if the target surface is clean and proper process conditions (base
pressure, process pressure, and power) are applied. However, many applications in various industrial fields also need
oxides, nitrides, or carbides. Table 1 gives an overview of the most important compound films. Many of them are highly
insulating.
As mentioned in Section 4.03.3, such dielectrics can always be deposited by RF sputtering. However, there are drawbacks such
as high cost for RF power, low deposition rates, and high thermal load to the substrate. Furthermore, RF sputtering is very difficult
to scale.
Table 1 Some important compound thin films
Oxide Nitride Carbide
Al2O3 AlN B4C

Bi2O3 CrN SiC
HfO2 GaN TaC
In2O3–SnO2 Si3N4 TiC
NiO TaN WC
SiO2 TiN
SnO2 ZrN
Ta2O5
TiO2
VO2
ZnO
www.iran-mavad.com
In a reactive sputter process, the metal part of the compound is sputtered and the oxide, nitride, or carbide film is formed by
addition of the corresponding reactive gas. It is an inherent problem of reactive sputtering that the compound film is formed not
only on the substrate surface (where we like to have it) but also, with increasing process time, on all inner surfaces of the process
chamber (where we do not like to have it), in particular in the noneroding areas of the target surface (where we absolutely do not
like to have it). If, for example, Al2O3 is deposited by use of an Al target and an Ar/O2 gas mixture, parts of the target will get
poisoned by thin Al2O3 films. It is quite obvious that the degree of target poisoning will increase with the oxygen partial pressure.
Depending on this degree of target poisoning, the reactive sputter process is governed by two different stable target modes
l the metallic mode
l the reactive mode
Both modes are separated by an unstable transition regime.
In the metallic mode, the erosion zone of the target is free of reaction products. The deposition rate is close to the rate of the pure
metal, but the growing film is substoichiometric and useless. In the transition regime, at a critical reactive gas flow fRa, a sudden
coverage of the complete target with reaction products occurs, followed by a significant drop of sputter yield and an increase of
secondary electron emission (plasma impedance decreases). Since less-reactive gas for film formation is needed, the reactive gas
partial pressure (and thus the total pressure) in the coating chamber increases. In the reactive mode, the target is poisoned and only
compound material is sputtered. Stoichiometric films are deposited, however, at very low deposition rates (often a factor of 10–15
lower than in the metallic mode). The interaction of target surface with reactive gas results in the well-known hysteresis behavior of
reactive gas partial pressure, voltage, and deposition rate during reactive sputtering. The reactive gas flow has to be reduced to
a critical value fRb < fRa to turn back to the metallic mode. The hysteresis is shown schematically in Figure 18. Figures 19 and 20
show hysteresis curves for DC-reactive Al2O3 and TiO2 deposition.
To obtain stoichiometric films at high rate, it is indispensable to stabilize the process in the transition regime. The basic approach
for process stabilization is a constant reactive gas partial pressure. This can be accomplished by control of either the cathode power
or the reactive gas flow into the chamber. Several concepts have been developed. The most important ones are:
l Measurement of reactive gas partial pressure by a mass spectrometer.
l Only for oxygen: measurement with lambda probe.
l Optical emission spectroscopy: sputtered particles emit characteristic radiation in the visible spectrum. The intensity decreases
with increasing reactive gas partial pressure.
If the control is successful, any working point may be stabilized in the transition regime, as indicated in Figure 21.
While the hysteresis problem can be solved by suitable means of process control, challenges remain due to the growth of
insulating films.
The most severe disadvantage of reactive magnetron sputtering is ‘arcing,’ since it may limit the process stability in a drastic way.
A typical magnetron discharge is operating at 500 V and a power density of 10–20 W cm2. However, the discharge may spon-
taneously collapse into a small thread of ions and electrons called an arc. Such an arc discharge is characterized by low voltage and,
since the total power is concentrated on a very small target spot, by a power density in the 105 W cm2 range. Arcs will always
generate particulates, which may contaminate the growing film. Severe arcing may even give rise to local melting of the target or
local destruction of thin film and substrate. Simple models have been developed to explain arcing phenomena and their origin (10).
It is quite clear that arcs are mainly caused by the thin dielectric layers accumulated in the noneroding areas of the conductive target.
These layers steadily charge up due to the bombardment with positive ions from the plasma. The system negative-biased target-thin
dielectric film-positive surface charge can be described as a capacitor. An electrical breakdown occurs if the field strength in the
Figure 18 Schematic hysteresis behavior of pressure, voltage, and deposition rate for reactive sputtering.
www.iran-mavad.com
Figure 19 Hysteresis for Al2O3.
Figure 20 Hysteresis for TiO2.
capacitor exceeds the breakdown field strength, the latter being a constant for a dielectric material. Such breakdown phenomena
(they are called ‘microarcs’) dominate in the transition zones close to the metallic race track of the target. In these areas, the dielectric
coating is very thin. The electrical breakdown can inject sufficient charge into the plasma to cause it to collapse into a macroscopic
self-sustaining arc discharge.
Based on this understanding of mechanisms for arc formation, different strategies may be developed to achieve long-term stable
reactive sputter processes with high deposition rate. One attempt is the minimization or elimination of redeposited target areas. For
circular targets, this can be realized by plasma scanning of the entire surface using rotating magnet sets (11). Also, the target tubes of
the ‘C-MAG’ described in Section 4.03.5 remain almost free of redeposition. A second attempt may be the periodical neutralization
www.iran-mavad.com
Figure 21 Hysteresis with process control.
of the said surface charges. It has to be fast enough in order to avoid the breakdown. This implies the use of pulsed power supplies or
AC power instead of DC.
Besides arcing, the DC reactive sputter deposition of insulating materials is accompanied by another substantial problem, which
was often described by the term ‘disappearing anode.’ The sputter system requires an electrode that collects electrons from the
discharge (the anode), the resulting electron flow closes the electrical circuit given by power supply, sputter cathode, and the plasma.
This electrode may be a separate anode or metallic surfaces in the vicinity of the cathode such as shielding, chamber walls, etc.
During reactive deposition of insulating films, all inner surfaces finally will be covered with the insulator, and the anode will
disappear. The impedance of the discharge may increase; an undefined glow distribution, drifting potentials, unstable process
parameters, and, consequently, undefined film properties are observed. In a worst case, the discharge may be extinguished.
4.03.7 Pulse Magnetron Sputtering
Patents teaching that it might be advantageous to pulse a plasma range back to the late 1960s. However, substantial development
work started around 1990, driven by the needs of large-area coating for architectural glass and flat displays. If a single-sputter
cathode is driven by a rectangular pulse, a unipolar and a bipolar shape is possible according to Figure 22. The pulse frequency is in
the range of 1–100 kHz. Compared to DC power, in the case of the unipolar pulse also arcing is remarkably reduced since many arcs
are extinguished during the pulse-off time. In the bipolar version, the cathode is switched periodically to a small positive potential,
thus acting as an anode. Positive surface charges having been built up may be neutralized, and the process stability is further
improved. The problem of the disappearing anode, however, still remains. Furthermore, both solutions suffer from a loss in
deposition rate dependent on the duty cycle.
The optimum solution turned out to be a double magnetron arrangement powered by midfrequency in the range of 10–100 kHz
according to Figure 23. At any time, one of the magnetrons is on negative potential and acts as a sputter cathode, while the second
one acts as an anode. The temporary cathode is generating secondary electrons, which are accelerated toward the anode and
neutralize positive surface charges that have been built up in insulating areas during the negative half cycle. It is obvious that such
a system can also solve the disappearing anode problem. Independent of the environment, the temporary anode is able to collect
the electrons at any time. The result is a substantially improved process stability for reactive sputter deposition of materials such as
SiO2, Si3N4, or TiO2 even on larger areas and with higher power densities. Compared to DC reactive sputtering, often the deposition
rate can be improved by a factor of 2–5.
Figure 22 Pulses for a single magnetron.
www.iran-mavad.com
Figure 23 The principle of pulse magnetron sputtering from a dual unit.
Figure 24 Density of DC- and midfrequency (MF)-sputtered Si3N4 layers.
In general, for films sputtered in midfrequency discharges improved properties are found (13). This relates to
l higher density
l higher refractive index
l higher hardness
l reduced surface roughness
These properties are accompanied by a higher compressive stress. Jäger et al. (12) reported on measurement of ion current and
ion energy during midfrequency sputtering. They found that compared to DC much more energetic ions hit the substrate, which
results in a densification of the film structure. Figures 24 and 25 show some results for Si3N4 films (density and hardness vs process
pressure).
With the transfer of pulsed or midfrequency-driven magnetrons to industrial coaters, novel applications became possible. Some
of them are:
l improved low emissivity and solar control coatings on architectural glass
l sputtered antireflective coatings on large glass panes
l coatings for flat panel displays
l transparent conductive films for solar cells
l tribological coatings for components and tools
l coatings for sensors and precision optics
www.iran-mavad.com
Figure 25 Hardness of DC- and MF-sputtered Si3N4 layers.
Figure 26 C-MAG for pulsed or MF operation. Source: BOC Coating Technology.
Midfrequency sputtering using rectangular or sine wave-shaped pulses nowadays is standard for many applications in the
coating industry (see Figure 26).
4.03.8 Special Developments
Improved film properties through highly ionized plasmas and higher deposition rates are ongoing challenges in the development of
magnetron sputter sources. In this section, some special developments are presented.
4.03.8.1 The Unbalanced Magnetron

For a standard magnetron, the plasma is confined close to the target, while the substrate is located in an area with a low-density
plasma. The ion current to the substrate is small. However, often, higher ion current densities are required in order to deposit
dense compact films and control their microstructure. A controlled film growth can be achieved by use of a negative substrate bias
(a few 100 V) or an unbalanced magnetron, as outlined in Figure 27. Such an unbalanced magnetron is generated by an additional
magnetic field round the circumference of the cathode created either by permanent magnets or an electromagnetic coil (13–15). It is
possible to expand the discharge from the target and to immerse the substrate in a plasma of high density. An ion density at the
substrate that is higher than that of the coating material flux can be achieved, and much denser films are possible. Unbalanced
magnetrons are mainly used for deposition of hard coatings. For a standard magnetron, the ion current density often is considerably
lower than 1 mA cm2.
www.iran-mavad.com
Figure 27 Unbalanced magnetron.
Figure 28 Sputter yield amplification (SYA).
4.03.8.2 Sputter Yield Amplification and Serial Cosputtering

In 1992, Berg et al. (16) reported an effect they called ‘atom-assisted sputtering yield amplification.’ If light materials such as carbon
are bombarded by argon ions, sputtering is not very effective, since the momentum transfer is low and the collision cascade may
spread deep into the bulk of the target. If atoms of a heavier material are located a few monolayers below the surface of the sputter
target, the collision cascade is reflected by these heavier atoms and the sputter yield increases (see Figure 28). For Pt atoms
embedded below the surface of a C target, the authors found an enhancement by a factor of 3.
4.03.8.3 Hot Target Sputtering

As indicated in Section 4.03.1, around 85% of the electric power delivered to a sputter target is transferred to heat. As a consequence,
intensive target cooling is necessary to avoid melting. If the flow of cooling water is reduced, the target operates at elevated
Figure 29 Circular target operating close to the melting point.
www.iran-mavad.com
Figure 30 Deposition rate as function of power for SiO2 deposition from a hot target.
temperature. At temperatures close to the melting point (approximately T ¼ 0.8 Tm), so-called thermal spikes are formed in the
target and sputtering is supported by evaporation, leading to a strong increase of the yield. ‘Hot target sputtering’ may be an
interesting alternative to evaporation processes, if very high deposition rates are required (for example, for coatings on metal strips
running at speeds in a range of 100 m min1). Figure 29 shows the intensive plasma of a circular magnetron sputtered close to the
melting point. Figure 30 demonstrates the increase of the deposition rate for sputtering of SiO2.
4.03.9 Concluding Remarks and Outlook
The industrial implementation of magnetron sputtering has been accompanied by various milestones. The 1980s may be described
as the decade of reactive DC sputtering and the 1990s as the decade of pulsed sputtering and higher target utilization. If we take into
account the enormous work on HIPIMS discharges during the past 10 years, the 2000s may be seen as the decade of high ionization.
Many challenging tasks remain for the future. High ionized plasmas at further improvement of deposition speed are desirable, and
the complexity of reactive magnetron sputter processes and their control has to be reduced. The role of computer simulation in the
field of plasma sources and processes is steadily increasing. Simulation often saves much time and cost for expensive test equipment.
Finally, sputtering is more and more employed to deposit film stacks used in precision optics, where conventional evaporation
processes are at their limit. Optical filters nowadays require up to 1000 individual layers with thickness deviations less than 0.5% for
each layer.
References
1. Sigmund, P. Phys. Rev. 1969, 184, 383.

2. Thompson, M. W. Philos. Mag. 1968, 18, 377.
3. Hippler, R.; Kersten, H.; Schmidt, M.; Schoenbach, K.-H. Low Temperature Plasmas; Wiley-VCH, 2008. ISBN: 978-3-572-40673-9.
4. Kukla, R.; Krug, T.; Ludwig, R.; Wilmes, K. A highest rate self-sputtering magnetron source. Vacuum 1990, 41 (7–9), 1968.
5. Krug, T.; Beißwenger, S.; Kukla, R. In 34th Annual Technical Conference Proceedings of the Society of Vacuum Coaters 1991, Vol. 83.
6. De Bosscher, W.; Lievens, H. Advances in magnetron sputter sources. Thin Solid Films 1999, 351, 15.
7. Adam, R.; Krempel-Hesse, J.; Bähr, M. US Patent 6,139,706, 2000.
8. McKelvey, H. E. Magnetron Cathode Sputtering Apparatus. US Patent 4,356,073, 1982.
9. Wright, M.; Beardow, T. J. Vac. Sci. Technol. 1986, A4 (3), 388.
10. Scholl, R. A. In 36th Annual Technical Conference Proceedings of the Society of Vacuum Coaters 1993, Vol. 405.
11. Bräuer, G.; Dicken, W.; Müller, S. Optical Thin Films III: New Developments. Proc. SPIE 1990, 1323, 151.
12. Jäger, S.; Szyszka, B.; Szczyrbowski, J.; Bräuer, G. Surf. Coat. Technol. 1996, 98, 1304.
13. Window, B.; Savvides, N. J. Vac. Sci. Technol. 1986, A4 (2), 196.
14. Musil, J.; Kadlec, S.; Münz, W.-D. J. Vac. Sci. Technol. 1990, A8, 1318.
15. Pauleau, Y. Materials Surface Processing by Directed Energy Techniques; Elsevier Science: Oxford, UK, 2006. ISBN: 9780080444963.
16. Berg, S.; Barklund, A. M.; Gelin, B.; Nender, C.; Katardjiev, I. J. Vac. Sci. Technol. 1992, A10 (4), 1592.
www.iran-mavad.com
www.iran-mavad.com
4.04 High Power Impulse Magnetron Sputtering – HIPIMS
R Bandorf, V Sittinger, and G Bräuer, Fraunhofer-Institut für Schicht- und Oberflächentechnik IST, Braunschweig, Germany
Published by Elsevier Ltd.
4.04.1 Introduction to High Power Impulse Magnetron Sputtering 75

4.04.2 HIPIMS Generation 75
4.04.2.1 HIPIMS Operation Modes 76
4.04.2.1.1 Unipolar HIPIMS Mode 77
4.04.2.1.2 Bipolar HIPIMS Mode 80
4.04.2.1.3 Pulse Packages (Pulse Trains, Oscillatory Pulses) 81
4.04.2.1.4 Superimposed HIPIMS 82
4.04.3 HIPIMS Process Technology 83
4.04.3.1 Influence of Magnetic Field 84
4.04.3.1.1 Magnetic Configuration of the Cathodes 84
4.04.3.1.2 Guiding the Material Flux by Magnetic Field 86
4.04.3.2 Reactive Process Control 86
4.04.4 HIPIMS Applications 88
4.04.4.1 HIPIMS Pretreatment 89
4.04.4.2 Coating Systems 90
4.04.4.2.1 Hard Coatings 91
4.04.4.2.2 Electrical Coatings 92
4.04.4.2.3 Insulating Layers 93
4.04.4.2.4 Transparent Conductive Oxides 94
4.04.4.2.5 Antibacterial Coatings 96
4.04.5 Patent Situation 97
4.04.6 Outlook 98
References 98
4.04.1 Introduction to High Power Impulse Magnetron Sputtering
High power impulse magnetron sputtering (HIPIMS) or high power pulse magnetron sputtering (HPPMS) is a relatively new
technology in the transition from academic research to industrial applications. By using ions instead of neutrals, or at least
a significant fraction of ionized species, for thin film deposition, the resulting film properties can be altered significantly.
There are different approaches for ionized sputtering, and HIPIMS is the most recent development (28). The basic devel-
opment can be traced back to the 1960s in Russia. It took time to develop the technology and apply it for coating
deposition. Some of the early papers on thin film deposition using high power pulse plasmas were published in the mid-
1990s (12,33). Commonly, the paper by Vladimir Kouznetsov et al. in 1999 is cited as the seminal paper on HIPIMS (31).
Ever since, the technology has been further developed and improved by different research teams (15–19,28,38).
In line with the development of this pulse technique, a new generation of pulse power supplies, different from
conventional ones, was developed. The HIPIMS power supply operates at average power comparable to DC or MF power
supplies but the discharge is pulsed with a low duty cycle, resulting in peak power and current up to some megawatts or
some kiloamperes, respectively. The high peak power or current densities at the cathode are required to get a significant
amount of target material ionized (2,13,26). The ionized target material results in a significant modification or improve-
ment of the properties of the growing films regarding film density, hardness, roughness, refractive index, and so on
(7,17,29,47). An example is shown in Figure 1 where the film structure is significantly different for DC and HIPIMS-
sputtered NiCr films.
4.04.2 HIPIMS Generation
The general principle for generating HIPIMS plasmas is well known, with small modifications to all of today’s commercially
available power supplies. The typical HIPIMS power supply consists of a bank of capacitors that is charged at a constant voltage
Ucharge by a DC power source. These capacitors are then unloaded into the plasma through an inductor, L. The energy stored in the
capacitor can be calculated by:
1
Epulse ¼ C$UCharge 2 [1]
2

www.iran-mavad.com
76 High Power Impulse Magnetron Sputtering – HIPIMS
Figure 1 SEM cross-section of NiCr films. Left: conventional DC-sputtering (columnar growth); right: HIPIMS (dense, glassy structure).
Using a specific duty cycle or repetition frequency f of the pulse, the average power on the cathode can be calculated easily.
Usually both target voltage and current are monitored during a single pulse using an oscilloscope. From the course of U(t) and I(t),
the delivered energy per peak can be calculated:
Z
Epulse ¼ UðtÞ$IðtÞdt [2]
The typical setup for an HIPIMS generator is shown in Figure 2. There might be an additional transformer between the capacitor
bank and the plasma, resulting in modified current–voltage behavior of the generator. Additionally, there could be a matching
network and some sophisticated approaches for arc handling.
Depending on the setup of the power supply, HIPIMS processes operate in unipolar, bipolar, or burst mode (pulse packages,
pulse trains). Additionally, the HIPIMS discharge can be superimposed by a second plasma generated at the same cathode or
a second source. A general model describing the HIPIMS process and especially the processes of the species involved in the sput-
tering process was proposed as target material pathway model by Christie (16) and further improved by Vlcek (45). This model is
shown in Figure 3. An arriving gas ion Gþ t hits the target and sputters an atom Mtot of the target material with a sputtering yield SMG.
There is a probability b that the target atom Mtot becomes ionized. With a probability 1 b the atom remains neutral. The
probability that the atom becomes ionized by collisions in the vapor phase is g. The neutral atom can get lost with a probability of
1 xn or will arrive at the substrate as neutral atom MS with a probability xn. The ionized target material Mþ t in front of the target can
be redirected to the target with a probability of s and sputter the target with a sputtering yield of SMM. 1 s is the probability that
the ion from the target will travel toward the substrate. 1 xi is the probability that the ion get lost on the way to the substrate.
xi finally is the probability that the ionized target atom Mþ S will reach the substrate (45).
4.04.2.1 HIPIMS Operation Modes

When using HIPIMS technology, different operation modes are possible (Figure 4), depending on the used power supply. Typical
pulse on-times in any case of the HIPIMS processes are ranging from a few microseconds up to some milliseconds. The basic
principle is a unipolar pulse. Furthermore like in conventional pulse magnetron sputtering also bipolar operation is possible.
Especially for reactive processes this might be beneficial. Instead of using a single pulse there might be also pulse packages or pulse
trains used. This can be either in repeating the single pulses or using a transformer leading to the so-called modulated pulse power
(MPP) sputtering. Furthermore there can be voltage oscillations applied to the cathodes. Finally the HIPIMS discharge can be
superimposed by additional discharges on the same cathode (8,43,44) or a second material source (34).
Figure 2 Schematic principle of an HIPIMS power supply.
www.iran-mavad.com
High Power Impulse Magnetron Sputtering – HIPIMS 77
Figure 3 Improved target pathway model. Reproduced from Vlcek, J.; Kudlacek, P.; Burcalova, K.; Musil, J. J. Vac. Sci. Technol. A 2007, 25 (1), 42.
4.04.2.1.1 Unipolar HIPIMS Mode

The most used and by far most communicated operation mode is unipolar HIPIMS. Depending on the setup of the power supply,
there might be a high initial applied voltage that is subsequently dropping after ignition with onset of the current rise (Figure 5(a)).
Alternatively, today mostly a quasiconstant voltage is applied to the sputtering cathode and the current rises in a triangular shape
(Figure 5(b)). Depending on the pulse duration, the current shows a triangular shape, or for longer pulses, reaching a maximum
followed by a steady state or decay of the pulse current.
Comparing DC, MF, and HIPIMS operation of a nonmagnetic target, an interesting effect on target utilization can be observed.
Keeping the average power constant, pulse discharges have a larger average power in the pulse. This becomes extreme in HIPIMS.
Figure 4 Different modes of HIPIMS: (a) unipolar; (b) bipolar; (c) pulse packages/pulse trains; and (d) superimposed.
www.iran-mavad.com
Figure 5 Voltage, current, power curves of a single HIPIMS pulse using different settings of power supply (different L, C). (a) High ignition voltage
with delayed current rise and (b) “rectangular voltage shape” correlated with linear current increase. Reproduced from Sittinger, V. et al. Thin Solid
Films 2008, 516, 5847–5859.
Therefore, the racetrack is broadened using pulse sputtering, especially HIPIMS. Figure 6 shows the erosion profiles of nonmagnetic
targets sputtered at the same cathode using DC, MF, and HIPIMS. It is obvious that with increasing peak power of the discharge the
racetrack is broadened.
Due to the higher voltage applied in HIPIMS and the nonlinear correlation of the sputtering yield, the deposition rate in
nonreactive HIPIMS is reduced, as reported by Emmerlich et al. (23). In reactive HIPIMS, there are working conditions in which
a stoichiometric compound is formed at a higher rate using HIPIMS compared to the reference process as shown by Sarakinos
et al. (37).
Increasing the peak power is also correlated to increasing the peak current in the discharge. Furthermore, the ionization of the
sputtered target material is correlated to the current in the discharge. Optical emission spectroscopy (OES) is used to gather
information on the existence and the amount of ionization in the discharge. Using OES, a pulse time mode was used to scan for
specific emission lines. Figure 7 shows the voltage and current evolution of a single HIPIMS pulse, as well as the correlated optical
emission for neutral and ionized titanium species. For neutral titanium, there is just a linear increase in the emission intensity
observed. The emission of the ion lines follows the shape of the current and shows a more triangular shape. This means that with
increasing current, more titanium atoms become ionized. At a certain threshold the emission line for Tiþ starts to increase. The
emission of Ti2þ follows at higher current levels of the discharge current. Plotting the emission intensities over the pulse current
offers the definition of threshold current for the excitation of specific species (6). On the other hand, it seems that the number of
ions, or in general the ion to neutral ratio, is correlated with the discharge current. Anyway, it has to be taken into account that the
www.iran-mavad.com
Figure 6 Erosion profile of nonmagnetic sputtering targets in DC, MF, and HIPIMS operation. Reproduced from Sittinger, V.; Szyszka, B.; Bandorf, R.;
Vergoehl, M.; Pflug, A.; Christie, D. J.; Ruske, F. 51st SVC TechCon Proceedings; 2008, 35.
measurement was performed with a fiber facing the target. For the propagation of the sputtered species, there are measurements
published for different distances to the target showing that the composition of the plasma is changing (22).
As suggested by the OES data, the electron density during the pulse is significantly changing. Hecimovic et al. showed that the
electron density in a unipolar HIPIMS pulse is a function of the operating pressure and the distance to the target (Figure 8) (27). For
80 ms pulses the electron density was measured at 2.5, 5, 10, and 15 cm distance from the target. It can be seen that the electron
Figure 7 Single HIPIMS pulse; voltage and current evolution in correlation with emission lines for Ti0, Tiþ, Ti2þ. Reproduced from Bandorf, R.; et al.
49th SVC TechCon Proceedings 2006, 21.
www.iran-mavad.com
Figure 8 Electron density as function of time at different distances to the target at (a) 1 Pa and (b) 3 Pa (27). Reproduced from Hecimovic, A.;
Ehiasarian, A. P. 52nd SVC Annual TechCon Proceedings; 2009, 240.
density is highest at closest distance to the target and at lowest pressure. Increasing the distance leads to some drop of electron
density. More severe is the increase of the operating pressure. The maximum in the electron density at 15 cm distance nearly fully
vanishes (one order of magnitude smaller). The electron density at the large distance shows only a small peak and remains at the
density of the closer distances after 300 ms time.
4.04.2.1.2 Bipolar HIPIMS Mode

Similar to midfrequency sputtering, especially when using reactive processes forming insulating coatings, bipolar operation offers
the chance to discharge the target surface and avoid arcing. In the bipolar mode a dual magnetron configuration is used. The two
cathodes are pulsed alternatingly. In the first half of a bipolar pulse cycle, a negative voltage is applied to the first cathode while the
second cathode is slightly positive. Then in the second half, the second target serves as cathode while the first is on positive potential
with reference to the second target. By alternatively pulsing, a charging of the target surface is avoided and thereby arcing can be
reduced or fully eliminated. One example is given for TiO2 coatings (voltage and current curves are shown in Figure 9).
As shown by Vergöhl et al., bipolar HIPIMS was used to improve refractive index and density of titania coatings (44). Figure 10
shows a collection of density and refractive index values published by several groups, as well as results of films deposited by HIPIMS.
The highest values for refractive index and density are reported for bipolar HIPIMS.
Using a symmetric pulse shape shows similar peak current and power at the target for each pulse. There might be a difference in
peak current when using a dual magnetron setup with different magnetic field strength and different target erosion. Figure 11(a)
shows a symmetric bipolar pulse on a dual magnetron setup. The charge voltage of the bipolar power supply was fixed but the
resulting peak current differed. The reason was the difference in the magnetic field strength (Figure 11(b)). Cathode B with the
stronger field showed higher values of the peak current.
Therefore, it is essential to adjust the settings properly when using more than one cathode in HIPIMS mode. Additionally the
off-time between the first and the second pulse is expected to show an influence on the resulting peak current in each pulse. In the
case of remaining charged particles in the chamber, the current rise of the second pulse will be different since the ignition process
is faster.
www.iran-mavad.com
Figure 9 Voltage and current curves of bipolar HIPIMS.
Figure 12 shows the current shape of bipolar HIPIMS pulses with modified off-time between first and second pulse for fixed
charging voltage and repetition frequency of the dual pulse.
4.04.2.1.3 Pulse Packages (Pulse Trains, Oscillatory Pulses)

Using pulse packages gives the benefit of creating a preionization with the first pulse for the following ones. Furthermore, by
interrupting a long single pulse into different segments, the probability of arcing is also reduced. Figure 13 shows a pulse package of
six pulses. The first pulse shows a delayed current rise due to ignition, while the following pulses show simultaneous current rise,
starting already from a defined current level.
By modifying the on- and off-times of the pulses within a sequence (also called macropulse), the resulting current level of the
discharge is modified. Depending on the setup of the power supply, the voltage is fixed or can be modified itself (in the case of
a transformer between pulsing unit and sputtering target). The latter case is also known as modulated pulse power sputtering MPP.
In MPP the overall pulse time is usually longer than in conventional HIPIMS. The peak current in contrast is mostly lower,
depending on the cathode size resulting in a current density of a few 0.1 A cm2, typically (Figure 14).
A recent modification within the field of HIPIMS uses voltage oscillations to build up pulse files with a high discharge current.
Especially by using the oscillating voltage (Figure 15), reaching zero level between the oscillations is another approach to avoid
arcing in the process.
Figure 10 Refractive index and density of TiO2 coatings prepared with different methods. Reproduced from Vergöhl, M.; Werner, O.; Bruns, S. Proc. of
SPIE 2008, 2008; Vol. 7101, p 71010B-1.
www.iran-mavad.com
Figure 11 Target current for constant charging voltage for a dual magnetron with different magnetic field strength: (a) current curves and (b) mapping
of the magnetic field strength.
Using Langmuir measurements, it can be seen that there is a significant difference in the floating potential at the sub-
strate position. In the conventional HIPIMS case, there is a constant negative potential during the pulse of a few volts.
When using the plasma oscillations, there is a significant influence of the off-time between the oscillations. If the off-time
is small, the plasma potential stays negative during all the oscillations with small modulations, comparable to single
HIPIMS pulses. When increasing the off-time between the oscillations the floating potential reaches ground potential
(Figure 16).
4.04.2.1.4 Superimposed HIPIMS

One of the reported drawbacks of HIPIMS is the reduction of the deposition rate (23). Due to differences in the sputtering yield as
a function of the discharge voltage and due to back attraction of ionized target material toward the target, an effectively lower
deposition rate is observed for nonreactive HIPIMS processes. To overcome this drawback, different approaches were made
matching different discharges. In general, a DC or MF discharge is superimposed with an HIPIMS discharge. This superposition can
either be on the same cathode (8,43,44) or using different cathodes (34) (Figure 17).
Combining DC and HIPIMS allows for increasing the resulting deposition rate. When combining DC and HIPIMS at the same
source, the total average power must not exceed the limits of the used cathode. Thus with increasing the DC amount the rate will be
increased, but the increase might be at the expense of the HIPIMS portion. When combining DC and HIPIMS also the resulting film
properties will be a mix of the pure DC and HIPIMS properties. At first glance, it looks like the resulting properties are corre-
spondingly adjusting to the portion of the DC and HIPIMS fraction applied to the cathode. Figure 18 shows the increase in the
deposition rate when adding DC to an HIPIMS discharge. In this case the average power is increased from 2 kW pure HIPIMS in
0.5 kW steps adding DC up to a total average power of 4 kW. The resulting SEM cross-section shows that in a similar matter the
morphology changes from very dense to more DC-like columnar structure (9).
Therefore combining HIPIMS and DC can lead to an increase of the deposition rate, but it effectively reduces the peak current in
pure HIPIMS discharge, which in turn also will increase the deposition rate.
www.iran-mavad.com
Figure 12 Current shape of bipolar HIPIMS pulses with modified off-time between first and second pulse for fixed charging voltage and repetition
frequency of the dual pulse: (a) long off-time of 4900 ms: no remaining charge; (b) short off-time of 20 ms: higher current, instantaneous current rise
with higher peak current.
4.04.3 HIPIMS Process Technology
Besides the pulse pattern used for the discharge, there are additional important factors influencing the deposition process (32). On
the one hand, the magnetic setup, both at the cathode side and for guiding the material flux, plays an important role. On the other
hand, in the case of reactive HIPIMS, there are reports on hysteresis free sputtering as well as an adapted process control to stabilize
an HIPIMS discharge also on a large scale.
Figure 13 Pulses package of six pulses. Current rise of the first pulse is delayed; the following pulses have an instantaneous current rise.
www.iran-mavad.com
Figure 14 Voltage, current, and power evolution in the case of MPP for a macropulse of 1 ms; the pulse is subdivided in a low ionized phase (first half)
and a high ionized phase (second phase). Reproduced from Papa, F.; et al. Thin Solid Films 2011, http:dx.doi:10.1016/j.tsf.2011.09.004.
4.04.3.1 Influence of Magnetic Field

There are two different sites where the magnetic setup can influence the HIPIMS process. The plasma impedance is influenced
mainly by the magnetic field at the cathode, and also by the use of magnetic coils between the target and the substrate to guide the
flux of the ionized species.
4.04.3.1.1 Magnetic Configuration of the Cathodes

As shown by Anders (3) for a given electron temperature, the gyration radius of the electrons and therefore the cross-section for
ionization in front of the cathode are depending both on the applied voltage and the strength of the magnetic field in front of the
cathode. Figure 19 shows that the electron gyration radius is increasing with increasing electron energy, i.e., increasing applied
voltage at the cathode. Additionally, the gyration radius decreases with increasing magnetic field strength of the cathode. Therefore,
increasing the magnetic field strength will trap more electrons close to the cathode. With increasing target voltage, which is usually
the case when increasing the ionization in HIPIMS, the gyration radius increases. The confinement of the electrons in front of the
cathode is improved with increased magnetic field strength. As a result, the same peak current can be reached in HIPIMS at lower
charging voltages when increasing the magnetic field strength.
The deposition rate can be significantly increased by weakening the magnetic field strength and releasing more ionized material
in the vicinity of the cathode (21). In turn also the resulting structure is modified. Figure 20 shows the cross-section of chromium
films deposited by HIPIMS using different magnetic configurations of the cathode. The maximum tangential component of the
Figure 15 Oscillating voltage plasma; voltage and current are a function of the used frequency.
www.iran-mavad.com
Figure 16 Floating potential at substrate position with oscillatory excitation. ton: 16 ms, toff: 50 ms, Ipeak: 100 A (w0.5 A cm2).
Figure 17 DC superimposed HIPIMS on the same cathode. Initial current rise in the HIPIMS pulse when running a constant DC discharge in the
HIPIMS off-time. Reproduced from Bandorf, R.; Falkenau, S.; Schmidt, V. 50th SVC Annual TechCon Proceedings; 2007, 477.
Figure 18 Increase in deposition rate by superimposing DC on HIPIMS. Reproduced from Bandorf, R.; Falkenau, S.; Schiffmann, K.; Gerdes, H.;
Heckmann, U. 51st SVC Annual TechCon Proceedings; 2008, 59.
www.iran-mavad.com
Figure 19 Electron gyration radius as a function of the magnetic field strength of the sputtering cathode. Reproduced from Anders, A.; Andersson, J.;
Ehiasarian, A. P. J. Appl. Phys. 2007, 102, 113303.
magnetic field is decreasing from (a) 50 mT, (b) 40 mT, (c) 32 mT, (d) 25 mT to (e) 17 mT. It can be clearly seen that the resulting
structure also changes from a dense structure to a columnar DC-like structure by reducing the magnetic field strength.
Furthermore open or closed field configuration affects the number of electrons and ions. For an open field configuration, the
electron density is lower. Using the same process parameters in closed field configuration, the resulting structure is densified. Also
the single grains are close together, while in open field configuration there are numerous voids visible in top view (Figure 21).
4.04.3.1.2 Guiding the Material Flux by Magnetic Field

Besides the properties of the magnetic field directly at the cathode, the flux of ionized material can be influenced by using coils
between cathode and substrate. Since in HIPIMS a significant fraction of the target material is ionized, the current applied to the
coils will influence the fraction of ionized species in their sputter direction. Figure 22 (left) shows the experimental setup used to
study the influence of the magnetic field for guiding the flux of target material (a) without and (b) with the coils. For the setup with
the coils two different magnetic configuration (low and high field strength) were used. For several positions perpendicular and
parallel to the target surface, the deposition rate was measured (Figure 22, right). While the deposition rate for the chosen system
was between 2 and 4 nm min1 without additional field for all the measurement points the deposition rate was almost doubled for
the high field parallel to the target surface. The deposition rate perpendicular is correspondingly reduced showing, as expected that
the flux, because ionized can be guided by magnetic fields. Depending on the application, e.g., for static deposition a significant
improvement of the deposition rate and also an increased ion to neutral ratio by focusing the ionized flux can be realized.
4.04.3.2 Reactive Process Control

The use of reactive gases like oxygen or nitrogen for sputtering reactive HIPIMS processes has attracted growing interest in recent
years. For laboratory-scale systems, a reduction of the typical hysteresis of reactive oxide processes (36), and even a hysteresis free
Figure 20 SEM images of cross-section of the coating deposited at five different magnetic field settings of the cathode: decreasing field strength
from (a) 50 mT, (b) 40 mT, (c) 32 mT, (d) 25 mT to (e) 17 mT. Reproduced from Ehiasarian, A. P.; Vetushka, A. 52nd SVC Annual TechCon Proceedings;
2009, 265.
www.iran-mavad.com
process, are reported (46). Sarakinos et al. studied reactive HIPIMS of TiOx films. The results indicated that under specific conditions
the deposition rate can be higher than during DC using the same average current. Furthermore, the change in target voltage
associated with the oxidation of the target occurs at higher O2 flows in the HIPIMS case.
Wallin and Helmersson reported on sputtering Al2O3 from a 200 cylindrical target with current densities in the range of 0.3–
1.3 A cm2 in HIPIMS mode without observing any hysteresis. For the investigations, pulses of 35 ms and a repetition frequency
of 1 kHz was used. For DC operation, the highest deposition rates for stoichiometric oxide films are reported at 70% of the metal
deposition rate; for HIPIMS, the deposition rate was increased to 140% (46).
For large-area cathodes, Sittinger et al. still observed a hysteresis and the need for active process control. Since HIPIMS is using
a charged capacitor bank for pulsing, a power control as used in reactive sputtering is not directly applicable for HIPIMS. Hence,
a frequency modulation in reactive HIPIMS was used. The pulse duration was kept constant for all the HIPIMS pulses. With
sufficient off-time between the pulses, the shape of the HIPIMS pulse and the corresponding peak current is not changed, as shown
in Section 4.04.2.1.2. Therefore the modification of the off-time results in an effective modulation of the applied voltage without
changing the charging voltage and therefore the pulse shape of the single HIPIMS pulse. In the DC or pulsed-DC case, again the
power can be modified according to the feedback signal from a lambda probe or optical emission, by modulating the pulsing
frequency, i.e., the off-time (39). Figure 23 shows the process characteristics for reactive HIPIMS of aluminum-doped zinc oxide at
(b)
Figure 21 Open (a, b) and closed field (c, d) configuration. Left: OES spectra; (a, c), right: SEM top view (b, d); 2.5 kW average power.
www.iran-mavad.com
(d)
Figure 21 (Continued).
different charge voltages UCh. The hysteresis was examined by recording the oxygen partial pressure for increasing and decreasing
frequency. The process curve has been recorded with activated control and continuously changing set point for oxygen partial
pressure. With increasing charge voltage the deviations of the repetition frequency for a fixed pulse time occurred. Nevertheless, any
set point of the hysteresis curve could be stabilized (39).
Wallendorf et al. showed for reactive ZnOx deposition that for the coating system used, a hysteresis occurred. Furthermore, using
an active feedback system, every point in the hysteresis could be stabilized by changing the average power using the off-time
modulation (47). Figure 24 shows two hysteresis curves for fixed oxygen flow of 2 and 3 sccm in dependence of the average power.
Beneath the different stabilized working points the average off-time used between the pulses to maintain the oxygen partial pressure
of the process is indicated.
4.04.4 HIPIMS Applications
From the seminal work of Kouznetsov (31) significant process development was made in HIPIMS technology. Today, several
industrial applications or applications close to industrialization are available. This section highlights different available solutions
www.iran-mavad.com
Figure 22 (a) Setup for measurement of influence of magnetic coils for measuring material flux. (b) Deposition rate as function of applied field
to magnetic coils ((a), ii). Reproduced from Bohlmark, J.; Östbye, M.; Lattemann, M.; Ljuncrantz, H.; Rosell, T.; Helmersson, U. Thin Solid Films 2006,
515, 1928.
for improved coating pretreatment and realization of significantly improved coatings as well as unique solutions that would not be
possible without HIPIMS.
4.04.4.1 HIPIMS Pretreatment

One of the fundamental issues of thin film coatings is the adhesion of the coating for nearly all applications. Conventionally, Ar
glow discharge etching is used. Depending on the application, problems arise with the desired adhesion strength. The improvement
of the adhesion is realized by, e.g., cathodic arc discharge pretreatments. The drawback of this method is additional droplet
formation from the arc source at the interface. HIPIMS offers the chance to use the benefit from arc etching, i.e., the use of target ions
for bombardment without the disadvantage of droplet formation. Ehiasarian et al. compared the adhesion of CrN/NbN coatings on
HSS substrates using Ar plasma etching, arc discharge pretreatment, and HIPIMS etching (Figure 25) (20). The adhesion was
quantified using the critical load in a scratch test for the different methods. While Ar plasma etching showed poor adhesion,
a significant improvement was realized using arc discharges. The best adhesion was realized in the case of HIPIMS etching.
One reason for the excellent adhesion in the case of HIPIMS is the absence of droplets and macroparticles compared to arc.
Furthermore, Ehiasarian et al. showed in earlier works that using HIPIMS etching creates a metal implantation zone of several
nanometers on the substrate surface (18). Scanning transmission electron microscope cross-sections showed that the interface
between substrate and film was clean without additional phases except the film and substrate ones. Compared to conventional
www.iran-mavad.com
Figure 23 Process characteristics for reactive HIPIMS of Al-doped zinc oxide at different charge voltages (800 V, 1000 V, 1200 V). Reproduced from
Sittinger, V.; Ruske, F.; Werner, W.; Jacobs, C.; Szyszka, B.; Christie, D. J. Thin Solid Films 2008, 516, 5847.
sputtering, the grains were larger for HIPIMS. This indicates a high adatom mobility. Furthermore, no intercolumnar voids were
observed. Figure 26 shows the scanning transmission electron microscopy – energy dispersive spectrometry (STEM-EDS) analysis of
the chemical composition of the coating-substrate interface after etching and CrN deposition using HIPIMS (18).
4.04.4.2 Coating Systems

The main focus of the development in HIPIMS technology was the realization of coatings with completely new or at least significant
improved properties. The seminal paper from Kouznetsov (31) highlighted the potential for trench filling and via contacts by
HIPIMS due to the high ionization. Several publications on the topic of trench filling followed. A large number of publications focus
Figure 24 Hysteresis curve with stabilized working points in the transition regime. Reproduced from Wallendorf, T.; Vergöhl, M.; Werner, O.; Bandorf, R.
51st SVC Annual TechCon Proc.; 2008, 3.
www.iran-mavad.com
Figure 25 Adhesion of CrN/NbN coatings on HSS substrate pretreated by Ar glow discharge, cathodic arc, and HIPIMS discharges. Reproduced from
Ehiasarian, A. P.; Wen, J. G.; Petrov, I. J. Appl. Phys. 2007, 101, 054301.
Figure 26 STEM-EDS analysis of the chemical composition at the coating–substrate interface after pretreatment and CrN coating deposition by
HIPIMS. Repoduced from Ehiasarian, A. P.; Münz, W.-D.; Hultmann, L.; Helmesson, U.; Petrov, I. Surf. Coat. Technol. 2003, 163–164, 267.
on hard coatings. Also papers on oxides and reactive processes with and without feedback control are published. The applications
span a wide range of areas from tools, microelectronics, photovoltaic, and optics. A small number of examples follow in this
subsection.
4.04.4.2.1 Hard Coatings

For hard coatings, nitrides are very common. Utilizing HIPIMS, there are a significant number of publications on the synthesis and
properties of nitrides.
Ehiasarian et al. reported on the synthesis of CrN films using HIPIMS. The prepared smooth films showed a low sliding friction
coefficient of 0.4 against Al2O3. This low value was attributed to the absence of droplets. The hardness was reported as being higher
than conventional physical vapor deposition (PVD) CrN films (UBM, cathodic arc) (18).
Bobzin et al. reported a combined DC-HIPIMS process for deposition of Ti–Al–N films. A box coater with multicathodes was
used. Some of the cathodes were running in conventional sputtering mode, while others were operated in HIPIMS mode. For cutting
inserts it was shown that the deposition rates for both flank and rake side, being parallel and perpendicular to the target surface,
respectively were similar. Especially the typically low deposition rate at the rake face (perpendicular to the target) was significantly
improved. Furthermore, for both rake and flank face, the film growth was perpendicular to the surface (Figure 27) (10).
Besides nitrides, diamond-like carbon films (DLC) are widely used in tribological applications. Improvement by using ionized
deposition is also expected. On the other hand, carbon presents some challenges, since the deposition rate and the ionization
probability are quite low (1,3). There are some publications on using HIPIMS for deposition of carbon coatings, however. DeKoven
et al. reported on using HIPIMS for carbon deposition from a graphite target. The density of the films was with 2.7 g cm3,
significantly higher than with other processes. The reported hardness of the films was 7 GPa, which is surprisingly low (17).
In our experiments, the focus was on synthesizing DLC films with hardness as high as possible. Therefore a graphite target was
sputtered using HIPIMS in an Ar/C2H2 atmosphere. The resulting maximum plastic hardness exceeded 60 GPa. The corresponding
www.iran-mavad.com
Figure 27 Cross-sectional SEM images of HIPIMS Ti–Al–N coating deposited on the flank and the rake side of a cutting insert. Reproduced from
Bobzin, K.; Bagcivan, N.; Immich, P.; Bolz, S.; Alami, J.; Cremer, R. J. Mater. Process. Technol. 2009, 209, 165.
indentation hardness was higher than 40 GPa. Figure 28 shows the cross-section of an HIPIMS DLC film with 53 GPa plastic
hardness.
4.04.4.2.2 Electrical Coatings

In microelectronics, vertical interconnects in chip stacks are realized by through-silica via (TSV) as an alternative to wire bonding.
The deposition of the seed layer is realized by PVD technology. Weichart et al. reported on improvement of the sidewall and bottom
coverage using HIPIMS. Aspect ratios (ARs) of up to 30:1 with vertical sidewalls were coated. For Ti bottom coverage of 20% for
trenches with AR of 10:1 and still 7% for AR of 30:1 is reported (48). Focused ion beam scanning electron microscope (FIB SEM)
pictures of Ti/Cu layers in a via with AR 10:1 are shown in Figure 29.
HIPIMS offers an efficient technology to coat trenches and vias with good bottom coverage and acceptable rate. Comparing
HIPIMS with conventionally used post-ionization by an RF coil showed in HIPIMS a twofold higher degree of ionization that was
Figure 28 SEM cross-section of a DLC film using HIPIMS in Ar/C2H2 (5%), Hpl: 53.0 GPa, indentation hardness: 35.7 GPa, 3376 HV.
www.iran-mavad.com
constant over the target substrate distance. For the RF coils, a significant ionization is found only for distances more than 10 cm from
the target (35). Therefore, HIPIMS allows for smaller target substrate distances and higher ionization for improving the coating
properties for vias and trenches.
4.04.4.2.3 Insulating Layers

Sproul et al. were the first to report on the reactive HIPIMS deposition of dielectric films. The target peak power reached up to 3 MW.
Two different settings of the cathode were investigated. In both cases some arcing occurred, but the power supply was able to handle
the arcing. For an equivalent average power deposition rates were lower in the HIPIMS case. This lower rate is attributed to back
attraction of ionized target material to the target. Both DC and HIPIMS processes showed similar hysteresis curves (42).
For a novel approach using modulated pulse power, extremely high deposition rates for alumina films of 12 mm h1 were
reported (14). The films showed a dense and glassy structure. Therefore, it seemed very attractive to use these films as insulating
layers for electrical applications. The reported high rates were due to a very specific sputtering configuration. For reactive sputtering
without feedback control, it was reported that very high breakdown voltages, i.e., excellent insulating properties, can be achieved
(25). A comparison of reactive MF-sputtered films and HIPIMS films showed that the HIPIMS films exhibit a significantly higher
(a)
(b)
Figure 29 FIB SEM pictures of Ti/Cu layers in a via with AR 10:1: (a) upper sidewall; (b) lower sidewall. Reproduced from Weichart, J.; Elghazzali, M.;
Kadlec, S.; Ehiasarian, A. P. 52nd SVC Annual TechCon Proc.; 2009, 201.
www.iran-mavad.com
(c)
Figure 29 (continued).
breakdown voltage at reduced coating thicknesses. Figure 30 shows the breakdown voltage as function of the coating thickness for
HIPIMS and MF-sputtered films. For the determination of the breakdown voltage, a drop of saturated sodium chloride was applied
to the surface and a voltage was applied. The leakage current was measured while the voltage was ramped up. When the leakage
current exceeded 0.5 mA, the actual voltage was defined as breakdown voltage for that film.
4.04.4.2.4 Transparent Conductive Oxides

Indium tin oxide (ITO) is the most prominent transparent conductive oxide today. Nevertheless, efforts are made to replace ITO
with aluminum-doped zinc oxide (AZO). Both materials can be synthesized by HIPIMS with unique or superior properties to state
of the art.
4.04.4.2.4.1 Indium–Tin-Oxide
Sittinger et al. showed that using HIPIMS for the ITO deposition will change the resulting morphology and texture. It is reported that
the spike formation can be adjusted by charge voltage and oxygen flow (39). Horstmann et al. focused on investigating a HIPIMS-
ITO process for industrial production. Therefore, the requirements of the coat and bend process should be fulfilled. The focus was
the application of low emissivity coating based on ITO for exterior use, e.g., curved architectural or automotive glazing (30). The
deposition rate of the HIPIMS-ITO process was 5.5 nm*m min1*kW. Figure 31 shows the results of the taber test (DIN standard
52347) and the sand trickling test (DIN standard 52348). Both tests are commonly used for investigation of mechanical durability.
Figure 30 Comparison of breakdown voltage for reactive MF-sputtered and HIPIMS films as function of coating thickness.
www.iran-mavad.com
Figure 31 HIPIMS ITO films: (a) performance in the taber test: lowest wear – lowest haze for HIPIMS-ITO and (b) results of impact resistance using
sand trickling test: lowest haze for HIPIMS-ITO. Reproduced from Horstmann, F.; Sittinger, V.; Szyszka, B. Thin Solid Films 2009, 517, 3178.
Figure 31(a) shows the haze as function of the number of rotations of the different test samples. The haze of the HIPIMS ITO is even
after 1000 cycles lower than the haze of uncoated glass. This means that there is a better wear resistance in this case. Figure 31(b)
confirms the results of the taber test with the sand trickling test. Here the haze is plotted as function of the mass of the falling sand.
With increasing mass of falling sand the haze linearly increases for both coatings, DC and HIPIMS. Again the HIPIMS film shows the
lowest haze (30).
Future applications of this process are passive ice-free windows for cars or glass bodies for homogenous heating of liquids.
Figure 32(a) shows the direct comparison of a conventional car window and a window coated with HIPIMS-ITO. Figure 32(b)
shows a coated distillation column. Using HIPIMS-ITO visual monitoring without optical distortions due to heating tapes becomes
possible. The right-hand side of this figure shows an IR image of the homogeneous heating area without the formation of hotspots
at the glass tube (41).
4.04.4.2.4.2 Aluminum-Doped Zinc Oxide

Aluminum-doped zinc oxide (AZO) was investigated for replacing ITO, e.g., in solar cells, due to the concern of indium shortage
and increasing price. For HIPIMS deposition of AZO, a reactive process from a metallic target was used. The deposition
temperature was 200 C. Sittinger et al. showed that HIPIMS can produce films with lower resistivity compared to DC sputtering.
The best film was deposited at a charge voltage of 1100 V and reached a minimum resistivity of 536 mU cm. With DC sputtering,
a resistivity of 758 mU cm was reached for optimized oxygen partial pressure. The visual transmittance of both DC and HIPIMS
was comparable (39).
Using the improved coatings in mini-modules, different processes were compared regarding their efficiency. Figure 33 shows the
efficiency of mini-modules from ZSW Stuttgart. Beside the good damp heat stability, also the efficiency for the HIPIMS sputtered
ZnO:Al shows the best performance compared to the other deposition techniques (40).
www.iran-mavad.com
(a)
(b)
Figure 32 (a) Ice-free effect on a windshield in a test bed at Fraunhofer IST and (b) heatable distillation column. Reproduced from Sittinger, V.;
Horstmann, F.; Boentoro, W.; Werner, W.; Szyszka, B. Proceedings of 1st Engineered Transparency, International Conference on Glasstec; 2010, 565–574.
4.04.4.2.5 Antibacterial Coatings

Also in the field of antibacterial behavior significant improvement by the use of HIPIMS is reported. Baghriche et al. report on the
deposition of silver coatings on polyester substrates. The inactivation of Escherichia coli was investigated for both DC and HIPIMS
coatings (4,5). Using HIPIMS the inactivation of E. coli was significantly improved.
Figure 34 shows the inactivation time of E. coli as function of the nominal coating thickness for both processes, DC and HIPIMS.
The comparison shows that DC sputtering saturates at an inactivation time of 5 h for films of 100 nm and more. HIPIMS reaches an
inactivation time of 2 h for coatings of more than approximately 50 nm (5).
Figure 33 The efficiency of the mini-modules made at ZSW Stuttgart. Beside the good damp heat stability also the efficiency for the HIPIMS
sputtered ZnO:Al shows the best performance compared to the other deposition techniques. Reproduced from Sittinger, V.; Szyszka, B.; Bandorf, R.;
Vergoehl, M.; Pflug, A.; Christie, D.J.; Ruske, F. 51st SVC TechCon Proceedings; 2008, 35.
www.iran-mavad.com
Figure 34 Inactivation time of E. coli on Ag-coated polyester substrates as function of the coating thickness; comparison of DC and HIPIMS.
Reproduced from Baghriche, O.; Zertal, A.; Ehiasarian, A. P.; Sanjines, R.; Pulgarin, C.; Kusiak-Nejman, E.; Morawski, A. W.; Kiwi, J. Thin Solid Films
2012, 520, 3567.
4.04.5 Patent Situation
HIPIMS has matured in the last decade from academic research to industrial production. The new technology offers completely new
coatings or coatings with significant improved quality for a very broad range of applications. Since HIPIMS can be applied to
existing coating machines by changing the power supply and adjusting the cathodes, retrofitting of the existing resources offers a fast
technology change.
The growing number of filed patents, both on processes and equipment, documents the relevance and potential of the new
technology for industrial applications. A recent patent study by Fraunhofer IST gives an overview on the legal situation (24). The
filed and granted patents furthermore point the direction toward future development of the technology as well as expected
applications with high impact. The study focuses on patent families, i.e., a collective of claims filed in only one country or even
worldwide.
All in all, 71 patent families (by July 2012) were classified as HIPIMS patent family applications. These patent applications were
filed by more than 30 different companies and institutions. The most designated states are the United States and Germany; 60% of
all granted patents have legal rights in these two countries. The patent families were subdivided into patent families on methods and
equipment and patent families on coatings and applications (Figure 35).
The patents on coatings and applications are mostly granted in Europe. The most patent activities are found for transparent
coating systems, hard coatings, and tool coatings. In the case of the patents on methods and equipment, most patent activities are
found in the field of methods for coating and pretreating, plasma generation, and electronic devices. The patents on methods and
equipment are mostly granted in the United States.
Figure 35 Classification of today’s HIPIMS patent families filed/granted in the field of method and equipment/coatings and applications. Reproduced
from Fraunhofer IST, HIPIMS Patent Study, Braunschweig, DE, 2012.
www.iran-mavad.com
4.04.6 Outlook
Starting from the late 1990s, HIPIMS has matured to a promising and attractive modification of magnetron sputtering for the
synthesis of coatings with significantly improved or completely new properties. Still, there are a number of challenges remaining for
this relatively new technology.
The theoretical description of HIPIMS processes is still in early stages, and today a transition from simple modeling of exper-
imental results to the generation of codes for the prediction of process and plasma properties is taking place. In combination with
expertise from plasma physicists from the field of fusion or space plasma, further understanding and a better theoretical description
as well as new approached for modeling are expected.
Basing on the promising results reported from academia, a scaling up toward industrial-size coating plants is required. Today, the
number of publications for results achieved with industrial-size machines is increasing. Additionally, with the scaling of the cathode
size, to some extent a scaling of the existing power supplies is required. On the other hand, there are already results reported on
cathodes of 1.5 m length and even longer showing the applicability of HIPIMS processes on a large scale.
Regarding commercial processes, there are already HIPIMS processes available and in production. Today most reported
processes are in the field of hard coatings. Still the so-called ‘killer application’ is missing to boost the commercialization and
broad-scale introduction of industrial HIPIMS. A number of promising results with the potential for breakthrough are reported
such as HIPIMS etching, which is already commercially used to some extent, reactive HIPIMS processes due to reducing the
hysteresis or even full avoiding it (depending on the working conditions), densification of coatings, realization of ultraclean
coatings, or realization of totally new properties like the scratch-resistant ITO coatings for, e.g., use in ice-free windows. Therefore,
starting with some applications today, HIPIMS will become part of industrial production. The speed of implementation and
acceptance will depend on finding that killer application and on the successful process development on industrial-scale
machines.
References
1. Anders, A. Surf. Coat. Technol. 2004, 183, 301.

2. Anders, A.; Andersson, J.; Horwat, D.; Ehiasarian, A. P. J. Appl. Phys. 2007, 102, 113303.
3. Anders, A.; Andersson, J.; Ehiasarian, A. P. J. Appl. Phys. 2007, 102, 113303.
4. Baghriche, O.; Ehiasarian, A. P.; Kusiak-Nejmana, E.; Pulgarin, C.; Sanjines, R.; Morawski, A. W.; Kiwi, J. J. Photochem. Photobiol. A: Chem. 2012, 227, 11.
5. Baghriche, O.; Zertal, A.; Ehiasarian, A. P.; Sanjines, R.; Pulgarin, C.; Kusiak-Nejman, E.; Morawski, A. W.; Kiwi, J. Thin Solid Films 2012, 520, 3567.
6. Bandorf, R.; Vergöhl, M.; Schiffmann, K.; Wallendorf, T. 49th SVC Annual TechCon Proceedings; 2006, 21.
7. Bandorf, R.; Vergöhl, M.; Giesel, P.; Wallendorf, T.; Mark, G. 50th SVC Annual TechCon Proceedings; 2007, 160.
8. Bandorf, R.; Falkenau, S.; Schmidt, V. 50th SVC Annual TechCon Proceedings; 2007, 477.
9. Bandorf, R.; Falkenau, S.; Schiffmann, K.; Gerdes, H.; Heckmann, U. 51st SVC Annual TechCon Proceedings; 2008, 59.
10. Bobzin, K.; Bagcivan, N.; Immich, P.; Bolz, S.; Alami, J.; Cremer, R. J. Mater. Process. Technol. 2009, 209, 165.
11. Bohlmark, J.; Östbye, M.; Lattemann, M.; Ljuncrantz, H.; Rosell, T.; Helmersson, U. Thin Solid Films 2006, 515, 1928.
12. Bugaev, S. P.; Koval, N. N.; Sochugov, N. S.; Zakharov, A. N. Proceedings of the XVIIth International Symposium on Discharges and Electrical Insulation in Vacuum;
Berkeley; CA, USA, July 21–26, 1996; p 1074.
13. Burcalova, K.; Hecimovic, A.; Ehiasarian, A. P. J. Phys. D: Appl. Phys. 2008, 41, 115306.
14. Chistyakov, R.; Abraham, B. 49th SVC Annual TechCon Proceedings; 2006, 88.
15. Christie, D. J.; Tomasel, F.; Sproul, W. D.; Carter, D. C. J. Vac. Sci. Technol. A 2004, 22 (4), 1415.
16. Christie, D. J. Vac. Sci. Technol. A 2005, 23, 330.
17. DeKoeven, B. M.; Ward, P. R.; Weiss, R. E.; Scholl, R. A.; Sproul, W. D.; Tomasel, F.; Anders, A. 46th SVC Annual TechCon Proceedings; 2003, 158.
18. Ehiasarian, A. P.; Münz, W.-D.; Hultmann, L.; Helmesson, U.; Petrov, I. Surf. Coat. Technol. 2003, 163–164, 267.
19. Ehiasarian, A. P.; Hovsepian, P. E.; Hultmann, L.; Herlmersson, U. Thin Solid Films 2004, 457, 270.
20. Ehiasarian, A. P.; Wen, J. G.; Petrov, I. J. Appl. Phys. 2007, 101, 054301.
21. Ehiasarian, A. P.; Vetushka, A. 52nd SVC Annual TechCon Proceedings; 2009, 265.
22. Ehiasarian, A. P.; Andersson, J.; Anders, A. J. Phys. D: Appl. Phys. 2010, 43, 275204.
23. Emmerlich, J.; Mráz, S.; Snyders, R.; Jiang, K.; Schneider, J. M. Vacuum 2008, 82, 867.
24. Fraunhofer IST, HIPIMS Patent Study, Braunschweig, DE, 2012.
25. Gerdes, H.; Bandorf, R.; Loch, D.; Bräuer, G. IOP Conf. Ser.: Mater. Sci. Eng. 2012, 39, 012009.
26. Gudmundsson, J. T.; Alami, J.; Helmersson, U. Appl. Phys. Lett. 2001, 78, 3427.
27. Hecimovic, A.; Ehiasarian, A. P. 52nd SVC Annual TechCon Proceedings; 2009, 240.
28. Helmersson, U.; Lattemann, M.; Bohlmark, J.; Ehiasarian, A. P.; Gudmundsson, J. T. Thin Solid Films 2006, 513, 1.
29. Helmersson, U.; Alami, J.; Eklund, P.; Andersson, J. M.; Lattemann, M.; Wallin, E.; Bohlmark, J.; Persson, P. Thin Solid Films 2007, 515, 3434.
30. Horstmann, F.; Sittinger, V.; Szyszka, B. Thin Solid Films 2009, 517, 3178.
31. Kouznetsov, V.; Macák, K.; Schneider, J. M.; Helmersson, U.; Petrov, I. Surf. Coat. Technol. 1999, 122, 290.
32. Lundin, D. Linköping studies in science and technology; Dissertations, ISBN 978-91-7393-419-0 2010, 1305.
33. Mozgrin, D. V.; Fetisov, I. K.; Khodachenko, G. V. Plasma Phys. Rep. 1995, 21, 400.
34. Paulitsch, J.; Mayrhofer, P. H.; Mitterer, C.; Münz, W.-D.; Schenkel, M. 50th SVC TechCon Proceedings; 2007, 150.
35. de Poucques, L.; Imbert, J.; Boisse-Laporte, C.; Bretagne, J.; Ganciu, M.; Teule-Gay, L.; Touzeau, M. Plasma Sources Sci. Technol. 2006, 15 (4), 661.
36. Sarakinos, K.; Alami, J.; Wuttig, M. J. Phys. D: Appl. Phys. 2007, 40, 2108.
37. Sarakions, K.; Alami, J.; Klever, C.; Wuttig, M. Surf. Coat. Technol. 2008, 202, 5033.
38. Sarakinos, K.; Alami, J.; Konstantinidis, S. Surf. Coat. Technol. 2010, 204, 1661.
39. Sittinger, V.; Ruske, F.; Werner, W.; Jacobs, C.; Szyszka, B.; Christie, D. J. Thin Solid Films 2008, 516, 5847.
40. Sittinger, V.; Szyszka, B.; Bandorf, R.; Vergoehl, M.; Pflug, A.; Christie, D. J.; Ruske, F. 51st SVC TechCon Proceedings; 2008, 35.
www.iran-mavad.com
41. Sittinger, V.; Horstmann, F.; Boentoro, W.; Werner, W.; Szyszka, B. Proceedings of 1st Engineered Transparency, International Conference on Glasstec; 2010,
565–574.
42. Sproul, W. D.; Christie, D. J.; Carter, D. C. 47th SVC TechCon Proceedings; 2004, 96.
43. Vergöhl, M.; Werner, O.; Bruns, S.; Wallendorf, T.; Mark, G. 51st SVC TechCon Proceedings; 2008, 307.
44. Vergöhl, M.; Werner, O.; Bruns, S. Proc. of SPIE 2008; 2008; Vol. 7101, 71010B-1.
45. Vlcek, J.; Kudlacek, P.; Burcalova, K.; Musil, J. J. Vac. Sci. Technol. A 2007, 25 (1), 42.
46. Wallin, E.; Helmersson, U. Thin Solid Films 2008, 516, 6398.
47. Wallendorf, T.; Vergöhl, M.; Werner, O.; Bandorf, R. 51st SVC Annual TechCon Proc.; 2008, 3.
48. Weichart, J.; Elghazzali, M.; Kadlec, S.; Ehiasarian, A. P. 52nd SVC Annual TechCon Proc.; 2009, 201.
www.iran-mavad.com
www.iran-mavad.com
4.05 Atomic Layer Deposition
M Leskelä, J Niinistö, and M Ritala, University of Helsinki, Finland

4.05.2 Characteristic Features of ALD 102
4.05.2.1 Benefits 103
4.05.2.2 Limitations 104
4.05.3 ALD Chemistry 105
4.05.3.1 Metal Precursors 105
4.05.3.1.1 Elements 106
4.05.3.1.2 Metal Halides 106
4.05.3.1.3 Metal Complexes 107
4.05.3.1.4 Organometallics 108
4.05.3.2 Nonmetal Precursors 108
4.05.4 ALD Reactors 109
4.05.4.1 Conventional ALD Reactors 110
4.05.4.1.1 Reaction Chamber 110
4.05.4.1.2 Batch Reactors 111
4.05.4.1.3 Reaction Chamber Wall Temperature 111
4.05.4.1.4 Precursor Sources and Valving Systems 111
4.05.4.2 Plasma-Enhanced ALD Reactors 112
4.05.4.3 Spatial ALD Reactors 114
4.05.4.4 Other Reactor Configurations 116
4.05.5 ALD Applications 117
4.05.5.1 Industrial Applications 117
4.05.5.1.1 Magnetic Heads 118
4.05.5.1.2 Microelectronics 118
4.05.5.1.3 Protective Coatings 119
4.05.5.2 Other Applications 119
4.05.5.2.1 Energy Technologies 119
4.05.5.2.2 Optical Applications 119
4.05.5.2.3 Coatings on Powders and Polymers 120
4.05.5.2.4 Photocatalysts 120
4.05.5.2.5 Micro-electro-mechanical Systems 120
4.05.5.2.6 Nanotechnology 120
References 120
4.05.1 Introduction
Atomic layer deposition (ALD) is a technique for growing thin films for a wide range of applications. ALD is a special variant
of the chemical vapor deposition (CVD) technique where gaseous reactants (precursors) are introduced into the reaction
chamber for forming the desired material via chemical surface reactions. A characteristic feature of ALD is that the precursors
are pulsed alternately, one at a time, and separated by inert gas purging in order to avoid gas phase reactions (Figure 1). The
successive, self-terminated surface reactions of the reactants enable controlled growth of the desired material. The unique self-
limiting growth mechanism results in perfect conformality and thickness uniformity of the film even on complicated 3D
structures.
Thanks to its superior conformality, uniformity, and atomic level control, ALD has made a breakthrough in various applications
in modern technology. Especially in the microelectronics industry, ALD has become a mainstream technology enabling the
continuous shrinking of semiconductor devices and adoption of increasingly demanding high-aspect-ratio structures. Other
emerging areas of wide interest toward the ALD technology include photovoltaics, nanotechnology, photonics, and biotechnology
as well as energy-related applications (1).
On the one hand, a key issue in successful application of ALD is continuous development of the precursor chemistry. On the
other hand, ALD technology is also strongly dependent on the reactor design. These two factors are emphasized in this chapter. In
addition, basic features of ALD as well as application areas are described.

www.iran-mavad.com
102 Atomic Layer Deposition
Figure 1 Schematics of an ALD growth cycle. Reproduced from Säynätjoki, A. Atomic-Layer-Deposited Thin Films for Silicon Nanophotonics; SPIE
Newsroom. http://dx.doi.org/10.1117/2.1201204.004218.
4.05.2 Characteristic Features of ALD
ALD processes and their potential as well as already existing industrial applications have been reviewed numerous times in the past,
and some very extensive reviews have been published too (2–6). ALD is a special variant of the well-known CVD method. However,
the differences between the ALD and CVD processes are obvious. In CVD the precursors react at the same time on the surface or in
the gas phase and precursors may also decompose, whereas in ALD the highly reactive precursors react separately via alternate
saturating surface reactions without self-decomposition. The ALD method is therefore controlled strictly by surface chemistry while
in CVD there are several elementary processes that can affect the growth. In ALD, the growth process proceeds in a cyclic manner and
can be described as follows and is schematically exemplified by Figure 1.
One ALD cycle consists of the following steps:
l pulse of the first gaseous precursor and its chemisorption onto the substrate,
l inert gas purging of the excess of the precursor and formed reaction by-products,
l pulse of the second gaseous precursor and its surface reaction with the adsorbate formed by the first precursor, and
l inert gas purging of the excess of the precursor and formed reaction by-products.
In order to achieve the surface saturative, self-limiting ALD process, the precursor dose needs to be large enough so that all the
available surface sites become occupied (Figure 2). Above the dose required for saturation the growth rate stays constant no matter
how large a dose is supplied. Naturally, the precursor self-decomposition would lead to a CVD-type growth mechanism and thus
Figure 2 Growth per cycle as a function of the pulse time (dose). The inset shows the growth per cycle with different growth modes over deposition
temperature.
www.iran-mavad.com
Atomic Layer Deposition 103
sufficient thermal stability of the precursor is required. Ideally the ALD growth cycle produces one atomic layer of the material.
However, due to steric hindrances or a limited number of reactive surface sites only a fraction of a monolayer is deposited per cycle.
Typical growth rates are less than half a monolayer per cycle.
There is usually a certain temperature range where the ALD process is self-limiting. In this temperature range the growth rate is
either temperature independent or only weakly dependent on the deposition temperature, which is controversial to CVD processes.
It is desirable to have a broad temperature range for the self-limiting ALD growth and as small as possible temperature dependence
within this range because these ensure the reproducibility of the film growth. Especially when a ternary material is to be deposited,
overlapping temperature ranges of the constituent binary processes offer a good starting point for the development of a ternary or
nanolaminate process. The observed growth rates vs. temperature in ALD processes are shown in Figure 2, inset.
The unique growth mode of ALD results in many advantages for film growth and properties as already briefly stated. However, it
is useful to also take a critical look at the limitations of the method. Detailed discussion about the pros and cons is presented in the
following section.
4.05.2.1 Benefits
Conformality and large area uniformity: The self-limiting growth mode of ALD leads to excellent conformality and uniformity of the
grown films. As long as the precursor dose is large enough to saturate all the available surface sites, perfectly conformal films can be
deposited on complex structures, like high-aspect-ratio trenches and nanostructures (Figure 3). Large area uniformity can be easily
achieved; even 1.2 m 0.6 m substrates have been coated with Al2O3 with thickness nonuniformity of 1% (7,8).
Precise thickness control: Straightforward thickness control is obtained by varying the number of deposition cycles. Ultrathin films
with subnanometer thickness as well as thick films of 1 mm and even more can be grown.
Dense and homogeneous films: Films deposited with ALD are generally dense, continuous, homogeneous, and pinhole-free.
Sometimes defects may be detected in the case of ultrathin films, usually caused by problems in nucleation because of a lack of
reactive surface sites on the starting surface.
Composition control, multilayer deposition, and doping: As in the case of thickness control, composition can be easily controlled.
Dopants can be imported by adding cycles of dopant material in between the host material growth cycles. In a similar way,
deposition of nanolaminates and other multilayers is straightforward.
Reproducibility: The self-limiting growth mechanism ensures good reproducibility and straightforward scale-up of the processes.
Especially for industrial applications, these factors make ALD an appealing technique.
Low thermal budget: For many applications, a low deposition temperature is sought after. Growth temperatures in ALD processes
can be considerably lower than in CVD processes. Typical reaction temperature in an ALD process is in the range of 200–350 C.
Figure 3 Examples of conformality of the grown Sb films achieved by ALD (a) cross-section field-emission scanning electron microscope (FESEM)
image of a Sb film deposited on a high-aspect-ratio trench structure. (b)–(d) FESEM images of Sb nanotubes prepared using a porous alumina
membrane as a template. The deposition temperature was 100 C, and the number of ALD cycles was 1000. Pulse/purge times were 2/10 s for SbCl3 and
4/10 s for (Et3Si)3Sb. Et ¼ ethyl. Reproduced from Pore, V., Hatanpää, T., Ritala, M., Leskelä, M. J. Am. Chem. Soc. 2009, 131, 3478. Reprinted with
permission. 2011, American Chemical Society.
www.iran-mavad.com
Lower temperatures as well as higher temperatures are, however, often applied. Plasma-enhanced ALD can lower the growth
temperature of a thermal ALD process substantially.
4.05.2.2 Limitations
Slowness: The most important limiting factor of ALD is the slowness of the method. Precursor pulses and purging in between takes
time, causing the effective deposition rate to be lower than in the case of CVD, for example. Typical deposition rates of ALD
processes are in the range of 40–130 nm h1 (7). The growth rate in an ALD process is affected by both the process chemistry and
reactor configuration. The chemistry defines the thickness increment per cycle and how fast the saturation can be reached, while the
reactor configuration is critical for how rapidly the gas phase composition can be changed from one precursor to the other. With fast
chemistry and single wafer traveling-wave reactors the cycle time can be squeezed even below 0.5 s (e.g., 350 nm h1 for Al2O3).
Batch reactors, on the other hand, can accommodate tens of substrates and thereby increase the throughput significantly (7,9).
Further increase of throughput is looked for with spatial ALD reactors that are under active development (Section 4.05.4).
Limited materials selection: Another limitation often considered in ALD is the materials selection that can be effectively grown by
ALD. Several technologically interesting materials, such as Si, Ge, Cu, multicomponent oxide superconductors, and many transition
metals, are lacking effective and production-worthy processes. However, the list of available processes (5) is continuously
expanding. In addition, with plasma-enhanced ALD (PEALD) the materials selection is likely to be increased with an increasing
rate (10).
Impurities: Impurities from precursor molecules are always incorporated to some extent in ALD films as characteristic of chemical
techniques. The actual residual concentrations are highly process dependent, and their significance is application dependent. Good
ALD processes result in impurity contents below the detection limits of most surface analytical tools, i.e., in the range of 0.1 at.%,
while some other processes may leave a few at.% of impurities, yet the films show reasonably good materials properties.
ALD nonidealities: There are also a number of unwanted characteristics that are more or less frequently observed in ALD processes.
As these are not common to all ALD processes, they should not be considered as limitations of the method itself but as process
specific deviations from the ideal ALD growth. One may define the ideal ALD process to be completely saturating and self-limiting,
depositing an exact monolayer in every cycle over a reasonably wide temperature range, fully preserving the substrate surface
smoothness, and releasing by-products that are readily removed from the reactor.
Usually the growth rate in ALD is not an exact monolayer per cycle even if the process otherwise fully follows the ALD principle.
Still, the film build-up can take place layer by layer; it just takes more than one cycle to complete one monolayer. Therefore the less
than a monolayer per cycle growth rate alone is rarely a problem.
During the very first cycles, ALD film is deposited on the starting surface, which is either a substrate or an underlying film, while
during the rest of the process the growth occurs on the film material itself. If the starting surface differs chemically from the film
surface, the growth rate is not necessarily constant but may change quite substantially as the surface becomes covered by the film.
Usually the growth rate is slower at the beginning, during the nucleation period, and it takes a certain number of cycles before the
linear correlation between the film thickness and the number of ALD cycles is reached. This kind of retarded nucleation has been
observed, for example, in the technologically important application of ALD of high-k oxides on hydrogen terminated silicon from
which the native oxide has been removed.
The deposition rate may also be somewhat temperature dependent. This is due to temperature dependent density of precursor
molecules in the saturated chemisorption layer which, in turn, is affected by either temperature dependent density of reactive sites or
temperature dependent structure of chemisorbed precursor molecules. Anyhow, the temperature dependency of the growth rate is
usually so weak that requirements on temperature control and uniformity of the reaction chamber are not as strict as in CVD.
The growth is often not truly self-limiting because the precursors also thermally self-decompose. The higher the temperature, the
higher the contribution arising from the self-decomposition. In fortunate cases, like when using metal alkoxides for ALD of oxides,
the decomposition product is the same as that to be deposited by ALD. As long as the self-decomposition remains reasonably slow,
it is likely to proceed in a surface reaction rather than a diffusion limited manner, and thus maintain film uniformity and con-
formality reasonably well.
Instead of exactly reproducing the substrate surface morphology ALD growth has often been observed to cause roughening of the
surface at the nanometer level while still showing excellent conformality on larger scales. The roughening appears characteristic of
polycrystalline films, whereas amorphous and epitaxial films usually retain the substrate surface smoothness. The dominant factor
with respect to the surface morphology of amorphous and epitaxial films is the minimization of the surface area, which favors
atomically smooth films because either the surface energy is isotropic (amorphous films) or the crystal orientation is dictated by the
substrate (epitaxial films). With polycrystalline films the competitive crystal growth of the originally more or less randomly oriented
nuclei leads to roughening of the surface as the film grows thicker. The roughening is further emphasized by the retarded nucleation
discussed above: the lower the nucleation density, the rougher the film becomes because the crystallites have a longer time to grow
independently before coalescing to a continuous film (11).
Many reaction by-products are polar molecules and as such show reactivity toward the surface. If the by-products readsorb on the
surface, they can block adsorption sites from the precursor molecules and thereby decrease the growth rate. Furthermore, if the by-
product readsorption does not occur uniformly across the substrate, it can lead to thickness nonuniformity. Therefore it is likely to
have a more severe effect in the cross-flow reactors than in the perpendicular flow reactors (see Section 4.05.4) because in the former,
different substrate areas can experience substantially different exposures to by-products. If the film growth begins only at the vicinity
www.iran-mavad.com
of the leading edge of the substrate (closest to the precursor inlet), this part becomes exposed to only a small amount of by-product
molecules while the trailing edge (farthest away from the precursor inlet) will receive a substantially higher amount. The effect of by-
product adsorption is further emphasized by the fact that in the cross-flow reactors the by-products travel in front of the precursor
pulse that is producing them (2). By contrast, in properly designed perpendicular flow reactors, the film growth begins everywhere at
about the same time, and hence the effect of by-product readsorption should be more uniform too.
In the most severe cases reactive by-products can etch the film material or the underlying material. Another possible reason for
film nonuniformity is slow etching of the film material by the precursor molecules themselves. In the worst case the etching may
fully prevent the film growth. It is also possible that the precursor molecule reacts directly with the underlying material. For
a thorough discussion of factors affecting film uniformity, see Ref. (12).
4.05.3 ALD Chemistry
As a chemical vapor deposition technique, ALD relies on the successful precursor chemistry. Table 1 summarizes the requirements
set to the ALD precursors.
Volatile precursors with sufficient vapor pressure are needed – a common estimate for the required volatility is 0.1–1 Torr at
temperatures that can be reached with the sources in use. ALD has special requirements for the precursors compared to CVD since in
ALD the precursors are introduced one at the time onto the substrate surface. Besides volatility, reactivity and thermal stability are
crucial prerequisites for ALD precursors. Reactivity means that the incoming precursors should react fast and completely with the
species left on the surface by the other precursor; the higher the reactivity, the faster the surface is saturated and the shorter the ALD
cycle can be made. Complete reaction is also essential for film purity. Decomposition of the precursor is not allowed because then
the surface control is lost, and consequently homogeneity and conformality – the important properties of ALD films – cannot be
achieved. Unfortunately, reactivity and thermal stability are often controversial properties and often there is a need to balance
between these properties.
There are several other properties that are desirable for ALD precursors, such as stability in storage and in the heated source,
low toxicity, low price, and ease of handling. Precursors can be solids, liquids, or gases. In CVD solid precursors are problematic
since the evaporation rate of solids is often not stable, resulting in nonstable precursor fluxes into the CVD reactor. Due to the
self-limiting film growth mechanism, the precursor flux stability is not as important in ALD as in CVD but still, especially in
industrial use, liquid and gaseous precursors are desired. Therefore one driving force in ALD precursor development is searching
for precursors that are liquid at the source temperature. The small size of the precursor molecule is one benefit, since it enables
higher growth rate.
Low reactivity of the precursors can be enhanced by using plasma assistance. PEALD is useful also when temperature sensitive
substrates are employed since with plasma assistance the deposition temperatures can be significantly lowered compared to thermal
ALD. Plasma assistance also gives high growth rate and film purity at these low deposition temperatures. The use of plasma has
expanded the ALD process and thin film material selection to non-noble metal and certain nitride films (10).
4.05.3.1 Metal Precursors

According to recent reviews more than 300 metal precursors have been studied in ALD to deposit films of about 160 different
compositions. The number of processes studied is rather high, but many of the processes do not work very well and suffer from the
precursor problems described above: thermal unstability and reactivity.
The volatile metal precursors can be divided in four categories: elements, inorganic compounds, complex compounds, and
organometallics (Figure 4). In addition, a recent trend in tuning the precursor properties has involved heteroleptic complexes,
i.e., complexes that contain two or more different ligands and therefore do not directly fall into these main categories but will be
viewed as their modifications. Interestingly, heteroleptic compounds show different properties than their homoleptic counter-
parts. In the heteroleptic complexes one type of ligand can provide stability and the other reactivity. One good example is
heteroleptic Zr(Cp)(NMe2)3, where the cyclopentadienyl ligand gives thermal stability and alkylamides reactivity (13). The
homoleptic alkylamide complexes are thermally quite unstable; the decomposition temperature of Zr(Cp)(NMe2)3 is around
325 C while that of Zr(NMe2)4 is below 300 C. In some cases the heteroleptic compounds can show good properties although
Table 1 Summary of ALD precursor requirements
Essential properties Desirable but not necessary
Volatility Unreactive volatile byproducts

Fast and complete reactions Inexpensive
No self-decomposition Easy to synthesize and handle
No etching of the film or substrate material Nontoxic and environmentally friendly
No dissolution into the film substrate
Sufficient purity
www.iran-mavad.com
Figure 4 Examples of ALD precursor types.
the parent homoleptic compounds are both thermally unstable. Two examples from this are Ti(NiPr-amd)(iOPr)3 (NiPr-
amd¼di-N,N-isopropylacetamidinate) and Ti(NMe2)2(iOPr)2, which show ALD-type growth characteristics at 375 and 325 C,
respectively, while the homoleptic counterparts already show decomposition at 300 C (14).
4.05.3.1.1 Elements
The only metallic elements that have been extensively explored in ALD are Zn and Cd, which have been used for depositing II–VI
semiconductors (15). The use of other metallic elements, apart from Hg, is limited by their low volatility. Zn and Cd react efficiently
with elemental chalcogens, and the lack of sterically demanding ligands may enable a full one monolayer per cycle growth as
exemplified by some II–VI compounds on single crystal substrates (15–17).
4.05.3.1.2 Metal Halides

Metal halides are stable molecules; many of them are volatile and highly reactive, making them almost ideal for ALD. Good oxide
processes with water as an oxygen source have been reported for many chlorides such as AlCl3, TiCl4, ZrCl4, and HfCl4. Processes
using TaCl5, SnCl4, and InCl3 have also been reported, but the growth rate has been relatively low. The success in the use of
nonmetal chlorides (BCl3, PCl3, SiCl4, GeCl4) with water for oxide processes has been poor. Some metals have volatile oxo-
chlorides (VOCl3, WOCl4) that are usable precursors for ALD of oxides. Volatile metal chlorides are good precursors to be used
with H2S and alkyl silyl selenium and tellurium precursors. From sulfides, the most well-known example is ZnS and its doped
form ZnS:Mn, which is used in electroluminescent flat-panel displays as a yellow-emitting layer (15,18,19). Another example is
In2S3, a potential material for thin film solar cells, which can be made from chlorides and hydrogen sulfide (20). GeTe, Sb2Te3,
and Ge2Sb2Te5 represent chalcogenides made from metal chlorides and alkyl silyl chalcogenides (21,22). Metal chlorides (AlCl3,
GaCl3, TiCl4, NbCl5, TaCl5, MoCl5) react with ammonia and form nitrides (23,24). Due to the lower reactivity of ammonia
compared to water, higher temperatures have to be used in nitride ALD processes, but despite that the chloride residues in nitride
films tend to be higher than those in oxide films. In transition metal nitrides the desired conductive nitride has a 1:1 stoichi-
ometry, which requires reduction of the þ4 and þ5 metal ions to the þ3 oxidation state. Several reducing agents (zinc, dimethyl
hydrazine, allyl amine) have been studied (25). Trimethylaluminum (TMA) serves as an interesting reducing agent but leaves
aluminum and carbon residues. Anyhow, for example in the case of Ta(Al)N(C), the material properties are good enough for
conducting barrier use (26).
Metal chlorides are criticized as ALD precursors because of the formation of HCl as a by-product. HCl may cause problems in the
form of corrosion. It potentially can also readsorb and react with the surface and block the surface sites, causing growth rate decrease
(27). Chlorine residues in the film may affect the electrical properties and may make the film unstable. If deposition temperatures
are high enough, the chlorine contents of the oxide films remain well below 1 at.%, however, and the weakening of the film’s
properties remains minimal. Due to their many positive properties (price, stability, reactivity) chlorides are used in ALD in
large scale.
Many metal bromides are also volatile. Benefits in their applications in ALD compared to chlorides are limited. They are usually
less stable, more expensive, and larger molecules than chlorides. Metal iodides, on the other hand, have been studied in ALD more
than the bromides. Deposition of oxide films has been reported for TiI4, ZrI4, HfI5, NbI5, TaI5, CoI2, SnI4 (28), nitride for TiI4, and
sulfide for ZnI2 and PbI2. In oxide processes both water, hydrogen peroxide and molecular oxygen have been employed as the
oxygen precursor. The use of O2 is interesting but requires high deposition temperatures indicating partial decomposition of the
metal iodide.
A few volatile transition metal fluorides exist for group 4–6 metals. From those WF6 is by far the most studied. It can serve as
a precursor in metal, nitride and nitride carbide ALD. The reduction of WF6 to metal by silanes and the application of metal films
have been thoroughly studied (29,30). Reports on the use of metal fluorides in ALD oxide processes are scarce. TaF5 has been used
www.iran-mavad.com
together with water for the ALD of Ta2O5 (31), but for example tungsten hexafluoride cannot be used for oxide deposition because
of an etching problem. Instead, oxyfluoride WOF4 has been utilized successfully in oxide ALD. On the other hand, transition metal
fluorides have been used as innovative fluoride precursors in ALD of MgF2, CaF2, YF3, and LaF3 thin film (32).
4.05.3.1.3 Metal Complexes

The most important volatile metal complex types are alkoxides, b-diketonates, alkylamides, alkyl silyl amides, and amidinates
(guanidines). By number the largest group is formed by b-diketonates, acetylacetones (acac) and 2,2,6,6-tetramethyl-
3,5-heptanedione (thd) complexes being the most common. The chelate structure formed by the two oxygen atoms from each
ligand coordinated to the metal stabilizes the metal b-diketonates. The acac ligand is suitable only for small metal ions while the thd
complexes studied in ALD expand from Li to alkaline earth metals, rare earths, several transition metals, lead, and bismuth (4,5).
The metal b-diketonates have been used in deposition of oxide, fluoride, sulfide, and metal films. The advantages of the b-diket-
onates are simple synthesis, stability, and modest price. Some of them are air stable, and the rest can be exposed for a short time to
air. The disadvantages are oligomeric structure in case of large cations (alkaline earth and rare earth metals), which may limit the
volatility (33), and low reactivity toward water. The reactivity and thermal stability have been balanced with heteroleptic complexes
such as Ti(OiPr)2(thd)2, which has allowed the use of water as an oxygen source at 370 C in deposition of SrTiO3 films (34). If the
temperatures needed for reactions with water exceed the decomposition temperature of the thd-complex, the oxygen precursor has
to be changed to ozone or oxygen plasma. Those efficient oxidation agents may however decompose the b-diketonate ligand to
carbonate, and for example in the case of alkaline earth and rare earth metals, the resulting film is not oxide but carbonate (35).
Although the reaction between thd complexes and water is not very favorable, the reaction with hydrogen sulfide to sulfide films is
efficient. Many metal sulfides, but in particular alkaline earth sulfides, which have been studied as potential host materials for thin
film electroluminescent displays, have been deposited by this process (36). The thd complexes are also good precursors for metal
fluoride films in reaction with volatile transition metal fluorides (32). In ALD of noble metal and noble metal oxide films,
b-diketonates are possible alternative precursors (37). The deposition process in formation of noble metal films with O2 is based on
oxidative decomposition of the metal precursor where oxygen atoms activated by the noble metal film itself combust the ligands
and the mechanism is quite independent of the ligand-type (38). When ozone is used instead of O2, the resulting film is noble metal
oxide (39).
Metal alkoxides are less important ALD than CVD precursors because of their limited thermal stability. The lower stability of
alkoxides compared to b-diketonates originates from the single metal–oxygen bond. In CVD of oxide films the decomposition is
beneficial but in ALD an unwanted side-process. Several well behaving alkoxide based ALD oxide water processes are known,
however, such as Al2O3 using ethoxide or propoxide, TiO2 using methoxide, ethoxide, or iso-propoxide, and Nb2O5 and Ta2O5
using ethoxides (5). Besides the simple alkoxo ligands branched ligands such as tert-pentoxo, 2,3-dimethyl-2-butoxo (dmp), and
2-methyl-3-pentoxo (mp) have been used. Stabilization of alkoxides has been studied by adding donor functional group such as
keto (1-methoxy-2-methyl-2-propoxo, mmp) and amino groups (dimethylaminoethoxo, dmae) in order to produce a chelate
complex. A lot of studies have been carried out with these functionalized alkoxides to deposit Zr, Hf and rare earth oxide films, and
results are better than those obtained with simple alkoxo complexes (40,41). One of the few successful nickel oxide processes
utilizes functionalized alkoxide (1-dimethylamino-2-methyl-2-propanole) where a dimethylamino group has been added to the
end of the alkyl chain of the alkoxo ligand. Water cleaves the ligand and NiO film is formed (42). Precursors containing two metals
are rare in ALD, but one successful example is SrTa2(OEt)10(dmae)2 which has been used in deposition of SrTa2O6 and SrBi2Ta2O9
films (43).
Alkylamides form a large group of potential ALD precursors. Most attention has been given for group 4 metal dimethyl, diethyl,
and methylethyl amides, which show thermal stability up to 250–300 C (44,45). Alkyl amides react with both water and ozone,
and good results have been obtained with ZrO2 using ozone as the oxygen source. Permittivity values around 40 and CET
(capacitance equivalent thickness) of 0.8 nm have been reported for these high-k oxides (46). As mentioned above, use of mixed
ligand complexes, in this case cyclopentadienyl-alkylamide, improves the precursor properties of Zr and Hf when compared to the
homoleptic complexes (13,47,48). Many other metal alkylamides (Al, Ti, Ga, Ge, Sb, Ta, W) have been studied for deposition of
oxides, but thermal stability of the metal precursor has been an issue. Since the alkylamides contain a metal–nitrogen bond they
have attracted interest as precursors for nitride films (AlN, TiN, ZrN, HfN, TaN, MoNx, WNx) (4,5). A substantial amount of studies
have been carried out to deposit TaN films from alkyl amides and ammonia. Thermal stability of the metal precursor and low
reactivity of ammonia below 300 C have been the main limitations, and thermal ALD processes have resulted in amorphous as-
deposited films with plenty of impurities (49). The use of plasma (ammonia, nitrogen or hydrogen) improves the quality of the
nitride films (50).
Alkyl silyl amides form volatile metal complexes that have been studied in ALD of Zr, Hf, Bi, and rare earth oxides. The role of
silicon is to add size of the alkyl amide ligand and thereby protect the metal ion and stabilize the complex. In the case of rare earth
metals and bismuth the results with water are promising (41,51), but thermal stability of the precursors is usually a concern. Li
forms an exceptionally stable complex with HMDS (hexamethyldisilazide), which can be employed even at 400 C (52). When
ozone is used in the oxide process, silicon remains in the film. This has been confirmed in rare earth and bismuth cases, and
LiHMDS forms with ozone pure lithium silicates.
Amidinates have two metal–nitrogen bonds, and they relate to amides analogously to how b-diketonates relate to alkoxides. The
chelate effect should make the amidinates thermally more stable than amides, while the metal–nitrogen bonds should provide
amidinates higher reactivity compared to the b-diketonates with metal–oxygen bonds. Properties of the amidinates can be varied by
www.iran-mavad.com
changing the alkyl groups in the two nitrogen atoms. Too-small alkyl groups may result in oligomeric complexes, and monomeric
complexes are obtained with bulkier groups like t-butyl (53). Amidinates have been used in deposition of many oxides, including
challenging rare earth oxides (53,54). More interestingly, amidinates are the first precursors allowing the deposition of the first series
transition metals (Fe, Co, Ni, Cu) by a simple thermal process with hydrogen (55). In addition, ruthenium films have been grown
from amidinate (56). Closely related to amidinates are the guanidinate anions of the general type [R2NC(NR0 )2] (R ¼ alkyl, SiMe3;
R0 ¼ alkyl, cycloalkyl, aryl, SiMe3) having an additional amino group in the mid-carbon atom. Guanidinates have shown similar
promises as amidinates as ALD precursors (57).
4.05.3.1.4 Organometallics
Alkyl and cyclopentadienyl compounds are the most important organometallic (contain metal–carbon bond) ALD precursors.
Carbonyl, phenyl, 1,5-cyclo-octadienyl, 1,3-cyclohexadienyl, and cycloheptadienyl compounds have been studied only for
isolated metals (4,5). The number of volatile alkyl compounds is limited, but group 12–14 metals form volatile, reactive and
often liquid alkyl compounds very suitable for ALD. Methyl and ethyl compounds are the most often used alkyl compounds.
TMA is by far the most studied and used ALD precursor. It shows an almost ideal ALD reaction with water depositing Al 2O3 over
a wide temperature range (4,58). The TMA–water process is extensively utilized for dielectric and protective ALD coatings in
a controlled manner. Versatility of TMA is shown in its good reactivity with ozone and oxygen radicals to oxide. Alkyl compounds
of the other group 13 metals (Ga, In) are very extensively used in MOCVD of III–V semiconductors, but in ALD their role has been
much smaller. ALD (ALE) of III–V compounds was extensively studied in the late 1980s and early 1990s, but the processes
required temperatures where the decomposition of the alkyl compounds begun resulting in carbon contamination in the films
(59). Diethyl zinc (DEZ) is, after TMA, the second most widely used ALD precursor. It, as well as dimethyl zinc, reacts readily with
water to zinc oxide and with hydrogen sulfide to zinc sulfide (60–62). Polycrystalline films can be grown in a wide temperature
range from room temperature to 500 C (5).
Cyclopentadienyls (Cp, C5H 5 ) form the other important group of organometallic ALD precursors. In Cp compounds the metal
is coordinated to all five carbon atoms in the Cp ring and this gives shielding and stability while still keeping the reactivity. The Cp
rings can be substituted to increase the bulkiness of the ligand, and it is also possible to bridge two rings together to an ansa-
compound. Cp compounds of alkaline earth metals (Mg, Ca, Sr, Ba) are very important because they enable the growth of oxide
films at reasonably low temperatures using water as an oxygen source, which is not possible with the alternative thd-precursors. Cp
compounds of alkaline earth metals are utilized especially for multicomponent oxides SrTiO3 and BaTiO3 (63,64). The suitability of
Cp ligands for electropositive large metal ions is also seen with rare earth compounds. Several rare earth Cp compounds have been
used for deposition of oxide films (65). Cyclopentadienyls of Zr and Hf have played an important role in studies of high-k oxides for
microelectronics. Both Cp2MCl2 and Cp2M(CH3)2 have shown good ALD behavior in reactions with water and ozone (48,65,66).
As mentioned above, heteroleptic compounds containing both Cp and alkyl amide ligands have been even more promising ALD
precursors for ZrO2 and HfO2 (13,47). Cp compounds have also been used in deposition of noble metal films when reacting with
O2 and noble metal oxide films when ozone is used. In the case of ruthenium the simple Cp (RuCp2) or substituted Cp
(Ru(CpEt)2), Et ¼ ethyl, Ru(CpEt2)2, Ru(CpMe)(CpEt) compounds have been employed (67,68) but different heteroleptic
compounds such as Pt(CpMe)Me3 (Me ¼ methyl), Ir(CpEt)(cod) (Et ¼ ethyl, cod ¼ cyclo-octatriene) are also applicable (69). Even
related pyrrole compounds, where one of the carbon atoms in the five-member ring has been replaced by nitrogen, have shown
promising results (70).
4.05.3.2 Nonmetal Precursors

Metal oxides are the materials most often deposited by ALD. Water vapor is the most often employed oxygen precursor. The reaction
relies on surface hydroxyl groups which serve as adsorption sites for the metal precursor. The ligand exchange reaction with water
vapor recovers the OH groups on the surface and releases the ligands from the metal ion in their protonated form without any
fragmentation. Hydrides in general are useful precursors for a nonmetal element. Those studied in ALD include H2S, H2Se, H2Te,
NH3, PH3, AsH3, and HF (4,5). The toxic selenium, tellurium and fluorine hydrides have had only limited use and phosphine and
arsine have been needed in ALE of III–V semiconductors. Instead ammonia has been very widely used for deposition of different
nitride films. In thermal processes the reactivity of ammonia below 300 C is limited, and films may contain high amounts of H, C,
and O contaminants. Alkylation of hydrides makes them less toxic, and several alkylated compounds like tBuNH2, tBuAsH2, and
fully alkylated compounds like (Me2N)3P and (Me2N)3As have been tried but they may leave carbon contaminations. H2S is also
toxic but much less so than H2Se. H2S is widely used in ALD of sulfide films. It reacts effectively with many metal precursor types
(halides, metal complexes, organometallics). The reaction mechanisms in sulfide and nitride ALD processes have been scarcely
studied. It is obvious though that the surface S–H and N–H groups play a vital role. Alkyl silyl compounds (R3Si)nX offer
a convenient alternative for tellurides, selenides (21), and antimony (71), and possibly also some other nonmetals. They are
moderately volatile liquids or solids and thus safer to use than the gaseous and toxic hydrides. Alkyl silyl nonmetals react efficiently
with metal chlorides, releasing (R3Si)nX as a by-product (72), but with other metal precursors the reactivity has been found to be
poor so far.
H2, N2 and O2 molecules are quite inactive in normal ALD conditions. There are a few examples where reduction of metal
precursor to metal films can occur, amidinates being the most well-known case. In growth of noble metal films O2 is used as
a reactant, but there it is catalytically activated by dissociative adsorption to atomic oxygen on the noble metal surface. N2 is inactive
www.iran-mavad.com
at temperatures used in ALD, and nitrogen is commonly used as a carrier and purge gas. While H2, N2 and O2 molecules cannot be
generally used in thermal ALD processes they are common reactant gases in radical- and plasma-enhanced ALD of metals, nitrides
and oxides, respectively.
Ozone has become a popular oxygen precursor, since in some cases avoiding water in oxide processes may be useful. Ozone is
a strong oxidizer and is beneficial when oxidizing reactions are needed. Hydrogen peroxide is another alternative for oxidation.
Ozone is made from O2 and is therefore accompanied by an excess of O2. The difference in oxidation power between O2 and O3 can
be seen in reactions with noble metal precursors where O2 produces metal and O3 metal oxide films. In ALD processes ozone seems
to oxidize the carbon and hydrogen in the ligands to CO, CO2, and H2O, and the water formed may play a role in the film growth
process.
Some solid nonmetal elements can be used as ALD precursors. These include chalcogens (S, Se, Te), which have been used to
deposit II–VI semiconductors (15–17).
4.05.4 ALD Reactors
ALD reactors naturally have many similarities to CVD reactors (Figure 5). The main difference arises from the need to expose the
substrates alternately to the precursor vapors hundreds or even thousands of times in each ALD process while at the same time
keeping the precursors carefully separate from each other. In most ALD reactors this is done by keeping the substrates stationary
Figure 5 Reactor chamber designs in flow-type ALD reactors. On the right, substrate placement in a batch ALD reactor is shown. Reproduced from
Ritala, M., Niinistö, J. Atomic Layer Deposition. In Chemical Vapour Deposition: Precursors and Processes; Jones, A. C., Hitchman, M. L., Eds., Royal
Society of Chemistry, 2009; pp 158–206.
www.iran-mavad.com
and pulsing the precursors alternately into the chamber. The precursor pulses are separated by purging or evacuation steps.
An alternative approach, which was demonstrated already in the very beginning of ALD and gained renewed interest recently, moves
the substrates from one continuous precursor flow to another with a purge gas in between keeping the precursor flows separate. As
opposed to the conventional ALD reactors, the latter are called spatial ALD reactors.
Productivity considerations require completion of an ALD cycle as fast as possible, and this has been a key criterion in designing
especially industrial ALD tools. Some relaxation of the ALD reactor design, temperature uniformity requirements in particular,
comes from the weaker temperature dependence of the ALD processes as compared with CVD processes. The self-limiting growth
mechanism in turn means that in ALD the precursor flows do not need to be as uniformly distributed over the substrates as in CVD.
However, these relaxations apply only to the true ALD processes; if there is any CVD component involved, this must be taken into
account in the reactor design, which thereby approaches CVD reactors.
While most ALD processes are solely thermally activated, additional activation by plasma discharges has been gaining
substantial importance for both lowering the deposition temperatures and for depositing materials that are not achievable with
purely thermally activated chemistry. Many alternatives exist for adding plasma sources to the ALD reactors, leading to a variety of
PEALD reactor designs. This chapter mostly focuses on the conventional ALD reactors, also including plasma-enhanced ALD
reactors, but it also gives an overview of the spatial ALD tools. Detailed discussions of ALD reactors can be found for example in
Refs. (2,9,60,73–76).
4.05.4.1 Conventional ALD Reactors

In the conventional ALD reactors the substrates are fixed in the chamber, and the gas phase composition of the chamber is
repeatedly changed: first precursor/purge or evacuation/second precursor/purge or evacuation. ALD processes have been run over
a wide range of operation pressures from ultrahigh vacuum to atmospheric, and even supercritical fluid has been demonstrated
(77). However, most reactors seem to be of the flow-type, where a continuous flow of inert gas, nitrogen or argon, is used for
precursor transportation and purging, and the pressure is commonly in the range of 1–10 Torr. Purging is usually preferred over
evacuation because it is faster.
Because inert gas is used with high flow rates from 0.5 slm in small research reactors to some tens of slm in the largest production
reactors, the inert gas serves potentially as the largest source of impurities in the process. Reactive residual contents must be at least
on a ppm level, in the most sensitive processes even far below.
4.05.4.1.1 Reaction Chamber

Depending on the number of substrates, ALD reactors are divided to single wafer and batch reactors. The single wafer reactors are
usually of either the cross-flow (also called as a flow-channel or traveling-wave reactor) or perpendicular-flow type (Figure 5). Either
a showerhead or a more simple gas dispersing device is used in the perpendicular-flow reactors to distribute the precursors
uniformly. In the simplest case there may just be precursor inlets on top of the chamber, and the uniformity is ensured simply by the
self-limiting ALD growth mechanism. If the top wall with the inlet port is brought just above the substrate and susceptor (Figure 5),
the reactor becomes converted to a radial flow-channel reactor with the same benefits and concerns as in the normal cross-flow
reactor (see below), but in a radial symmetry.
The cross-flow reactor allows faster cycling and thereby higher throughput than the perpendicular-flow reactor. The faster
cycling is made possible by the geometry of the flow channel between the substrate and the top wall, or between a pair of
substrates as is often the case in the batch reactors. The height of the flow channel can be made very small, down to a millimeter
range, by bringing the opposing surfaces very close to each other. When being transported along such a channel, precursor
molecules make multiple hits with the surfaces. The multiple-hit conditions increase both the probability of a molecule to find
an open adsorption site and the rate with which the substrate surface becomes saturated with the precursor. The former
increases precursor utilization efficiency and the latter makes exposure times shorter as compared to alternative reactor
geometries. The small volume of the flow channel, together with the plug-like flow conditions, also enables very rapid purging
of the reactor. With good chemistry, high enough precursor partial pressures for efficient transportation and high enough
temperature for rapid desorption, 0.1 s pulses can be enough to saturate the surface of a large silicon wafer; the purge may be
completed within 0.1 s, thereby resulting in less than 0.4 s cycle times. With large substrates larger amounts of precursor need to
be transported, but still cycle times below half a second have been reported, e.g., 475 ms for Al2O3 deposition from Al(CH3)3
and O3 onto 300 mm wafers (78). The largest single substrate tool reported is an application-specific system designed for
industrial in-line production of buffer layer coatings on copper-indium-gallium-diselenide thin film solar cells with
1200 mm 1200 mm substrates (7,8).
The disadvantage of the cross-flow reactor is its sensitivity to various kinds of nonidealities in ALD chemistry like precursor self-
decomposition and by-product readsorption (see above), both leading to thickness nonuniformity. These effects are pronounced in
the cross-flow reactors because different parts of the substrates receive different doses of precursors and by-products. Essentially all
nonideal features of ALD chemistry can be described as CVD contributions, and thus it is of no surprise that their harmful effects can
be better tolerated in reactors resembling CVD reactors, i.e., in the perpendicular-flow reactors. The downside of the perpendicular-
flow reactors is slower purging because the showerhead or another kind of injector piece increases the reaction chamber volume to
be purged. The volume to be purged can be minimized by bringing the injector piece closer to the substrate than commonly used in
CVD, though.
www.iran-mavad.com
4.05.4.1.2 Batch Reactors

While single wafer reactors are preferred in certain applications in semiconductor industry, they have limited productivity due to the
moderate deposition rates of ALD. Therefore multiple parallel single wafer ALD reactors integrated to a single tool as well as semi-
batch reactors (e.g., 4 wafers) have been developed. Also large-batch tools with up to 150 wafers with 200 or 300 mm diameter (9)
or 50 wafers with 300 mm diameter (79) are commercially available, but these particular tools require significantly longer cycle
times. In non-semiconductor applications large-batch reactors are common and they can accommodate, e.g., 36 substrates
measuring 24 50 cm2. So that a double side coating results in 8 m2 of film (7). Substrates can be also packed back-to-back so that
the growth surfaces from the adjacent substrate pairs form a flow channel similar to the single wafer cross-low reactors (Figure 5).
Even larger surface areas in the batch tools are possible. In protective coating applications, like those applied on jewelry, quite
massive batch processing is necessary to make the process cost-effective. Furthermore, there is also a tool that integrates four coating
chambers, each taking in 500 crystalline silicon solar cell wafers (156 156 mm2), with a single precursor delivery system (80).
A throughput of 3000 wafers h1 for 20 nm film is specified for this four chamber tool.
4.05.4.1.3 Reaction Chamber Wall Temperature

ALD reactors must be of the hot wall type to exclude any slow adsorption–desorption processes on the walls of the reaction chamber
and the source lines. This is necessary for rapid change of the gas composition in the reaction chamber as needed for an efficient ALD
process. Often the walls are at the same temperature as the substrates and film grows also on the walls, thereby requiring their
frequent cleaning. As an alternative, warm wall configuration has been suggested so that the walls would be at a temperature that is
below the onset temperature of the film growth but still high enough to ensure rapid desorption. Ideally, in such a reactor there
would be no growth on the walls and accordingly no need of cleaning the walls, but it is probably hard to find proper temperatures
for all processes of an interest.
Many ALD reactors are built so that the reaction chamber and its heating elements are enclosed in a larger vacuum chamber with
cold or warm walls (60). This is convenient especially with large substrates and batches, and adds freedom to the design of the
reaction chamber that does not need to withstand atmospheric pressure.
4.05.4.1.4 Precursor Sources and Valving Systems

An ALD source should supply in a short time, preferably fractions of a second, a large enough dose of a precursor vapor to saturate
all the surfaces in the reaction chamber. Therefore there is a minimum vapor pressure the precursor must have in the source, and
often the compound needs to be heated for reaching that vapor pressure. The actual value of the required vapor pressure depends on
the reactor, including the source itself, and is largely affected by the surface area to be coated. For the alternate supply of the
precursor pulses, the precursor source and the reaction chamber must be separated by a mechanical valve or some other fast valving
system. A rapid change from one precursor to the other requires careful elimination of cold spots, dead spaces, and turbulences in
the source lines and the reaction chamber.
There are numerous source designs in use in ALD that differ from each other in heating capacity, location with respect to the
reaction chamber and the valving mechanism. The main factor affecting the source design is the vapor pressure of the precursor – the
higher the volatility, the simpler the source that suffices. Gases need just a valve for pulsing and some device for controlling the flow
from the cylinder into the reactor. High vapor pressure liquids are placed outside the reaction chamber, and a mechanical valve is
used to pulse the precursor into the reaction chamber. If the precursor needs to be heated to reach the desired vapor pressure, source
lines and valves must also be heated to avoid condensation. One option is to enclose the whole source – valve assembly into
a separate oven.
The simplest way for transporting the high vapor pressure precursor from the source into the reactor is vacuum draw. This occurs
when the pressure of the precursor in the source is higher than the total pressure in the reactor on the other side of the pulsing valve.
Vacuum draw is therefore applicable for all gases and for those liquids that have, at the source temperature, a vapor pressure higher
than the reactor pressure. No carrier gas is used for driving the precursor into the reactor, but further transportation inside the flow-
type reactors usually employs carrier gas. An alternative for the precursor transportation from the source is to use bubblers or other
vessels where an inert carrier gas flows through or over the source compound. This adds complexity but makes the precursor
transportation more efficient compared to the vacuum draw and is preferred especially in large surface area applications where large
precursor doses are needed. Importantly, bubblers and over-flow sources can also be used with precursors whose vapor pressure is
not high enough for the vacuum draw but still high enough so that the source does not need to be heated to high temperatures.
Efficient transfer of the solid to the overflowing transport gas can be ensured by maximizing the contact area using, for example,
a multiple tray assembly through which the transport gas percolates (81). When loaded with 500 g of HfCl4, such a source is
expected to run for an impressive 6 months in deposition of 2 nm gate oxides.
Pulsing of the high vapor pressure sources is relatively simple with fast solenoid or pneumatic valves. Requirements for ALD
valves include short response times below 20 ms, reliable operation for millions to tens of millions of cycles, no particle or other
kind of contamination, and consistent flow performance (82). Pneumatically driven ultrahigh-purity diaphragm valves are the
preferred choice for semiconductor applications and therefore have become the dominant ALD valves, though solenoid valves are
used too. The high operation temperature of the valve becomes important with less volatile precursors. Current valves can be
operated up to 300 C, which thereby sets the upper limit for the mechanically valved sources.
A low vapor pressure precursor can be defined as a compound that cannot be delivered with the mechanically valved sources
because its vapor pressure does not reach a high enough level at temperatures tolerated by the mechanical valves. A good example is
www.iran-mavad.com
ZnCl2, a precursor used in the first commercial application of ALD, i.e., thin film electroluminescent displays. ZnCl2 must be heated
around 350 C to reach a sufficient vapor pressure. Even the newest high temperature valves fail to reach this temperature, and
pulsing of this kind of precursor was not possible before the invention of the inert gas valving system (73–75) that constituted
a major breakthrough in ALD technology.
In the inert gas valved low vapor pressure sources the precursors are placed either inside the reactor in a region which is heated to
a temperature giving the desired vapor pressure, typically 0.1–1 Torr, or in a furnace outside the reactor and connected with hot
tubes to the chamber. In either case, the inert gas valving mechanism separates the source from the reaction chamber. The valving
action is realized by directing two inert gas flows into each source tube: one is a transport gas flowing over the source and the other
one a valving gas. The valving gas is fed into the source line at a point between the source and the reaction chamber. When the source
is in an off-state, the transport gas flow is off and the valving gas on. Upon entering the source line the valving gas divides into two
parts: one part purges the reaction chamber and the other part flows toward the source, setting a diffusion barrier that prevents
precursor molecules from entering the reaction chamber. The source is turned to the on-state by switching off the valving gas flow
and turning on the transport gas, thereby breaking the diffusion barrier and letting a pulse of precursor molecules enter the reaction
chamber. While the valves controlling the two gas flows operate at room temperature, the valving action itself takes place at the
reaction chamber temperature. Proper operation of the inert gas valving system relies on careful design, in particular a balance of
conductances around the valving point.
The inert gas valved sources were originally developed to the urgent need of pulsing truly low vapor pressure precursors, but
since that they have been used extensively also with precursors that from the vapor pressure point of view could be supplied from
the mechanically valved sources too. This is because once available, the inert gas valved sources are very convenient to operate at
least in research scale. It has also been found useful to combine the inert gas valving with mechanical valves to back-up the latter for
small leakage (83).
Liquid injection offers an alternative for ALD precursor delivery. Though the technology has been developed for CVD, many
liquid injection delivery systems operate in the pulsed, drop-on-demand mode, and as such meet the ALD requirement of pulsed
delivery of precursors. Both liquid precursors and solutions containing a solid precursor dissolved into an appropriate, preferably
inert solvent can be delivered this way. The liquid is injected into a hot evaporation zone where it is immediately vaporized
completely and transported further to the reaction chamber with an inert gas.
One advantage of the liquid injection delivery is that the precursor experiences elevated temperature only for a very short time of
a few seconds in maximum. This is critical with precursors that do not fully meet the requirement of thermal stability against self-
decomposition and degrade when kept for long times in a conventional heated source. One can also control accurately the volume
of liquid injected in each pulse. Potential disadvantages of the liquid injection delivery systems include reliability issues, like
clogging of the injector, residual particle formation and participation of the solvents, when used, into the growth reactions.
Avoidance of particles on the films is the most critical in semiconductor applications but also elsewhere. Some solid precursors
consist of very fine particles and are therefore highly potential particle sources, though particles may form from liquid residues too.
Effective particle filtering may therefore become necessary. Mechanical, electrostatic, and inertial trap filters have been suggested
(83,84). Mechanical filters, such as sinters and membranes, may suffer from low flow conductances that increase the precursor pulse
rise and decay times and thereby lengthen the ALD cycle time. These problems may be avoided by doing the filtering before the
valving point.
Precursors are usually transported from the different sources to the reaction chamber along separate lines, one for each precursor.
This avoids deposition on the walls of the source lines and the consequent cleaning needs. The closer to the substrate the lines
merge, the less precursor is wasted. It is worth noting, however, that often various inlet effects are observed right after the merging
point of the source lines where the film growth starts. Therefore some dummy area should be left between the merging point and the
substrate. In the cross-flow reactors the inlet effects are often clear, whereas in the perpendicular-flow reactors they may remain
hidden.
4.05.4.2 Plasma-Enhanced ALD Reactors

In principle, a conventional thermal ALD reactor can be converted to a PEALD reactor simply by adding a plasma source keeping the
other things the same. In practice, several issues are calling for attention and causing the need for further modifications. First, plasma
sources need space that must be taken from either inside or outside the reaction chamber. Second, the reactive species generated by
the plasma discharge have limited lifetimes, during which they need to be transported to the substrate surface. Not only ions and
electrons recombine faster than radicals but also the latter recombine rapidly in collisions with each other and other gas phase
species, and especially upon colliding with surfaces. Therefore the plasma source location, operation pressure of the reactor, and
flow path from the plasma source to the substrate are important parameters affecting chemical character and flux of reactive species
arriving on the substrate. For these reasons application of plasma in batch reactors is difficult though this has been studied too (85).
Yet another important issue is whether a film is growing in the plasma discharge region because it can greatly affect the plasma
discharge ignition and stability. In the worst case conductive films can short circuit the electrodes or prevent inductive coupling
through the originally dielectric discharge tube.
Plasma sources have been located either remotely, upstream of the substrates (86), or so that the substrates become immersed in
the plasma (Figure 6). Accordingly, different processes have been distinguished and named (87), the main difference being in the
magnitude of ion flux received by the substrate.
www.iran-mavad.com
Figure 6 (a) Schematics of a remote plasma ALD reactor with an inductively coupled RF plasma source. Only the nonmetal precursor gases are fed
through the discharge region while metal precursors are introduced directly into the growth chamber. One high vapor pressure source is shown.
The ellipsometer allows real-time measurement of thickness and optical properties of the growing film. (Copyright Heil, S. B. S; Kessels, W. M. M.
Eindhoven University of Technology.) (b) Schematics of a remote plasma ALD reactor with a capacitively coupled RF plasma source. Grid between the
substrate and plasma source prevents ion bombardment and thus only radicals reach the substrate. (Reproduced from Kariniemi, M.; Niinistö, J.; Hatanpää,
T.; Kemell, M.; Sajavaara, T.; Ritala, M.; Leskelä, M. Plasma-Enhanced Atomic Layer Deposition of Silver Thin Films. Chem. Mater. 2011, 23, 2901.).
In the direct plasma ALD the substrate is in intimate contact with plasma. This is usually accomplished by using a capacitively
coupled RF plasma discharge and placing the substrate on a grounded electrode with the opposite surface serving as the powered
electrode. Also a showerhead may serve as the powered electrode. Because only the nonmetal precursor can be activated with the
plasma, this needs to be pulsed. The intense ion bombardment experienced by the substrate from the intimate plasma may be
beneficial or detrimental as it supplies energy to many surface processes. At the same time a high flux of radicals is ensured in this
configuration. Limitations include film growth in the plasma discharge region and the space taken by the electrodes.
In both radical-enhanced ALD (REALD) and remote plasma ALD the plasma source is placed within a distance of the substrate
(Figure 6), and the distinction between the two should be made based on the species reaching the substrate. In radical-enhanced
ALD only radicals but no ions and electrons reach the substrate. The intermediate case where the plasma is remote but still close
enough so that some charged species survive to reach the substrate is called remote plasma ALD. For example, in a configuration
where an inductively coupled H2–N2 plasma source was as far as 30 cm from the substrate, an ion flux of about 2 1014 cm2 s1
was still estimated to arrive on the substrate when the reactor was operated at 10 mTorr pressure (88). Close to the substrate an
electron temperature of 3 eV was measured, indicating that an active plasma extended to the downstream region above the
substrate. The remote location of the plasma allows more space for its integration, and when properly protected by either purge gas
flows or gate valves, film growth in the plasma region can be completely eliminated for reliable and consistent operation. The
plasma can be either pulsed or continuously on so that the gas fed into the plasma is cycled between argon and a mixture of argon
and the reactant. Commonly inductively coupled RF discharges have been used as truly remote sources, whereas a capacitively
coupled RF source can be brought closer but still in a remote position by separating the discharge from the substrate by a grounded
metal grid (Figure 6). Microwave discharges offer a good alternative but require more space and are in this respect similar to the
inductively coupled RF discharges.
Because of the radical recombination in gas phase collisions, it was originally thought that plasma ALD reactors should be
operated at lower pressures than commonly used in thermal ALD reactors. Radical recombination upon surface collisions was in
turn thought to limit the conformality of plasma-enhanced ALD. It has been demonstrated, however, that remote plasma ALD
oxide processes run at around 5 mbar can produce conformal films in trenches with aspect ratios as high as 60:1 (89). These
successful results were interpreted to indicate that at higher pressures the increased loss of radicals in gas phase collisions is
compensated by their higher absolute density in the plasma. Also, in the reactor used in that study, the plasma discharge is only
a few centimeters from the substrate, behind a metal grid. As a consequence, a high flux of radicals can reach the substrate despite
the high operation pressure. Inside the small trenches the radical diffusion is molecular and thereby pressure independent.
Surface collisions dominate the recombination losses and make these highly dependent on chemical composition of both the
radicals themselves as well as the surface (90). Hydrogen atoms recombine on the surfaces more efficiently than oxygen atoms
and, accordingly, conformality of PEALD Ag films deposited with hydrogen plasma was poorer compared to the PEALD oxides
deposited with oxygen plasma (89).
As a final remark on plasma-enhanced ALD reactors, it is noted that in most cases the reactant gases (H2, N2, O2) are reactive
toward the metal precursor only when activated with plasma. Therefore, after switching off the plasma, these gases do not need to be
purged out of the reactor before the metal precursor pulse, hence allowing one to skip the purge step after the plasma pulse, thereby
shortening the cycle time. If so desired, the reactant gas flows can be kept constantly on also.
www.iran-mavad.com
4.05.4.3 Spatial ALD Reactors

In spatial ALD reactors the ALD cycle is realized by moving the substrate alternately in two or more continuous precursor flows
(Figure 7). The precursors and the ALD half-reactions are thus divided spatially rather than temporarily as is done in the
conventional ALD reactors. Inert gas flows and exhaust connections between the precursor flows prevent the precursors from mixing
and purge the substrates. Spatial ALD reactors with various designs have been developed as reviewed comprehensively in Ref. (91).
Here only a concise overview of the various reactor types is made.
Spatial ALD was in fact used already in the very first experiments on the ALD growth of ZnS. The elemental precursors were
evaporated from vertical ovens enclosed in a high vacuum chamber (Figure 8). The openings of these ovens were covered with
a rotating plate to which the substrates were fixed and became thus exposed alternately to Zn vapor, vacuum, Sx vapor and vacuum
again. In this reactor no purge or shield gas was used, but in their following patent Suntola et al. also described a spatial ALD reactor
with shield gases separating the precursors and reaction zones (Figure 8) (92). In the late 80s and early 90s Bedair and co-workers
studied extensively ALE of GaAs and other III–V semiconductors using a reactor where the substrates are exposed with the aid of
a rotating susceptor alternately to organometallic precursor and nonmetal hydride flows that are kept separated with a H2 purge gas
directed with a properly shaped wedge piece (Figure 9) (93). Cycle times as short as 0.5 s were reached with this tool.
The renewed interest toward spatial ALD reactors arises from the limited throughput of ALD. The main drivers have been on one
hand solar cells and on the other hand flexible electronics and other applications like packaging requiring coating of web shaped
substrates in a roll-to-roll manner. Many of the recently reported spatial ALD reactors can be viewed as modifications of the idea of
Suntola et al. with shield gases (Figure 8) (92). The precursor and shield gas zones are most often organized in a linear array under
which the substrates are moved (Figure 7(a)), but they can also placed around a drum directing the flexible web substrate
(Figure 10). For depositing a thicker film, two options exist: the substrate is moved back and forth under a limited number of
precursor zones, or the substrate makes a single pass along a long row of reaction zones. In the former case, the number of back and
forth passages determines the ALD cycle number and film thickness, while in the latter case these are determined by the number of
Figure 7 Schematics of two spatial ALD reactors. In (a) the substrate is moved under an injector with separate gas inlets of precursors A and B,
separated by inert gas inlets and exhaust connections. In (b) the flexible substrate is woven in a serpentine configuration back and forth between
two precursor zones separated by an inert gas zone. Note that without plasma activation O2 is inert toward TiCl4. Reproduced from Poodt, Paul;
Cameron, David C; Dickey, Eric; George, Steven M.; Kuznetsov, Vladimir; et al. Spatial Atomic Layer Deposition: A Route towards Further Industrialization
of Atomic Layer Deposition. J. Vac. Sci. Technol. A 2012, 30, 010802, http://dx.doi.org/10.1116/1.3670745. Copyright JVST.
Figure 8 (a) A high vacuum spatial ALD reactor used by Suntola et al. in their first ALD experiments. Left side shows a cross-section of the vacuum
chamber and right side a transverse section along line II – II on the left side image, in the direction of the arrows. Items 13a and 13b are heated
vapor sources for elemental zinc and sulfur, over which the substrates 14 are alternately rotated with the aid of disc 12. (Reprinted from Suntola, T.;
Antson, J. US Patent 4,508,430.) (b) A flow-type spatial ALD reactor design presented in 1983. Top image shows a cross-section and bottom
image vertical section along line XIV – XIV. The substrate 110 is reciprocally mover a fixed array of openings of sources 75, inert gas 73 and exhausts 74.
(Reprinted from Suntola, T.; et al. US Patent 4,389,973.).
www.iran-mavad.com
Figure 9 Schematics of a rotating susceptor ALD reactor. The susceptor consists of fixed and rotating parts. The precursor streams are directed
into windows in the fixed part below which the substrate rotates and cuts through the streams. A wedge and a large H2 flow from the center tube prevent
mixing of the two precursors. Reprinted from Tischler, M. A.; Bedair, S. M. Self-Limiting Mechanism in the Atomic Layer Epitaxy of GaAs. Appl.
Phys. Lett. 1681(1988), 48, 1681.
Figure 10 Schematics of a double-sided gas bearing spatial ALD concept. A gas bearing above and below the substrate, forming a narrow slit, directs
the substrate to float virtually frictionless through the reactor with a linear arrow of alternate precursor zones. Alternatively, the substrate can also
be moved back and forth in a reciprocating manner under a spatial ALD injector with a limited number of precursor zones. Reproduced from Poodt,
Paul; Cameron, David C; Dickey, Eric; George, Steven M.; Kuznetsov, Vladimir; et al. Spatial Atomic Layer Deposition: A Route towards Further
Industrialization of Atomic Layer Deposition. J. Vac. Sci. Technol. A 2012, 30, 010802, http://dx.doi.org/10.1116/1.3670745. Copyright JVST.
the precursor zones in the array. Both options are already in use in the first industrial spatial ALD reactors aimed for deposition of
passivation layers for silicon solar cells. A throughput of 3600 wafers h1 is achievable with the single pass tool, and a tool length of
10 m is needed for deposition of 10 nm Al2O3 (91). With the back-and-forth movement and a single TMA zone, a throughput
of 100 wafers h1 is obtained for 10 nm Al2O3, and scale-up is envisioned by multiplying the precursor zones and further on by
arranging 10–15 modules in parallel (91). A critical parameter in these designs is the gap between the precursor injector and the
substrate. A close proximity (<100 mm) of the two ensures not only efficient separation of the precursors but the gas shields can also
act as gas bearings, maintaining the substrate in close proximity to the injector without the need for extremely high tolerance
mechanical fixtures. When another gas bearing with inert gases only is added on the backside of the substrate, the substrates can be
levitated and moved frictionless using gas flows only (Figure 10). In the roll-to-roll coater in Figure 11 a drum directs the web close
to the curved injector head, which is in a rocking motion of about 1 Hz along the lower half of the drum. The web width can be
500 mm and speed up to 2 m min1 at 25 nm film thickness, giving a throughput of >200 000 m2 per year.
The reaction zones can also be incorporated into a round reactor head under which the substrate is rotated (Figure 12). An Al2O3
growth rate of 1.2 nm s1 has been achieved with this tool (94). However, because of the rotational symmetry, the central part of the
rotating susceptor cannot be coated and therefore this tool perhaps serves mostly as a research tool for the above mentioned linear
array tools.
Quite a different approach to spatial ALD is shown in Figure 7(b) where the web shaped substrate is woven in a serpentine
configuration back and forth between two precursor zones separated by an inert gas zone. The nonmetal precursor may be activated
either thermally or plasma (91).
Recent efforts and progress in spatial ALD reactors have been impressive, but many challenges remain. In particular, all the
processes run so far with spatial ALD reactors (Al2O3, ZnO, TiO2) have used high vapor pressure precursors. Getting high enough
www.iran-mavad.com
Figure 11 Roll-to-roll ALD reactor. Left: An overview showing the feeding and collecting rolls on top and the web directing drum and the curved
coating head at the bottom part of the chamber. Right: A closeup of the drum – web – coating head assembly. The large picture shows only the
precursor feed slots while the magnification shows also the nitrogen feed and exhaust slots between the precursor slots. Copyright Beneq.
Figure 12 Schematics of a rotating spatial ALD reactor. (a) Bottom side of the reactor head under which the substrate is rotated as shown in (b).
Reproduced from Poodt, Paul; Cameron, David C; Dickey, Eric; George, Steven M.; Kuznetsov, Vladimir; et al. Spatial Atomic Layer Deposition:
A Route towards Further Industrialization of Atomic Layer Deposition. J. Vac. Sci. Technol. A 2012, 30, 010802, http://dx.doi.org/10.1116/1.3670745.
Copyright JVST.
fluxes from lower vapor pressure precursors will be a critical issue for expanding the spatial ALD materials selection toward that
covered by conventional ALD reactors. In addition, coating of more irregularly shaped substrates with spatial ALD reactors will form
another challenge. With the highest aspect ratio structures the speed benefits of these tools probably cannot be enjoyed because
cycle times become dominated by diffusion of the precursors into the trenches and holes of the substrate.
4.05.4.4 Other Reactor Configurations

Several special reactor designs have been developed for substrates that have demanding high-aspect-ratio pores or are in a powder
form. For example, a flow-through reactor was constructed for coating such substrates that have holes running through from one
face to the other (95). In this reactor configuration all the precursor molecules are forced to flow through the substrate, which both
makes the process faster and improves the precursor utilization as compared with the other reactors where the precursor trans-
portation into the pores occurs only by diffusion of some molecules from the main stream flowing along the substrate surface.
Another reactor type for coating very high-aspect-ratio structures, applicable to close-ended pores as well, uses a static vacuum
that is created by isolating the chamber with valves from the sources and the vacuum pump. This allows running the ALD process in
a fill-in – hold – pump-down mode where the hold period can be set long enough for the precursors to reach and saturate all the
surface sites also at the bottom of the deepest pores. Thanks to the static vacuum, no precursor is wasted into the exhaust while
waiting for the saturation to be completed. Due to the long exposure times, batch processing is preferred for productivity reasons.
Also this reactor type was suggested already in the 1970s (92) but is only now making it to the wider use.
www.iran-mavad.com
Powders represent an extreme challenge for ALD both in terms of surface area to be coated and conformality already when
processed in amounts like 1 cm3. The first studies on ALD on powders dealt with silica and alumina catalyst supports that consist of
porous powders with specific surface areas of several hundreds of m2 g1 (96,97). However, as the aim of these studies was in
modifying the surface composition, only a few ALD cycles needed to be applied and the requirements to the reactor were not as
severe as in complete coating of powders. The powder was placed into a container with a porous bottom plate, and the gases were
flown through the powder bed without any stirring. Saturation of all the surfaces required exposure times up to hours even with
small amounts of powder. By contrast, coating of nonporous powders with continuous films sets more challenges to the reactors
even if the surface area is often substantially smaller than with the porous powders. There is an obvious risk of binding the particles
together with the coating if the process is done without movement of the powder bed. In one reactor, the reactor vessel is constantly
shaken by a mechanical vibration (98), while another design uses a rotary porous tube as a holder for the powder particles that are
agitated by balancing gravity and centrifugal forces (<1G) (99). Fluidized bed is a common approach for powder processing in
general but does not allow the use of static exposures, which might be beneficial especially when combining high surface areas and
moderately reactive precursors. The high surface areas of powders consume lots of precursors and thereby emphasize the impor-
tance of the precursor utilization efficiency.
4.05.5 ALD Applications
Application areas where ALD is considered and studied have increased in number rapidly during the past years along with the rapid
growth of ALD research in general. Yet the commercial applications of ALD, reported at least, are still quite limited. This is largely
due to the fact that ALD is a relatively slow method and as such often also more expensive than the alternative methods. Conse-
quently, ALD appears to become adopted into commercial use only when the other methods fail. This will be well exemplified by
the existing and published commercial applications to be discussed below.
This section begins with a brief survey of applications where ALD is already in production and is followed by an overview of areas
where ALD is under active research, with a possibility to see production use in the near future.
4.05.5.1 Industrial Applications

Thin film electroluminescent displays (TFEL) (Figure 13) served as the original motivation for developing the ALD technology,
including both chemistry and reactors (100–104). In the mid-1970s the concept of the TFEL display was known, but the existing
Figure 13 Schematics and photos of TFEL displays. The luminescent ZnS:Mn (about 1 mm in thickness), insulating Al2O3 or AlxTiyO (about 200 nm
each) as well as protective and passivating Al2O3 layers are produced by ALD. In the transparent display (lower photo) the metal electrode is replaced
with a transparent electrode and the black background is left out. TFEL display is operated by applying about 200 V ac voltage to the electrodes
crossing a chosen pixel. The high ac field forces electrons to move back-and-forth across the ZnS:Mn luminescent layer. During their travel the electrons
gain high energy from the electric field and impact excite the manganese ions. De-excitation results in light emission. Copyright Planar Systems Inc.
www.iran-mavad.com
thin film deposition methods did not allow their production in a reliable and cost-effective manner. This was largely due to the high
electric field (1–2 MV cm1) to which the dielectric-luminescent-dielectric three-layer structure is exposed in operation of the
display. The high electric field makes the large area device sensitive to breakdown, causing defects like pinholes. ALD turned out to
enable high yield in TFEL display manufacturing and has been in commercial use since the early 1980s. The thin film materials
produced by ALD for the TFEL display are ZnS:Mn as the luminescent layer, Al2O3 or AlxTiyO as the insulators, and Al2O3 as the
passivation and protective layers. It should be noted that the ALD grown layers for this application are rather thick, at least 200 nm,
the luminescent layer even exceeding 1 mm. A variety of other ALD materials and processes were also developed and explored over
the years in TFEL display research; this has laid ground for bringing ALD into new application areas.
4.05.5.1.1 Magnetic Heads

The second commercial application of ALD appeared around the turn of the millennium in thin film magnetic heads that are used to
read and write data into hard disks. Because of the aggressive downscaling of the magnetic heads, the so far used sputtering method
was rapidly approaching its capability limit in depositing thin insulating gap layers on nonplanar surfaces. This created an urgent
need for a low temperature deposition process of conformal insulator with good dielectric strength. ALD technology could luckily
reply to this request by the already existing nearly ideal Al(CH3)3–H2O process for Al2O3 (105) and became rapidly adopted into
production (106,107). It could even be stated that ALD saved the magnetic head industry (107).
4.05.5.1.2 Microelectronics
Microelectronics has been the major driving force for the ALD technology for the past 15 years. This development started in the late
1990s when it became obvious that the continuation of Moore’s law would critically depend on an introduction of new materials
and their deposition methods to the integrated circuit technology. Key requirements for the deposition methods are atomic level
accuracy in thickness control, uniformity over the increasing wafer sizes and conformality over the increasingly complex device
structures, all factors favorable for ALD. At the same time, the layer thicknesses have been shrinking to the nanometer level, thereby
decreasing though not completely eliminating the greatest drawback of ALD, i.e., slowness. High-k dielectric materials as both
alternative gate oxides in metal-oxide-semiconductor field effect transistors (MOSFETs) and capacitor dielectrics in dynamic
random access memories (DRAMs) have been in the central role in microelectronics applications of ALD. DRAMs require
continuously more complex 3D capacitor structures to ensure the required storage cell capacitance despite the ever decreasing
memory cell areas (Figure 14). Currently the dielectric material is a trilayered ZrO2–Al2O3–ZrO2 nanolaminate structure
(Figure 14) but intense research is focused on other higher-k value materials like SrTiO3 and doped TiO2 (108). Besides the
capacitor dielectric, ALD may be used for the electrode deposition. TiN is the currently used electrode material, but in the future for
leakage reasons it may need to be replaced with the higher work function ruthenium.
Intel was the first to report ALD in production of microprocessors when they made the historic change of materials into the
high-k dielectric – metal gate electrode stacks in MOSFETs (109). The gate dielectric was reported to be hafnium oxide based
material grown by ALD. The traditional MOSFETs were planar devices but the replacement gate approach and more complex
transistor structures like FinFETs (Figure 15) exploited already by Intel have also made MOSFET processing a 3D task. In addition to
the gate dielectric, the gate electrode is an obvious application to ALD and extensive research is ongoing.
Third semiconductor application area where ALD has been extensively studied is interconnects. Apparently the first real
application was tungsten seed layers for tungsten CVD contact plug fill for which several companies announced production worthy
solutions already in the early 2000s (110). These processes are based on a reduction of WF6 with either B2H6 or silanes and have
exceptionally high growth rates. For copper interconnects ALD has been explored for years but its implementation has been delayed
Figure 14 (a) Schematics and evolution of the stack DRAM memories (Copyright Wiley-VCH). (b) ZrO2/Al2O3/ZrO2 (ZAZ) dielectric film in a stacked
capacitor. Kim, S. K.; et al. Adv. Mater. 2008, 20, 1429. Copyright IEEE.
www.iran-mavad.com
Figure 15 Cross-section of Intel’s FinFET structure. Copyright Chipworks.
because innovations in the traditionally used PVD technologies have extended their use from device generation to another. ALD
metal nitrides have been explored as diffusion barriers and ALD copper as a seed layer for copper electroplating. Unfortunately
copper appears to be a challenging material to find good ALD chemistry. Noble metals like ruthenium have been evaluated as
potential combined barrier-seed layers. There are also many other application areas in microelectronics for which ALD is explored
and maybe already used. These include high-aspect-ratio RF and decoupling capacitors and nonvolatile memories such as FeRAM,
flash and phase change memories.
A different but increasingly important application for ALD in interconnect technology has emerged from the self-aligned double
patterning technology that is an efficient way for preparing repetitive narrow line width patterns (111). In this application high
quality, conformal films need to be deposited on patterned photoresists at low temperatures as required by the photoresist
materials.
4.05.5.1.3 Protective Coatings

A quite surprising commercial application of ALD is in the field of jewelry (112). ALD Al2O3 proved effective in preventing silver
tarnishing at thicknesses as low as about 10 nm. This low thickness is important because it leaves the visual appearance of the silver
objects essentially unchanged. Robust process, inexpensive precursors, and batch processing with 2000 valuable pieces ensure the
cost effectiveness of the process. Obviously the effective sealing by conformal ALD coatings could be used in providing chemical
protection for also many other objects. Promising results have been obtained in preventing corrosion of both stainless steel and low
alloy steel, for example using ALD films alone or combined with PVD coatings (113,114).
4.05.5.2 Other Applications

Besides the reported commercial applications, ALD has been studied in a number of other areas. In the following some potential
and emerging application areas are briefly described.
4.05.5.2.1 Energy Technologies

Photovoltaics (PVs) are in a key role in solving the future energy needs. For ALD there is certainly room for further development
of nanostructured solar cells. During the past 20 years ALD has been studied for various applications, e.g., ALD Al2O3 as
a passivation layer in crystalline Si or encapsulation layer in CIGS (115). For dye-sensitized solar cells various ALD grown oxide
layers have been intensively studied especially for barrier and blocking layers. For industrial breakthrough of ALD in solar cells,
the key issue is how to adapt the most potential applications, passivation layer in c-Si solar cells and buffer layer in CIGS, in a cost
effective manner for high volume manufacturing. Here the development of spatial and roll-to-roll ALD reactors is of utmost
importance.
Other emerging energy technologies for ALD are solid oxide fuel cells (SOFCs) and lithium-ion batteries (LIBs). For SOFCs,
electrolyte fabrication by ALD has especially been studied; for a more comprehensive review, see Ref. (116). For lithium-ion
batteries, a fully ALD manufactured LIB (anode, electrolyte, cathode) is under research. Another approach is to enhance the
properties of conventional LIB electrodes by coating them with ultrathin Al2O3 films by ALD. A more detailed description of the use
of ALD in LIBs can be found in Ref. (117).
4.05.5.2.2 Optical Applications

In optics ALD has the potential for a wide range of applications, but as in the case of PVs, the cost effectiveness has limited high
volume manufacturing. However, in complex 3D components ALD has clear advantages and opportunities for industrial
production. Film thickness control requirements in certain areas of optics are extremely strict and in addition, the conformality
capability clearly favors ALD. Multilayer interference coatings have been made with ALD for wavelength ranges from IR to
www.iran-mavad.com
visible and soft and hard x-rays (30,118–120). In 3D structured optical components the development is still ongoing and ALD
has been explored in preparation of novel optical devices from, for example, nanogratings (121) and porous silicon (122).
Inverse opal (123–125) photonic crystals, grid filters (126), erbium-doped waveguides (127), and ultrahigh resolution Fresnel
zone plates for x-ray microscopy (128) are further examples of the exploitation of the conformality of ALD in complex optical
components.
4.05.5.2.3 Coatings on Powders and Polymers

Besides surface modification of porous powders, ALD has also been used in encapsulating compact powders with continuous films
(98,99,129). If the ongoing work on upscaling the coating process to large batches turns out to be successful and economically
feasible, potential application areas are numerous.
ALD films, Al2O3 in particular, have been deposited also on polymers at temperatures below 100 C (130). With plasma-
enhanced ALD Al2O3 has been deposited even at room temperature (131). One of the main motivations for the studies on low
temperature deposition on polymers has been the barrier coatings for organic light emitting diodes. Al2O3 films seem to degrade
over time with exposure to water, however (132). Combination of Al2O3 with SiO2 into multilayer structures improves the stability
and results in a lower water vapor transmission rate than either of the multilayer constituents alone (132).
4.05.5.2.4 Photocatalysts
ALD has been shown to produce photocatalytically active TiO2 films (133). Titanium dioxide is the most widely used pho-
tocatalyst in air and water purification as well as in self-cleaning and sterilizing coatings. Lots of methods exist for the prep-
aration of TiO2 photocatalysts, making it difficult for ALD to compete. However, due to the excellent conformality and
possibility to vary the composition at the atomic layer level ALD may find use in preparation of tailored 3D photocatalyst
structures (134).
4.05.5.2.5 Micro-electro-mechanical Systems

Micro-electro-mechanical systems (MEMS) contain nearly always 3D structures whose conformal coating is an obvious application
area for ALD. On the other hand, ALD is a most attractive technique when nanometer scale coatings are desired, and thus ALD in
MEMS technology is suited for modification of structural layers. Publications on ALD in MEMS applications have so far remained
limited (135). Nevertheless, for example wear resistant (136), lubricating (137), anti-stiction (138), and charge dissipating (139)
coatings have been deposited by ALD on MEMS components.
4.05.5.2.6 Nanotechnology
Fabrication of complex nanostructures via templates and modification of nanostructure surfaces are emerging application areas
for ALD. Potential applications in the field of nanotechnology are numerous. However, as general in the field of nanotech-
nology, commercial applications often take a considerable amount of time. The research efforts have expanded rapidly due to
the fact that ALD by its characteristic features, like perfect conformality and accurate thickness control, is an ideal tool for
nanotechnology. A long list of nanofibers and nanotubes have been coated by ALD (140), like carbon nanotubes (141,142),
semiconductor nanowires (143–145), self-assembled polymer fibers (146), cellulose fibers (147), metal wires made by tem-
plating with porous membranes (134), and electrospun ceramic and polymer nanowires (148,149). Also many nanoporous
membranes have been coated by ALD, like anodic alumina (134,150), g-Al2O3 (151) and track-etched polycarbonate (152).
Other nanostructured objects coated with ALD include porous silicon (153,154), porous anodic alumina films (155,156),
aerogels (157), and synthetic opal (123–125). Application of ALD to biomaterials has been of interest too: viruses (158),
proteins, natural fibers like DNA and spider silk and patterned biomaterials like nanostructured butterfly wings (159). If the
substrate is removed after the coating, a hollow replica directly following the original structure is formed from the ALD grown
film. Such a templating process has been used to prepare nanotubes with electrospun nanofibers (148), anodic alumina
(155,156), and cellulose fibers (147). Several reviews (140,160–162) have been published for a more comprehensive overview
of using ALD in nanotechnology.
References
1. Ritala, M.; Niinistö, J. ECS Trans. 2009, 25 (8), 641.

2. Ritala, M.; Leskelä, M. In Handbook of Thin Film Materials, Nalwa, H. S., Ed.; Academic Press: New York, 2002; Vol. 1, pp 103–159.
3. Ritala, M.; Niinistö, J. In Chemical Vapour Deposition: Precursors, Processes and Applications; Jones, A. C., Hitchman, M. L., Eds.; Royal Society of Chemistry: Cambridge,
2008; pp 158–206.
4. Puurunen, R. L. J. Appl. Phys. 2005, 97, 121301.
5. Miikkulainen, V.; Ritala, M.; Leskelä, M.; Puurunen, R. L. J. Appl. Phys. 2013, 113, 021301.
6. George, S. M. Chem. Rev. 2010, 110, 111.
7. Maula, J.; Alasaarela, T.; Sneck, S. Vacuum Technol. Coat. October 2009, 54.
8. Putkonen, M. ECS Trans. 2009, 25 (4), 143.
9. Granneman, E.; Fischer, P.; Pierreux, D.; Terhorst, H.; Zagwijn, P. Surf. Coat. Technol. 2007, 201, 8899.
10. Profijt, H. B.; Potts, S. E.; van de Sanden, M. C. M.; Kessels, W. W. M. J. Vac. Sci. Technol. A 2011, 29, 050801.
www.iran-mavad.com
11. Copel, M.; Gibelyuk, M.; Gusev, E. Appl. Phys. Lett. 2000, 76, 436.
12. Elers, K.-E.; Blomberg, T.; Peussa, M.; Aitchison, B.; Haukka, S.; Marcus, S. Chem. Vap. Deposition 2006, 12, 13.
13. Niinistö, J.; Kukli, K.; Kariniemi, M.; Ritala, M.; Leskelä, M.; Blasco, N.; Pinchart, A.; Lachaud, C.; Laaroussi, N.; Wang, Z.; Dussarrat, C. J. Mater. Chem. 2008, 18, 5243.
14. Blanquart, T.; Niinistö, J.; Longo, V.; Pallem, V. R.; Dussarrat, C.; Ritala, M.; Leskelä, M. Chem. Mater. 2012, 24, 3420.
15. Suntola, T. Mater. Sci. Rep. 1989, 4, 261.
16. Yao, T.; Takeda, T. Appl. Phys. Lett. 1986, 48, 160.
17. Suntola, T. Appl. Surf. Sci. 1996, 100/101, 391.
18. Lahtinen, J. A.; Lu, A.; Tuomi, T.; Tammenmaa, M. J. Appl. Phys. 1986, 58, 1851.
19. Goodman, C. H. L.; Pessa, M. J. Appl. Phys. 1986, 60, R65.
20. Asikainen, T.; Ritala, M.; Leskelä, M. Appl. Surf. Sci. 1994, 82/83, 122.
21. Pore, V.; Hatanpää, T.; Ritala, M.; Leskelä, M. J. Am. Chem. Soc. 2009, 131, 3478.
22. Sarnet, T.; Pore, V.; Hatanpää, T.; Ritala, M.; Leskelä, M.; Schrott, A.; Zhu, Y.; Raoux, S. J. Electrochem. Soc. 2011, 158, D694.
23. Ritala, M.; Leskelä, M.; Rauhala, E.; Haussalo, P. J. Electrochem. Soc. 1995, 142, 2731.
24. Alen, P.; Ritala, M.; Arstila, K.; Keinonen, J.; Leskelä, M. J. Electrochem. Soc. 2005, 152, G361.
25. (a) Juppo, M. Ph.D. Thesis, University of Helsinki, 2001.
(b) Alen, P. Ph.D. Thesis, University of Helsinki, 2005.
26. Alen, P.; Juppo, M.; Ritala, M.; Sajavaara, T.; Keinonen, J.; Leskelä, M. J. Electrochem. Soc. 2001, 148, G566.
27. Ritala, M.; Leskelä, M.; Nykänen, E.; Soininen, P.; Niinistö, L. Thin Solid Films 1993, 225, 288.
28. (a) Schuisky, M. Ph.D. Thesis, University of Uppsala, 2000.
(b) Sundqvist, J. Ph.D. Thesis, University of Uppsala, 2003.
29. Klaus, J. W.; Ferro, S. J.; George, S. M. Thin Solid Films 2000, 360, 145.
30. Fabreguette, F. H.; Wind, R. A.; George, S. M. Appl. Phys. Lett. 2006, 88, 013116.
31. Hill, C. W.; Derderian, G. J.; Sandhu, G. J. Electrochem. Soc. 2005, 152, G386.
32. Pilvi, T. Ph.D. Thesis, University of Helsinki, 2008.
33. Leskelä, M.; Mölsä, H.; Niinistö, L. Supercond. Sci. Technol. 1993, 6, 627.
34. Lee, S. W.; Kwon, O. S.; Han, J. H.; Hwang, C. S. Appl. Phys. Lett. 2008, 92, 222903.
35. Kosola, A.; Putkonen, M.; Johansson, L.-S.; Niinistö, L. Appl. Surf. Sci. 2003, 211, 102.
36. Tiitta, M.; Niinistö, L. Chem. Vap. Deposition 1997, 3, 167.
37. Aaltonen, T.; Ritala, M.; Arstila, K.; Keinonen, J.; Leskelä, M. Chem. Vap. Deposition 2004, 10, 215.
38. Aaltonen, T.; Rahtu, A.; Ritala, M.; Leskelä, M. Electrochem. Solid State Lett. 2003, 6, C130.
39. Hämäläinen, J.; Kemell, M.; Munnik, F.; Kreissig, U.; Ritala, M.; Leskelä, M. Chem. Mater. 2008, 20, 2903.
40. Jones, A. C.; Aspinall, H. C.; Chalker, P. R.; Potter, R. J.; Kukli, K.; Rahtu, A.; Ritala, M.; Leskelä, M. Mater. Sci. Eng. B 2005, 118, 97.
41. Jones, A. C.; Aspinall, H. C.; Chalker, P. R.; Potter, R. J.; Kukli, K.; Rahtu, A.; Ritala, M.; Leskelä, M. J. Mater. Chem. 2004, 14, 3101.
42. Yang, T. S.; Cho, W. T.; Kim, M.; An, K. S.; Chung, T. M.; Kim, C. G.; Kim, Y. J. Vac. Sci. Technol. A 2005, 23, 1238.
43. Vehkamäki, M.; Ritala, M.; Leskelä, M.; Jones, A. C.; Davies, H. O.; Sajavaara, T.; Rauhala, E. J. Electrochem. Soc. 2004, 151, F69.
44. Kukli, K.; Ritala, M.; Sajavaara, T.; Keinonen, J.; Leskelä, M. Chem. Vap. Deposition 2002, 8, 199.
45. Hausmann, D. M.; Kim, E.; Becker, J.; Gordon, R. G. Chem. Mater. 2002, 14, 4350.
46. Kim, S. K.; Hwang, C. S. Electrochem. Solid State Lett. 2008, 11, G9.
47. Niinistö, J.; Mäntymäki, M.; Kukli, K.; Costelle, L.; Puukilainen, E.; Ritala, M.; Leskelä, M. J. Cryst. Growth 2010, 312, 245.
48. Niinistö, J.; Kukli, K.; Heikkilä, M. J.; Ritala, M.; Leskelä, M. Adv. Eng. Mater. 2009, 11, 222.
49. van der Straten, O.; Zhu, Y.; Dunn, K.; Eisenbraun, E. T.; Kaloyeros, A. E. J. Mater. Res. 2004, 19, 447.
50. Maeng, W. J.; Park, S.-J.; Kim, H. J. Vac. Sci. Technol. B 2006, 24, 2276.
51. Vehkamäki, M.; Hatanpää, T.; Ritala, M.; Leskelä, M. J. Mater. Chem. 2004, 14, 3191.
52. Hämäläinen, J.; Munnik, F.; Hatanpää, T.; Holopainen, J.; Ritala, M.; Leskelä, M. J. Vac. Sci. Technol. A 2012, 30, 01A106.
53. Lim, B. S.; Rahtu, A.; Gordon, R. G. Inorg. Chem. 2003, 42, 7951.
54. De Rouffignac, P.; Park, J. S.; Gordon, R. G. Chem. Mater. 2005, 17, 4808.
55. Lim, B. S.; Rahtu, A.; Gordon, R. G. Nature Mater. 2003, 2, 749.
56. Wang, H. T.; Gordon, R. G.; Alvis, R.; Ulfig, R. M. Chem. Vap. Deposition 2009, 15, 312.
57. Whitehorne, T. J. J.; Coyle, J. P.; Mahmood, A.; Monillas, W. H.; Yap, G. P. A.; Barry, S. T. Eur. J. Inorg. Chem. 2011, 21, 3240.
58. Higashi, G. S.; Flemming, C. G. Appl. Phys. Lett. 1989, 55, 1963.
59. Tischler, M. A.; Bedair, S. M. In Atomic Layer Epitaxy; Suntola, T., Simpson, M., Eds.; Blackie: Glasgow, 1990; p 110.
60. Skarp, J. I.; Soininen, P. J.; Soininen, P. T. Appl. Surf. Sci. 1997, 112, 251.
61. Hunter, A.; Kitai, A. H. J. Cryst. Growth 1988, 91, 111.
62. Stuyven, G.; DE Visschere, P.; Hikavyy, A.; Neyts, K. J. Cryst. Growth 2002, 234, 690.
63. Vehkamäki, M.; Hatanpää, T.; Hänninen, T.; Ritala, M.; Leskelä, M. Electrochem. Solid State Lett. 1999, 2, 504.
64. Leskelä, M.; Kukli, K.; Ritala, M. J. Alloy. Compd. 2006, 418, 27.
65. Niinistö, J.; Putkonen, M.; Niinistö, L.; Kukli, K.; Ritala, M.; Leskelä, M. J. Appl. Phys. 2004, 95, 84.
66. Niinistö, J.; Putkonen, M.; Niinistö, L.; Arstila, K.; Sajavaara, T.; Lu, J.; Kukli, K.; Ritala, M.; Leskelä, M. J. Electrochem. Soc. 2006, 153, F39.
67. Aaltonen, T.; Alen, P.; Ritala, M.; Leskelä, M. Chem. Vap. Deposition 2003, 9, 45.
68. Kwon, O.-K.; Kim, J.-H.; Park, H.-S.; Kang, S.-W. J. Electrochem. Soc. 2004, 151, G109.
69. Aaltonen, T.; Ritala, M.; Sajavaara, T.; Keinonen, J.; Leskelä, M. Chem. Mater. 2003, 15, 1924.
70. Kukli, K.; Aarik, J.; Aidla, A.; Jogi, I.; Arroval, T.; Lu, J.; Laitinen, M.; Sajavaara, T.; Kiisler, A.-A.; Ritala, M.; Leskelä, M.; Peck, J.; Natwora, J.; Geary, J.; Spohn, R.;
Meiere, S.; Thompson, D. M. Thin Solid Films 2012, 520, 2756.
71. Pore, V.; Knapas, K.; Hatanpää, T.; Sarnet, T.; Kemell, M.; Ritala, M.; Leskelä, M.; Mizohata, K. Chem. Mater. 2011, 23, 247.
72. Knapas, K.; Hatanpää, T.; Ritala, M.; Leskelä, M. Chem. Mater. 2010, 22, 1386.
73. Suntola, T. S.; Pakkala, A. J.; Lindfors, S. G. U.S. Patent 4389973, 1983.
74. Suntola, T. S.; Pakkala, A. J.; Lindfors, S. G. U.S. Patent 4413022, 1983.
75. Suntola, T. In Handbook of Crystal Growth; Hurle, D. T. J., Ed.; Elsevier: Amsterdam, 1994; p 601.
76. Ylilammi, M. J. Electrochem. Soc. 1995, 142, 2474.
77. Barua, D.; Gougousi, T.; Young, E. D.; Parsons, G. N. Appl. Phys. Lett. 2006, 88, 92904.
78. Haukka, S. ECS Trans. 2007, 3, 15.
79. Okuyama, Y.; Barelli, C.; Tousseau, C.; Park, S.; Senzaki, Y. J. Vac. Sci. Technol. A 2005, 23, L1.
80. Putkonen, M. In IMEC Workshop Atomic Layer Deposition for Applications in Nanotechnology; 2011. www.beneq.com.
www.iran-mavad.com
81. Verghese, M.; Battle, S.; Roeder, J. Semicond. Int. 2008, 31 (4), 24.
82. Glime, W.; Seidel, T. Semicond. Int. 2005, 28 (9), 76.
83. Aitchison, B. J.; Maula, J.; Lang, T.; Kuosmanen, P.; Härkönen, K.; Sonninen, M. U.S. Patent Appl. 0124131, 2004.
84. Kesälä, J. U.S. Patent 6699524, 2001.
85. Lehnert, W.; Ruhl, G.; Gschwandtner, A. J. Vac. Sci. Technol. A 2012, 30, 01A152.
86. Heil, S. B. S.; van Hemmen, J. L.; Hodson, C. J.; Singh, N.; Klootwijk, J. H.; Roozeboom, F.; van de Sanden, M. C. M.; Kessels, W. M. M. J. Vac. Sci. Technol. A 2007,
25, 1357.
87. van Hemmen, J. L.; Heil, S. B. S.; Klootwijk, J. H.; Roozeboom, F.; Hodson, C. J.; van de Sanden, M. C. M. J. Electrochem. Soc. 2007, 154, G165.
88. Heil, S. B. S.; Langereis, E.; Roozeboom, F.; van de Sanden, M. C. M.; Kessels, W. M. M. J. Electrochem. Soc. 2006, 153, G956.
89. Kariniemi, M.; Niinistö, J.; Vehkamäki, M.; Kemell, M.; Ritala, M.; Leskelä, M.; Putkonen, M. J. Vac. Sci. Technol. A 2012, 30, 01A115.
90. Knoops, H. C. M.; Langereis, E.; van de Sanden, M. C. M.; Kessels, W. M. M. J. Electrochem. Soc. 2010, 157, G241.
91. Poodt, P.; Cameron, D. C.; Dickey, E.; George, S. M.; Kuznetsov, V.; Parsons, G. N.; Roozeboom, F.; Sundaram, G.; Vermeer, A. J. Vac. Sci. Technol. A 2012, 30,
010802.
92. Suntola, T.; Antson, J. U.S. Patent 4058430, 1977.
93. Tischler, M. A.; Bedair, S. M. Appl. Phys. Lett. 1988, 48, 1681.
94. Poodt, P.; Lankhorst, A.; Roozeboom, F.; Spee, K.; Maas, D.; Vermeer, A. Adv. Mater. 2010, 22, 3564.
95. Ritala, M.; Kemell, M.; Lautala, M.; Niskanen, A.; Leskelä, M.; Lindfors, S. Chem. Vap. Deposition 2006, 12, 655.
96. Lindblad, M.; Haukka, S.; Kytökivi, A.; Lakomaa, E.-L.; Rautiainen, A.; Suntola, T. Appl. Surf. Sci. 1997, 121/122, 286.
97. Haukka, S.; Lakomaa, E.-L.; Suntola, T. Stud. Surf. Sci. Catal. 1998, 120, 715.
98. Hakim, L. F.; George, S. M.; Weimer, A. W. Nanotechnology 2005, 16, S375.
99. McCormick, J. A.; Clouter, B. L.; Weimer, A. W.; George, S. M. J. Vac. Sci. Technol. A 2007, 25, 67.
100. Suntola, T.; Antson, J.; Pakkala, A.; Lindfors, S. SID 80 Digest 1980, 11, 108.
101. Suntola, T. Mater. Sci. Rep. 1989, 4, 261.
102. Leskelä, M.; Niinistö, L. In Atomic Layer Epitaxy; Suntola, T., Simpson, M., Eds.; Blackie: Glasgow, 1990; p 1.
103. Törnqvist, R. Displays 1992, 13, 81.
104. Leskelä, M.; Li, W.-M.; Ritala, M. In Electroluminescence, Semiconductors and Semimetals; 1999; Vol. 4, p 413.
105. Paranjpe, A.; Gopinath, S.; Omstead, T.; Bubber, R. J. Electrochem. Soc. 2001, 148, G465.
106. Devasahayam, A. Solid State Technol. 2005, 48 (9), S16.
107. Kautzky, M.; Lamberton, R.; Chakravarty, S.; Stearns, L.; Kumar, A.; Dolejsi, J.; McKinlay, S.; Colaianni, L.; Perrey, C. AVS 3th International Conference on Atomic Layer
Deposition. San Jose, United States, 2003 (presentations available on CD-ROM).
108. Hwang, C. S. In Atomic Layer Deposition of Nanostructured Materials; Pinna, N., Knez, M., Eds.; Wiley-VCH: Weinheim, 2011; p 161.
109. Bohr, M. T.; Chan, R. S.; Ghani, T.; Mistry, K. IEEE Spectrum 2007, 44, 29.
110. Braun, A. E. Semicond. Int. 2001, 24 (10), 52.
111. Shibata, T.; Satoh, J.; Ogawa, J.; Abe, T.; Ishii, T.; Hasebe, K. Int. Symp. Semicond. Manuf. 2010, 1.
112. Sneck, S. Helsinki. In Nanotechnology in Northern Europe Conference; 2007.
113. Matero, R.; Ritala, M.; Leskelä, M.; Salo, T.; Aromaa, J.; Forsen, O. J. Phys. IV 1999, 9, Pr8–493.
114. Díaz, B.; Swiatowska, J.; Maurice, V.; Seyeux, A.; Härkönen, E.; Ritala, M.; Tervakangas, S.; Kolehmainen, J.; Marcus, P. Electrochim. Acta 2013, 90, 232.
115. van Delft, J. A.; Garcia-Alonso, D.; Kessels, W. M. M. Semicond. Sci. Technol. 2012, 27, 074002.
116. Cassir, M.; Ringuede, A.; Niinistö, L. J. Mater. Chem. 2010, 20, 8987.
117. Meng, X.; Yang, X.-Q.; Sun, X. Adv. Mater. 2012, 24, 3589.
118. Kumagai, H.; Toyoda, K.; Kobayashi, K.; Obara, M.; Iimura, Y. Appl. Phys. Lett. 1997, 70, 2338.
119. Ishii, M.; Iwai, S.; Kawata, H.; Ueki, T.; Aoyagi, Y. J. Cryst. Growth 1997, 180, 15.
120. Riihelä, D.; Ritala, M.; Matero, R.; Leskelä, M. Thin Solid Films 1996, 289, 250.
121. Wang, J. J.; Deng, X.; Varghese, R.; Nikolov, A.; Sciortino, P.; Liu, F.; Chen, L.; Liu, X. J. Vac. Sci. Technol. B 2005, 23, 3209.
122. Kochergein, V.; Foell, H. Mater. Sci. Eng. R 2006, 52, 93.
123. King, J. S.; Gaillot, D. P.; Graugnard, E.; Summers, C. J. Adv. Mater. 2006, 18, 1063.
124. Rugge, A.; Becker, J. S.; Gordon, R. G.; Tolbert, S. H. Nano Lett. 2003, 3, 1293.
125. King, J. S.; Neff, C. W.; Summers, C. J.; Park, W.; Blomquist, S.; Forsythe, E.; Morton, D. Appl. Phys. Lett. 2003, 83, 2566.
126. Jefimovs, K.; Laukkanen, J.; Vallius, T.; Pilvi, T.; Ritala, M.; Meilahti, T.; Kaipiainen, M.; Bavdaz, M.; Leskelä, M.; Turunen, J. Microelectron. Eng. 2006, 83, 1339.
127. Solehmainen, K.; Kapulainen, M.; Heimala, P.; Polamo, K. IEEE Photon. Tech. Lett. 2004, 16, 194.
128. Jefimovs, K.; Vila-Comamala, J.; Pilvi, T.; Raabe, J.; Ritala, M.; David, C. Phys. Rev. Lett. 2007, 99, 264801.
129. Ferguson, J. D.; Weimer, A. W.; George, S. M. Appl. Surf. Sci. 2000, 162–163, 280.
130. Wilson, C. A.; Grubbs, R. K.; George, S. M. Chem. Mater. 2005, 17, 5625.
131. Niskanen, A.; Arstila, K.; Ritala, M.; Leskelä, M. J. Electrochem. Soc. 2005, 152, F90.
132. Dameron, A. A.; George, S. M.; Garcia, P. F.; McLean, R. S. AVS 7th International Conference on Atomic Layer Deposition. San Diego, United States, 2007 (presentations
available on CD-ROM).
133. Pore, V.; Rahtu, A.; Leskelä, M.; Ritala, M.; Sajavaara, T.; Keinonen, J. Chem. Vap. Deposition 2004, 10, 143.
134. Kemell, M.; Pore, V.; Tupala, J.; Ritala, M.; Leskelä, M. Chem. Mater. 2007, 19, 1816.
135. Stoldt, C. R.; Bright, V. M. J. Phys. D: Appl. Phys. 2006, 39, R163.
136. Mayer, T. M.; Elam, J. W.; George, S. M.; Kotula, P. G.; Goeke, R. S. Appl. Phys. Lett. 2003, 82, 2883.
137. Scharf, T. W.; Prasad, S. V.; Dugger, M. T.; Kotula, P. G.; Goeke, R. S.; Grubbs, R. K. Acta Mater. 2006, 54, 4731.
138. Hoivik, N. D.; Elam, J. W.; Linderman, R. J.; Bright, V. M.; George, S. M.; Lee, Y. C. Sens. Actuators A 2003, 103, 100.
139. Herrmann, C. F.; DelRio, F. W.; Miller, D. C.; George, S. M.; Bright, V. M.; Ebel, J. L.; Strawser, R. E.; Cortez, R.; Leedy, K. D. Sens. Actuators A 2007, 135, 262.
140. Pinna, N., Knez, M., Eds. Atomic Layer Deposition of Nanostructured Materials; Wiley-VCH: Weinheim, 2011; p 435.
141. Javey, A.; Kim, H.; Brink, M.; Wang, Q.; Ural, A.; Guo, J.; McIntyre, P.; McEuen, P.; Lundstrom, M.; Dai, H. Nature Mater. 2002, 1, 241.
142. Farmer, D. B.; Gordon, R. G. Nano Lett. 2006, 6, 699.
143. Xiang, J.; Lu, W.; Hu, Y.; Yan, H.; Lieber, C. M. Nature 2006, 441, 489.
144. Hwang, J.; Min, B.; Lee, J. S.; Keem, K.; Cho, K.; Sung, M.-Y.; Lee, M.-S.; Kim, S. Adv. Mater. 2004, 16, 422.
145. Keem, K.; Jeong, D.-Y.; Kim, S.; Lee, M.-S.; Yeo, I.-S.; Chung, U.-I.; Moon, J.-T. Nano Lett. 2006, 6, 1454.
146. Ras, R. H. A.; Kemell, M.; de Wit, J.; Ritala, M.; ten Brinke, G.; Leskelä, M.; Ikkala, O. Adv. Mater. 2007, 19, 102.
147. Kemell, M.; Pore, V.; Ritala, M.; Leskelä, M.; Linden, M. J. Am. Chem. Soc. 2005, 127, 14178.
148. Santala, E.; Kemell, M.; Pilvi, T.; Ritala, M.; Leskelä, M. Nanotechnology in Northern Europe Conference, Helsinki, 2006; Abstracts, p 104.
149. Leskelä, M.; Kemell, M.; Kukli, K.; Pore, V.; Santala, E.; Ritala, M.; Lu, J. Mater. Sci. Eng. C 2007, 27, 1504.
www.iran-mavad.com
150. Elam, J. W.; Routkevitch, D.; Mardilovich, P. P.; George, S. M. Chem. Mater. 2003, 15, 3507.
151. Pan, M.; Cooper, C.; Lin, Y. S.; Meng, G. Y. J. Membr. Sci. 1999, 158, 235.
152. Triani, G.; Evans, P. J.; Attard, D. J.; Prince, K. E.; Bartlett, J.; Tan, S.; Burford, R. P. J. Mater. Chem. 2006, 16, 1355.
153. Dücsö, C.; Khanh, N. Q.; Horváth, Z.; Bársony, I.; Utriainen, M.; Lehto, S.; Nieminen, M.; Niinistö, L. J. Electrochem. Soc. 1996, 143, 683.
154. Kemell, M.; Ritala, M.; Leskelä, M.; Ossei-Wusu, E.; Carstensen, J.; Föll, H. Microelectron. Eng. 2007, 84, 313.
155. Sander, M. S.; Côte, M. J.; Gu, W.; Kile, B. M.; Tripp, C. P. Adv. Mater. 2004, 16, 2052.
156. Daub, M.; Knez, M.; Goesele, U.; Nielsch, K. J. Appl. Phys. 2007, 101, 09J111.
157. Biener, J.; Baumann, T. F.; Wang, Y.; Nelson, E. J.; Kucheyev, S. O.; Hmza, A. V.; Kemell, M.; Ritala, M.; Leskelä, M. Nanotechnology 2007, 18, 055303.
158. Knez, M.; Kadri, A.; Wege, C.; Gösele, U.; Jeske, H.; Nielsch, K. Nano Lett. 2006, 6, 1172.
159. Huang, J.; Wang, X.; Wang, Z. L. Nano Lett. 2006, 6, 2325.
160. Knez, M.; Nielsch, K.; Niinistö, L. Adv. Mater. 2007, 19, 3425.
161. Liu, M.; Li, X.; Karuturi, S. K.; Tok, A. I. Y.; Fan, H. J. Nanoscale 2012, 4, 1522.
162. Marichy, C.; Bechelany, M.; Pinna, N. Adv. Mater. 2012, 24, 1017.
www.iran-mavad.com
www.iran-mavad.com
4.06 Laser Ablation
D Zhang, Huazhong University of Science and Technology, Wuhan, China
L Guan, Hebei University, Baoding, China
4.06.1 Laser Ablation Technology: Introduction and Physical Fundamentals 125

4.06.1.1 Laser Ablation 125
4.06.1.2 History of Laser Ablation Technology 126
4.06.1.3 Long-Pulsed Laser Ablation 128
4.06.1.4 Femtosecond Laser Ablation 130
4.06.2 Applications of Laser Ablation in Material Processing I 130
4.06.2.1 Laser Drilling and Cutting 130
4.06.2.2 Laser Welding and Laser Modification of Physical Properties of Functional Materials 133
4.06.2.3 Laser Surface Modification 133
4.06.2.3.1 Surface Hardening 134
4.06.2.3.2 Laser Cladding 134
4.06.2.3.3 Alloying 134
4.06.2.4 Femtosecond Laser-Induced Surface Periodic Structure and Nanogratings 135
4.06.3 Applications of Laser Ablation in Material Processing II 137
4.06.3.1 Microimage of Femtosecond Laser Interaction with Transparent Material 137
4.06.3.2 Applications of Femtosecond Laser Ablation of Transparent Materials 139
4.06.3.3 Applications of Femtosecond Pulsed Laser in Nanoparticle Formation 139
4.06.4 Pulsed Laser Ablation and Pulsed Laser Deposition Technology 143
4.06.4.1 Physical Picture of Pulsed Laser Deposition 143
4.06.4.2 Introduction to Plasma Expansion in Pulsed Laser Deposition 144
4.06.4.3 Introduction to Film Growth in Pulsed Laser Deposition 146
4.06.5 Thermodynamics of Laser Ablation 147
4.06.5.1 Theoretical Framework of the Thermodynamics of Long-Pulsed Laser Ablation – The Basic Equation
and Plasma Shielding Effect 148
4.06.5.2 Theoretical Framework of the Thermodynamics of Long-Pulsed Laser Ablation – Dynamic Physical
Parameters and the Vaporization Effect 150
4.06.5.3 Main Theoretical Results of the Thermodynamics of Long-Pulsed Laser Ablation 152
4.06.5.4 Femtosecond Laser Ablation Models: Classic and Improved Two-Temperature Equations 156
4.06.5.5 Femtosecond Laser Ablation Models: Unified Two-Temperature Equations and Density of State Effect 160
References 165
4.06.1 Laser Ablation Technology: Introduction and Physical Fundamentals

4.06.1.1 Laser Ablation
Laser ablation is the thermal or nonthermal process of removing atoms from a solid by irradiating it with an intense continuous
wave (CW) or pulsed laser beam. As one of the most important techniques for material processing, laser ablation can be used for
drilling extremely small, deep holes through very hard materials such as metals or diamonds, producing thin films or nanoparticles,
preparing material surface in a micro- and nano-controlled fashion, and so on.
The CW laser beam refers to a continuous output once the laser system is powered on, while the pulsed laser refers to a short time
(e.g., milliseconds to femtoseconds) output, as illustrated in Figure 1. The peak output power of a pulsed laser beam is much higher
than that of a CW laser beam, giving the same average output power.
When a solid surface is irradiated by a CW laser beam or long-pulsed (e.g., nanoseconds pulsed) laser beam, the material is
heated by the absorbed laser energy. The thermal motion of some particles is accelerated. Once the absorbed energy exceeds the
sublimation energy, these particles evaporate or sublimate and become vaporized particles; that is, part of the target is ablated. The
laser ablation rate N_ is defined as the vaporized particles per unit area per second:
_ ¼ rd=sm
N [1]
where s is the duration time of the laser pulse, r is the density of the target, d is the thickness of the ablated material, and m is the
average mass of ablated atoms.
However, for a short-pulsed laser (e.g., femtosecond laser), the laser-material interaction time is very short. Thus, the heat energy
has no time to diffuse in lattice. The irradiated zone of the material quickly reaches vaporization temperature and the ablated

www.iran-mavad.com
126 Laser Ablation
Peak output power
Average output power Average
Output power
Output power
& Peak output power output power
Time Time
(a) Continuous wave (CW) (b) Pulsed laser beam
laser beam
Figure 1 Illustration of the peak output power and the average output power of a CW laser beam and pulsed laser beam.
Figure 2 The characteristics of different laser ablation processes. Reproduced from Yang, J. J. Laser Optoelect. Prog. 2004, 41, 44 (in Chinese).
particles evaporate from the surface. The duration time of the pulse is much less than the time taken by excited molecules, atoms,
and electrons to release heat energy by moving or rotating motion (i.e., electron–phonon relaxation time), which is too short for
linear absorption processes such as the single-photon process, electron–phonon interaction process, and thermal diffusion process
to happen. Consequently, nonlinear absorption of laser energy (e.g., multiple-photon process) does occur during short-pulsed laser
ablation. The short-pulse laser process is a nonthermal melting process. The characteristics of different laser ablation processes are
shown in Figure 2 (1).
4.06.1.2 History of Laser Ablation Technology

The thermal effect of laser ablation is utilized in thin-film preparation, laser welding, laser surface treatment and modification, laser
cladding, drilling, cutting, and so on. In 1960, the first working laser, a ruby laser, was realized by Theodore Maiman at Hughes
Research Laboratories (2). In 1965, it was discovered that a laser could be used for diamond drilling using laser ablation mecha-
nisms (3). In 1967, British scientists started research on cutting titanium metal with a laser-assisted oxygen jet, which was later
widely used in aerospace projects. Due to its longer wavelength, CO2 laser was only used for nonmetal cutting (3). In the 1960s and
1970s, laser equipment consisted mainly of CW or long-pulsed lasers such as CO2 and Nd:YAG, which are typical high-power laser
systems. Figure 3 presents the main process of laser evaporation by thermal effect. Once the incident laser energy density exceeds the
threshold, the thermal effect leads to evaporation.
In the middle 1960s, it was found that in the process of laser beam irradiation, electrons, ions, and neutral atoms removed from
the solid surface form a glowing plasma near the surface with the temperature of 103–104 K. Then the ablated material is condensed
on the substrate, and the film is finally obtained. This is the initial principle of laser deposition (4,5). In 1965, Smith and Turner
deposited optical thin films by a ruby laser for the first time (6). This is one of the earliest pulsed laser deposition (PLD) techniques
for thin-film preparation. Limited to low laser energy density at this stage, the high quality of the deposited films was not achieved.
The long wavelength of the laser beam led to a deep melted liquid layer, where sputtering easily occurred. The sputtering produces
many droplets in the deposition process and strongly affects the quality of films.
In the 1980s to 1990s, with the development of lasers with high-energy density and short wavelength, the sputtering effect was
gradually reduced in film deposition. The PLD technique became one of the most successful film fabrication techniques (7,8). PLD
is a physical vapor deposition technique. Figure 4 gives a typical setup of PLD equipment. Under laser irradiation, the target is
www.iran-mavad.com
Laser Ablation 127
Figure 3 The process of laser ablation by thermal effect.
Figure 4 A typical experimental setup of pulsed laser deposition technique.
ablated and vaporized to form plasma. The plasma plume expands in vacuum or a background gas, and finally arrives at a substrate
surface to deposit a film.
In the 1970s, since lasers with sufficient power and energy density to melt metals were realized, laser welding technology was
developed (9). The Avco Everett Research Laboratory brought the world a laser surface modification machine for heat treatment of
metals in 1973 (10). The laser beams produced in this era had a relatively long wavelength and duration time. The interaction
between the laser and the target is mainly the linear absorption of energy via electron photon resonance, resulting in melting,
vaporization, and removal of material. Therefore, the ablation essentially depends on the optical and thermal properties of the
material. A heat effect zone is unavoidable around the processing zone, which causes heat stress and other defects. In the 1980s and
1990s, various laser techniques were developed for many applications. As an example, laser cladding and PLD technology were fully
developed and made breakthrough progress in this period.
As laser power density is enhanced and pulse duration is shortened, a nonthermal effect was observed in laser ablation for
material processing. In the 1980s, ultraviolet excimer lasers with nanosecond or picosecond pulse width could accomplish many
precise processing tasks, such as drilling, etching, and surface heat treatment, on nonmetallic materials such as polymers and
ceramics (11,12). Each photon of an excimer laser has a large energy, which is enough to break the bonds of molecules or atoms to
generate plasma near the target surface. This is actually a photochemistry reaction rather than a thermal melting process, as shown in
Figure 5. Due to its short wavelength, the laser beam can easily focus on a tiny area, which greatly improves the processing precision.
In 1976, the dye laser for the first time was mode locked to produce pulses with durations as short as 0.3 ps (13). In 1981, Fork
et al. reported a colliding pulse mode-locking technique to produce a continuous and stable train of pulses shorter than 0.1 ps (14).
Asaki et al. utilized a self mode-locked technique to obtain an 11 fs pulse laser in 1993 (15). In 2000, the emergence of photonic
crystal fibers indicated a new stage of femtosecond laser development. The femtosecond laser has two prominent characteristics: (1)
the pulse width can be as short as a few femtoseconds, which is 10 trillion times smaller than the resolution limit of the human eye,
and (2) the transient power is extremely high (16). 200 TW peak power has been achieved experimentally by chirped pulse
amplification (17). The focused peak intensity of a femtosecond laser can run up to 1021 W cm2, or even higher (16,18).
The emergence of femtosecond laser technology further expanded the application field of nonthermal laser ablation. As
a nonthermal and ‘cold’ treatment technique, the femtosecond laser started a new era of laser material processing, showing many
advantages such as superprecision, high spatial resolution, versatility, and universality.
www.iran-mavad.com
128 Laser Ablation
Figure 5 The process of laser ablation by photochemistry effect.
Femtosecond laser technique is an effective way of fabricating nanoparticles and large-area regular nanogratings (19,20).
Femtosecond laser-induced periodic surface structures (LIPSSs) have potential applications in the photoelectronics, thermal radi-
ation source, and bio-optic devices fields (21). For example, the so-called black metal prepared by the femtosecond laser ablation
technique can absorb light of any wavelength.
4.06.1.3 Long-Pulsed Laser Ablation

When a laser irradiates the target surface, several physical processes including reflection, dispersion, absorption, and transmission
are involved, as shown in Figure 6. P0, Pb, PR, and Pz are the whole incident energy, the absorbed energy, the reflected energy, and
the transmitted energy, respectively. When laser energy absorbed by the target reaches a specific value (ablation threshold), the
ablation phenomenon emerges. Lasers with long wavelength are not suitable for metal ablation, as most of the laser energy will be
reflected by the metal surface. UV short-pulsed lasers with short wavelength, such as frequency multiplied Nd:YAG laser or excimer
laser, are good candidates for metal ablation. The infrared laser with long wavelength can be effective when ablating some special
materials (22).
According to the energy conservation rule, the energy components in Figure 6 can be related as:
P0 ¼ PR þ Pb þ Pz [2]
or
PR Pb Pz
þ þ ¼1 [3]
P0 P0 P0
R ¼ PR/P0, b ¼ Pb/P0, and z ¼ Pz/P0 are reflectivity, absorbance, and transmittance, respectively. Hence, eqn [3] can be rewritten as
R þ b þ z ¼ 1.
Laser ablation is a special interaction between laser radiation and matter, which depends not only on the laser parameters (e.g.,
the output power, the wavelength, and the radius of irradiation spot) but also on the physical properties of the material (e.g., optical
parameters, such as reflectivity, absorbance, and transmittance, and the thermal properties, such as thermal conductivity and specific
heat capacity). With a low power density laser, the target under irradiation will heat up. When the power density exceeds the
ablation threshold energy, at which evaporation starts at the target surface, the target is ablated.
Here is a detailed physical image of long-pulsed laser ablation (LPLA). When pulsed laser irradiates a solid target, most of the
laser energy is accumulated on the surface of the nontransparent target. Part of the laser power is absorbed by this thin layer under
the irradiated surface, which results in a continuous increase of the surface temperature. Simultaneously, part of the energy is
transported into the inner layer. Thus the thickness of the heated region increases. As the thickness increases, the temperature
Figure 6 During laser irradiation of the target surface, several physical processes, including reflection, absorption, and transmission, occur.
www.iran-mavad.com
Laser Ablation 129
gradient gets smaller and smaller, leading to a gradually decreased heat conduction rate until zero. That is, the laser heat energy can
only penetrate a thin layer of target. The thickness of this thin layer is defined as the thermal penetration depth.
If the laser energy density is high enough, the excited electrons transfer energy to the lattice by collisions and the target heats up.
Thus, the thermal motion of some of the atoms in the lattice is accelerated. These high-energy atoms can overcome the attraction of
their neighbors. Then, corresponding physical phenomena such as melting, vaporization, and boiling occur. Finally, a complicated
layer structure emerges near the target surface, which indicates the beginning of ablation. The vapor above the surface (including
atoms, molecules, some clusters, micrometer-size droplets, and solid particles) continuously absorbs laser irradiation, until it is
almost ionized. Thus, the ablated material in plasma state ejects from target. Subsequently, a plume is produced near the surface.
The temperature of the surface is dominated by the vaporization mechanism.
The particle density of the plasma plume near the target surface, called the corona region, is very high. The corona region absorbs
about 98% of the laser energy through the inverse bremsstrahlung (IB) absorption effect and photoionization (PI) effect, thus
shielding the target surface from laser irradiation. The heat conduction region is the region outside the corona region where the
particle density of the plasma is relatively low. The laser energy is not effectively absorbed in this region. Intense energy transport
phenomena occur near the target surface, that is, the thermal penetration layer, which is mainly solid-phase, together with liquid-
phase and gas-phase materials. This layered structure due to pulsed laser irradiation will extend deep into the target with time.
The actual physical process of the interaction between the laser and the target is much more complicated than the physical
processes mentioned above. There are complex physical processes such as electron excitation effects (such as, induced electron–hole
pairs), photoelectronic effects, atom or cluster emission, and so on (23,24).
When laser power density increases to 1010 W cm2, the energy of the atoms in the melted target surface is dramatically
increased, which usually leads to the boiling phenomenon. Due to the absence of a vaporization nucleus coupled with the
extremely rapid melting process, the temperature of the melting layer rises suddenly. The target does not boil, although the
temperature exceeds the boiling point. This is the so-called superheating phenomenon. The superheated melt is in a metastable
state. Any small disturbance, such as density perturbation or impurity defect formation, can cause an explosive boiling. This
abnormal boiling phenomenon is called a phase explosion, which is an important topic in pulsed laser ablation (PLA) research
(25–30).
The ablation evaporation is intrinsically different from normal evaporation, since a Knudsen layer is generated in the corona
region during laser ablation (31,32). If the particle density of the vapor is low and the collision between the particles is negligible,
only a normal evaporation phenomenon occurs. However, if the laser power density increases to 109 W cm2 (a typical parameter
for ablation), the density of the vapor particles can reach 1016–1027 cm3. In this case, frequent collisions between the ablated
particles result in the highly preferential distribution of the particles along the perpendicular direction of target surface. The
collisions take place within a few mean free paths away from the surface, which is defined as the Knudsen layer. The velocities of
various particles tend to be the same in this layer, as the particles frequently collide with each other due to the high density. In fact,
this is why the PLD technique can achieve the stoichiometric deposition of films. The presence of the Knudsen layer makes it
possible for approximately the same flight time of different particles in the atmosphere. The related physical process is illustrated in
Figure 2(a).
The physical process of LPLA can be briefly described as followed. When the high-power pulsed laser irradiates the target, the
laser energy is absorbed by the target surface. The surface is then melted and vaporized, which forms a high-temperature, high-
pressure plasma. Therefore, the target under interaction with a pulsed laser can be roughly divided into three separate parts: the
high-temperature and high-pressure plasma, the liquid-phase region, and the solid-phase region, as illustrated in Figure 7.
For various materials, irreversible ablation requires that the laser energy density reaches or exceeds the ablation threshold. For
the interaction between the long-pulsed laser and the material, avalanche ionization brings about a final ablation, which is
mainly determined by doped impurities and various defects. Therefore, the ablation threshold for LPLA is different for different
materials.
Figure 7 Illustration of the regions in a laser-irradiated solid. (a) Unaffected region, (b) laser ablating region, (c) high-temperature and high-density
plasma, and (d) the transparent region for laser.
www.iran-mavad.com
130 Laser Ablation
4.06.1.4 Femtosecond Laser Ablation

The physical image of femtosecond laser ablation is quite different from nanosecond laser ablation. When the pulse duration time is
in the order of a femtosecond, the impact of the relaxation time of electron–phonon interaction, which is in the order of picosecond,
should be considered. When the femtosecond laser focuses on the surface of a bulk, the photon energy is first absorbed by electrons,
leading to an accelerated thermal motion of the electrons and a dramatic temperature increase of the electron subsystem. However,
during the short pulse, the electrons have no time to transfer the obtained energy to the lattice (or the ions). At this time, the
temperature of the electron gas is very high, while the lattice subsystem’s temperature keeps relatively low. Thus, a ‘cold’ ablation
process occurs. Usually, the thermal equilibrium time of the metal is roughly a few femtoseconds, compared to the order of ns for
semiconductors or insulators. This is in fact the relaxation time of electron–phonon interaction. The relaxation time difference is
a result of the energy gap difference. The larger the energy gap, the longer it takes for the excited electron–hole pairs to reach thermal
equilibrium through generation-recombination (34). In the short period of pulse, there are two temperature subsystems in the
target: the electronic subsystem and the lattice subsystem. These two subsystems can be described using two electron–phonon
coupling thermal conduction equations, named as the two-temperature model for femtosecond laser ablation.
In contrast to LPLA, electrons produced in femtosecond laser ablation depend mainly on multiphoton ionization or tunneling
ionization mechanisms (see Section 4.06.3.1). The dopants and defects in the lattice only have a very small influence on generation-
recombination of electron and vacancy pairs. Therefore, the ablation thresholds for different materials are similar for femtosecond
laser ablation (35). Assuming the same energy density and repetition frequency, the pulse laser with shorter pulse width and hence
higher peak power easily satisfies the ablation threshold of target.
Figure 2(b) demonstrates the main physical processes of femtosecond laser ablation. The features are summarized here:
1. Nonthermal processing. This reduces or even eliminates some defects (the roughness of the ablation edge and large crack
formation) in the traditional long-pulsed laser material processing.
2. High-precision processing. Suppression of the heat conduction effect and the hydrodynamics effect dramatically improves the
working precision of femtosecond laser ablation.
3. Accurate three-dimensional spatial processing. When a femtosecond laser with a focused intensity close to the ablation threshold
irradiates transparent materials, the low-beam intensity at the positions other than the focus point cannot meet the requirement
for the nonlinear absorption of multiphotons. Thus, the laser beam arrives at the focal point inside the material without any
attenuation. The high energy accumulated here induces multiphoton absorption and ionization. Consequently, femtosecond
laser ablation can be precisely controlled at any focused position inside a three-dimensional transparent material, leading to
a highly accurate spatial selection and positioning ability.
4. Universality of femtosecond laser material processing. The extremely high peak intensity induces only multiphoton absorption.
The ablation is dominated by laser intensity, suggesting an invariable ablation threshold. The multiphoton absorption and
ionization threshold mainly depend on the atoms rather than on the density of the electrons in the material. Therefore,
a femtosecond pulse laser can theoretically accomplish precision processing, repair, and treatment for any kind of
materials.
4.06.2 Applications of Laser Ablation in Material Processing I

4.06.2.1 Laser Drilling and Cutting
The two remarkable applications of laser ablation are laser drilling and cutting. The use of lasers for drilling and cutting was first
developed in the 1960s. These two techniques have the longest history and the most mature technology for laser material
processing.
Figure 8 represents the main steps for laser drilling. Two physical processes are involved: (1) The temperature of irradiated
material exceeds the melting point, and surface vaporization occurs; (2) the internal pressure of the vapor is so great that the ablated
material is sputtered. Laser drilling requires incident laser energy above the ablation threshold. Thus, high-power pulsed lasers are
extensively used in industry. The pulsed laser has many advantages, such as short interaction time, small heat effect zone, and
precise dimension control.
The factors that affect the quality of laser drilling are the laser parameters (laser power density, irradiation spot area, incident
angle, pulse width, and pulse repetition frequency), the parameters of the desired hole (diameter, shape, and depth), and the
physical properties of the material (absorption coefficient, reflectivity, and thermal conductivity). To accomplish a high-quality
hole, the material with low-heat conductivity (ensuring less dissipation around the drilled hole), low reflectivity, and high
absorbance (ensuring much laser energy absorbed by target) is preferred. Figure 9 shows micrographs of the laser-drilled holes using
a pulsed laser with the same wavelength and different pulse widths. It is obvious that as the pulse width reduces, the quality of the
holes improves, while the power required decreases (36). The process of Figure 9(c) is the best. This is processed by femtosecond
laser drilling, and it keeps a sharp, clean edge and a regular shape. The worst is the one drilled by nanosecond laser, as shown in
Figure 9(a).
Nowadays, many kinds of materials including metals, semiconductors, ceramic, and polymers dielectric materials can be drilled
by laser. As mentioned earlier, metals are suitable for drilling by short-wavelength lasers, such as excimer lasers (36). For most
ceramic materials with a low reflectivity for infrared lasers, CO2 laser is a good candidate for ceramic drilling (37,38). However, for
www.iran-mavad.com
Laser Ablation 131
(a) (b)
(c) (d)
Figure 8 Schematic diagram of laser drilling processing: (a) unstructured, (b) Cu opened, (c) dielectric opened, and (d) hole formation.
Figure 9 Micrographs of the laser-drilled holes on steel. The wavelength is 780 nm. The pulse widths are (a) 3.3 ns, (b) 80 ps, and (c) 200 fs,
respectively. Reproduced from Chichkov, B. N.; Momma, C.; Nolte, S.; Alvensleven, F. V.; Tunnermann, A. Appl. Phys. A 1996, 63, 109.
polymers, phase transitions (e.g., solid / melt / vapor) and chemical degradation often occur at high temperatures (39). Unlike
most metals, the polymer materials under laser irradiation can produce a deeper thermal penetration. Lawrence et al. confirmed that
for CO2 laser, the absorption length for Al2O3/SiO2-based refractory material is three orders of magnitude greater than most of
metals (38,40). Generally, both polymers and ceramics have a deeper heat penetration, so the heat transport theoretical model
considering a volumetric heating for drilling is more accurate than a model only considering surface heating (41–43).
Laser cutting is a technology that uses a high-power laser to scan material surface, and the material is heated in an extremely short
time, leading to the temperature rises up to 103 or 104 K, then the melting and vaporization phenomena occur. The ablated material
is blown away by an assisted jet gas, and finally a cut edge with a high quality is left. Figure 10 is a schematic diagram of the
Figure 10 A typical experimental setup of laser cutting technique.
www.iran-mavad.com
132 Laser Ablation
Figure 11 Configuration of the laser-cutting system with the dual-laser-beam method. Reproduced from Jiao, J. K.; Wang, X. B. Opt. Laser Eng.
2009, 47, 860.
laser-cutting technique. The experimental system usually includes a laser device, a laser delivery system, an assisting gas output
device, and a computer-assisted work-piece or laser movement system.
Initially, laser technology was utilized to cut metals, and gradually its applications extended to the cutting of various materials
such as glass and wood, as well as polymers. A typical laser beam used for cutting needs a power intensity of 1000–2000 W. In the
cutting process, the quality of the cutting edge depends on some parameters at different extent, such as the output mode (pulsed or
CW) of laser, power density, pulse repetition frequency, cutting speed, type and pressure of assistant gas, as well as the type and
thickness of the workpieces (44).
Compared to conventional methods, the laser-cutting technique relies on the thermal rather than mechanical properties of
materials, avoiding mechanical damage of the workpiece. Laser cutting is noncontact, which can deliver a finer finish without
cracks. Therefore, computer-assisted laser cutting is widely applied in industry to continuously and precisely accomplish any
desired shape, although the machining process sensitively depends on the physical properties of materials and laser
parameters.
The most commonly used lasers for cutting are Nd:YAG laser and CO2 laser (45). CO2 lasers with high average power and
better beam quality are suitable for cutting thick materials except for metal workpieces with high reflectivity for the longer
wavelength. Nd:YAG lasers have a shorter wavelength, which is able to process highly reflective materials, such as titanium and
titanium alloy (46,47). Lv. Shanjin et al. (47) investigated the influence of the laser-cutting parameters, for example, pulse
energy, pulse rate, cutting speed, and the type and pressure of gas on the heat-affected zone (HAZ) and surface morphology of
titanium alloy.
The CO2 laser is frequently applied to cut glass. There are two main methods for glass cutting: the controlled-fracture laser-
cutting technique and the evaporative laser-cutting technique. The key problems in glass cutting involve how to improve the control
precision, how to reduce the thermal stress generated during glass cooling, and how to avoid the fracture propagation along an
unexpected path. Recently, a dual-laser-beam method was proposed to cut glass substrates in order to further improve the cutting
quality. In this method, a focused laser beam was used to scribe a straight line on the substrate, and a defocused laser beam was used
to irradiate on the scribing line to generate a tensile stress and separate the substrate. With this method, the fracture propagates
stably, and the glass substrate can be separated along the desired path (48,49). The configuration of the laser-cutting system with the
dual-laser-beam method is illustrated in Figure 11.
The important procedures in laser cutting are: (1) controlling the focus position to obtain an accurate focus and small spot
diameter, and (2) adjusting the nozzle parameters to ensure an appropriate direction and flux of the assisting gas. To avoid the
oxidizing reaction and molten dross formed in the process, high-pressure and high-purity nitrogen or inert gases as assisting cutting
gases have been used in laser cutting. Hong Lei et al. (50) used a CO2 laser with a power of 1800–4000 W to cut silicon steel sheets.
In their experiments, a cyclone slag separator was used to form dross-free cutting kerfs (Figure 12).
Figure 12 Photography of the cutting kerfs: (a) by using a cyclone slag separator and (b) by traditional laser-cutting technology. Reproduced from
Hong, L.; Zhang, Y.; Mi, C. L. Opt. Laser Technol. 2009, 41, 328.
www.iran-mavad.com
Laser Ablation 133
4.06.2.2 Laser Welding and Laser Modification of Physical Properties of Functional Materials
As a high-efficiency and inexpensive technology, lasers are widely used to weld metals or alloys in industry. In 1970, the CO2 laser
was first utilized to weld metal, but thermal damage was easily generated during welding due to the long wavelength of the CO2
laser and the low absorbance of the material (9). In the following years, the Nd:YAG laser with shorter pulse width and higher peak
power was employed to weld metals. The metal surface can effectively absorb the laser energy to achieve a deeper heat penetration,
and thus the strength of the welded joint is remarkably improved (51). It was found that the Nd:YAG laser could produce a minimal
HAZ and less thermal damage. With the introduction of fibers into laser technology, a more flexible delivery system can produce
a higher-efficiency and better-quality welding.
When a laser beam hits a metal surface, the material at the irradiation spot is heated up to the melting and vaporization
temperature, and then vaporization occurs. As a result, several unexpected blowholes may be left on the metal surface. How to
obtain a desired ‘deep-weld effect’ is a key problem in laser welding technology. Because metals have a temperature-dependent
absorption coefficient and a high reflectivity (40,41), laser energy cannot be effectively absorbed by the workpiece, resulting in
unexpected thermal damage on the surface. Recently, CW laser matched with an appropriate pulsed laser can perform dual-beam
laser welding (51) shown in Figure 13. It was found that dual-beam laser welding could markedly reduce or avoid the formation of
the blowholes in the welded joints.
The welding efficiency 3 can be expressed as follows:
3 ¼ ½ydWDHm =P [4]
where P is the incident laser power, y is the speed of welding, d is the thickness of weldment, W is the laser beam width, and DHm is
thermal enthalpy at the melting point. For thermal penetration welding, the welding efficiency 3 is 0.48, while for heat conduction
welding, 3 is 0.37 (52). The parameters, such as pulse width, power density of laser, irradiation spot area, absorption coefficient, and
the thickness of weld, substantially influence the quality of weld. Compared to the conventional arc welding technique, laser
welding can lead to a deeper penetration, a smaller HAZ, and a rapid cooling rate. However, the same problem in these methods is
still that the internal residual stress caused by heat penetration cannot be fully released, affecting the physical properties of the
workpiece.
In the 1990s, it was found that the physical properties of materials can be modified by laser irradiation. Table 1 lists the recent
progress on the modification of the electrical, optical, and magnetic properties of functional materials, including semiconductors,
conductors, superconductors, and magnetic materials.
4.06.2.3 Laser Surface Modification

Conventional surface modification techniques, for example, ion carburizing, boriding, and nitriding, are conditioned by the
incident energy of ions. The common disadvantages of these techniques are a shallow strengthened layer, a long processing period,
a high required temperature, and an easy transformation of the workpiece. Laser surface modification is noncontact processing by
high-power laser beam heating the material surface, and the subsequent cooling takes place through the heat conduction of the
material itself. The characteristics of laser surface modification are as follows: (1) the convenient delivery of laser energy, which
easily allows selective positioning on the surface for strengthening, and (2) a concentrated laser energy acting on the material, giving
Figure 13 Schematic of the experimental setup of the dual-beam laser welding technique. Reproduced from Yan, S.; Hong, Z.; Watanabe, T.; Tang, J. G.
Opt. Laser Eng. 2010, 48, 732.
www.iran-mavad.com
134 Laser Ablation
Table 1 Applications of the modification of physical properties of materials by laser irradiation
Irradiated films Modification of properties Wavelength (nm) References
Al2O3 Increasing the resistivity of surface 10640 (53)

YBCO Enhancement of electrical conductivity 248 (54)
ITO Enlarged grain size and reduced resistivity 248 (55)
n-GaAs Decrease of photoconductivity and resistivity 694 (56)
SnO2 Improved Hall mobility and refractive index 1064 (57)
Mg-doped GaN Enhanced fluorescence intensity of blue light 10640 (58)
Doped ferrite Improved magneto-optical effect 632.8 (59)
ZnO Reduced UV emission intensity 248 (60)
n-type ZnO Improved electrical conductivity 248 (61)
Ta2O5 ceramic Enhanced dielectric permittivity 10640 (62)
La0.67Ca0.33MnO3 Improved electrical and ferromagnetic properties 10640 (63)
ZnO Improved UV emission intensity 248 (64)
short processing time, a small HAZ, and a small deformation of the workpiece. This technique is suitable for dealing with some
sheet metals but not with thicker sheets. The laser surface modification process includes surface hardening, laser cladding, and
alloying.
4.06.2.3.1 Surface Hardening

Surface hardening, also known as surface heat treatment, includes transformation hardening and shock hardening. Transformation
hardening is a heat-treatment process by high-power laser beam irradiation on the surface of metals and other materials. The
temperature of the heated material reaches the phase transition point, but melting does not occur. After the end of the laser pulse,
the metal begins to quench due to the cooling of the material itself. Compared to conventional heat treatment, the laser modifi-
cation process induces an increase of 15–20% in the hardness of the material. For shock hardening, a high-strength shock wave or
stress wave is generated by the pulsed laser acting on the material surface, and then a strong plastic deformation occurs on the metal
surface. In the laser-impacted zone, the microstructure of the material shows a network of dislocation tangles, which is similar to the
substructure produced by an explosive shock or fast plane shock, significantly improving the surface hardness, yield strength, and
fatigue life of the material. In fact, transformation hardening relies on the thermal properties of materials, while shock hardening
relies on the dynamics properties of materials.
4.06.2.3.2 Laser Cladding

Under laser beam irradiation, a coating material on the substrate surface will melt. Then, when the beam is removed, the molten
coating material will rapidly cool (the cooling rate up to 103–108 K s1) and then be solidified and combined with the matrix
material. In the cladding process, the crystal grain is refined, the composition segregation decreases, and the defect formation is
reduced. A higher compressive stress exists in the solidified structure, which greatly leads to surface wear resistance and antifatigue
properties. The laser cladding process is used mainly for cast iron and alloy materials. For example, Ni-based powder was selected as
laser-surface-modifying material to improve the wear resistance of copper by CW CO2 laser (65).
4.06.2.3.3 Alloying
When a stoichiometric alloy coating on a metal substrate surface is irradiated by a high-energy laser beam, then a mixed melting
occurs within a very short time to form a new surface alloy. Because of a fast heating, the composition, structure, and properties of
the workpiece only have a slight difference in the melted and heat-affected zone, and the deformation of the workpiece also is small.
Finally, the elements of the alloy coating are fully dissolved in the substrate surface layer. Hence, the modified layer has a uniform
composition, low porosity, and dense structure. Experiments indicate that the energy and the moving speed of the laser beam on the
workpiece surface can influence the composition of the alloy (66).
An excimer laser having a high photon energy is often used for surface modification of various materials, such as pure metals or
alloys, glass, ceramics, plastics, and many wide-gap materials (67,68). For many wide-gap materials, an ultraviolet laser is more
suitable for surface modification than other lasers. Reference (66) reported that titanium nitride was modified by the laser beams,
with an energy density of 20.0 J cm2 (TEA CO2) and 2.4 J cm2 (XeCl laser), respectively. The results show that the energy absorbed
from the CO2 laser is mainly converted into thermal energy, causing melting, vaporization, and shock waves in the vapor. Energy
absorbed from the XeCl laser leads primarily to a quick and intense target evaporation. Thus, XeCl laser-induced target effects are
localized and confined to the target surface and its vicinity. With the use of the excimer laser treatment, the corrosion resistance of
the Ti-6Al-4V alloy is significantly improved (69). Experiments show that the reflectivity of metal or metal alloy is affected by laser
wavelength, leading to a different ablation threshold. Figure 14 indicates that the reflectance spectrum of the Ti-6Al-4V alloy surface
and a comparative sum-up of effects produced by the three laser wavelengths 1064, 534, and 266 nm of Nd:YAG laser is given in
Table 2 (70). According to the results, it can be seen that laser surface modification of metals significantly depends on laser
wavelength.
www.iran-mavad.com
Laser Ablation 135
65
60
1064 nm
55
50
Reflectance (%)
45 532 nm
40
35
266 nm
30
25
1000 800 600 400 200
Wavelength (nm)
Figure 14 Reflectance spectrum of the nonirradiated Ti-6Al-4V surface, with laser wavelengths indicated. Reproduced from Milovanovic, D. S.;
Radak, B. B.; Gakovic, B. M.; Batani, D.; Momcilovic, M. D.; Trtica, M. S. J. Alloy Compd. 2010, 501, 89.
Table 2 A comparative sum-up of effects produced by the three laser wavelengthsa
Reflection Damage threshold Periodic surface structures

.
Laser at 1064 nm 57.00% 0.9 J cm2 Nanometer scale: parallel waves, tto .
E , period: 800 nm after 30 pulses
Laser at 532 nm 42.70% 0.25 J cm2 Nanometer scale: parallel waves, tto .
E , period: 400 nm after 50 pulses
Laser at 266 nm 29.00% 0.12 J cm2 Nanometer scale: parallel waves, tto E , period: 200 nm after 8 pulses
a
Reproduced from Milovanovic, D. S.; Radak, B. B.; Gakovic, B. M.; Batani, D.; Momcilovic, M. D.; Trtica, M. S. J. Alloy Compd. 2010, 501, 89
The integrated surface modification technology by laser combined with other modified methods is growing rapidly. Reference
(71) reported that titanium alloy matrix can be treated by glow discharge plasma nitriding and laser remelting processing at the
same time. Composite gradient coatings containing TiN and Ti2N were prepared, and thus the surface morphology of titanium was
remarkably improved and cracks were reduced.
4.06.2.4 Femtosecond Laser-Induced Surface Periodic Structure and Nanogratings

Femtosecond LIPSSs have become a focus in this field (72–74). The ripples or nanogrooves have a period on the scale of the incident
laser wavelength, and the orientation of these periodic structures is perpendicular to the polarization of the incident laser. The
mechanism of LIPSSs has been extensively studied by many scientists in the past several decades, and some theoretical models have
been established, such as the uneven energy distribution model (75,76), general surface scattering (GSS) model (77,78), Coulomb
explosion model (79), photon resonance absorption model (80), Boson condensation model (81), and energy accumulation effect
model (82).Of these models, the most famous one is the GSS model.
The GSS model to explain the nature of the observed periodic structure was proposed by M. Oron and co-workers (77,78). The
incident laser light is partially scattered by surface defects and produces a tangential wave propagating along the surface. The formation
mechanism of LIPSSs is the interference at the air-material (or vacuum-material, etc.) interface between the incident laser light and the
matter polarization associated with laser-induced surface excitation. Predictions given by the GSS model were in good agreement
with experimental observations up to nanosecond pulse duration, but for the femtosecond pulse laser, a morphology emerged that
was not predicted by the theory (74), implying that other processes were taking place. In 1992, M. Bonch-Bruevich et al. (83) modified
the GSS model and showed that LIPSS is induced by interference between the incident femtosecond laser with the excited plasmons
on the surface. In the modified model, the separation of period structure for linear polarization laser can be expressed as:
l
d¼ [5]
h sin q
where h ¼ Re[3/(3 þ 1)]1/2 is the real part of the effective refractive index of the surface plasmon, and 3 is the dielectric constant
of the metal.
www.iran-mavad.com
136 Laser Ablation
In 2002, Jurgen Reif et al. (79) studied femtosecond laser ablation of wide band gap insulators and presented a Coulomb
explosion model. They observed a complex structure of fine ripples at the bottom of the ablated crater, and they assumed that the
ripples structures were due to self-organized structure formation during the relaxation of the highly nonequilibrium surface after
explosive positive ion emission. This model explained that for dielectric material, LIPSS with a period less than laser wavelength
scale was produced by multipulse ablation with pulse energy less than the threshold energy.
Three-dimensional, periodic nanowriting on diamond clusters was reported (84). Periodic ripples have been observed on single-
crystal and polycrystalline diamond surfaces, shown in Figure 15. Further, it has been experimentally shown that the periodicity of
these corrugated two- and three-dimensional structures is shorter than that of the laser wavelength used (248 nm for the excimer
femtosecond laser and 825 nm for the Ti: sapphire femtosecond laser). In 2006, B. Tan et al. (85) found that a femtosecond LIPSS
was formed on the surface-polished crystalline silicon. Unlike the patterns formed by a continuous or nanosecond pulsed laser, the
spacing of the ripple formed by femtosecond pulses was not affected by the incident angle of the laser beam. Also, the pulse
repetition rate had no impact on the ripple formation on crystalline silicon. The orientation of these periodic structures was
perpendicular to the vector of the electric field of the laser beam.
Wang and Guo et al. (86,87) investigated periodic structures on the surfaces of three different noble metals, Cu, Ag, and Au,
following femtosecond laser radiation. Under identical experimental conditions, laser-induced surface patterns show a higher
morphological clearness on the metal with a larger electron–phonon energy coupling coefficient. Angular dependence of the period
of these structures was studied, and the results indicated that these structures were formed by the interference between the incident
light and the excited surface plasmons.
A. Y. Vorobyev et al. (88) performed a detailed study of the formation of LIPSSs on tungsten at near-damage threshold fluences
and found a unique type of LIPSS entirely covered with nanostructures, illustrated in Figure 16.
In Figure 17, periodic nanostructures were observed on the surface of ZnSe after irradiation by a femtosecond Ti: sapphire laser,
which was aligned perpendicularly to the laser polarization direction (89). With the laser polarization parallel to the moving
direction, long-range gratings by slowly moving the crystal were produced. Huang and Xu et al. reported that by the simple scanning
technique with appropriate irradiation conditions, arbitrary size of uniform nanograting could be produced on wide band-gap
materials as well as graphite (90).
In 2009, Eric Mazur et al. induced periodic linear grooves with 40 nm wide, 500 nm deep, and up to 0.3 mm long synthetic
single-crystal diamond with femtosecond pulses at 800 nm (20). L. Sudrie et al. (91) prepared permanent birefringent
5 μm 2 μm
Figure 15 Three-dimensional periodic ripples induced by a femtosecond laser on single-crystal and polycrystalline diamond surfaces. Reproduced
from Ozkan, A. M.; Malshe, A. P.; Railkar, T. A. Appl. Phys. Lett. 1999, 75, 3716.
(a) (b)
542 nm
2 μm 500 nm
λ =800 nm λ =800 nm
Figure 16 Femtosecond laser-induced periodic surface structures on W surface. Reproduced from Vorobyev, A. Y.; Guo, C. J. Appl. Phys. 2008,
104, 063523.
www.iran-mavad.com
Laser Ablation 137
(a) (b) (c)
NONE SEI 10.0kV X 20.000 1μm WD 8.2mm
Figure 17 SEM images of gratings at the angle 0 (a), 45 (b) and 90 to the moving direction of laser and its polarization direction.
(a) (b)
d = 120 μm d = 430 μm
Figure 18 Photographs of the titanium samples. (a) The sample with a groove period of 120 mm. (b) The sample with a groove period of 430 mm.
Reproduced from Vorobyev, A. Y.; Topkov, A. N.; Gurin, O. V.; Svich, V. A.; Guo., C. L. Appl. Phys. Lett. 2009, 95, 121106.
structures with controllable microscopic dimensions inscribed in pure fused silica platelets. The birefringence properties of
transmission gratings and of a quasi-uniform layer have been established. S. H. Cho et al. fabricated an internal diffraction
grating with photo-induced refractive index modification in planar silica plates by a high-intensity femtosecond (150 fs) Ti:
0
sapphire laser (lp ¼ 790 nm). The low-density plasma formation causes the refractive index modification with a SiE center
defect (92).
Self-organized nanograting periodical structures were induced by an ultrashort intense pulsed laser, which effectively changed
the absorption and reflective properties of metal surfaces. Guo et al. investigated femtosecond laser-induced grating structures with
different periods on titanium surfaces (72–74,93). They found that by using a femtosecond laser structuring technique, near-
perfectly reflective metals are transformed to highly absorptive over an ultrabroad electromagnetic spectrum, ranging from ultra-
violet to terahertz, indicated as Figure 18. The so-called ‘black metal’ phenomenon can be applied to the metal shell surface of
military weapons, allowing weapons to achieve effective stealth.
4.06.3 Applications of Laser Ablation in Material Processing II

4.06.3.1 Microimage of Femtosecond Laser Interaction with Transparent Material
From a microscopic point of view, laser–solid interaction is achieved via the electrons absorbing energy from photons and being
excited from the initial equilibrium state. The absorption mechanisms include single-photon resonance transition, two-photon and
higher-order multiphoton transitions, tunneling ionization, and above-barrier ionization (94). Typically, the single-photon tran-
sition mechanism plays a dominant role due to a larger cross-sectional area.
The interaction between femtosecond laser pulses and transparent materials has a unique nature. The key lies in the transparent
dielectric material usually having a wider band gap, which is much larger than the visible and near-infrared single-photon energy, so
the linear absorption of solid from a low-intensity laser is weak. For example, a wavelength of an 800 nm fs laser, single-photon
www.iran-mavad.com
138 Laser Ablation
energy is 1.5498 eV, far less than the band gap of some transparent materials, such as glass. It is not possible for electronic transition
from the valence band to the conduction band by single-photon absorption, that is, linear absorption, and consequently color
centers cannot be produced in glass. Femtosecond laser-induced color centers in glass are attributed to nonlinear effects caused by
multiphoton absorption, which requires the intensity of a laser exceeding w1 GW cm2.
When a high-intensity femtosecond laser interacts with transparent dielectrics, the bound electrons in solid absorb incident
multiphoton energy and their kinetic energy exceeds the ionization potential energy to become free electrons by jumping from the
valence band to the conduction band; this process is called multiphoton ionization, namely, multiphoton absorption, as shown in
Figure 19. Multiphoton ionization is a multi-order and nonlinear process, and the reaction cross section is very small. Thus, only an
intense laser with a high-photon density could give rise to such a reaction. Furthermore, if this ionization process shows avalanche
characteristics, this process will be called avalanche ionization.
The mechanisms of femtosecond laser interaction with transparent materials can be described by several theoretical models,
such as, the micro-explosion model (95), the Coulomb explosion model (96–98), the avalanche breakdown model (99), and the
threshold model (100). Early studies suggest that avalanche ionization is the main reason for ablation (101). However, femto-
second laser ablation of dielectric materials relies on multiphoton ionization or avalanche ionization, and there is some dispute
about the mechanism (102,103). Most researchers believe that the multiphoton ionization process first provides seed electrons in
the ablation process and subsequently leads to avalanche ionization to accelerate ablation (104). This is the physical meaning of the
so-called multiphoton collision ionization.
In 2001, Chris B. Schaffer of Harvard University presented that in femtosecond laser irradiation, the nonlinear ionization of
transparent materials includes three processes: multiphoton absorption, tunneling effect, and avalanche ionization. The three
processes can be determined by the Keldysh parameter (eqn [6]). They measured the ablation threshold of transparent materials
with a 110 fs pulse laser, and the damage threshold was found to depend on the laser wavelength and band gap of the material itself:
For the center wavelength of 400 nm, multiphoton ionization plays a dominant role in ablation, while for the wavelength of
800 nm, the tunneling effect does; for wide gap materials, the generation of free electrons is largely due to the role of avalanche
ionization; for narrow gap materials, PI provides a large number of free electrons (105). The Keldysh parameter g is defined as (106):
uð2m
Eg Þ1=2
g¼ [6]
eE
where m and e are effective mass and charge of electron, respectively, u and E are angular frequency and amplitude of the incident
electric field, respectively, and Eg is the band gap of solid dielectric. If the width of the band gap is narrow and the incident electric
field is strong, the Keldysh parameter g w1.5 and then tunneling ionization dominates the ablation process; if g w1.5, multi-
photon ionization does; if g w1.5, a mixture mechanism of tunneling ionization and multiphoton ionization. Three ionization
modes of femtosecond laser ablation of transparent material are indicated in Figure 20.
Figure 19 Schematic diagram of multiphoton ionization.
Figure 20 Photon ionization mechanism with different Keldysh parameters.
www.iran-mavad.com
Laser Ablation 139
4.06.3.2 Applications of Femtosecond Laser Ablation of Transparent Materials

A variety of nonlinear effects produced by ultrashort pulse laser interaction with transparent dielectric material led to many
microstructures induced in materials, suggesting that the ultrashort laser has a broad prospect for development of high-density
three-dimensional data optical storage, optical waveguides, photonic crystals, and so on.
The Mazur group at Harvard University first studied the micro-explosion process and the scale of the microcavity formed in
femtosecond laser interaction with transparent materials (94). Intense laser beam is focused inside the transparent material, and
then ultrahigh-temperature, high-pressure plasma is produced near the focus, causing micro-explosions and microcavity in the body
of the material, and the surrounding material becomes compact because of compression. The diameter of the microcavity formed in
explosions is less than the optical diffraction limit (94). Through multiphoton absorption, this femtosecond laser process induced
a microstructure with an ultra-diffraction limit, mainly applicable in optical storage, as shown in Figure 21.
The femtosecond laser can achieve three-dimensional precise positioning in the transparent material ablation. The principle is as
follows: The femtosecond laser with the intensity near-damage threshold is focused at any internal position of transparent material.
Due to the low-energy beam intensity, it cannot meet the material requirements for multiphoton nonlinear absorption, so the laser
beam has almost no attenuation before it reaches the focal point. Therefore, an accurate adjustment of the focal point is necessary
for more energy absorption to complete multiphoton absorption and ionization, accomplish ultra-precision machining of any parts
of the internal three-dimensional space of the transparent material, and induce the expected microstructure. For example,
a femtosecond laser can engrave a ‘micro bull’ inside transparent materials, with approximately 7 106 pixels with a spatial
resolution of about 150 nm, as shown in Figure 22 (107). Femtosecond laser-engraved ‘micro spider’ and ‘m-dragon’ models are
given in Figure 23 (108,109). Figure 24 illustrates that micro-tweezers with submicron probe tips are fabricated by a two-photon
microstereolithography system developed with a Ti: sapphire femtosecond laser (110).
When the laser beam is focused on amorphous materials, such as quartz glass and a variety of doped glasses, the material will
absorb more laser energy at the focal point, causing changes of lattice structure and the refractive index. This process is helpful for the
successful preparation of three-dimensional optical waveguides (105,111,112).
4.06.3.3 Applications of Femtosecond Pulsed Laser in Nanoparticle Formation

Droplets are often generated in the laser ablation stage of the PLD technique, leading to a reduced surface smoothness of films,
which is not expected in the film growth process. However, in the preparation of granular materials, especially for nanoparticles, the
PLA technique has great development potential. Since the mid-1990s, nanoparticles have been successfully synthesized in a gas
medium using laser ablation in the laboratory (113–119). This preparation technology has rapidly developed since the advent of the
femtosecond laser (120–123). Currently, according to the type of medium, the laser methods for the laser preparation of nano-
particles can be classified into two categories: one is the synthesis of nanoparticles in the atmosphere, caused by condensation of the
vapor during laser ablation of targets, and the other is the preparation of nanoparticles in a liquid medium. Many theoretical
mechanisms have been discussed to make the formation of nanoparticles clear, and most of them show that the nanoparticles are
formed during the plasma expansion and flight in gas or liquid medium. Of course, it is possible to synthesize particles after the
plasma reaches the substrate surface, and even directly produce nanoparticles in the jet of the ablation process (117,124,125). In
2000, F. Mafuné in the Clusters Research Institute of Japan’s Toyota investigated in detail the dynamics of metal nanoparticle
formation and consequently presented a theoretical model (126,127).
The synthesis process of nanoparticles by the laser technique can be simply described as follows. In the atmosphere or liquid
medium, the plasma generated during laser ablation continuously absorbs the laser energy and expands in the medium. Subse-
quently, various ions, atoms, or clusters interact with the atmospheric species in the deposition chamber, leading to the combi-
nation of the larger particles under certain technical conditions (Figure 25). The characteristic parameters of plasma such as
temperature and density are dependent on the thermophysical properties of the target material and the laser parameters, such as
laser fluence and ambient atmosphere (128,129).
The earliest technology for nanoparticle preparation using the laser is accomplished in the gas phase. When a pulsed laser
irradiates the target, a high-temperature and high-density plasma plume is formed in the target surface, and the plume rapidly
Figure 21 Diagram of femtosecond laser direct writing in transparent solid materials within the layered three-dimensional optical data storage
medium.
www.iran-mavad.com
140 Laser Ablation
Figure 22 Application examples of femtosecond laser ablation of transparent materials. (a) A micro-gearwheel, (b) a micro-chain, (c) and (d) illustrate
two different scanning modes: raster scanning and profile scanning, and (e) a microbull. Reproduced from Tanaka, T.; Sun, H. B.; Kawata, S. Appl.
Phys. Lett. 2002, 80, 312.
expands due to continuous energy absorption from the laser in the gas medium. Ashfold et al. estimated that the density of neutral
species nn ¼ 1018 cm3, the temperature of plasma T ¼ 4500 K, and the density of ions ni ¼ 1013 cm3; if the expansion speed of
plasma is 20 km s1, there will be 1015 atoms in a 0.13 mm3 plume and the generated pressure will be several times atmospheric
pressure (130).
Figure 23 A scanning electron microscopic image of a microspider-array (Reproduced from Chichkov, B. N.; Fadeeva, E.; Koch, J., et al. Proc. SPIE 2006,
6106, 610612.) and m-dragon model (b) (Reproduced from Juodkazis, S.; Nishimura, K.; Misawa, H. Chin. Opt. Lett. 2007, 5, 198.) fabricated by
femtosecond laser machining of transparent materials.
www.iran-mavad.com
Laser Ablation 141
Figure 24 SEM image of the micro tweezers, with submicron probe tips fabricated by two-photon microstereolithography. Reproduced from
Maruo, S.; Ikuta, K.; Korogi, H. Appl. Phys. Lett. 2003, 82, 133.
Figure 25 Illustration of nanoparticle formation in plasma.
After a laser pulse, the plasma persistently and rapidly expands, and temperature and pressure sharply decline, resulting in
collisions and aggregations between the atoms, electrons, and ions, and then new condensed matter, or a phase transition to form
a new substance, is generated. The experiments show that neutral species play a leading role in the phase transition process: The
condensation between the species helps to form the nuclei, and subsequently the plasma rapidly cools and the condensation occurs
in a different way. If the condensed process is completed on the substrate surface, this is just the principle of PLD of thin films. If the
particles of the plasma freely coalesce, nanoparticles or other shaped material will be formed, accompanied by a large number of
complex physical and chemical processes. The properties of the substrate (such as surface structure and temperature) as well as the
surrounding gas medium will affect the quality of the final synthesis.
In 1998, Geohegan et al. (115) reported the first time-resolved measurements of photoluminescence from gas-suspended
nanoparticles and utilized gated intensified CCD-array imaging of this PL to reveal dramatically different Si-nanoparticle formation
and propagation dynamics in He and Ar. Ar (1.0 Torr) stops and reflects the Si plume, resulting in a stationary, uniformly distributed
nanoparticle cloud. He (10 Torr) slows the silicon plume, angularly segregating most of the nanoparticles to a turbulent smoke ring
that propagates at w10 m s1 through the chamber. This experiment confirmed that the nanoparticles can be condensed in the gas
phase.
The first essential factor is the type and pressure of the atmosphere. The background gas pressure is one of the most important
parameters affecting the size of the nanoparticles (131,132). The experiments (131), in the laser ablation system filled with inert
buffer gas, show that the uniform and dispersed nanoparticles can be successfully prepared. As the buffer gas pressure increases, the
size of the nanoparticles is increased accordingly. Grigoriu et al. discussed the relation between the average size and the ambient
pressure by an inertia fluid model (133). In 2000, Ozawa et al. synthesized nanometer-size particles of tungsten W using a Q-switch
Nd:YAG laser (134). Particle size with the maximum particle generation increased from less than 10 nm to more than 80 nm,
depending on the increase of the ambient pressure. In 3.8 J cm2, the number of nanoparticles generated larger than about 80 nm
does not depend on the pressure. On the other hand, the number of particles less than 80 nm in size depends on the ambient
pressure. The measurement results are shown in Figure 26.
The type of atmosphere has a direct impact on the preparation of nanoparticles. N. Okubo et al. (135) have fabricated nano-
particles of titanium oxide by ablating a Ti target with pulsed CO2 laser in an Ar-diluted oxygen environment. In the Ar gas envi-
ronment, the Ti atoms collide with gaseous atoms and create the ions and electrons, that is, a plasma plume. The formation of the
plasma indicates that Ti atoms have chances to form TiO2 nanoparticles in the O2 as reactive gas. The mean size of NPs fabricated at
www.iran-mavad.com
142 Laser Ablation
Figure 26 The comparison of size distributions with different fluences, (a) 1.9 J cm2 and (b) 3.8 J cm2. Reproduced from Ozawa, E.;
Kawakami, Y.; Seto, T. Scr. Mater. 2001, 44, 2279.
the lower pressure are small (5 nm) and cohesive to each other, and the major part of the deposit is amorphous, suggesting that the
oxygen supply is not high enough to grow crystalline oxide particles. On the contrary, at the higher pressure, the average size of the
independent NPs significantly increases (28 nm). Under this condition, the oxygen is plentiful, and the generated nanoparticles of
anatase TiO2 crystal and amorphous hybrid structure. It should be emphasized that the NPs are spherical crystals of single anatase, if
sufficient environmental gas is supplied. No rutile phases were observed in any of the conditions studied (Figure 27).
In a liquid medium, pulsed laser synthesis of nanoparticles is a more effective technique. The first attempt to prepare metal
nanoparticles in liquid medium by laser ablation technique was made by German scientists A. Henglein and A. Fojtik. In 1993,
a glass-supported gold and nickel film in the solution was irradiated by the ruby laser with 694 nm wavelength, and gold and nickel
nanoparticles were prepared. The results show that the average particle size was dependent on laser intensity (136). In the same year,
T. M. Cotton and his colleagues developed this technology. In their experiments, a 1064 nm Nd:YAG pulsed laser was used to
synthesize different sizes and distributions of gold, silver, copper, platinum, and other metal nanoparticles, that is, nano-metal
colloids, by controlling the experimental conditions in different solutions (137).
In liquid medium, the better preparation efficiency of NPs is not accidental. In liquid medium, plasma was rapidly formed on
the surface zone of a solid target irradiated by laser. Some researchers (138–142) show that the liquid medium has a greater
influence on the evolution of the plasma than does gas. As the laser ablation proceeds, plasma continues to absorb the laser energy,
and simultaneously the ablated target continues to supply new ablation products for the plume, which prompts the rapid expansion
of plasma at supersonic speed, thus forming a shock wave. The shock wave generated in the plasma produces an additional pressure,
leading to an increase in the temperature of the plasma. Therefore, the liquid medium limiting the shock waves elevates the
temperature, pressure, and density of the plasma.
Figure 27 TEM and SAD images of nanoparticles fabricated by PLD at a total pressure of 1.5 104 Pa (a) and 6.7 104 Pa (b), where flow rates of Ar
and O2 gas for both cases are 83 cm3 s1 and 8.3 cm3 s1, respectively. Reproduced from Okubo, N.; Nakazawa, T.; Katano, Y.; Yoshizawa, I. Appl.
Surf. Sci. 2002, 197–198, 679.
www.iran-mavad.com
Laser Ablation 143
Figure 28 TEM images of nanoparticles at fluence: (a) 10, (b) 20, and (c) 35 J cm2 and the particle distribution. Reproduced from Kumar, B.;
Yadav, D.; Thareja, R. K. J. Appl. Phys. 2011, 110, 074903.
For example, Berthe et al. investigated the pressure (2–2.5 109 Pa) of laser-induced Al-plasma in a water-confinement regime by
a XeCl excimer laser with the intensity of 1–2 109 W cm2 and the pulse width of 50 ns (139–141). Peyre et al. found that higher-
pressure plasma (1010 Pa) was induced by short pulse laser (3 ns) (140–142). Laser wavelength and pulse width will affect the
pressure of induced plasma. In Sakka’s experiments (143,144), a 532 nm Nd:YAG pulsed laser was used to produce graphite plasma
in water, and the density 1022–1023 cm3, the temperature 4000–5000 K, and the pressure 1010 Pa of C-plasma could be obtained.
The presence of a high-temperature, high-pressure, and high-density plasma zone is advantageous to growing high-temperature
and high-pressure nuclei, and thus the metastable phase is formed at room temperature. A variety of chemical reactions may occur
during the evolution of the plasma. The last stage of the evolution of plasma in the liquid limit is cooling and condensation. Different
types of nucleation have different applications in materials processing. After plasma quenching, part of the plasma is deposited on the
surface of a solid target due to fluid restriction, which will lead to the coating of the target surface, applicable to the surface treatment
of materials (145,146). Another part of the plasma is condensed and dispersed into the liquid medium, and small particles can
generally be suspended or floating on the liquid surface. This can be used for the preparation of nanoparticles themselves.
The average size and size distribution of laser-fabricated nanoparticles in liquid medium are affected by the physical properties of
the target material, laser parameters (laser power density, pulse number, and pulse duration time) and liquid concentration (147).
Figure 28 (148) shows the TEM image of NPs synthesized using PLA in deionized water and their particle size distribution at laser
fluence of 10, 20, and 35 J cm2, respectively. NPs have a spherical shape with a narrow size distribution. The particles are mainly
distributed in the range w5–35 nm with an average of 10, 13, and 15 nm at the fluence of 10, 20, and 35 J cm2, respectively.
Laser ablation synthesis of nanoparticles in liquid medium technology continues to develop, succeeding in the preparation not
only of elemental nanoparticles, but also the alloy nanoparticles. I. Lee et al. (149) fabricated nanometer-sized gold-silver alloy
nanoparticles by ablating gold-silver alloy with an ideal ratio in solution. Nanometer-sized alloy particles have unique physical,
chemical, and catalytic properties, and thus the laser ablation preparation method has a great market potential in the fields of
microelectronic materials, catalysts, and biological engineering applications.
4.06.4 Pulsed Laser Ablation and Pulsed Laser Deposition Technology

4.06.4.1 Physical Picture of Pulsed Laser Deposition
The PLD technique is now widely used in film preparation, which is one of the most important applications in the field of laser
ablation. In the 1970s, with the advent of the excimer laser and the popularization of the electronic Q-switched laser, the PLD
technique began to attract a significant amount of attention (150). The excimer laser has a high-power density and a very short pulse
www.iran-mavad.com
144 Laser Ablation
Absorption
Thermal Conduction
Surface Melting
Vaporization
Plasma Production
Plasma Shielding
Inverse Bremsstrahlung
Absorption
Figure 29 Illustration of the pulsed laser deposition process.
duration, which could lead to a lower ablation depth and hence the droplets caused by the melting effect are reduced greatly during
the ablation process. Therefore, the excimer laser deposition technique can remarkably improve the quality of the film. In 1987,
D. Dijkkamp et al. at Bell Laboratory first successfully prepared the high-temperature superconducting thin films by using the high-
energy KrF excimer laser (151). Also, it was discovered that PLD can be used for fabrication of epitaxial thin films and multi-
component ceramic oxides, nitrides, metallic multilayers, as well as a variety of superlattice materials (152). Therefore, the PLD
technique has undergone rapid development in the following years (153).
In the late 1990s, femtosecond laser was utilized in the PLD technique, and diamond-like carbon (DLC) films were successfully
prepared (154,155). The power density of the femtosecond laser is extremely high (up to 1021 W cm2) within a very short pulse
duration time (1014–1015 s). The laser ablation process shows several new features, such as nonthermal effects and high-precision
micro-machining to achieve clean ablation during preparation of the targets, which can greatly improve the quality of the films and
promote the development of PLD technology. Research on the mechanisms of femtosecond laser ablation and deposition is
increasing, and some theoretical work has been carried out. For example, the ultrashort pulse laser interaction with the target should
obey the non-Fourier heat conduction rule; the high-energy photon colliding with the target will significantly affect the physical
parameters of the target; the electron–electron collision effect and electronic density of states (DOSs) effect should be considered in
heat transport process; and so on (33,156).
The experimental setup is shown in Figure 4, and the whole preparation process includes three regimes (7) as shown in Figure 29:
(1) Laser ablation and plasma formation
A focused high-energy pulsed laser irradiates the target surface, and then the target absorbs the laser energy, resulting in surface
temperature up to the vaporization point. Subsequently, the vapor is ionized, and local high-density plasma is formed.
(2) Isothermal and adiabatic expansion of plasma
Plasma generated on the target surface will continue to absorb the laser energy and gain further ionization, resulting in a rapid
rise of temperature and internal pressure of the plasma. The elliptical plasma, that is, the plasma plume, normal to the target surface
is formed, and the plasma plume expands and travels in vacuum or atmosphere.
(3) Deposition of thin film
When the expanding plasma plume arrives at the substrate surface, the gaseous species meet each other, combine, and finally
aggregate together.
4.06.4.2 Introduction to Plasma Expansion in Pulsed Laser Deposition

Figure 30 indicates that under the high-energy laser irradiation, a high-temperature and high-pressure plasma is sputtered from the
target surface. Figure 31 shows experimentally the plasma plume observed under conditions of different times and different oxygen
pressures (157).
www.iran-mavad.com
Laser Ablation 145
For nanosecond laser ablation, the expansion evolution of plasma plume in the space can be roughly divided into two stages: (1)
isothermal plasma expansion. When t < s (pulse duration time), the temperature of the plasma tends to decrease due to expansion,
but at the same time the plasma still absorbs the incident laser energy. The two opposing processes roughly cancel each other out;
thus, the plasma temperature remains constant in this stage (158). (2) Adiabatic plasma expansion. After the termination of the
laser pulse, the rapid expansion of the plasma gives rise to a temperature decrease. At this stage, the heat exchange of the plasma with
its surroundings is negligible, so the expansion is in an adiabatic regime.
Regarding the plasma as an ideal gas and using fluid mechanics, the dynamic equation of plasma expansion is established
according to the overall mass and momentum conservation law (7,159,160). However, modified dynamics models have been
developed based on the local mass and momentum conservation law (161–164), and the theoretical results obtained are in
agreement with the experimental results. The models can naturally describe the characteristics of plasma expansion confirmed by
experiments (165–168): (1) the velocity distribution of the plasma is self-similar, (2) the plasma density at the ablation surface is
approximately constant, and (3) the degree of ionization has an effect on the evolution of the plasma.
The laser ablation process determines that the number density distributions in radial and longitudinal directions have distinct
characteristics. In the direction perpendicular to the target surface, a large-density gradient and different initial velocities of sputtered
atoms lead to the maximal number density of plasma existing in the region adjacent to the target surface rather than on the surface.
Substrate
Thin film
Plasma flow
Focus lense
Target
Figure 30 A sketch of plasma expansion.
Figure 31 Photos of plasma plume at different oxygen pressures. Reproduced from Harilal, S. S.; Bindhu, C. V.; Tillack, M. S.; Najmabadi, F.;
Gaeris, A. C. J. Appl. Phys. 2003, 93, 2380.
www.iran-mavad.com
146 Laser Ablation
In the radial direction parallel to the target surface, the plasma expansion is mainly caused by the large density gradient, so the
maximal number density locates just on the target surface. Near the surface, the number density in several mean free paths
approximately retains a constant value in the whole ablation process.
Studies have shown that the plasma generated by each pulse may be considered to expand independently and finally reaches
the substrate surface. The state of the plasma is critical for the thin-film growth process of PLD. Based on the above mentioned,
the whole plasma as a single pulse eventually arrives at the substrate. Because the plasma concentration in the spatial distri-
bution is inhomogeneous, the incident species flux changes with time, which provides a theoretical basis for investigation of
thin-film growth in PLD. The evolution of a femtosecond laser-produced plasma is a highly nonequilibrium and nonlinear
process, and the physical conclusions based on nanosecond laser deposition cannot be simply applied to the femtosecond laser
deposition.
4.06.4.3 Introduction to Film Growth in Pulsed Laser Deposition

After isothermal and adiabatic expansions, the plasma plume is collected on a substrate on which thin films grow. The deposition
process has a direct impact on the quality of the thin films. First, the species impinge on the substrate and begin to gather together,
and then the nuclei are continuously formed. In the following deposition process, a number of so-called island structures appear on
the substrate. The existing islands gradually capture diffusing and later-arriving incident species so that the sizes of islands increase
quickly. Finally, the coalescence of existing islands completes a layer of continuous film. The deposition conditions can be
reasonably controlled to achieve a continuous and layer-by-layer growth, until the required film thickness is achieved. The film
growth is a very complex process, and some factors, including the interaction between the ablated species (atoms, molecules, ions,
clusters, etc.) and the substrate surface, the interaction between various species, the substrate temperature, and the incident kinetic
energy of species, play an important role in the growth process.
For the evolution of film growth, the formation of surface defects, adatom migration and diffusion on the substrate, thermo-
dynamics of the adatoms gathered into a nucleus as well as the thin-film growth mode, and so on, are very important research
topics. The island nucleation in the early stage is particularly important for thin-film growth. Two methods are commonly used to
study the film growth process: the continuity equation theory and numerical simulation (169,170). Considering thin-film growth as
a random dynamics process, the Monte-Carlo method is widely used in this field.
Compared with other deposition techniques, the thin-film growth process of PLD has some distinct characteristics: (1) energetic
deposition. The kinetic energy of the species generated by PLA is in the order of 100–103 eV. High-energy species reach the substrate
surface and have a bombardment effect on the deposited film and substrate so that the adatoms obtain an instantaneous mobility
that enhances their diffusion speed on the surface. Also, energetic particle bombardment can cause atomic bond breaking and
structural reconstruction (171). The energetic species impinge from the plume to a surface and transfer their kinetic energies to
surface atoms, leading to the formation of point defects such as vacancies or interstitials. This defect structure is likely to attract the
following incident species to form nuclei. Consequently, the density of islands increases, and the layer-by-layer growth of thin film
is accelerated. (2) Pulsed deposition. The production of PLA reaches the substrate in pulsed mode. After the termination of one
pulsed laser, a large number of species almost simultaneously impinge on the substrate. Thus, a high instantaneous concentration of
species increases the chance for the adatoms to meet each other and aggregate, which promotes layer-by-layer film growth.
Therefore, the interval of two pulses needs to ensure that the incident species have sufficient time to find a suitable growing point.
Namely, pulse repetition frequency for the PLD process is an important parameter (172,173). (3) Stoichiometric deposition. The
most significant advantage of the PLD technique is to keep identical stoichiometry between the deposited film and the target (174–
176). The reason is that a large number of energetic species reach the substrate surface, freely diffuse, and condense to a film on the
substrate. The principle can be found in Section 4.06.2.
The quality of the films prepared by PLD is closely related to laser parameters (laser energy density, wavelength, pulse repetition
frequency, and pulse width), type and pressure of atmosphere in the chamber, the geometric parameters of the experimental setup
(the position of on-axis and off-axis, the distance between target and substrate), as well as substrate parameters (material type,
temperature, and quality of surface).
Here, the influences of laser parameters on film growth are emphasized. Laser energy density can directly affect the type and
concentration of the species in the plasma plume (177,178), which can change the stoichiometric ratio of films. The higher the laser
energy, the greater the bombardment intensity of plasma, the more the number of point defects on the substrate surface (179,180)
and the higher the concentration of the active adatoms. Adjusting the laser energy density sometimes makes it possible to change the
film growth mode (181), for example, such that vertical growth changes to two-dimensional growth (182).
It must be pointed out that laser energy density should be near the ablation threshold in order to avoid the generation of droplets
and improve the quality of the films (183). The experiments show (132,184) that, when the laser energy density exceeds the ablation
threshold, the ablation plume with a large number of droplets or particles reaches the substrate surface, and the surface roughness of
the film remarkably increases. Figure 32 (185) shows an SEM micrograph of an Mg:Nb film deposited at low-energy density of the
laser beam, 8 J cm2 (Figure 32(a)), and at high-energy density, 47 J cm2 (Figure 32(b)). In Figure 32(a), the left and right sides of
the micrographs show a 2500 and 5000 magnification, respectively. In Figure 32(b), the left and right sides of the micrographs
show a 1000 and 4000 magnification, respectively. The most important difference between the low- and high-energy deposited
samples is the size distribution of the spheres. At low-energy densities of the laser beam, these spheres have a diameter lower than
2 mm while at high-energy densities diameters larger than 10 mm can be observed.
www.iran-mavad.com
Laser Ablation 147
Figure 32 SEM micrograph of the Mg:Nb thin-film samples deposited at different energy densities 8 J cm2 (a) and 47 J cm2 (b) of laser beam.
With an increase in the laser irradiation energy density, the grain size becomes larger (186,187). An explanation of this
phenomenon could be that by increasing the energy density, both the plasma density and the plasma kinetic energy increase. When
the species of the plasma arrive at the substrate surface, the kinetic energy of the deposited adatoms creates higher diffusion, which
results in the coalescence of grains and particles (186).
At fixed pulse intensity, pulse repetition frequency has a great influence on film growth and surface morphology (188,189).
Hydroxyapatite coating on a metal substrate has been prepared by the ArF PLD technique (190). When the pulse repetition rate
increases from 20 to 80 Hz, the structural order of the coatings decreases, and, eventually, at 100 Hz the coatings become
amorphous. CeO2 films at pulse repetition frequency 1, 3, and 5 Hz were prepared, and it was found that with decreasing pulse
frequency, the island density on the surface was significantly reduced and characteristically smoother films were achieved using
a low deposition rate (191). Ag films were deposited at a pulse repetition frequency of 10–100 Hz, and the experimental results
showed that the average size of the island decreased but the island density increased with the pulse frequency increase (192). In the
growth process, a slower deposition rate, that is, a lower pulse frequency at fixed pulse intensity, means that the atoms on the
island can much more easily reach the island edge and the nuclei will be given more time to ripen, avoiding a three-dimensional
island growth (193,194). In a word, adjusting the frequency of pulse repetition has a significant influence on the preparation of
films (195).
Experimental and theoretical simulations show that there are several scaling laws in the thin-film growth process (196–199).
Reference (196) reported that Ag thin films were deposited on an insulator substrate with substrate temperatures of 93 C and
135 C, respectively. Both experimental and theoretical results confirmed that the electrical percolation thickness of the films
depended on pulse repetition frequency, with a scaling law with a single exponent of 0.31 and 0.34, as shown in Figure 33.
Many theoretical models have been developed to describe the thin-film growth process, including the diffusion-limited aggregation
(DLA) model (200), the improved DLA model (201,202), the Bruschi model considering substrate temperature (203), the Taylor
model (204), the Kuzma model (205), and the pulsed kinetic Monte-Carlo model, considering pulse frequency and incident particle
energy (170).
4.06.5 Thermodynamics of Laser Ablation
Laser ablation includes some complex physical phenomena: heat transfer process, plasma effects, various radiation effects,
mechanical effects, and so on. The physical nature is determined by the thermodynamics of the pulse ablation process. Namely, the
study of the laser ablation mechanism is based on thermodynamics in order to study the quantitative transfer law and the related
physical effects in the laser ablation process and to analyze the relationship between various technical parameters and ablation
www.iran-mavad.com
148 Laser Ablation
135 °C
60
Thickness dperc at Percolation (nm)

50
40 f -0.31
30
(a)
50 93 °C
40
f -0.34
30
20 (b)
2 5 10 20 50
Pulse frequency (Hz)
Figure 33 Effect of pulse repetition frequency on Ag film thickness at the electrical percolation transition. The data are well fit by a power law with
a single exponent of about 0.33 for (a) and (b). Reproduced from Warrender, J. M.; Aziz, M. J. Phys. Rev. B 2007, 76, 045414.
results. Investigation of the laser ablation mechanism can advance the development of laser ablation technology and improve the
quality of ablated workpieces or fabricated materials (such as films and nanoparticles).
4.06.5.1 Theoretical Framework of the Thermodynamics of Long-Pulsed Laser Ablation – The Basic Equation
and Plasma Shielding Effect
The thermodynamics of LPLA is physically based on the local equilibrium heat transport process. In this process, the interaction of
laser and target depends mainly on the electron–phonon interaction. The electron–phonon relaxation time is tens of a picosecond,
that is to say, the time taken by the electrons to transfer their kinetic energy, obtained from the laser, to the lattice. For a long-pulse
laser such as the nanosecond laser, the electrons have sufficient time to deliver their energy to the lattice, and then thermal equi-
librium is easily established between the electrons and the lattice subsystem before the termination of the laser pulse.
After the target absorbs the laser energy, the temperature of the target surface will increase up to a phase transition point. If the
required heat energy for phase change is satisfied, the target will melt. The heating and melting processes of pulsed laser irradiation
on material constitute a three-dimensional heat flow problem. However, even if a long-pulsed laser is utilized in the ablation
process, that is, nanosecond or picosecond, the thermal diffusion distance is short, too. The dimension of the laser irradiation area
(in millimeters) is usually several orders of magnitude larger than the thermal diffusion distance in the vertical direction of the target
surface. Thus, if the target surface is regarded as an infinite plane, the corresponding heat conduction problem can be approximately
treated as the transport of laser energy along the direction perpendicular to the target surface. This is why the laser ablation thermal
conduction phenomenon is theoretically simplified into a one-dimensional heat flow model.
In Figure 34, a schematic of the laser–solid interaction is given. During the time for which a long-pulsed laser irradiates a target
surface, the target can be roughly divided into three separate regions and five parts (33,206): (1) Solid phase region A without
A B C D E
Laser
(1) (2) (3)

x 0
Figure 34 Illustration of long-pulsed laser ablation of target.
www.iran-mavad.com
Laser Ablation 149
melting, but absorbing the laser energy, (2) liquid-phase region generated by the melted target (including normal liquid phase
region B and superthermal liquid phase region C), and (3) high-temperature and high-pressure plasma cloud D and plume E.
A basic theoretical framework of the LPLA is based on the basic heat conduction equation and the corresponding boundary and
initial conditions.
Only considering the coexistence of a solid and liquid phases in the ablation stage, a basic thermodynamics equation, that is,
a classical Fourier conduction law, can be expressed as:

vTi ðx; tÞ v vTi ðx; tÞ
ri ðTÞcpi ðTÞ ¼ ki ðTÞ ð0 < t s; 0 < x dÞ [7]
vt vx vx
where i ¼ 1, 2 represents the solid and liquid phases, respectively, r(T) is the temperature-dependent density of target, and cp(T) is
the temperature-dependent specific heat capacity under constant pressure, s is pulse width, d is the target thickness, and x refers to
the vertical direction of the target. The right side of eqn [7] presents the transferred heat flow caused by the existing temperature
gradient in the target, and the left side presents target temperature changes due to the heat conduction. This equation does not take
into account the heat source term, that is, the laser energy radiation during the heat conduction process. During laser irradiation, the
formed plasma plume can absorb 98% of the laser energy by IB and PI mechanisms, and the remainder (only 2%) can transmit to
the target surface. As an approximation, the heat diffusion equation, without considering a heat source term, can be used to describe
the laser–target interaction.
For long-pulse ablation, thermal conductivity ki is assumed to be a constant due to the temperature only having a small effect on
it. The thermal diffusion coefficient ai ¼ ki/(rcp), so eqn [7] can be changed into:
v2 Ti ðx; tÞ 1 vTi ðx; tÞ

¼ ð0 < t s; 0 < x dÞ [8]
vx2 ai vt
However, to adequately describe the physical nature of the laser–material interaction, the heat diffusion equation with the heat
source term should be considered. Here are two evolution stages of plasma: (1) one is that at the beginning of ablation, the vapor is
not significantly ionized and plasma cannot be formed, (2) the other is that plasma rapidly expands and consequently its density
decreases, which makes plasma transparent to the incident laser beam. For the two stages, the pulsed laser can directly illuminate the
target surface, and the laser energy absorbed by the corona zone will be transferred to the target surface by plasma radiation. In order
to deeply investigate the temperature evolution of the target, the heat diffusion equation with a heat source term can be expressed as:

vTi ðx; tÞ v vTi ðx; tÞ
ri ðTÞci ðTÞ ¼ ki ðTÞ þ Sðx; tÞ ð0 < t s; 0 < x dÞ [9]
vt vx vx
where S(x,t) is the heat source term, usually regarded as the Gaussian profile (207–209), that is, S(x,t) ¼ (1 R)I0(x,t)$bebx. Here,
I0(x,t) is the output power density of laser, R is the reflectance, and b is the absorption coefficient of the target.
In addition to considering the source term, the shielding effect of plasma and the evaporation effect should be considered (210).
The shielding effect of plasma is caused by two methods – IB radiation and PI – which absorb laser energy. High-temperature and
high-pressure plasma formed on the target surface absorbs the incident laser radiation and prevents the laser energy from reaching
the surface. This blocking effect is called the plasma shielding effect. For infrared laser ablation, the shielding effect is caused mainly
by IB radiation. For UV laser ablation, the shielding effect of plasma is based on the PI mechanism (211). It is assumed that the
plasma radiation can be regarded as blackbody radiation.
Assuming that plasma is an ideal gas, the ionization process of the vapor is accomplished through IB and PI absorption
mechanisms. The IB process is usually described by the inverse absorption length aIB (cm1) (212):

N0
aIB z 1:25 1046 l3 Te0:5 þ Ne Ne ¼ sIB Ne [10]
200
where l is the laser wavelength, Te is the electron temperature, N0 and Ne are the density of the neutral atom and electron,
respectively, and sIB is the cross section of IB absorption. For the excimer laser ablation of metals, such as iron, the plasma particle
density is in the range of 5 1018 cm3, and the electron temperature is 3 eV (211). In eqn [10] sIB represents the total cross section
of the IB process, and the contributions of both neutral and ion in IB processes are included (213,214), that is, sIB ¼ sIB,neu þ sIB,ion.
The cross section of neutral sIB,neu equals approximately 2.75 1023 cm2. A laser fluence of 1–6 J cm2 is equivalent to
0.125 1019–0.75 1019 photons cm2, thus, 0.0034–0.02% photons can be absorbed by IB due to neutrals, which is too small to
significantly attenuate the laser or heat plasma.
In the UV laser ablation process, direct PI of excited atoms plays an important role in the shielding effect of the vapor. The cross
section of PI, sPI can be expressed by (212):
ðEI E
Þ2:5 IH
0:5
sPI z7:9 1018 cm2 [11]
ðhvÞ3
where IH is the ionization potential of hydrogen, EI is the ionization potential of the atom, and E* is the energy of the excited
level that can be photoionized. For example, the photon energy of the 248 nm excimer laser is 5 eV, making
sPI z 1.3 1017 cm2, which are several orders of magnitude higher than the IB cross section. If a laser fluence of 1–6 J cm2 is
www.iran-mavad.com
150 Laser Ablation
equivalent to 0.125 1019–0.75 1019 photons cm2, the probability of an excited atom absorbing an UV photon equals
approximately 100%; that is, the ionized condition is formed. As a consequence, the vapor ionization degree is readily
enhanced, and the PI is the most important absorption mechanism in this case.
During the pulse duration time, the total energy E absorbed by the target can be expressed as:
Z s
E¼ E0 ðtÞdt [12]
0
where E0 (t) is the laser energy transiently absorbed by the target. For convenience, the time tth is defined as the time taken by the
target to absorb laser energy to reach the ablation threshold Eth. So, when 0 < t < tth, the transient power density absorbed by the
target I(t) is as follows:
IðtÞ ¼ ð1 RL ÞI0 ðtÞ ð0 < t < tth Þ [13a]
Here RL is the target reflectivity for the laser. For tth < t < s, I(t) can be given as:
IðtÞ ¼ ð1 RL ÞI0 ðtÞexp½ðaIB þ sPI nz ÞH þ ð1 RP ÞIem ðtth < t < sÞ [13b]
z
Here, I0(t) is the laser pulse intensity profile, n is the number density of the excited neutrals, and H is the dimension of the plasma
perpendicular to target. RP is the target reflectivity for the plasma emission. Iem is the plasma-emitted intensity. The second term in
the right side of eqn [13a] indicates the absorbed laser power density by the target after plasma irradiation. Equation [13] presents
the heat source term in the heat conduction equation considering the shielding effect of the plasma.
It is assumed that the target ablation threshold Eth is a constant. According to the Lambert–Beer law (215):
EðxÞ ¼ Einc expðbxÞ [14]
where E(x) is the energy density at target depth x and Einc is the total incident energy density. Substituting E(x) ¼ Eth and eqn [12]
into eqn [14], then the maximum ablation depth x can be expressed as:
Z s
1
x ¼ ln E0 ðtÞdt=Eth [15]
b 0
4.06.5.2 Theoretical Framework of the Thermodynamics of Long-Pulsed Laser Ablation – Dynamic Physical
Parameters and the Vaporization Effect
Normally, the thermophysical parameters of the target such as the absorbance and the absorption coefficient are all temperature
dependent, which affects the heat conduction process (216,217). For simplicity, the ablation of transparent material and the plasma
radiation effect are not considered in the following discussions.
The absorption by the material of energy from the laser can be regarded as a ‘secondary’ energy source on the surface and a variety
of physical effects occur due to this energy source (218,219). The experiments show that for ordinary metals, the resistivity of the
melted metal linearly increases with temperature (220) and is two times larger than that of solid metal below the melting point.
Thus, the absorption coefficient rises, and the reflectivity declines after the melting. Reference (221) shows that the absorbance of
material is proportional to the resistivity. For semiconductors, the number of carriers increases with the temperature, so the lattice
scattering effect is enhanced and then the mobilities of electrons and holes decrease. As a result, the resistivity of such material also
increases with temperature; consequently, the absorbance also does (222).
According to the time-dependent surface temperature in Refs. (216,223), the dynamic absorptivity can be written in the
following explicit form:
b ¼ b0 þ A1 ½Ct þ ðC2 t 2 þ DtÞ1=2 [16]
where b0 is the absorbance at room temperature, and A1 is a constant varying with the material too. The value of A1 for Fe target is
0.85 105 K1 (223). Based on Maxwell’s equations and the Lambert–Beer law, the temperature-dependent absorption coefficient
is given as:
rffiffiffiffiffiffiffiffiffiffiffi sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
4pk 4ps 4ps0
bðTÞ ¼ ¼ ¼ [17]
l0 30 l0 c 30 l0 c½1 þ aðT T0 Þ
and the temperature-dependent absorptivity can be expressed by:

rffiffiffiffiffiffiffiffiffiffiffiffiffi sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
4py30 4pc30 ð1 þ aðT T0 ÞÞ
bðTÞ ¼ 2 ¼ [18]
s l0 s0
where s0 is the target conductance at initial temperature T0, 30 is the permittivity of vacuum 30 ¼ 8.85 1012 C2 N1 m2, c is the
laser propagation velocity in vacuum, and a is the target temperature coefficient of resistance. For a general metal, a is in the range
400 105–700 105 K1 (224).
The following discussion is about the effect of vaporization on laser ablation. Vaporization is the species (atoms or
molecules) emitting from the surface of target, due to the phase transition from condensed matter to vaporized
www.iran-mavad.com
Laser Ablation 151
phase. Vaporization phenomena can occur at any laser influence and pulse duration time, which does not need to satisfy
the temperature threshold, but energy threshold (if E < Eth, the vaporization cannot occur). For the low-power laser ablation,
the vaporization effect on laser ablation is not obvious. However, for the high-power laser, the vaporization
effect is significant due to the sufficient absorption of laser energy, which should be considered in the heat diffusion
equation (225):

vTðx; tÞ vT v Tðx; tÞ
rl ðTÞCl ðTÞ Cl rl yr ¼ kl ðTÞ þ Sðx; tÞ ðsm < t < s; 0 < x < sðtÞÞ [19]
vt vx vx vx
where the subscript l means all parameters are under the condition of liquid state. The heat loss of the vaporized species per area is
vT
Cl rl yr , and the velocity of vaporization (or the velocity of surface recession) is yr, which is determined by the Hertz–Knudsen
vx
equation (29):
p
yr ¼ Cs [20]
rl ð2pkB T=ml Þ1=2
where p is the gas pressure, Cs is the sticking coefficient (226), and ml is the particle average mass. Under the limit Vliq << Vgas, Vlip
and Vgas are the molar volumes of liquid and gas; the relation between the equilibrium vapor pressure p and the temperature T may
be computed from the Clausius–Clapeyron equation (227):

DHv ðTb Þm 1 1
p ¼ pb exp [21]
kB Tb T
where T is the temperature of the target surface, Pb is 1 atm, DHv(Tb) is the latent heat of vaporization, and Tb is the boiling
temperature at 1 atm.
It is assumed that for the time sv taken by the target to fulfill the requirement for the phase transition of vaporization, the
ablation energy threshold is:
Eth ¼ ð1 RÞI0 sv [22]
According to the vaporization relaxation time sv (228), eqn [22] can be changed into:
Eth ¼ 3½1 2expð1Þk1 cl rl ðTv =ð1 RÞÞ2 =I0 [23]

where rl is mass density of liquid material, cl is specific heat capacity per mass, and Tv is the vaporization temperature at
1 atm. Equation [23] indicates that Eth depends on target material and laser parameters. The evolution of ablation interface
(229) is:
ð1 RÞ2 I0 ðt sv Þ
xt ¼ ðsv < t < sÞ [24]
ðrl cl $DT þ Lv Þ
DT is the biggest temperature difference and Lv is the latent heat of ablation. For sv < t < s, the ablation thickness, which is
proportional to time, increases with the incident laser power density. The ablation surface moves forward at an even speed,
which is determined by the parameters of the laser and the target (159,229,230). The maximum ablation thickness xt can be
expressed as:
ð1 RÞðE Eth Þ
xt ¼ [25]
r1 c1 $DT þ Lv
When pulse width is reduced on the order of a picosecond, this is roughly equal to the electro–phonon interaction relaxation
time. For the nanosecond laser ablation process, the heat conduction of laser energy is a diffusion process, obeying the Fourier
conduction law. In this case, pulse width is much greater than the time required for electrons to deliver energy to the lattice, which is
effectively instantaneous relative to the pulse duration time. However, for picosecond laser ablation, the heat conduction is not
a diffusion process but a wave-type propagation, which can be described by a hyperbolic heat process model. Different from the
conventional heat transfer law, ultrashort pulse laser ablation has the characteristics of nonequilibrium heating and nonequilibrium
phase transition. For example, when a pulsed laser irradiates the target with an ultrashort pulse width of 150 ps and energy density
of more than 1013 W m2, a very high-temperature gradient will be produced on the target surface, and the heat penetration depth is
less than several nanometers, which is equivalent to the mean free path of electrons. For uniform material, the thermal equilibrium
relaxation time of the target in laser heating (electron–phonon interaction time s0) is approximately between 108 and 1014
(231–233). In other words, the heat conduction process does not obey Fourier law, but rather non-Fourier heat conduction law
(231,234–236). At the end of the 1940s, Cattaneo first presented the non-Fourier heat conduction law, which is called the general
Fourier law (237):
vq
q þ s0 ¼ kVT [26]
vt
www.iran-mavad.com
152 Laser Ablation
It is assumed that the target is irradiated by an Nd:YAG laser with power density 109 W cm2 and a pulse width of 150 ps (238,239).
The non-Fourier heat conduction considering the heat source term is:
vT v2 T v2 T Sðx; tÞ
þ s0 2 ¼ a 2 þ [27]
vt vt vx rc
When pulse duration time decreases on the order of a few picoseconds, the classical Fourier law will be transformed into non-
Fourier law as eqn [27]. In Section 4.06.5.3, for femtosecond laser ablation, two-temperature systems exist in the target, suggesting
that the above-mentioned heat conduction theory is not applicable.
4.06.5.3 Main Theoretical Results of the Thermodynamics of Long-Pulsed Laser Ablation

To solve the thermodynamic equation, suitable boundary and initial conditions are necessary. Using Si as an example, the ther-
modynamics of LPLA is investigated. The heat conduction eqn [9] can be transformed into:
v2 Ti ðx; tÞ ð1 RÞI0 ðx; tÞbebx 1 vTi ðx; tÞ

þ ¼ ð0 < t s; 0 < x dÞ [28]
vx2 ki ai vt
Based on the energy conservation rule and an adiabatic approximation, the boundary conditions before melting can be given as
follows:
vTs ðx; tÞ
ks x¼0
¼ ð1 RÞI0 ð0 < t < sm Þ [29a]
vx
vTs ðx; tÞ
ks x¼d
¼0 ð0 < t < sm Þ [29b]
vx
where d is the heat diffusion distance, which is time-dependent d ¼ 3.37(ait)0.5 (228). After melting, there are solid and liquid
phases in target; thus the boundary conditions at the solid–liquid interface can be satisfied by an energy balance equation and
continuous-temperature condition:
vT1 vTs vs
k1 ks ¼ ð1 RÞIðtÞ x¼s Lm [30]
vx x¼s vx x¼s vt
Tl ðx; tÞ ¼ Ts ðx; tÞ ¼ Tm ðx ¼ sðtÞÞ [31]

The boundary conditions at the back surface can be given by:

Ts ¼ T0 [32]
x¼xn
During pulse duration time, the temperature at the frontier of the liquid target remains at the vaporization point Tq, so:

T1 x¼0 ¼ Tq ðsm < t sÞ [33]
When a pulse is finished, we assume the adiabatic hypothesis at the ablation surface, that is:
vT1
k1 ðs < t < s þ t0 Þ [34]
vx x¼0
where s þ t0 is the pulse period.
Based on the above conditions, the time sm taken by the target to reach the melting point is obtained (33).
3ðTm T0 Þ2 Cs rs Ks
sm ¼ [35]
4Is2
Based on the above-mentioned heat conduction equation and the boundary and initial conditions, the space- and time-
dependence of target temperature before target melting is represented in Figure 35.
For the melted target, the temperature and the dynamic interface position can be given analytical solutions. The evolutions of
temperature in liquid and solid phases are:
ð1 RÞI0
Tm Tq ð1 ebs Þ
bKl pffiffiffiffiffiffi ð1 RÞI0
Tl ðx; tÞ ¼ Tq þ erf ðx=2 al t Þ þ ð1 ebx Þ [36]
erf ðlÞ bKl
ð1 RÞI0 bs
Tm T0 þ e pffiffiffiffiffiffiffi
bKs ð1 RÞI0 bx
Ts ðx; tÞ ¼ T0 þ 1=2 ! erfcðx=2 as t Þ
bKs
e [37]
a
erfc l l
as
www.iran-mavad.com
Laser Ablation 153
Figure 35 Temperature distribution of target at different times and positions. Reproduced from Zhang, D. M.; Li, Z. H.; Zhong, Z. C.; Li, X. G.;
Guan, L. Dynamic Principle of Pulsed Laser Deposition, 66; Science Press: Beijing, 2011 (in Chinese).
where al and as are the thermal diffusivity of the liquid and solid phases, respectively. l depends on the liquid–solid interface
pffiffiffiffiffiffi
l ¼ sðtÞ=2 al t ; here s(t) satisfies the differential equation:
2
s2 Is s Is
exp $ Kl ðTm Tq Þ ð1 ebs Þ exp $ Ks ðTm T0 Þ ebs
4al t b vs 4as t b
þ Lm ¼ Is $expðbsÞ [38]
pffiffiffiffiffiffiffiffiffiffi s vt pffiffiffiffiffiffiffiffiffiffi s
pal t $erf pffiffiffiffiffiffi pas t $erfc pffiffiffiffiffiffiffi
2 al t 2 as t
Usually, if the corresponding physical properties of the target and technique parameters are given, one can directly solve the
heat conduction equation by analytical or numerical methods to obtain accurate results of target temperature at different times
and positions, and the evolution of the dynamic interface. This research method for LPLA is quite universal.
According to eqns [9], [13], and [19], the shielding effect of plasma, the vaporization effect, and dynamic absorptivity can be
investigated for the thermodynamics of laser ablation (207,208,217). In Ref. (217), according to the Saha equation and for the
example of Fe, the temperature evolution of the target is obtained from three different models: (1) neglecting the effects of
vaporization and plasma shielding, (2) taking only the vaporization effect into account, and (3) considering the two effects of
vaporization and plasma shielding. The results of three models are represented by curve (1), curve (2), and curve (3), respectively.
The temperature dependence of the absorption coefficient and absorptivity are taken into account in all curves (Figure 36).
In Figure 37, the calculated and measured ablation depths depend on the laser fluence shown, where the experimental
data come from Ref. (211). Only considering the dynamic absorptivity, Lunney et al. simulated the blue dashed curve.
The red dot curve is simulated to describe the thermal phenomena before and after melting (217). The satisfactory agreement
between model predictions and experimental data confirms that the improved thermal model is correct and reasonable.
The Saha equation is usually used to study the plasma-shielding effect in the ablation process. However, for a multielemental
target, the mean ionization of crystals is difficult to define. Reference (208) presents a theoretical model to describe the high-power
nanosecond LPA of multielemental oxide superconductors YBa2Cu3O7 by considering both the vaporization effect and the plas-
ma-shielding effect. In this model, a mean ionization of crystals hU0 i is reasonably defined to analyze the plasma-shielding effect.
hU0 i is the weighted mean of the first ionization energy of each atom (240):
P
Xi U0i
hU0 i ¼ i P [39]
Xi
i
www.iran-mavad.com
154 Laser Ablation
Figure 36 Comparison of the evolution of temperature at d ¼ 0.00 mm and d ¼ 0.05 mm obtained from three different models: (1) neglecting the effects
of vaporization and plasma shielding, (2) taking only the vaporization effect into account, and (3) considering the two effects of vaporization and
plasma shielding. Reproduced from Fang, R. R.; Zhang, D. M.; Li, Z. H.; Yang, F. X.; Li, L.; Tan, X. Y.; Sun, M. Solid State Commun. 2008, 145, 556.
Figure 37 Fluence dependence of the calculated and measured ablation depths. Reproduced from Fang, R. R.; Zhang, D. M.; Li, Z. H.; Yang, F. X.;
Li, L.; Tan, X. Y.; Sun, M. Solid State Commun. 2008, 145, 556. The experimental data is from Lunney, J. G.; Jordan, R. Appl. Surf. Sci. 1998,
127–129, 941.
where the subscript i refers to the four kinds of atom in YBa 2Cu 3O 7 , U0i is the first ionization energy of the ith type of
atom, and X i is the number of the ith type atom. From eqn [39], we can obtain the mean ionization hU0 i of YBa 2 Cu 3 O 7 as
10.398 eV.
The ratio of IB absorption and PI absorption in the shielding effect can be estimated. If each evaporated
atom in local thermal equilibrium has only one excited electron, the free electron density N e is given by the Saha
equation (241):
Ne2 ¼ 3 1021 ðkB TP Þ3=2 expðhU0 iÞ=kB TP [40]
www.iran-mavad.com
Laser Ablation 155
Figure 38 Mass removal per pulse as a function of laser fluence. Reproduced from Fang, R. R.; Zhang, D. M.; Li, Z. H.; Li, L.; Tan, X. Y.; Yang, F. X. Phys.
Status Solidi A 2007, 204, 4241. The experimental data is from Bulgakova, N. M. and Bulgakov, A. V. Appl. Phys. A 2001, 73, 199.
From eqn [40], Ne ¼ 1.44 1020, at which the IB by electron–ion collisions becomes the primary mechanism of photon
absorption (sIB,ion w1.25 1019 cm2). A laser fluence of 1–15 J cm2 is equivalent to 0.5 1019–8.1 1019 photons cm2; the
product of the cross section and photon density clearly indicates a probability of 62.5–100% for the incoming photons to be
absorbed by the IB mechanism.
According to the cross section of PI in eqn [11], the photon energy of an infrared laser with wavelength of 1064 nm is 1.16 eV;
thus the cross section sPI z 2.7 1022 cm2. The absorption from an infrared laser by excited atoms is 0.135–2.187%. Through
the IB mechanism, 2.5–40% of the incident energy is absorbed by electron–neutral collisions, while 65–100% by electron–ion
collision. Therefore, IB absorption plays a prominent role in infrared laser ablation.
Figure 38 presents the numerical results of two theoretical models: considering the effects of both vaporization and plasma
shielding and only taking the vaporization effect into account. We can see from Figure 38 that the two curves nearly overlap in the
range of the laser fluence from 1 to 3 J cm2 and begin to separate gradually when the laser fluence is above 3 J cm2. This
phenomenon can be explained by the fact that with the laser fluence increasing gradually, the plasma-shielding effect becomes more
and more marked. The numerical results considering the effects of vaporization and plasma shielding produce a good agreement
with the experimental data. It is clear that the modified model considering the vaporization effect and the plasma-shielding effect
can describe the ablation process more accurately.
Based on the non-Fourier heat conduction eqn [27], the picosecond laser ablation process was investigated using an Al target as
an example. The laser power density of 1013 W m2 and pulse duration time of 150 ps were used in the calculations. Table 3
provides the parameters of the Al target.
With the energy conservation law and adiabatic conditions, the boundary conditions can be depicted as:
vT
k ¼ bI0 ð0 < t < tm Þ
vx x¼0
vT [41]
¼ 0 ð0 < t < tm Þ
vx x¼N
lim T ¼ 0
x/N
where tm is the time at which the target begins to melt. The initial conditions are:

T x¼0 ¼ T0
vT [42]
¼0
vt t¼0
Table 3 Thermal and optical properties of Al targeta
ks (J s1 cm C) rs (g cm3) cs (J g1 C) s0(s) R
2.388 2.6991 0.86 1011 0.9

a
Reproduced from Rohsenow, W. M.; Hartnett, J. P.; Ganic, E. N. Heat Transfer Hand Book; McGraw-Hill: New York, 1985; p 236.
www.iran-mavad.com
156 Laser Ablation
700
Non-Fourier conduction model
Fourier conduction model
600
500
Temperature/°C 400
300
200
100
0
0 .0 5 .0 ×1 0–12 1 .0 ×1 0–11 1 .5 ×1 0–11 2 .0 ×1 0–11 2 .5 ×1 0–11 3 .0 ×1 0–11 3 .5 ×1 0–11
Time/s
Figure 39 The surface temperature evolution obtained from the Fourier and non-Fourier models. Reproduced from Zhang, D. M.; Li, L.; Li, Z. H.;
Guan, L.; Tan, X. Y. Phys. B 2005, 364, 285.
where T0 is the initial temperature of the target. Under the boundary and initial conditions, the target temperature depending on
depth and time was obtained (238,239).
Z
bI0 pffiffiffiffiffiffiffiffiffiffi t 1 pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Tðx; tÞ ¼ a=s0 pffiffiffiffiffiffiffiffi J0 t 02 s0 x2 =a expðt 0 =2s0 Þdt 0
k x s0 =a 2s0
Z t sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2bI0 ba 1 þ 4ab2 s0
þ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi expðbxÞ pffiffiffiffiffiffiffiffi expððt t 0 Þ=2s0 Þsin h ðt t 0 Þ [43]
k 1 þ 4ab2 x s0 =a 4s20

1 pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
J0 t 02 s0 x2 =a expðt 0 =2s0 Þdt 0
2s0
where J0(t) is the zero-rank Bessel function.
Under the same conditions, the dependence of Fourier temperature on depth x and time t can be expressed in the form:
Z t
pffiffiffiffiffiffiffiffiffi 1
Tðx; tÞ ¼ I0 a=p=k pffiffiffi expðx=ð4at 0 ÞÞdt 0
0 t0
Z t
pffiffiffiffiffiffiffiffiffi 1 x2
I0 ab a=p=k pffiffiffi expðab2 ðt t 0 ÞÞexp dt 0 [44]
0 t0 4at 0
Z
I0 a t
þ expðab2 t 0 Þdt 0
k 0
The calculation results based on eqn [43] and eqn [44] are shown in Figure 39 and Figure 40, respectively. The comparison
shows that the effect of non-Fourier heat conduction cannot be ignored. When t > 5 1012 s, the non-Fourier heat has an obvious
effect on the temperature of the target surface; that is, the non-Fourier temperature is distinctly higher than that of the Fourier heat
conduction model. There is a delay of heat conduction in the inner target, and the deeper the target is, the longer the delay time is.
4.06.5.4 Femtosecond Laser Ablation Models: Classic and Improved Two-Temperature Equations
When the pulse width exceeds 100 ps, the surface temperature of the material is determined by thermal conductivity; thus laser
ablation can be regarded as an equilibrium process, obeying the Fourier conduction rule. If the laser pulse width is reduced, the non-
Fourier heat conduction will be applicable.
However, if the pulse width is reduced on the order of a subpicosecond or femtosecond, the heat transfer process of laser
ablation shows entirely different characteristics. When the laser energy is absorbed by the electrons, their kinetic energy sharply
increases and then the temperature of the electron subsystem suddenly rises. In a short period of pulse, the energy cannot be
transferred from the hot electrons to the target ions (delivery time roughly from subpicosecond to tens of a picosecond). After the
termination of a pulse, the temperature of the ion subsystem in the target slowly rises, while the temperature of the electron
subsystem declines, until the equilibrium temperature is achieved.
www.iran-mavad.com
Laser Ablation 157
450
400 Fourier model

Non-Fourier model
350
300
Temperature/°C
250
200
150
100
50
0
0.0 2.0× 10–8 4.0× 10–8 6.0× 10–8 8.0× 10–8 1.0× 10–8 1.2× 10–7 1.4× 10–7 1.6× 10–7
Position/m
Figure 40 Temperature distributions for two models at t ¼ 2 1011 s. Reproduced from Zhang, D. M.; Li, L.; Li, Z. H.; Guan, L.; Tan, X. Y. Phys.
B 2005, 364, 285.
Soviet scientists first paid attention to the special heat conduction phenomenon. In 1957, Kaganov et al. found that an energy
exchange exists between electrons and lattice, and presented a theoretical model to describe the nonequilibrium heat transport
process. Also, they explained the electron–phonon coupling coefficient (243).
In 1974, Anisimov et al. put forward the classic two-temperature model (TTM) (244). Based on the one-dimensional nonsteady-
state heat conduction equation, the two different interaction processes, that is, photon and electron, and electron and lattice were
considered. The temperature evolvements of the electron and lattice subsystems can be given by two differential equations:
v v
Ce Te ¼ ke 2 Te gðTe Tl Þ þ Sðx; tÞ
vt vx
[45]
v
Cl Tl ¼ gðTe Tl Þ
vt
where Te and Tl are the electron and lattice temperature, respectively, Cl and Ce are the specific heat capacity of lattice and electron,
respectively, and g is the electron–phonon coupling coefficient. The first equation describes the electronic temperature,
and g(Te Tl) represents the transferred energy from electrons to lattice by electron–phonon collision; the heat source term S(x,t)
is the energy absorbed by electrons from the laser by photon–electron collision. The second equation indicates the temperature
evolvement of the lattice subsystem, and the temperature rises due to the electron–phonon interaction g(Te Tl).
In 1984, J. G. Fujimoto et al. found that when high-intensity 75 fs optical pulses illuminate a tungsten metal surface, an
anomalous heating mechanism appears and a transient nonequilibrium temperature difference occurs between the electrons and
lattice. Pump–probe measurements indicated an electron–phonon energy relaxation time of several hundred femtoseconds (245).
In 1987, Elsayed-Ali et al. (246) used 150–300 fs laser pulses to monitor the thermal modulation of the transmissivity of thin
copper films. The experimental results show that at the beginning of the laser pulse, a larger temperature difference between the
electron and lattice existed, which proved that TTM is reasonable. They also found that the process of electron–phonon energy
transfer (electron–lattice coupling relaxation time) was observed to be 1–4 ps, increasing with the laser fluence.
In 1987, Allen (247) presented an electron–phonon coupling coefficient equation. In 1990, Brorson et al. (248) reported the first
systematic femtosecond pump–probe measurements of the electron–phonon coupling constant in some metallic films, and their
experiments confirmed Allen’s theoretical predictions. In 1996, Chichkov et al. (249) studied the laser ablation of solid targets by
0.2–5000 ps Ti: sapphire laser system. They resolved the TTM equation to obtain analytical solutions.
Both before and after the 1990s, with the development of the femtosecond pulsed laser, the TTM equation has been extensively
investigated. In 1988, Corkum et al. (250) obtained an analytic solution of the TTM and compared the theoretical results with their
experimental measurements of ultrashort laser pulse ablation of molybdenum and copper targets. Optical damage caused by laser
pulses with a duration s ns can be understood only with a TTM of metals. In 1998, A. P. Kanavin (251) studied different regimes of
heat propagation in metals irradiated by subpicosecond laser pulses on the basis of TTM and obtained analytical solutions cor-
responding to the different temperature dependences of electron thermal conductivity.
In 2002, Gamaly et al. (252) investigated the transient reflectivity of gallium films induced by 150 fs laser pulses by the pump–
probe technique. The experimental results showed that the transient electron–phonon collision rate is a strong function of
www.iran-mavad.com
158 Laser Ablation
temperature and laser intensity, which is drastically different from that observed under equilibrium conditions. In 2005, Gamaly
et al. (253) carried out experiments on the ablation of metals such as Cu, Al, Fe, and Pb in both air and vacuum and revealed that the
presence of air results in a significant reduction in the ablation threshold. The temperature in the skin-layer during a single-laser
pulse is calculated using a two-temperature approximation while the heat conduction can be neglected. Thus, TTM equations can be
transformed into:
v
Ce Te ¼ gðTe Ti Þ þ ð1 ab ÞI0 ðtÞ expðab xÞ
vt
[46]
v
Ci Ti ¼ gðTe Ti Þ
vt
Additionally, the TTM model is combined with thermoelasticity dynamics and the electron blast model to investigate the
mechanisms of ultrashort pulse laser ablation. Hetnarski and Ignaczak (254,255) have summarized five nonclassical approaches
that include the relaxation time effects in dynamic thermoelasticity. Chen et al. (256) extended the dual-hyperbolic two-temperature
and hot-electron blast models to investigate the deformation in metal films subjected to ultrashort laser heating. The major driving
force for nonthermal damage is the so-called hot-electron blast force, which is generated by nonequilibrium hot electrons. Readers
can find related works on femtosecond laser ablation in a review paper in Ref. (257).
Here, an improved TTM model is introduced in order to illustrate the general theoretical method of investigating femtosecond
laser ablation of the target. Moreover, related theoretical work, such as a unified TTM model and electronic DOS effect, are discussed
(208,217,258,259).
When the electron temperature Te is higher than 4000 K, the electron–electron collision plays a very important role in the
femtosecond laser ablation (260–262), causing a change in the thermophysical parameters. In this case, the TTM equation cannot
reasonably describe the thermophysical mechanism. Also, it is necessary to take into account the electron–phonon and electron–
electron collisions affecting the thermophysical parameters, such as electronic thermal conductivity and the optical parameters of
materials.
The thermophysical parameters such as specific heat capacity and thermal conductivity are temperature dependent. When
4000 K < Te < 10 000 K, the electron heat capacity is usually a linear function of the electron temperature (263), given by:
ce ðTe Þ ¼ ce0 Te [47]
where c0e is a constant. The electron thermal conductivity can be described as:
kðTe ; T1 Þ ¼ ce vF2 s=3 [48]
where Tl is the lattice temperature, vF is the Fermi velocity, and s is the electron–phonon relaxation time.
For good conductors, such as metals, the electron–electron collision frequency may be determined by nee ¼ ATe2 ; whereas the
electron–phonon collision frequency is proportional to Tl, namely ne–ph ¼ BTl. Here A and B are constants, and both contribute to
the electron collision frequency n.
For 4000 K < Te < 10 000 K, a relationship between the electron–phonon coupling relaxation time s and the electron–electron
and electron–phonon collision frequency for electron temperature below the Fermi temperature is given by (264):
1
¼ n ¼ nee þ neph ¼ ATe2 þ BT1 [49]
s
Substituting eqn [49] into the expression of the electron thermal conductivity, we can obtain the following equation:
vF2 c0e Te 1 BTe

kðTe ; T1 Þ ¼ ¼ k0 2 [50]
3 ATe2 þ BT1 ATe þ BT1
where k0 ¼ vF2 c0e =ð3BÞ is the electron heat conductivity at room temperature (Te ¼ Tl ¼ 300 K).
Now, consider the variation of the optical parameters with temperature. The experiments confirmed that the optical parameters
of target directly influence the laser energy absorption, which has an important impact on femtosecond laser ablation (21). At lower
electron temperatures, it is allowed to assume that the optical parameter is constant (265–269). However, when the electron
temperature is rising, the optical parameters such as the absorption coefficient and the absorptivity will remarkably change with the
electron temperature.
According to classic Maxwell theory, the optical properties, such as the absorption coefficient and the absorptivity with electron
temperature, are obtained as shown in eqn [17] and eqn [18] (270). Substituting eqns [17], [18], [47], and [48] into the classic TTM
model, the following improved two-temperature model, which considers the effect of electron temperature on the heat capacity,
thermal conductivity, absorption coefficient, and absorptivity of the electron (270), is obtained.
v BTe v2
c0e Te Te ¼ k0 2 Te gðTe Tl Þ þ Sðx; tÞ [51]
vt ATe þ BT1 vx2
v
c1 T1 ¼ gðTe Tl Þ [52]
vt
www.iran-mavad.com
Laser Ablation 159
Table 4 The thermal and optical properties of copper
Coefficient of the electron heat capacity (Ce0 ) 96.6 J m3 K2

Lattice heat capacity (Cl) 3.5 106 J m3 K
e–ph coupling coefficient (g) 1 1017 W m3 K
Electron heat conductivity at room temperature (k0) 400 W m1 K
Coefficient of e–e collision frequency (A) 1.75 107 K2 S1
Coefficient of e–ph collision frequency (B) 1.98 1011 K1 S1
Target temperature coefficient of resistance (a) 4.3 103 K1
where the heat source S(x,t) is as follows:

!
4pI0 ð4ln 2Þt 2
Sðx; tÞ ¼ bbIð0; tÞ expðbxÞ ¼ exp
l0 ðsp Þ2
0:5
4ps0
exp x [53]
e0 l0 c½1 þ aðTe ðx; tÞ T0 Þ
" # 0:5
ð4ln 2Þt 2 D
¼ C exp exp x
ðsp Þ2 1 þ aðTe ðx; tÞ T0 Þ
where C ¼ 4pI0/l0, D ¼ 4ps0/(30l0c).

The initial condition is:
Te ðx; 0Þ ¼ T1 ðx; 0Þ ¼ T0 [54]
And the boundary conditions can be expressed as:
vTe
ke ¼ ð1 RÞI0 ðtÞ
vx x¼0
vTe
ke ¼0 [55]
vx d
where T0 ¼ 300 K is the initial temperature, which is uniform across the target, and d is the depth of the target. The thermal and
optical properties of copper in Table 4 are adopted (246,271).
Under the initial condition eqn [54] and the boundary conditions eqn [55], the improved TTM equation can be resolved and the
calculation results are indicated in Figure 41. Figure 41(a) and 41(b) show the temporal evolution of the electron and lattice
temperature at the surface for the copper target irradiated by a 100 fs, 800 nm pulse at 0.2, 0.3, and 0.4 J cm2, respectively.
In Figure 41(a), the maximum temperature of the electron temperature at the surface is in the range 5500–8000 K for the laser
fluence varying from 0.2 to 0.4 J cm2. The surface electron and lattice temperatures increase with increasing laser fluence. For a fixed
laser fluence, the electron temperature rapidly increases with the ablation time, while suddenly decreasing after it reaches the
maximum, as shown in Figure 41(a). The temperature of the lattice rises gradually; see Figure 41(b).
Figure 41 Temporal evolution of electron temperature (a) and lattice temperature (b) at the surface for the copper target.
www.iran-mavad.com
160 Laser Ablation
Figure 42 Dependence of the ablation rate on the laser fluence; numerical data 1 represents the improved model.
Figure 42 shows the dependence of the ablation rate on the laser fluence. Curve 1 represents how the ablation rate depends on
laser fluence based on the improved TTM model as eqns [51] and [52]. Curve 2 stands for the model without considering electron
temperature-dependent heat capacity, thermal conductivity of the electron, absorption coefficient, and absorptivity. The black dot
represents the experimental data (272). It can be seen that the results calculated based on the improved TTM model are in good
agreement with the experimental data. In a word, the effects of physical properties on femtosecond laser ablation should be
considered (246,248,272).
4.06.5.5 Femtosecond Laser Ablation Models: Unified Two-Temperature Equations and Density of State Effect
Pulsed laser ablation from nanosecond to femtosecond laser has been extensively investigated (273–275). The first stage of ablation
is the absorption of the laser energy through photon–electron interactions. It takes a few femtoseconds for the electrons to rees-
tablish the Fermi distribution. The second stage is the heat energy diffusion to the lattice through electron–phonon interactions, and
the characteristic time scale of this stage is called the electron–phonon coupling time, typically on the order of tens of picoseconds
(275). Experiments and theoretical research reveal that two ablation mechanisms exist in the laser ablation process: (1) equilibrium
ablation (208,209,225,276–278) and (2) nonequilibrium ablation (122,279–281). We define the electron–phonon coupling time
sR and the laser pulse width sL. When sR/sL >> 1, it is the nonequilibrium ablation; when sR/sL << 1, it is the equilibrium ablation;
and when sR/sL w1, it can be called a mixed ablation.
As for the picosecond laser ablation process, the ablation mechanism is very complex due to the coexistence of equilibrium and
nonequilibrium mechanisms. The general non-Fourier law mentioned above is suitable to describe the heat conduction law of
picosecond laser ablation. However, until now, no perfect heat conduction model exists that describes the thermophysical
phenomenon with the laser pulse width ranging from nanoseconds to femtoseconds. A unified model would be helpful for the
experimental investigations on laser ablation.
sR
Defining a factor exp a , the following equations give a unified thermal model, which can describe the thermophysical
sL
phenomenon of the laser ablation process with laser pulse width ranges from nanosecond to femtosecond:

v sR v2
ce Te ¼ ke exp a Te gðTe Tl Þ þ ð1 RÞbI0 ðtÞexp ðbxÞ [56]
vt sL vx2
v
c1 T1 ¼ gðTe Tl Þ [57]
vt
Equations [56] and [57] describe the heat conduction process of electrons and lattice subsystem, respectively. In eqn [56], a is an
undetermined coefficient, which varies for different materials and is obtained by fitting theoretical results to experimental data.
When Te is below 3000–4000 K, the influence of the thermal excitation of electrons on the ce and ke is very small (265). Therefore,
ce and ke can be regarded as constants.
www.iran-mavad.com
Laser Ablation 161
The ratio of the electron–phonon coupling time and laser pulse width is an important criterion. As sR >> sL, that is,
sR
exp a /0, the ablation is nonequilibrium ablation. The unified model changes into the simplified two-temperature model:
sL
v
Ce Te ¼ gðTe Tl Þ þ ð1 RÞbI0 ðtÞexpðbxÞ; [58]
vt
v
Cl T ¼ gðTe Tl Þ [59]
vt l
This is the classic TTM model.
For the long pulses, when sR << sL, the ablation is a thermal equilibrium process, that is, T ¼ Te ¼ Tl, eqns [56] and [57] can be
changed into one equation:
v v2
Cn T ¼ k 2 T þ ð1 RÞbI0 ðtÞexpðbxÞ; [60]
vt vx
where Cn ¼ Ce þ Cl, k, and T are the specific heat, the thermal conductivity, and the temperature of the target, respectively. Equation
[60] becomes the classical equation of one-dimensional heat conduction. The first term of the right side in eqn [60] is the heat
conduction term of the target.
sR
When 0 < exp a < 1, it is the mixed process, including equilibrium and nonequilibrium ablations. Therefore, the unified
sL
thermal model can well describe the whole thermal conduction phenomenon, with laser pulse width ranging from the nanosecond
to femtosecond time scale.
For eqns [56] and [57], the initial condition is:
Te ðx; 0Þ ¼ Tl ðx; 0Þ ¼ T0 [61]
and the boundary conditions can be expressed as follows:
vTe
ke ¼ ð1 RÞI0 ðtÞ; [62]
vx x¼0
vTe
ke ¼ 0; [63]
vx d
where T0 ¼ 300 K is the initial temperature, which is uniform across the target, and d is the thermal diffusion depth. In the low-
fluence ablation case, the number density of the hot electrons is considered low enough that the energy transfer occurs only within
the area characterized by the skin depth d ¼ 1/ab. However, if the laser fluence is very high, the electron thermal diffusion length
rffiffiffiffiffiffiffiffiffiffiffi
ke
should be d ¼ sR (282).
Ce
Under the initial condition eqn [61] and the boundary conditions eqns [62] and [63], the heat conduction eqns [56] and [57] are
numerically solved by a finite difference scheme to investigate damage threshold fluences versus the pulse width. The thermal and
optical properties of gold in Table 5 are adopted.
Assuming that pulse width 101–103 ps, the value of a can be determined as 0.67 through fitting the experimental results with
computed results. The melting point of gold is 1337.33 K (275). The computed modeling results and experimental results are shown
in Figure 43.
In Figure 43, for the pulse duration shorter than 2 102 ps, the threshold fluence slowly increases, and for the longer pulse
duration, the threshold fluence rapidly increases with pulse duration. The results are more reasonable than those obtained by the
simple one-dimensional heat conduction equation (274).
The absorbance b ¼ 1 R of a pure metal consists of two parts, b ¼ b1 þ b2. b1 is the intrinsic absorbance, and b2 is the
contribution due to surface roughness. For an optically smooth metal surface, the order of b2 is about 1.2% of b1, while the role of
b2 is enhanced as the surface roughness increases. In fact, the damage threshold influence is an average value obtained by multipulse
ablation. For low pulse duration (i.e., less than 10 ps), the modified surface is obscured and the value of b2 is very small. For pulse
Table 5 Thermal and optical properties of gold
Wavelength (l) 1053 nm

Thermal conductivity of electron (ke) 318 W m1 K
Specific heat of electron (Ce) 67.7 J m3 K
Specific heat of lattice (Cl) 2.30 106 J m3 K
Absorption coefficient (ab) 7.88 105 m1
Reflectance (R) 0.94
e–ph coupling coefficient (g) 2.1 1016 W m3 K
The electron–phonon coupling time (sR) 6 ps
www.iran-mavad.com
162 Laser Ablation
Figure 43 Damage threshold fluences versus pulse width for the gold target irradiated by 1053 nm pulse. The experimental data is from Stuart, B. C. and
Feit, M. D. J. Opt. Soc. Am. B 1996, 13, 459.
duration longer than 10 ps, the value of b2 becomes bigger (21). In our simulation, we only take into account the intrinsic
absorbance b1 without considering b2. So for pulse duration longer than 10 ps, the theoretical data are a little bit higher than the
experimental ones.
The rapid development of the femtosecond laser has increased the demand for quantitative and predictive modeling of the fast
and highly nonequilibrium processes occurring in the target material by laser excitation, especially for electron temperatures higher
than 104 K. When the electron temperature is higher than 104 K, the electron DOSs and the energy band structure will change,
leading to the corresponding thermophysical parameters (including electron–phonon coupling and electron heat capacity) of the
target varying with temperature. The specific heat capacity of the electron is small, so its temperature can suddenly increase up to the
Fermi temperature after laser irradiation. When the electron temperature reaches 104 K, the thermal excitation of the electron below
the Fermi surface can significantly influence the thermophysical properties of the target (258–260,263,265). Consequently, using
TTM equations to describe such a process, we must consider the temperature-dependent parameters, such as the electron–phonon
coupling coefficient, the specific heat capacity of the electron, the absorption coefficient, and absorptivity. Also, investigation of the
DOS effect on the physical parameters is essential to quantitatively study the ultrahigh-energy femtosecond laser ablation target
(265–268).
The electronic structure of the target directly determines the effect of thermal excitation of electrons on the thermal physical
parameters because the different electronic structures can induce different DOS effects. For noble metals such as Au, the thermal
excitation from the low-laying d band results in the positive deviation of the heat capacity from the linear dependence at sufficiently
high electron temperatures. However, for transition metals, such as Ni, d band electrons through the Fermi surface lead to a high
density of electrons at the Fermi surface. Therefore, the d band electrons of Ni metal with high-density electronic states can be excited
to the s band, which has a low density of electronic states. This different electronic structure produces smaller thermal physical
parameters than the commonly used parameters at higher temperatures.
When the electron temperature is lower, the optical parameters (absorption coefficient and absorptivity) can be considered as
constant (263,265,267–269). However, experiments have confirmed that both the absorption coefficient and the absorptivity are
temperature dependent. If electron temperature is over 4000 K, both the thermal parameters vary with the electron temperature,
while if the electron temperature exceeds 10 000 K, the temperature can much more significantly affect the parameters (283–285). In
a word, since the DOSs of electrons significantly determine the thermophysical and optical parameters, an improved TTM equation
describing the heat conduction process at the electron temperature 10 000 K is expected.
Several theoretical studies (265–268) show that for electron temperature 10 000 K, the heat capacity of electrons as a function of
temperature can be expressed as:
Z N
vf ð3; m; Te Þ
Ce ðTe Þ ¼ ð3 3F Þ gð3Þde [64]
N vTe
" # 1
ð3 mÞ
where g(3) is the DOS of target, m is the chemical potential at Te, and the Fermi distribution f ð3; m; Te Þ ¼ exp þ1 .
kB Te
Using the calculated electronic structure by Vienna Ab-initio Simulation Package (VASP) software (265–268), the DOS affecting the
www.iran-mavad.com
Laser Ablation 163
heat capacity of electrons and the electron–phonon coupling coefficient of Ni and Au can be investigated to obtain the relationship
between the parameters and electron temperature. In Figure 44, the great difference between the dashed and solid lines fully shows
that when the electron temperature is high, the DOS effect must not be ignored.
In Ref. (258), a simple method, that is, fitting the solid line calculated by VASP software in Figure 44, was used to obtain fitting
expressions of electron heat capacity with temperature. For the Ni target,

54:2
ce ðTe Þ ¼ 30:72 Te 0:0745 105 J m3 K1 [65]
1 þ e 0:4
and for the Au target,

38:27
ce ðTe Þ ¼ 5:05 Te 0:8 105 J m3 K1 [66]
1þe 0:38
Similarly, the electron–phonon coupling coefficient with temperature also can be obtained by the above-mentioned method
(258). For the Ni target,

183:6
Ge ðTe Þ ¼ 1:31 þ Te þ0:39 1017 W m3 K1 [67]
1 þ e 0:14
and for the Au target,

2:07
Ge ðTe Þ ¼ 2:08 þ Te 0:8 1017 Wm3 K1 [68]
1þe 0:28
Substituting eqns [65]–[68] into the classic TTM model, an improved TTM model can be given as:
v v
Ce ðTe Þ ¼ ke Te GðTe ÞðTe Tl Þ þ Sðx; tÞ [69]
vx vx
v
Cl T ¼ GðTe ÞðTe Tl Þ [70]
vt l
The initial condition is:
Te ðx; 0Þ ¼ Tl ðx; 0Þ ¼ T0 [71]

and the boundary conditions can be expressed as follows:
vTe
ke ¼ ð1 RÞI0 ðtÞ [72]
vx x¼0
vTe
ke ¼0 [73]
vx d
Figure 44 Electron temperature dependences of the electron heat capacity of Ni(a) and Au(b). Solid lines show the results of the calculations performed
with DOS obtained from VASP. Dashed lines show the commonly used approximations of the thermophysical material properties.
www.iran-mavad.com
164 Laser Ablation
Table 6 Physical parameters and optical parameters of Ni and Au
Metal ke (W m1 K1) Cl (J m3 K1) a (K1)
Ni 91 4.1 4.3 103

Au 318 2.5 6.5 103
Using the improved TTM model with the initial and the boundary conditions, the surface temperature evolutions of Ni and Au
targets during laser ablation are investigated. Table 6 gives the physical and optical parameters of Ni and Au. And the laser
parameters of pulse width 100 fs, wavelength 800 nm, and laser energy density 0.65 J cm2 are used in the calculations. Figure 45
indicates the time-dependent temperatures of the electron and the lattice of the Ni-target surface.
From Figure 45, it can be seen that the electron temperature suddenly increases to 10 000 K and then slowly declines. The lattice
temperature gradually increases to a final equilibrium temperature of about 2500 K. The results suggest the electron–phonon
coupling coefficient of laser ablation of Ni sep z 2 ps.
Figure 46(a) and 46(b) indicate that the obtained threshold fluence of Ni (a) and Au (b) in TTM simulations performed with
thermophysical parameters calculated with DOS is obtained from VASP (dot lines), and the commonly used approximations
(dashed dot lines) and the experimental data from (286) for the melting thresholds. The theoretical results from the new improved
TTM equation and the experimental data are more consistent than the results from the traditional two-temperature equation. For
12000
11000
10000
electron temperature at 0.65J cm–2
9000 lattice temperature at 0.65J cm–2
8000
Temperature(K)
7000
6000
5000
4000
3000
2000
1000
0 1000 2000 3000

Time (fs)
Figure 45 The time-dependent temperatures of the electron and the lattice of the Ni-target surface.
Figure 46 Threshold fluence of Ni(a) and Au(b) for surface melting as a function of film thickness. The dot lines are the results considering the
DOS effect.
www.iran-mavad.com
Laser Ablation 165
Au, the theoretical values of the new improved TTM equation at film thickness larger than about 350 nm tend to a constant; the
theoretical and experimental data of Ni and Au show that the melting threshold increases rapidly with the film thickness and finally
tends to a constant value. The saturation threshold fluence of Ni obtained in the TTM simulations considering the DOS effect,
230 J m2, is a little higher than the experimental data 220 J m2 (286). The calculated threshold fluence with the commonly used
parameters, 275 J m2, is not in good agreement with the experimental data. The saturation threshold fluence at large film thick-
nesses obtained in the TTM simulations performed with the commonly used parameters, 1100 J m2, is significantly lower than the
experimental value, 1800 J m2 (286), whereas the value predicted in TTM simulations performed with the new thermophysical
parameters, 1600 J m2, is in a better agreement with the experimental data. The comparisons confirm that the DOS effect on
femtosecond laser ablation is significant.
Ultrashort-pulsed laser ablation technology is developing in leaps and bounds, and the femtosecond laser ablation mechanism
is becoming more complex. It is a pity that we cannot give a deeper and more comprehensive introduction to this research field in
the present chapter. The interested reader can refer to the related references, such as a recent review paper in Ref. (257).
References
1. Yang, J. J. Laser Optoelect. Prog. 2004, 41, 44 (in Chinese).

2. Maiman, T. H. Nature 1960, 187, 493.
3. Bromberg, J. The Laser in America, 1950–1970; MIT Press: Cambridge, MA, 1991; p 202.
4. Hohig, R. H.; Woolston, J. R. Appl. Phys. Lett. 1963, 2, 138.
5. Ready, J. F. Appl. Phys. Lett. 1963, 3, 11.
6. Smith, H. M.; Turner, A. F. Appl. Opt. 1965, 4, 147.
7. Singh, R. K.; Narayan, J. Phys. Rev. B 1990, 41, 8843.
8. Norton, D. P. Pulsed Laser Deposition of Complex Material: Progress Towards Applications. In Pulsed Laser Deposition of Thin Films; Eason, R., Ed.; Wiley-Interscience: New
York, 2007; p 3.
9. Maddox, S. J. British Welding Inst. Report, 1970, p 36.
10. Locke, E. V. U.S. Patent 3848104.
11. Liu, G. Y.; Toncich, D. J.; Harvey, E. C. Opt. Laser Eng. 2004, 42, 639.
12. Lee, S. K.; Na, S. J. Appl. Phys. A 1999, 68, 417.
13. Ruddock, I. S.; Bradley, D. J. Appl. Phys. Lett. 1976, 29, 296.
14. Fork, R. L.; Greene, B. I.; Shank, C. V. Appl. Phys. Lett. 1981, 38, 671.
15. Asaki, M. T.; Huang, C. P.; Garvey, D.; Zhou, J.; Kapteyn, H. C.; Murnane, M. M. Opt. Lett. 1993, 18, 977.
16. Keller, U. Nature 2003, 424, 831.
17. Lozhkarev, V. V.; Freidman, G. I.; Ginzburg, V. N.; Katin, E. V.; Khazanov, E. A.; Kirsanov, A. V.; Luchinin, G. A.; Mal’shakov, A. N.; Martyanov, M. A.; Palashov, O. V.;
Poteomkin, A. K.; Sergeev, A. M.; Shaykin, A. A.; Yakovlev, I. V.; Garanin, S. G.; Sukharevi, S. A.; Rukavishnikov, N. N.; Charukhchev, A. V.; Gerke, R. R.; Yashin, V. E. Opt.
Express 2006, 14, 446.
18. Tajima, T.; Mourou, G. Phys. Rev. Spec. Top. Accel. Beams 2002, 5, 03130121.
19. Huang, M.; Zhao, F. L.; Cheng, Y.; Xu, N. S.; Xu, Z. Z. Opt. Express 2008, 16, 19354.
20. Shinoda, M.; Gattass, R. R.; Mazur, E. J. Appl. Phys. 2009, 105, 053102.
21. Vorobyev, A. Y.; Guo, C. L. Phys. Rev. B 2005, 72, 195422.
22. Bubb, D. M.; Papantonakis, M. R.; Horwitz, J. S.; Haglund, R. F., Jr.; Toftmann, B.; McGill, R. A.; Chrisey, D. B. Chem. Phys. Lett. 2002, 352, 135.
23. Bloembergen, N.; Kurz, H.; Liu, J. M.; Yen, R. In Laser and Electron-Beam Interactions with Solids; Appleton, B. R., Celler, G. K., Eds.; North-Holland: New York,
1982; pp 3–60.
24. Garrison, B. J.; Itina, T. E.; Zhigilei, L. V. Phys. Rev. E 2003, 68, 041501.
25. Perez, D.; Lewis, L. J. Phys. Rev. Lett. 2002, 89, 255504.
26. Yoo, J. H.; Jeong, S. H.; Mao, X. L.; Greif, R.; Russo, R. E. Appl. Phys. Lett. 2000, 76, 783.
27. Pakhomov, A. V.; Thompson, M. S.; Gregory, D. A. J. Phys. D Appl. Phys. 2003, 36, 2067.
28. Kudryashov, S. I.; Allen, S. D. J. Appl. Phys. 2003, 93, 4306.
29. Bulgakova, N. M.; Bulgakov, A. V. Appl. Phys. A 2001, 73, 199.
30. Yoo, J. H.; Borisov, O. V. Anal. Chem. 2001, 73, 2288.
31. Peterlongo, A.; Miotello, A.; Kelly, R. Phys. Rev. E 1994, 50, 4716.
32. Kelly, R.; Miotello, A. Phys. Rev. E 1999, 60, 2616.
33. Zhang, D. M.; Li, Z. H.; Zhong, Z. C.; Li, X. G.; Guan, L. Dynamic Principle of Pulsed Laser Deposition, 66; Science Press: Beijing, 2011 (in Chinese).
34. Allmen, M.v.; Blatter, A. Laser-Beam Interactions with Materials; Springer: Berlin, 1998, 187.
35. Liu, X. D.; Mourou, G. IEEE J. Quantum Electron. 1997, 33, 1706.
36. Chichkov, B. N.; Momma, C.; Nolte, S.; Alvensleven, F. V.; Tunnermann, A. Appl. Phys. A 1996, 63, 109.
37. Hector, L. G.; Hetnarski, R. B. In Thermal Stresses, Hetnarski, R. B., Ed.; Elsevier: Amsterdam, 1996; p 453.
38. Lawrence, J.; Minami, K.; Li, L.; Edwards, R. E.; Gale, A. W. Appl. Surf. Sci. 2002, 186, 162.
39. Godovsky, Y. K. Thermophysical Properties of Polymers, 28; Spring-Verlag: New York, 1992.
40. Lawrence, J.; Li, L. Appl. Surf. Sci. 2000, 168, 71.
41. Li, J. F.; Li, L.; Stott, F. H. Int. J. Heat Mass Transfer 2004, 47, 1159.
42. Zhang, C.; Salama, I. A.; Quick, N. R. J. Appl. Phys. 2006, 99, 113530.
43. Zhang, C.; Quick, N. R.; Kar, A. J. Appl. Phys. 2008, 103, 014909.
44. Choudhury, I. A.; Shirley, S. Opt. Laser Technol. 2010, 42, 503.
45. Sharma, A.; Yadava, V. Opt. Laser Technol. 2012, 44, 159.
46. Yilbas, B. S.; Zhuja, S. Z.; Hashmi, M. S. J. J. Mater. Process. Technol. 2003, 136, 12.
47. Shanjin, Lv; Yang, W. Opt. Laser Eng. 2006, 44, 1067.
48. Triantafyllidis, D.; Bernstein, J. R.; Li, L.; Stott, F. H. J. Laser Appl. 2003, 15, 49.
49. Jiao, J. K.; Wang, X. B. Opt. Laser Eng. 2009, 47, 860.
50. Hong, L.; Zhang, Y.; Mi, C. L. Opt. Laser Technol. 2009, 41, 328.
www.iran-mavad.com
166 Laser Ablation
51. Yan, S.; Hong, Z.; Watanabe, T.; Tang, J. G. Opt. Laser Eng. 2010, 48, 732.
52. Duley, W. W., Ed. Laser Welding; John Wiley & Sons: NewYork, 1999; p 1.
53. Stolz, B.; Backes, G.; Gillne, A.; Kreutz, E. W. Appl. Surf. Sci. 1997, 109–110, 242.
54. Konopka, J.; Lewandowski, S. J.; Okunev, V. D.; Samoilenko, Z. A.; Abal’oshevet, A. Phys. C 2000, 341–348, 2147.
55. Adurodija, F. O.; Izumi, H. Appl. Surf. Sci. 2001, 177, 114.
56. Gnatyuk, V. A. SPIE 2000, 3890, 464.
57. Dawar, A. L. J. Mater. Sci. 1993, 28, 639.
58. Lai, W. C.; Yokoyama, M. J. Appl. Phys. 2000, 39, L1138.
59. Katsnelson, E. J. Appl. Phys. 1995, 77, 4604.
60. Aohi, T.; Hatanka, Y.; Look, D. C. Appl. Phys. Lett. 2000, 76, 3257.
61. Oh, M. S.; Hwang, D. K.; Lim, J. H.; Choi, Y. S.; Park, S. J. Appl. Phys. Lett. 2007, 91, 042109.
62. Ji, L. F.; Jiang, Y. J. Appl. Phys. Lett. 2004, 85, 1577.
63. Chang, L.; Jiang, Y. J. Appl. Phys. Lett. 2007, 90, 082505.
64. Zhao, Y.; Jiang, Y. J. J. Appl. Phys. 2008, 103, 114903.
65. Trtica, M. S.; Tarasenko, V. F.; Gakovic, B. M.; Fedenev, A. V.; Petkovska, L. T.; Radak, B.; Lipatov, E. I.; Shulepov, M. A. Appl. Surf. Sci. 2005, 252, 474.
66. Gower, M. C. In Laser Processing in Manufacturing, Crafer, R. C., Oakley, P. J., Eds.; Chapman and Hall: London 1993. Vol. 189.
67. Cappelli, E.; Orlando, S.; Sciti, D.; Montozzi, M.; Pandolfi, L. Appl. Surf. Sci. 2000, 154, 682.
68. Yue, T. M.; Yu, J. K.; Mei, Z.; Man, H. C. Mater. Lett. 2002, 52, 206.
69. Yymada, K.; Moristita, S.; Kutsuna, M.; Miyamoto, I. First International Symposium on High-Power Laser Macroprocessing[C]; Isamu, Miyamoto, 2003, p 65.
70. Milovanovic, D. S.; Radak, B. B.; Gakovic, B. M.; Batani, D.; Momcilovic, M. D.; Trtica, M. S. J. Alloy. Compd. 2010, 501, 89.
71. Golebiewski, M.; Kluezl, G.; Major, R.; Mróz, W.; Wierzchon, T.; Ebner, R.; Major, B. Mater. Chem. Phys. 2003, 81, 315.
72. Vorobyev, Y.; Guo, C. Phys. Rev. B 2005, 72, 195422.
73. Vorobyev, Y.; Guo, C. Opt. Photonics News 2007, 18, 43.
74. Vorobyev, Y.; Guo, C. J. Appl. Phys. 2008, 103, 43513.
75. Sipe, J. E.; Young, J. F.; Preston, J. S.; Driel, H. M.v Phys. Rev. B 1983, 27, 1141.
76. Young, J. F.; Preston, J. S.; Driel, H. M.v.; Sipe, J. E. Phys. Rev. B 1983, 27, 1155.
77. Oron, M.; Sorensen, G. Appl. Phys. Lett. 1979, 35, 782.
78. Ehrlich, D. J.; Brueck, S. R. J.; Tsao, J. Y. Appl. Phys. Lett. 1982, 41, 1.
79. Reif, J.; Costache, F.; Henyk, M.; Pandelov, S. V. Appl. Surf. Sci. 2002, 197–198, 891.
80. Walser, R. M.; Baecker, M. F.; Ambrose, J. G.; Sheng, D. Y. In Laser and Electron Beam Solid Interactions and Materials Processing, Vol. 177; Elsevier: New York, 1981.
81. Vechten, J. A.v. Solid State Commun. 1981, 39, 285.
82. Li, Z. H.; Li, P. N.; Fan, J. Q.; Fang, R. R.; Zhang, D. M. Opt. Laser Eng. 2010, 48, 64.
83. Bonch-Bruevich, M.; Libenson, M. N.; Makin, V. S.; Trubaev, V. V. Opt. Eng. (Bellingham) 1992, 31, 718.
84. Ozkan, A. M.; Malshe, A. P.; Railkar, T. A. Appl. Phys. Lett. 1999, 75, 3716.
85. Tan, B.; Venkatakrishnan, K. J. Micromech. Microeng. 2006, 16, 1080.
86. Wang, J. C.; Guo, C. L. J. Appl. Phys. 2005, 87, 251914.
87. Wang, J. C.; Guo, C. L. J. Appl. Phys. 2006, 100, 023511.
88. Vorobyev, A. Y.; Guo, C. J. Appl. Phys. 2008, 104, 063523.
89. Ma, H. L.; Guo, Y.; Zhong, M. J.; Li, R. X. Appl. Phys. A 2007, 89, 707.
90. Huang, M.; Zhao, F. L.; Xu, Z. Z.; Xu, N. S. Opt. Express 2008, 16, 19354.
91. Sudrie, L.; Franco, M.; Prade, B.; Mysyrowicz, A. Opt. Commun. 1999, 171, 279.
92. Cho, S. H.; Kumagai, H.; Midorikawa, K. Nucl. Instrum. Methods Phys. Res. B 2002, 197, 73.
93. Vorobyev, A. Y.; Topkov, A. N.; Gurin, O. V.; Svich, V. A.; Guo., C. L. Appl. Phys. Lett. 2009, 95, 121106.
94. Mevel, E.; Breger, P.; Trainhan, R.; Petite, G.; Agostini, P. Phys. Rev. Lett. 1993, 70, 406.
95. Glezer, E. N.; Mazur, E. Appl. Phys. Lett. 1997, 71, 882.
96. Bitenskii, I. S.; Murakhmetov, M. N.; Parilis, E. S. Sov. Phys. Tech. Phys. 1979, 24, 618.
97. Cheng, H. P.; Gillaspy, J. D. Phys. Rev. B 1997, 55, 2628.
98. Henyk, M.; Mitzner, R.; Wolfframm, D.; Reif, J. Appl. Surf. Sci. 2000, 154–155, 249.
99. Yablonovitch, E.; Bloembergen, N. Phys. Rev. Lett. 1972, 29, 907.
100. Pronko, P. P.; Dutta, S. K.; Du, D.; Singh, R. K. J. Appl. Phys. 1995, 78, 6233.
101. Gamaly, E. G.; Rode, A. V.; Luther-Davies, B.; Tikhonchuk, V. T. Phys. Plasmas 2002, 9, 949.
102. Stuart, B. C.; Feit, M. D.; Rubenchik, A. M.; Shore, B. W.; Perry, M. D. Phys. Rev. Lett. 1995, 74, 2248.
103. Li, M.; Menon, S.; Nibarger, J. P.; Gibson, G. N. Phys. Rev. Lett. 1999, 82, 2394.
104. Pronko, P. P.; VanRompay, P. A.; Horvath, C.; Loesel, F.; Juhasz, T.; Liu, X.; Mourou, G. Phys. Rev. B 1998, 58, 2387.
105. Schaffer, C. B.; Brodeur, A.; Mazur, E. Meas. Sci. Technol. 2001, 12, 1784.
106. Keldysh, L. V. Sov. Phys. JETP 1965, 20, 1307.
107. Tanaka, T.; Sun, H. B.; Kawata, S. Appl. Phys. Lett. 2002, 80, 312.
108. Chichkov, B. N.; Fadeeva, E.; Koch, J., et al. Proc. SPIE 2006, 6106, 610612.
109. Juodkazis, S.; Nishimura, K.; Misawa, H. Chin. Opt. Lett. 2007, 5, 198.
110. Maruo, S.; Ikuta, K.; Korogi, H. Appl. Phys. Lett. 2003, 82, 133.
111. Schaffer, C. B.; Garcia, J. F.; Mazur, E. Appl. Phys. A Mater. Sci. Proc. 2003, 76, 351.
112. Schaffer, C. B.; Jamison, A. O.; Mazur, E. Appl. Phys. Lett. 2004, 84, 1441.
113. El-Shall, M. S.; Li, S.; Turkki, T.; Graiver, D.; Pernisz, U. C.; Baraton, M. I. J. Phys. Chem. B 1995, 99, 17805.
114. Burr, T. A.; Seraphin, A. A.; Werwa, E.; Kolenbrander, K. D. Phys. Rev. B 1997, 56, 48l8.
115. David, B.; Geohegan, A.; Puretzky, A.; Gerd, D.; Pennycook, S. J. Appl. Phys. Lett. 1998, 72, 2987.
116. Geohegan, D. B.; Puretzky, A. A.; Duscher, G.; Stephen, J. P. Appl. Phys. Lett. 1998, 73, 438.
117. Hata, K.; Fujita, M.; Yoshida, S.; Makimura, T.; Murakami, K.; Shigekawa, H.; Mizutani, W.; Tokumoto, H. Appl. Phys. Lett. 2001, 79, 692.
118. Belomoin, G.; Therrien, J.; Smith, A.; Rao, S.; Twesten, R.; Chaieb, S.; Nayfeh, M. H.; Wagner, L.; Mitas, L. Appl. Phys. Lett. 2002, 80, 841.
119. Nakata, Y.; Muramoto, J.; Okada, T.; Maeda, M. J. Appl. Phys. 2002, 91, 1640.
120. Eliezer, S.; Eliaz, N.; Grossman, E.; Fisher, D.; Gouzman, I.; Henis, Z.; Pecker, S.; Horovitz, Y.; Fraenkel, M.; Maman, S.; Lereah, Y. Phys. Rev. B 2004, 69, 144119.
121. Amoruso, S.; Bruzzese, R.; Spinelli, N.; Velotta, R.; Vitiello, M.; Wang, X.; Ausanio, G.; Iannotti, V.; Lanotte, L. Appl. Phys. Lett. 2004, 84, 4502.
122. Amoruso, S.; Ausanio, G.; Bruzzese, R.; Vitiello, M.; Wang, X. Phys. Rev. B 2005, 71, 033406.
123. Albert, O.; Roger, S.; Glinec, Y.; Loulergue, J. C.; Etchepare, J.; Boulmer –Leborgne, C.; Perriere, J.; Milton, E. Appl. Phys. A Mater. Sci. Proc. 2003, 76, 319.
www.iran-mavad.com
Laser Ablation 167
124. Li, S.; Silvers, S. J.; El-Shall, M. S. J. Phys. Chem. B 1997, 101, 1794.
125. Movtchan, I. A.; Marine, W.; Dreyfus, R. W.; Le, H. C.; Sentis, M.; Autric, M. Appl. Surf. Sci. 1996, 96–98, 251.
126. MafunÓ, F.; Kohno, J. Y.; Takeda, Y.; Kondow, T.; Sawabe, H. J. Phys. Chem. B 2000, 104, 9111.
127. MafunÓ, F.; Kohno, J. Y.; Takeda, Y.; Kondow, T. J. Phys. Chem. B 2001, 105, 9050.
128. Kumar, B.; Thareja, R. K. J. Appl. Phys. 2010, 108, 064906.
129. Kazakevich, P. V.; Simakin, A. V.; Shafeev, G. A.; Viau, G.; Soumare, Y.; Verduraz, F. B. Appl. Surf. Sci. 2007, 253, 7831.
130. Ashfold, M. N.; Claeyssens, F.; Fuge, G. M.; Henley, S. J. Chem. Soc. Rev. 2004, 33, 23.
131. Suzuki, N.; Makino, T.; Yamada, Y.; Yoshida, T.; Onari, S. Appl. Phys. Lett. 2000, 76, 1389.
132. Murugesan, S.; Kuppusami, P.; Parvathavarthini, N.; Mohandas, E. Surf. Coat. Technol. 2007, 201, 7713.
133. Grigoriu, C.; Nicolae, I.; Ciupinaa, V.; Prodan, G.; Suematsu, H.; Yatsui, K. J. Optoelectron. Adv. Mater. 2004, 6, 825.
134. Ozawa, E.; Kawakami, Y.; Seto, T. Scr. Mater. 2001, 44, 2279.
135. Okubo, N.; Nakazawa, T.; Katano, Y.; Yoshizawa, I. Appl. Surf. Sci. 2002, 197–198, 679.
136. Fojtik, A.; Henglein, A.; Bunsen-Ges, B. Phys. Chem. 1993, 97, 252.
137. Neddersen, J.; Chumanov, G.; Cotton, T. M. Appl. Spectrosc. 1993, 47, 1959.
138. Peyrer, P.; Berthe, L.; Scherpereel, X.; Fabbro, R. J. Mater. Sci. 1998, 33, 1421.
139. Beahe, L.; Fabbro, R.; Peyre, P.; Bartnicki, E. J. Appl. Phys. 1999, 85, 7552.
140. Beahe, L.; Sollier, A.; Peyre, P.; Fabbro, R.; Bartnicki, E. J. Phys. D Appl. Phys. 2000, 33, 2142.
141. Peyer, P.; Berthe, L.; Fabbro, R.; Sollier, A. J. Phys. D Appl. Phys. 2000, 33, 498.
142. Sollier, A.; Berthe, L.; Fabbro, R. Eur. Phys. J. Appl. Phys. 2001, 16, 131.
143. Sakka, T.; Takatani, K.; Ogate, Y. H.; Mabuchi, M. J. Phys. D Appl. Phys. 2002, 35, 65.
144. Saito, K.; Sakka, T.; Ogata, Y. H. J. Appl. Phys. 2003, 94, 5530.
145. Simakin, A. V.; Obraztsova, E. D.; Shafeev, G. A. Chem. Phys. Lett. 2000, 332, 231.
146. Inoue, W.; Okoshi, M.; Inoue, N. Appl. Phys. A 2004, 79, 1457.
147. Kazakevich, P. V.; Voronov, V. V.; Simakin, A. V.; Shafeev, G. A. Quantum Electron. 2004, 34, 951.
148. Kumar, B.; Yadav, D.; Thareja, R. K. J. Appl. Phys. 2011, 110, 074903.
149. Lee, I.; Han, S. W. Chem. Commun. 2001, 1782.
150. Basov, N. G.; Danilychev, V. A.; Popov, Y.; Khodkevich, D. D. Zh. Eksp. Fiz. i Tekh. Pis’ma. Red. 1970, 12, 473.
151. Dijkkamp, D.; Venkateasan, T.; Wu, X. D. Appl. Phys. Lett. 1987, 51, 619.
152. Chrisey, D. B., Hubler, G. K., Eds. Pulsed Laser Deposition of Thin Films; Wiley: New York, 1994.
153. Greer, J. A.; Tabal, M. D. J. Vac. Sci. Technol. A 1995, 13, 1175.
154. Qian, F.; Singh, R. K.; Dutta, S. K.; Pronko, P. P. Appl. Phys. Lett. 1995, 67, 3120.
155. Okoshi, M.; Higuchi, S.; Hanabusa, M. J. Appl. Phys. 1999, 86, 1768.
156. Eason, R., Ed., Vol. 100; Wiley-Interscience: New York, 2007.
157. Harilal, S. S.; Bindhu, C. V.; Tillack, M. S.; Najmabadi, F.; Gaeris, A. C. J. Appl. Phys. 2003, 93, 2380.
158. Harilal, S. S.; O’Shay, B.; Tillack, M. S.; Mathew, M. V. J. Appl. Phys. 2005, 98, 013306.
159. Zhang, D. M.; Li, Z. H.; Yu, B. M.; Huang, M. T.; Guan, L.; Zhong, Z. C.; Li, G. D. Sci. China Ser. A Math. 2001, 44, 1485.
160. Zhang, D. M.; Guan, L.; Li, Z. H.; Zhong, Z. C.; Hou, S. P.; Yang, F. X. Acta Phys. Sin. 2003, 52, 242 (in Chinese).
161. Tan, X. Y.; Zhang, D. M.; Li, Z. H.; Fang, R. R. J. Phys. D Appl. Phys. 2008, 41, 035210.
162. Tan, X. Y.; Zhang, D. M.; Feng, S. Q.; Li, Z. H.; Fang, R. R.; Liu, G. B.; Sun, M. Chin. Phys. Lett. 2008, 25, 198.
163. Chen, K. R.; King, T. C.; Hes, J. H.; Leboeuf, J. N.; Geohegan, D. B.; Wood, R. F.; Puretzky, A. A.; Donato, J. M. Phys. Rev. B 1999, 60, 8383.
164. Chen, K. R.; Leboeuf, J. N.; Wood, R. F.; Geohegan, D. B.; Donato, J. M.; Liu, C. L.; Puretzky, A. A. Phys. Rev. Lett. 1999, 75, 4706.
165. Shibkov, V. M.; Aleksandrov, A. F.; Ershov, A. P.; Timofeev, I. B.; Chernikov, V. A.; Shibkova, L. V. Plasma Phys. Rep. 2005, 31, 895.
166. Harilal, S. S.; O’Shay, B.; Tao, Y. Z.; Tillack, M. S. J. Appl. Phys. 2006, 99, 083303.
167. Zhigilei, L. V.; Garrison, B. J. J. Appl. Phys. 2000, 88, 1281.
168. Geohegan, D. B. Appl. Phys. Lett. 1992, 60, 2832.
169. Wang, E. G. Prog. Phys. 2003, 23, 1 (in Chinese).
170. Guan, L.; Zhang, D. M.; Li, L.; Li, Z. H. Nucl. Instrum. Methods Phys. Res. B 2008, 266, 57.
171. Mayr, S. G.; Moske, M.; Samwer, K.; Taylor, M. E.; Atwater, H. A. Appl. Phys. Lett. 1999, 75, 4091.
172. Troyan, V. I.; Nevolin, V. N.; Maximov, G. A.; Filatov, D. O.; Lægsgaard, E. Phys. Rev. B 2002, 65, 073406.
173. Ohresser, P.; Shen, J.; Barthel, J.; Zheng, M.; Mohan, Ch. V.; Klaua, M.; Kirschner, J. Phys. Rev. B 1999, 59, 3696.
174. Ingen, R. P. v.; Fastenau, R. H. J.; Mittemeijer, E. J. J. Appl. Phys. 1994, 76, 1871.
175. Antoni, F.; Fogarassy, E.; Fuchs, C.; Grob, J. J.; Prevot, B.; Stoquert, J. P. Appl. Phys. Lett. 1995, 67, 2072.
176. Riet, E.; van de, Kools, J. C. S.; Dieleman, J. J. Appl. Phys. 1993, 73, 8290.
177. Voevodin, A. A.; Donley, M. S. Surf. Coat. Technol. 1996, 82, 199.
178. Stevefelt, J.; Collins, C. B. J. Phys. D Appl. Phys. 1991, 24, 2149.
179. Chey, S. J.; Cahill, D. G. Surf. Sci. 1997, 380, 377.
180. Pomeroy, J. M.; Jacobsen, J.; Hill, C. C.; Cooper, B. H.; Sethna, J. P. Phys. Rev. B 2002, 66, 235412.
181. Schiechl, H.; Rauchbauer, G.; Biedermann, A.; Schmid, M.; Varga, P. Surf. Sci. 2005, 594, 120.
182. Shin, B.; Leonard, J. P.; McCamy, J. W.; Michael, J. A. Appl. Phys. Lett. 2005, 87, 181916.
183. Agura, H.; Suzuki, A.; Matsushita, T.; Aoki, T.; Okuda, M. Thin Solid Films 2003, 445, 263.
184. Blank, D. H. A.; IJsselsteijn, R. P. J.; Out, P. G.; Kuiper, H. T. J.; Floksura, J.; Rogalla, H. Mater. Sci. Eng. B 1992, 13, 67.
185. Mosaner, P.; Bazzanella, N.; Bonelli, M.; Checchetto, R.; Miotello, A. Mater. Sci. Eng. B 2004, 108, 33.
186. Tsoutsouvab, M. G.; Panagopoulosb, C. N.; Kompitsasa, M. Appl. Surf. Sci. 2011, 257, 6314.
187. Chen, H. J.; Jia, C. H.; Zhang, X. N.; Zhang, W. F. Vacuum 2010, 85, 193.
188. Warrender, J. M.; Aziz, M. J. Phys. Rev. B 2007, 75, 085433.
189. Dale, D.; Fleet, A.; Suzuki, Y.; Brock, J. D. Phys. Rev. B 2006, 74, 085419.
190. Arias, J. L.; Mayor, M. B.; Pou, J.; León, B.; Pérez-Amor, M. Vacuum 2002, 67, 653.
191. Develos, K. D.; Kusunoki, M.; Mukaida, M.; Ohshima, S. Phys. C 1999, 320, 21.
192. Warrender, J. M.; Aziz, M. J. Appl. Phys. A 2004, 79, 713.
193. Willmott, P. R. Prog. Surf. Sci. 2004, 76, 163.
194. Zhu, Z.; Zheng, X. J.; Li, W. J. Appl. Phys. 2009, 106, 054105.
195. Moholka, A. V.; Shinde, S. S.; Babar, A. R.; Sim, K.-U.; Kwon, Ye-bin; Rajpure, K. Y.; Patil, P. S.; Bhosale, C. H.; Kim, J. H. Sol. Energy 2011, 85, 1354.
196. Warrender, J. M.; Aziz, M. J. Phys. Rev. B 2007, 76, 045414.
www.iran-mavad.com
168 Laser Ablation
197. Hinnemann, B.; Hinrichsen, H.; Wolf, D. E. Phys. Rev. Lett. 2001, 87, 135701.
198. Hinnemann, B.; Hinrichsen, H.; Wolf, D. E. Phys. Rev. E 2003, 67, 011602.
199. Lam, P.; Liu, S. J.; Woo, C. H. Phys. Rev. B 2002, 66 (4), 045408–045413.
200. Witten, T. A.; Sander, L. M. Phys. Rev. Lett. 1981, 47, 1400.
201. Zhang, Z. Y.; Chen, X.; Lagally, M. G. Phys. Rev. Lett. 1994, 73, 1829.
202. Zhang, Z. Y.; Chen, X.; Lagally, M. G. Science 1997, 276, 377.
203. Bruschi, P.; Cagnoni, P.; Nannini, A. Phys. Rev. B 1997, 55, 7955.
204. Taylor, M. E.; Atwater, A. H. Appl. Surf. Sci. 1998, 127–129, 159.
205. Kuzma, M.; Bester, M.; Pyziak, L.; Stefaniuk, I.; Virt, I. Appl. Surf. Sci. 2000, 168, 132.
206. Groot, J. S. D.; Estabrook, K. G.; Kruer, W. L.; Drake, R. P.; Mizuno, K.; Cameron, S. M. Phys. Fluids B 1992, 4, 701.
207. Li, L.; Zhang, D. M.; Li, Z. H.; Guan, L.; Tan, X. Y. Phys. B 2006, 383, 194.
208. Fang, R. R.; Zhang, D. M.; Li, Z. H.; Li, L.; Tan, X. Y.; Yang, F. X. Phys. Status Solidi A 2007, 204, 4241.
209. Zhang, D. M.; Liu, D.; Li, Z. H.; Guan, L.; Tan, X. Y.; Li, L.; Fang, R. R.; Hu, D. Z.; Liu, G. B. Appl. Surf. Sci. 2007, 253, 6144.
210. Amoruso, S. Appl. Surf. Sci. 1999, 138–139, 292.
211. Lunney, J. G.; Jordan, R. Appl. Surf. Sci. 1998, 127–129, 941.
212. Zel’dovich, Ya. B.; Raizer, Yu. P. In Physics of Shock Waves and High-Temperature Hydrodynamics Phenomena; Academic Press: New York, 1966;. Vol. 1.
213. Amoruso, S.; Toftmann, B.; Schou, J. Phys. Rev. E 2004, 69, 056403.
214. Amoruso, S. Appl. Phys. A 1999, 69, 323.
215. Inam, A.; Wu, X. D.; Venkatesan, T.; Ogale, S. B.; Chang, C. C.; Dijkkamp, D. Appl. Phys. Lett. 1987, 51, 1112.
216. Li, L.; Zhang, D. M.; Li, Z. H.; Guan, L.; Tan, X. Y.; Liu, D.; Fang, R. R. Phys. Status Solidi A 2006, 203, 906.
217. Fang, R. R.; Zhang, D. M.; Li, Z. H.; Yang, F. X.; Li, L.; Tan, X. Y.; Sun, M. Solid State Commun. 2008, 145, 556.
218. Huang, Y. L.; Yang, F. H.; Liang, G. Y.; Su, J. Y. Chin. J. Lasers 2003, 30, 449 (in Chinese).
219. Yang, L. B.; Liu, X. F.; Zhang, K. X.; Zhang, G. R.; Wang, C. Y. High Power Laser Part. Beams 1994, 6, 99 (in Chinese).
220. Lu, Y. W.; Lu, Q. S. Laser Technology in Military Applications, 31; National Defence Industry Press: Beijing, 1999 (in Chinese).
221. Zheng, Q. G.; Gu, J. H. Laser Interaction with Matter, 13; Huazhong University Press: Wuhan, 1996 (in Chinese).
222. Lu, J. Laser Interaction with Matter, 18; China Machine Press: Beijing, 1996 (in Chinese).
223. Hassan, A.; Elnicklawy, M.; Eladawi, M.; Hemida, A. Opt. Laser Technol. 1993, 25, 155.
224. Billings, B. H.; Frederikse, H. P. R. American Institute of Physics Handbook. 3rd ed; McGraw-Hill: New York, 1972; p 346.
225. Zhang, D. M.; Tan, X. Y.; Li, Z. H. Phys. B 2005, 357, 348.
226. Xu, X. F.; Willis, D. A. J. Heat Transfer 2002, 124, 293.
227. Miotello, A.; Kelly, R. Appl. Phys. Lett. 1995, 67, 3535.
228. Harrach, R. J. J. Appl. Phys. 1976, 47, 2473.
229. Singh, R. K. J. Non-Cryst. Solids 1994, 178, 199.
230. Zhang, D. M.; Li, Z. H.; Huang, M. T.; Zhang, M. J.; Guan, L.; Zou, M. Q.; Zhong, Z. C. Acta Phys. Sin. 2001, 50, 914.
231. Saidane, A.; Pulko, S. H. Microelectron. Eng. 2000, 51–52, 469.
232. Luikov, A. V. Int. J. Heat Mass Transfer 1966, 9, 139.
233. Taitel, Y. Int. J. Heat Mass Transfer 1972, 15, 369.
234. Sieniutycz, S. Int. J. Heat Mass Transfer 1977, 20, 1221.
235. Jiang, R. Q. Heat Conduction, Mass Diffusion and Transient Impact Effect of Momentum Transfer; Science Press: Beijing, 1997, 154.
236. Jiang, F. M.; Liu, D. Y. Adv. Mechanics 2002, 32, 128.
237. Joseph, D. D.; Presiosi, L. Rev. Mod. Phys. 1989, 61, 41.
238. Zhang, D. M.; Li, L.; Li, Z. H.; Guan, L.; Tan, X. Y. Phys. B 2005, 364, 285.
239. Zhang, D. M.; Li, L.; Li, Z. H.; Guan, L.; Tan, X. Y.; Liu, D. Eur. Phys. J. Appl. Phys. 2006, 33, 91.
240. Villars, P.; Phillips, J. C. Phys. Rev. B 1988, 37, 2345.
241. Sun, C. W.; Lu, Q. S.; Fan, Z. X.; Chen, Y. Z. The Effect of Laser Irradiation, 93; National Defence Industry Press: Beijing, 2002 (in Chinese).
242. Rohsenow, W. M.; Hartnett, J. P.; Ganic, E. N. Heat Transfer Hand Book; McGraw-Hill: New York, 1985; p 236.
243. Kaganov, M. I.; Lifshita, L. M.; Tanatarov, L. V. Sov. Phys. JETP 1957, 4, 173.
244. Anisimov, S. I.; Kapeliovich, B. L.; Perelman, T. L. Sov. Phys. JETP 1974, 39, 375.
245. Fujimoto, J. G.; Liu, J. M.; Ippen, E. P. Phys. Rev. Lett. 1984, 53, 1837.
246. Elsayed-Ali, H. E.; Norrisal, T. B.; Pessot, M. A.; Mourou, G. A. Phys. Rev. Lett. 1987, 58, 1212.
247. Allen, P. B. Phys. Rev. Lett. 1987, 59, 1460.
248. Brorson, S. D.; Kazeroonian, A.; Moodera, J. S.; Face, D. W.; Cheng, T. K.; Ippen, E. P.; Dresselhaus, M. S.; Dresselhaus, G. Phys. Rev. Lett. 1990, 64, 2172.
249. Chichkov, B. N.; Momma, C.; Nolte, S.; Alvensleben, F. V.; Tünnermann, A. Appl. Phys. A 1996, 63, 109.
250. Corkum, P. B.; Brunel, F.; Sherman, N. K. Phys. Rev. Lett. 1988, 61, 2886.
251. Kanavin, A. P.; Smetanin, I. V.; Isakov, V. A.; Afanasiev, Y. V.; Chichkov, B. N.; Wellegehausen, B.; Nolte, S.; Momma, C.; Tünnermann, A. Phys. Rev. B 1998, 57,
14698.
252. Gamaly, E. G.; Rode, A. V.; Uteza, O.; Samoc, M.; Luther-Davies, B. Appl. Surf. Sci. 2002, 197–198, 730.
253. Gamaly, E. G.; Madsen, N. R. Laser Part. Beams 2005, 23, 167.
254. Hetnarski, R. B.; Ignaczak, J. J. Therm. Stresses 1999, 22, 451.
255. Hetnarski, R. B.; Ignaczak, J. Int. J. Solids Struct. 2000, 37, 215.
256. Chen, J. K.; Grimes, L. E. Int. J. Solids Struct. 2002, 39, 3199.
257. Gamaly, E. G. Phys. Rep. 2011, 508, 91.
258. Fang, R. R.; Zhang, D. M.; Wei, H.; Hu, D. Z.; Li, Z. H.; Tan, X. Y.; Sun, M.; Yang, F. X. Eur. Phys. J. Appl. Phys. 2008, 42, 229.
259. Fang, R. R.; Hua, W.; Li, Z. H.; Zhang, D. M. Solid State Commun. 2012, 152, 108.
260. Fann, W. S.; Storz, R.; Tom, H. W. K. Phys. Rev. B 1992, 46, 13592.
261. Sun, C. K.; Vallée, F.; Acioli, L.; Ippen, E. P.; Fujimoto, J. G. Phys. Rev. B 1993, 48, 12365.
262. Sun, C. K.; Vallée, F.; Acioli, L. H.; Ippen, E. P.; Fujimoto, J. G. Phys. Rev. B 1994, 50, 15337.
263. Wang, J. C.; Guo, C. L. Appl. Phys. Lett. 2005, 87, 251914.
264. Wang, X. Y.; Riffe, D. M.; Lee, Y. S.; Downer, M. C. Phys. Rev. B 1994, 50, 8016.
265. Lin, Z. B.; Zhigilei, L. V. Appl. Surf. Sci. 2007, 253, 6295.
266. Povarnitsyn, M. E.; Itina, T. E.; Sentis, M.; Khishchenko, K. V.; Levashov, P. R. Phys. Rev. B 2007, 75, 235414.
267. Lin, Z. B.; Zhigilei, L. V. Phys. Rev. B 2008, 77, 075133.
268. Lin, Z. B.; Zhigilei, L. V. Proc. SPIE 2006, 6261, 62610U–1.
www.iran-mavad.com
Laser Ablation 169
269. Tsuchiya, T.; Kawamura, K. Phys. Rev. B 2002, 66, 094115.

270. Fang, R. R.; Zhang, D. M.; Hua, W.; Li, Z. H.; Yang, F. X.; Gao, Y. H. Laser Part. Beams 2010, 28, 157.
271. Schafer, C.; Urbassek, H. M.; Zhigilei, L. V. Phys. Rev. B 2002, 66, 115404.
272. Hirayama, Y.; Obara, M. J. Appl. Phys. 2005, 97, 064903.
273. Amoruso, S.; Ausanio, G.; Bruzzese, R.; Gragnaniello, L.; Lanotte, L.; Vitiello, M.; Wang, X. Appl. Surf. Sci. 2006, 252, 4863.
274. Stuart, B. C.; Feit, M. D. J. Opt. Soc. Am. B 1996, 13, 459.
275. Jiang, L.; Tsai, H. L. J. Heat Transfer 2005, 127, 1167.
276. Lu, Q. M.; Mao, S. S.; Mao, X. L.; Russo, R. E. Appl. Phys. Lett. 2002, 80, 3072.
277. Lu, Q. M. Phys. Rev. E 2003, 67, 016410.
278. Amoruso, S.; Armenante, M.; Bruzzese, R.; Spinelli, N.; Velotta, R.; Wang, X. Appl. Phys. Lett. 1999, 75, 7.
279. Guo, X. D.; Li, R. X.; Hang, Y.; Xu, Z. Z.; Yu, B. K.; Ma, H. L.; Lu, B.; Sun, X. W. Mater. Lett. 2009, 62, 1769.
280. Zeng, X.; Mao, X. L.; Greif, R.; Russo, R. E. Appl. Phys. A 2005, 80, 237.
281. Crouch, C. H.; Carey, J. E.; Warrender, J. M.; Aziz, M. J.; Mazur, E.; Génin, F. Y. Appl. Phys. Lett. 2004, 84, 1850.
282. Furusawa, K.; Takahashi, K.; Kumagai, H.; Midorikawa, K.; Obara, M. Appl. Phys. A 1999, 69, S359.
283. Huang, C. H.; Shen, H. Y.; Zeng, Z. D. Opt. Laser Technol. 1990, 22, 345.
284. Huang, Y. L.; Yang, F. H.; Liang, G. Y.; Su, J. Y. Chin. J. Lasers 2003, 30, 449.
285. Fisher, D.; Fraenkel, M.; Zinamon, Z.; Henis, Z.; Moshe, E.; Horovitz, Y.; Luzon, E.; Maman, S.; Eliezer, S. Laser Part. Beams 2005, 23, 391.
286. Wellershoff, S. S.; Hohlfeld, J.; Güdde, J.; Matthias, E. Proc. SPIE 1998, 3343, 378.
www.iran-mavad.com
www.iran-mavad.com
4.07 Surface Processing Using Cold Atmospheric Pressure Plasmas
DP Dowling, University College Dublin, Dublin, Ireland
4.07.1 Classification of Atmospheric Plasmas by Electron Temperature 171

4.07.2 Classification of Atmospheric Plasmas Based on Discharge Type 172
4.07.2.1 Dark Discharges 172
4.07.2.2 Glow Discharges 173
4.07.2.3 Arc Discharges 173
4.07.2.4 Industrial Examples of Atmospheric Plasma Processing Equipment 174
4.07.3 Surface Activation and Coating Deposition Using Cold Atmospheric Plasmas 176
4.07.4 Surface Treatments 176
4.07.5 Coating Deposition 176
4.07.6 Organosilicon Precursors 177
4.07.7 Plasma Medicine 182
4.07.8 Summary 183
Acknowledgments 183
References 183
Cold atmospheric pressure plasmas (APPs) have been applied since the late 1980s for material processing applications. These
include the treatment of metal surfaces prior to painting, for organic contaminant removal, for the activation of polymers prior to
adhesive bonding, as well as for the deposition of functional coatings. Applications of the latter include the deposition of insect
repellent coatings on textiles and protective coatings on solar cells. A particular advantage of these APP treatments is that for some
applications they can overcome the batch processing limitations of vacuum processes. Sectors that can or have already benefitted
from APP technology processing include automotive, food packaging, textiles, electronics, and biomedical industries. The clear
processing benefits and potential commercial applications of atmospheric pressure have led to a dramatic rise in the number of
publications focused on cold atmospheric plasma technology, in particular since 2000, as illustrated in Figure 1 (1).
In addition to the use of APPs for surface activation, cleaning, and coating deposition, another rapidly growing area of research is
the use of APPs in medical treatments, where the technology is known as plasma medicine. These cold APPs have shown potential in
applications as diverse as sterilization, wound healing, and cancer treatments, all of which have contributed to the sharp increase in
publications since 2010.
4.07.1 Classification of Atmospheric Plasmas by Electron Temperature
Plasmas can be divided into either thermal or nonthermal, the thermal plasmas being characterized by high levels of ionization. As
a result of increased ionic and atomic decomposition, the gas temperature and plasma density are elevated. The increased move-
ment of charged particles of electrons and ions in thermal plasmas results in a high flux of radical species. With the resulting high
400
No. of publications with keywords in title
350
300
250
200
150
100
50
0
90' 92' 94' 96' 98' 00' 02' 04' 06' 08' 10'
Year of publication
Figure 1 Publications with titles including the words ‘atmospheric’ and ‘plasma’. Reproduced from Thomson Reuters. Web of Knowledge. [Online]
2012. Available at: http://apps.webofknowledge.com/UA_GeneralSearch_input.do?product1/4UA&search_mode1/4GeneralSearch&SID1/4Y1okAi4CkFg
Mbb4KpIJ&preferencesSaved (accessed Sep 4, 2012).

www.iran-mavad.com
172 Surface Processing Using Cold Atmospheric Pressure Plasmas
energy density, a thermal or equilibrium plasma is formed (2). Their electron temperature (Te) is regarded as being in thermo-
dynamic and chemical equilibrium with the neutral gas temperature (Tgas). Thermal plasmas are associated with joule heating and
thermal ionization, which enables the delivery of high power at high operating pressures. They heat the entire gas stream during
operation. Typical examples of thermal plasma sources include plasma torches, plasma spray, and arc jets (3).
Nonthermal plasmas, often referred to as ‘cold’ plasmas or nonequilibrium plasmas, are plasmas operating at or close to room
temperature and ambient conditions. They are regarded as nonequilibrium systems because their Te, usually above 10 000 K is not
in equilibrium with the Tgas, which is approximately 300 K (3). In nonthermal plasmas, the electrical energy is primarily used in the
production of free electrons. These energetic electrons then produce excited species, free radicals, and ions, as well as additional
electrons through electron-impact dissociation, excitation, and ionization of background gas molecules. The electrons do not have
the ability to heat large and heavy molecules. The electrons are, however, involved in ionization reactions and are also responsible
for atomic/molecular excitations necessary to sustain the plasma. In the following section, atmospheric plasma systems are clas-
sified based on discharge type and electrode geometry.
4.07.2 Classification of Atmospheric Plasmas Based on Discharge Type
Having provided an introduction to thermal and nonthermal plasmas, this section examines APPs in more detail demonstrating
how the classification of these plasma sources can be based on their current–voltage characteristics (Figure 2) (4). As the applied
voltage is increased, the plasma discharge changes from one operational regime to another. The observed current–voltage
characteristics will be highly nonlinear, hence the different types of discharge are regarded as the ‘discharge regimes.’ The three
basic types of discharge are as follows: dark, glow, and arc discharges (5). In this discussion, ‘regime’ and ‘region’ are used
interchangeably.
4.07.2.1 Dark Discharges

This regime is termed dark discharge because with the exception of corona discharges and the breakdown itself, the discharge may
be invisible to the human eye. In Figure 2, the regime between A and E on the voltage–current characteristic represents this dark
discharge regime (6). The background ionization of the gas occurs in the region A–B. As the voltage is increased, the current
subsequently saturates the electrodes. A further increase in the voltage does not give rise to an increase in the current. This region is
called the saturation regime (B–C). As the voltage increases further, the current rises exponentially, and this region is called the
Townsend regime (C–E). A typical example of an atmospheric plasma source operating in the dark discharge regime is the corona
discharge (7,8), which operates in region D–E. It is a nonarcing, nonuniform plasma that ignites in the region of high electric field
generated by the sharp points of the electrodes (9). This is shown schematically in Figure 3. The corona discharge usually operates at
currents below the onset of arcing, and the plasma density drops off rapidly with increasing distance from the electrode.
Figure 2 Current–voltage graph showing the regimes of the classical DC electrical discharge. Reproduced from Ramamoorthy, A.; Rahman, M.;
Mooney, D. A.; DonMacElroy, J. M.; Dowling, D. P. Thermal Stability Studies of Atmospheric Plasma Deposited Siloxane Films Deposited on Vycor TM
glass. Surf. Coat. Technol. 2008, 202 (17), 4130–4136.
www.iran-mavad.com
Surface Processing Using Cold Atmospheric Pressure Plasmas 173
Figure 3 Schematic of a corona discharge.
4.07.2.2 Glow Discharges

The glow discharge owes its name to a typical luminous glow. The plasma gas emits light because the electron energy and density are
high enough to generate excited gas atoms by collisions. These excited gas atoms, which eventually relax to their ground state, emit
photons. As the voltage is increased, the discharge transitions discontinuously from E to F in Figure 2 and enters the normal glow
region (F–G). In this region, the voltage is almost independent of the current. The plasma volume increases and occupies the surface
of the electrodes. With a further increase in the voltage, the discharge transforms into an abnormal glow regime (G–H), where the
glow is brighter than in the normal glow regime. An example of plasma types that can operate in this glow discharge regime are
dielectric barrier discharges (DBDs) (10) and plasma jet systems (4,11). An example of a DBD system is shown schematically in
Figure 4. This source utilizes a dielectric, covering one or both of the electrodes, while the other is grounded. The purpose of the
dielectric film is to restrict and rapidly terminate the arcs that can form in the potential field between the two electrodes. The
discharge consists of a multitude of statically random but rapidly forming and equally rapidly terminated arcs that fill the volume
between the electrodes.
4.07.2.3 Arc Discharges

The region H to K in Figure 2 is termed the arc discharge regime because the electrode is sufficiently hot to induce
thermionic emission of electrons. In the H–I region, the discharge undergoes a glow-to-arc transition as sufficient current is
applied to the electrodes and voltage decreases with an increase in current. Nonthermal arcs operate in the I–J region, where
the voltage continues to decrease with an increase in current. After attaining large currents at point J, the voltage increases
slowly with an increase in current in the J–K region, and thermal arcs operate in this regime. Cold arcs (12) and plasma
torches (13) operate in the arc discharge regime. These plasma systems operate at high currents with high power density
(14). A typical example of an atmospheric plasma source operating in the arc discharge regime is the plasma torch
(Figure 5). Plasma torches and arc jets are thermal plasmas characterized by a high ion and electron temperature, nearly all
of which goes into heating the neutral gas and the entire gas stream. The gas stream is highly ionized and conductive and is
useful in high-temperature applications.
Typical examples of atmospheric plasma sources classified according to the three basic discharge types – dark, glow, and arc
discharges – detailed earlier are given in Table 1. This table also provides the operating gas, temperature, and electrode gap distances
of these plasma sources.
Figure 4 Typical configuration of a dielectric barrier discharge (DBD) system.
www.iran-mavad.com
Plasma torch
Figure 5 Typical configuration of a microwave plasma torch.
Table 1 A selection of atmospheric plasma discharges
Type Plasma source Gas or gas mixtures Electrode gap distance (mm) Reference
Dark discharge Point-to-plane corona discharge O2 20 Madani et al. (7)

Wire-to-plane corona discharge Air 8–16 Carreno et al. (8)
Townsend DBD system N2 1 Massines et al. (15)
LabLine™ DBD system He, N2, O2, Ar 5 Tynan et al. (10)
Glow discharge PlasmaStream™ jet system He, N2, CO2, O2 1–80 Nwankire et al. (11)
OAUGDP™ DBD system N2, He, air 1–5 Roth et al. (4)
Cold arc-plasma jet Air, N2, O2 15 Toshifuji et al. (12)
Arc Microwave plasma torch Air, Ar – Bae et al. (13)
4.07.2.4 Industrial Examples of Atmospheric Plasma Processing Equipment

In this section, examples of cold atmospheric plasma jets and reel-to-reel DBD processing equipment are discussed. These sources
can be used for both plasma activation of surfaces and, in some cases, coating deposition.
Some cold atmospheric plasma jets use helium (He) only or He combined with other gases including nitrogen and oxygen as the
excitation gas. Examples of commercial helium jet systems include the PlasmaStream (11) and the Sur-Fx (16) systems
(Figure 6). The PlasmaStream system is configured with a dielectric head housing a dual pin electrode mounted on either side of
a pneumatic nebulizer (Burgener Mira Mist) through which a precursor aerosol can be introduced. Driven in the electrical frequency
range of 10–25 kHz, the resultant aerosol and He discharge flow through the reactor tube, which is fabricated from Teflon or quartz.
Unlike the PlasmaStream system, the glow discharge Sur-Fx source, which operates at 13.56 MHz, is only used for the activation of
surfaces. Figure 6 also shows a photograph of an example of an air plasma system (OpenAir system) manufactured by
PlasmaTreat (17–22). The PlasmaTreat is a blown arc source where dry compressed air at a pressure 100 and 300 kPa is forced
through a 4-mm-diameter nozzle to form the air plasma jet.
For these three atmospheric pressure jet systems, the ejected ionized gas is directed onto a substrate 5–30 mm from the exit
nozzle. The use of the PlasmaStream system for plasma activation of polymers prior to heat-sealing and adhesive bonding has
been successfully demonstrated (23,24). The application of the PlasmaTreat source for the plasma activation of polymers (25),
adhesive bonding (26,17), and textile polymer treatments (18) has also been reported.
In addition to the cleaning and activation of surfaces, plasma jets can be used for coating deposition via the introduction of
plasma polymerizable precursors into the plasma. For example, in the case of the PlasmaTreat system, liquid monomers are
evaporated in a heating chamber prior to being brought into the plasma. Plasma polymerized coatings from this system have been
evaluated for gas barrier applications (19) and corrosion protection (20). In the case of the PlasmaStream system, a liquid
precursor is fed into the plasma through a nebulizer (27). These nebulizers utilize a parallel path mechanism whereby a stream of
liquid precursor converges with a nebulizing gas stream resulting in the formation of aerosolized droplets approximately 5 microns
Figure 6 Examples of atmospheric plasma jet systems. The 5.0-cm-long orifice of the Sur-Fx He discharge (left). The 1.6 cm orifice of the
PlasmaStream™ He discharge (middle). The 0.8 cm orifice of the PlasmaTreat™ air plasma (right).
www.iran-mavad.com
Figure 7 Schematic and photograph of the LabLine atmospheric pressure plasma system for the reel-to-reel processing of polymers. The equipment
incorporates a system for the nebulization of liquid precursors as an aerosol into the plasma as well as process monitoring using optical emission
spectroscopy (OES).
in diameter. Downstream, these droplets combine with a highly charged process gas and undergo plasma polymerization largely at
the substrate surface. This process is known as atmospheric pressure plasma liquid deposition (APPLD), as the liquid precursor is
injected directly into the plasma.
A number of reel-to-reel processing systems have also been developed. These include a web system developed by the Belgium
research group Vito for the activation and coating (using liquid precursor) of polymer webs up to 0.6 m in diameter (28). Another
example is the LabLine reel-to-reel system developed by DowCorning. A photograph and schematic of this latter system is shown in
Figure 7. This reel-to-reel processing tool is capable of plasma activating or coating (liquid precursors) polymer films up to 320 mm
wide. It comprises of two vertical parallel-plate electrodes. The inner electrode of each pair is powered by the supply unit, while the
outer electrodes are grounded. Each electrode comprises a conductive liquid (salt and water solution) housed within a dielectric
perimeter. The electrode gap is 5 mm, allowing flexible materials up to approximately 3 mm thick to be treated. The precursor is
nebulized as an aerosol into the ‘top box’ using a Burgener nebulizer system. As shown in Figure 8, the LabLine system was scaled up
by DowCorning to treat webs up to 0.9 m wide (AP4 system) (10). Subsequently, this helium-based plasma/liquid delivery pro-
cessing technology was further developed by Triton to treat webs with widths up to 1.8 m diameter (Figure 8) (29). One application
of the Triton technology is the deposition of antimicrobial and insect repellent coatings on textiles.
An example of an industrial scale atmospheric plasma system for the removal of organic contaminants is the multiple rotary
nozzle system developed by PlasmaTreat (22). This has been used for the treatment of 3-m-diameter aluminum panels used in the
manufacture of refrigerated trucks, prior to adhesive bonding (Figure 8). One of the environmental advantages of the atmospheric
plasma treatment is the elimination of organic solvents, which were previously used for cleaning the panels.
Figure 8 Triton RC1000 helium-based atmospheric plasma system for coating textiles and polymers sheets/web up to 1.8 m wide (left) (reproduced
from www.tritonsys.com). PlasmaTreat™ system for the air plasma removal of contaminants from aluminum sheets web up to 3 m wide (right)
(reproduced from http://www.plasmatreat.com).
www.iran-mavad.com
4.07.3 Surface Activation and Coating Deposition Using Cold Atmospheric Plasmas
As detailed earlier, atmospheric plasma can be characterized as thermal and nonthermal based on the type of discharge. Among the
applications of thermal plasma are the deposition of hydroxyapatite coatings by plasma spray (30), chemical waste destruction,
silicon etching, and the treatments of vulcanized rubber (31,32). Some of the applications of nonthermal plasmas include the
generation of ozone for disinfection and the degradation of organic pollutants in water (33), air pollution control by the removal of
acidic (SOx, NOx) and greenhouse gases (CH4, CO2), as well as the decomposition of toxic volatile organic compounds (34). This
section is focused on applications of cold atmospheric plasma surface treatments and the deposition of coatings.
4.07.4 Surface Treatments
Surface treatment methods include plasma etching, which involves the selective removal of material from a substrate, a major
process requirement of semiconductor manufacturing (35), plasma cleaning, a process of removing contaminant from a material
surface (oil, dust, oxides, chemicals, and biological agents); and plasma activation, a process that involves both chemical and
physical changes of the material surface, where nonpolymer-forming plasma active species bombard a material surface in order to
give it specific properties such as modifying its surface energy (36). Both low and APP treatments are widely used for the surface
activation of polymers prior to coating, printing, bonding, or painting processes. Polymers generally have low surface energy (37)
leading to problems such as poor wettability, dyeability, and adhesion. A wide range of techniques have been used to overcome this
problem. These include the use of chemical treatments (38), flame (39), corona (40), as well as both low pressure and APPs (41,26).
A particular advantage of plasma treatments for activating polymers is the uniformity of the treatment (42). The depth of modi-
fication for plasma treatment is generally less than 10 nm (43). One of its effects is to remove weak boundary layers, hence
strengthening adhesive bonds (44). Chain scission of the long polymer molecules may also occur, thus reducing the level of
polymer crystallinity and generating chemical sites that are available for bonding with an adhesive. For example, the incorporation
of functional groups containing oxygen and nitrogen into the surface has been demonstrated after plasma activation (41). It has
been shown that even if only a few chemical sites are created, there will be a large increase in adhesive strength (45). Removal of
surface contaminants is also an important contribution of plasma treatment to polymer adhesion (31). A further advantage of the
treatment of polymers using plasmas is that the subsequent temperature required for heat-sealing the polymer has been demon-
strated to be reduced. For example, the atmospheric plasma pretreatment of polyethylene terephthalate polymer used in food
packaging, achieved using a helium or a helium oxygen discharge, has been shown to yield up to a 25-fold increase in polymer peel
strength compared to that obtained with the unactivated polymer (23).
4.07.5 Coating Deposition
Coatings deposited using nonthermal APP systems are usually achieved using plasma-enhanced chemical vapor deposition
(PECVD) processes, where the energy to initiate polymerization comes from the plasma species. The coating process is usually
preceded by both plasma cleaning and substrate activation stages, which can be performed together.
The process of plasma polymerization involves reactions between plasma species, between plasma and surface species, and
between surface species (46). It is generally considered to proceed via a mechanism that involves free radicals, along with two
separate mechanisms for polymerization of organic monomers (47). These are plasma-induced polymerization and plasma-state
polymerization. In the case of the former, the plasma initiates polymerization, which is then thought to proceed via a conven-
tional polymerization mechanism. Plasma-state polymerization, however, can only occur in the plasma glow discharge, where the
continuous reinitiation of oligomers (fragments) formed by the recombination of free radicals occurs in a repeated cycle and leads
to the formation of high molecular weight compounds (47). Both mechanisms can occur in PECVD processes, although it is
considered that plasma-state polymerization is the main step for polymer formation. This is due to the high concentration of free
radicals present in the plasma, which favors recombination of free radicals (termination) as the mechanism for plasma-state
polymerization rather than the propagation reaction, which occurs by addition reactions onto double bonds that comprise the
main step for conventional polymerization (47).
After interaction with plasma species, the monomer is fragmented by breaking one of the bonds present. Complete fragmen-
tation of the monomer in the plasma is avoided by keeping the plasma intensity relatively low and achieving short precursor dwell
times by maintaining a rapid throughput of the carrier gas. Processing conditions have thus to be optimized for each precursor type.
The precursor fragments that partially preserve the chemical structure of the monomer are highly reactive, and a highly cross-linked
coating is achieved by free radical polymerization mechanisms.
The plasma polymer formed by PECVD processes generally exhibits short chain lengths that are randomly branched and
terminated with a high degree of cross-linking. Due to this formation mechanism, nearly all organic compounds can be poly-
merized, though some groups may be absent in the resulting polymer. In some cases, a large number of free radicals can become
trapped within the network, which cannot recombine rapidly and as a result may react with oxygen and water in the atmosphere,
leading to increased aging of the plasma deposited polymer. The properties of these polymer coatings are strongly dependent on the
processing conditions, such as discharge power, frequency of excitation, gas flow rate, monomer type, and flow rate. The next section
helps to demonstrate the versatility of APPs to act as a processing technique for the deposition of plasma polymerized coatings. This
www.iran-mavad.com
section takes as an example organosilicon precursors and highlights how the resulting coatings can be tailored using the plasma
processing conditions to exhibit water contact angles from superhydrophilic to superhydrophobic as required.
4.07.6 Organosilicon Precursors
The deposition of silicon dioxide (SiO2) films by low pressure PECVD processes was a major contributor to the application of large-
scale silica deposition processes. Plasma deposited SiO2 films exhibit unique properties, such as low gas and moisture permeability
at very low thicknesses (<20 nm), thus producing effective barrier layers that could be applied to food or pharmaceutical packaging
(48) or as anticorrosion treatments (49). SiO2 layers also have applications as hard coatings on soft substrates and coatings for
ophthalmic lenses creating transparent antiscratch coatings, and more recently they have shown promising results in the area of
biomaterials engineering with respect to protein adsorption (50). Silanes (Si–H4) were originally used in PECVD deposition of SiO2
layers. However, the desire to treat large areas, develop in-line processes, or treat different substrate materials gave rise to problems
associated with adhesion, diffusion, thermal expansion, and limited step coverage on submicron level features when using these
silane precursors. The use of organosilicon precursors overcame many of these issues and enabled the deposition of SiO2 layers for
large-scale processes on different substrate materials. An organosilicon molecule describes any molecule consisting of one atom of
silicon and at least one organic group such as a methyl group (CH3). Organofunctional silanes consist of two different reactive
atoms on their silicon atom, which can react and couple with different materials. The presence of at least one organic group attached
to the silicon atom (e.g., Si–CH3) facilitates the transition between a ‘soft’ polymeric surface and SiO2 by controlling the deposition
process conditions (51). The carbon-silicon bond is very stable, nonpolar, and in the presence of an alkyl group it gives rise to low
surface energy and hydrophobic effects.
Siloxane monomers provide a large number of possible reactants for plasma polymerization reactions and are generally suffi-
ciently volatile near room temperature, nontoxic, nonflammable, commercially available, and relatively cheap. For example,
siloxane monomers are the constituents of many silicone-based materials and have been used in the biomedical industry for the
past 50 years (52). They are used in many different biomaterial applications, such as dental implants and catheter tubing, and they
are generally considered nontoxic. There are many different organosilicon monomer chemistries available, and they have been
reviewed by Smith (53) and others. Two of the most commonly researched precursors are hexamethyldisiloxane (HMDSO) and
tetraethyl orthpsilicate (TEOS). HMDSO exhibits an Si–O–Si bridge structure while TEOS exhibits a tetrahedral structure, both of
which are shown in Figure 9.
By selecting appropriate siloxane and fluorinated siloxane precursors, it has been demonstrated that coatings exhibiting water
contact angles (q) ranging from hydrophilic (q < 5 ) to superhydrophobic (q > 150 ) can be obtained. This is demonstrated in
Figure 10 (54). In this study, the siloxane coatings were deposited from TEOS and HMDSO precursors, while fluorinated siloxane
coatings (TCFS) were deposited from an equal volume mixture of tetramethylcyclotetrasiloxane (Tomcats (TC)) and perfluoro-
octyltriethoxysilane (FS). The coatings were deposited onto silicon wafer substrates using the PlasmaStream jet system (Figure 11).
The jet plasma was formed using a mixture of helium–nitrogen gas for the deposition of the hydrophobic and superhydrophobic
coatings. To achieve hydrophobic siloxane coatings, both the HMDSO and TCFS precursor monomers were nebulized into the
plasma at a flow rate of 5 ml min1. Superhydrophobic siloxane coatings were deposited by reducing the precursor flow rate to
3 ml min1, thus increasing the ratio of reactive species to a precursor monomer in the plasma discharge, yielding a coating that
exhibited substantially increased surface roughness. In order to obtain hydrophilic coatings, it is necessary to partially oxidize the
TEOS monomer by the addition of O2 into the He plasma during coating deposition. The SiO2 substrate in Figure 10 is that of the
uncoated silicon wafer substrate. Note in this figure that the increase in the water contact angle (increase in hydrophobic properties)
is associated with a corresponding decrease in surface energy.
The morphology of the plasma polymerized coatings can vary enormously depending on the plasma processing conditions in
addition to the type of monomer used and its flow rate. For example, in a study carried out with a TEOS precursor deposited using
a He plasma (PlasmaStream source), it was observed that either high flow rates or poor atomization can result in the formation of
large liquid precursor droplets, the molecules of which were not broken down by the plasma. This can give rise to a coating with
a dropletlike morphology. Figure 12 compares the latter morphology with the relatively homogeneous coating obtained under
optimized deposition conditions. It was also observed that in some cases unreacted TEOS can be incorporated into the coating. This
evaporates over time, resulting in a gradual decrease in thickness of the siloxane coating (55). Another issue with high precursor flow
rates was the formation of significant numbers of particulates. This may be due to the presence of larger droplets of liquid precursor,
which react in the gas phase and arrive at the substrate surface as larger solid particles. Even under optimized deposition conditions,
CH3 CH3 OC2H5
CH3 Si O Si CH3 H5C2O Si OC2H5
CH3 CH3 OC2H5

HMDSO TEOS
Figure 9 The chemical structure of HMDSO and TEOS.
www.iran-mavad.com
Figure 10 Water contact angle and surface energy of uncoated and plasma polymer coated silicon wafer substrates. Variation in monomer
chemistry and plasma processing conditions such as gas composition and input power result in plasma polymerized coatings with wetting properties
ranging from superhydrophilic to superhydrophobic. Reproduced from Stallard, C. P.; McDonnell, K. A.; Onayemi, O. D.; O’Gara, J. P.; Dowling,
D. P. Evaluation of Protein Adsorption on Atmospheric Plasma Deposited Coatings Exhibiting Superhydrophilic to Superhydrophobic Properties.
Biointerphases 2012, 7 (1–4), 31.
Figure 11 He atmospheric plasma jet formed using the PlasmaStream™ system used in the treatment of a polymer catheter tube.
Figure 12 Optical profilometry images of a coating deposited from TEOS on a polystyrene substrate at both a low flow rate, 10 ml min1 (left),
and a high flow rate, 50 ml min1 (right), scale 55.5 44.5 mm.
www.iran-mavad.com
a feature of atmospheric plasma deposited coatings can be the incorporation of particulates such as those shown in Figure 13, again
for the coating deposited from TEOS (55). This cross-sectional image of the coating deposited onto a microporous silica support
demonstrates how the particulates are embedded in the coating. Studies of particle formation in silane plasmas by Howling et al.
(56) have indicated that anions play a major role in plasma polymerization processes. DeBleecker et al. (57) also concluded, from
an investigation of silane plasmas, that it was SiH3 produced by dissociative electron attachment that was the starting point of anion
formation, and they were considered as the most important initiator of particulate formation.
Fluorinated Siloxane Precursors. Fluorocarbon, also known as fluoropolymer coatings, is a Teflon-like coating that has been
extensively investigated as a method of modifying material surfaces due to their surface energy, wettability, friction coefficient,
chemical inertness, low dielectric constant, and interactions with biological systems (58). Fluoropolymer coatings deposited in
most PECVD processes have CF3, CF2, CF-, and C moieties assembled in a random network rather than in ordered chains. They have
applications in microelectronics, textiles, and anticorrosion coatings. Due to their chemical inertness and low surface energy, they
have also been studied in biological applications and have been shown to exhibit nonthrombogenic properties, a resistance to
platelet adhesion, activation, a resistance to protein adsorption, and blood compatibility (59,60). While a detailed study of blood
compatibility of fluoropolymer films has not yet been carried out, it has been shown that the adsorption of fibrinogen, a protein
present in blood and involved in the coagulation cascade, is related to the fluorine–carbon (F/C) ratio, surface morphology, and
wettability (61). These properties can be controlled by varying the deposition process conditions. Fluorinated siloxane copolymer
blends have also been deposited by PECVD processes, and they exhibit properties that will help to reduce marine fouling (62).
Superhydrophobic (SH) coatings. Due to their antistick and self-cleaning properties, SH coatings have been shown to exhibit
potential in applications ranging from antibiofouling paints for boats, biomedical applications, corrosion resistance, self-cleaning
windshields and automobiles, stain resistant textiles, separation of water and oil, antifouling coatings, and so on (63). As high-
lighted earlier, APPs have also been used for the deposition of SH coatings, which generally exhibit a low polar chemistry in
conjunction with a high surface roughness, such as the needlelike morphology shown in Figure 14. The advantage of using plasmas
Figure 13 Plasma polymerized SiOx rich coating deposited from TEOS onto a microporous silica substrate. The particulates shown on the surface
of the coating are typically 140 nm (50 nm) in diameter. Reproduced from McCann, M. T. P.; Mooney, D. A.; Dowling, D. P.; MacElroy, J. M. D.
Silica Nanofilms Deposited by Atmospheric Pressure Plasma Liquid Deposition. Thin Solid Films 2012, 520 (7), 2619–2626.
nm
250
153°
200
150
100
120 m
50
91 m
–10
Figure 14 Optical profilometry image of the HMDSO superhydrophobic coating (contact angle 153 ) showing the needlelike morphology. Reproduced
from Nwankire, C. E.; Favaro, G.; Duong, Q.-H.; Dowling, D. P. Enhancing the Mechanical Properties of Superhydrophobic Atmospheric Pressure
Plasma Deposited Siloxane Coatings. Plasma Process. Polym. 2011, 8 (4), 305–315.
www.iran-mavad.com
for this application is first the relative speed of SH coating deposition as generally only a single-step deposition process is required.
Among the precursors used for these atmospheric plasma deposited SH coatings, there have been hexamethyldisilazane (HMDSN)
(64), HMDSO (65), and fluorocarbon gases (66). Yang et al. reported the deposition of SH films using O2/HMDSN as the precursor
in an Ar plasma jet system (64). There have been also very few reports on the study of mechanical abrasion/friction resistance of SH
surfaces, which is one of the most important practical limitations of these self-cleaning surfaces (67,63). Nature has already
addressed this mechanical robustness issue on the Lotus leaf. Epicuticular wax is secreted continuously to regenerate the micro and
nanorough structures responsible for superhydrophobicity on this leaf (68). The ability of plasma polymerized superhydrophobic
coatings to resist the adhesion of proteins and bacteria has also been demonstrated (54).
Primer Coatings. As detailed earlier, the APPLD technology has the ability to plasma polymerize precursor monomers, forming
coatings that retain the chemical functionality of the original precursor. One application of this is to deposit plasma polymerized
primer coatings that, for example, can enhance metal-polymer adhesion (69). O’Neill et al. of DowCorning was the first to propose
the use of APPLD technology for the deposition of a plasma polymerized primer (PPP) (27). In his paper, a siloxane-based adhesive
primer was used to form an interlayer, which, due to the energetic plasma treatment, formed a strong chemical ‘sandwich bond’
between a metal substrate and a silicone elastomer. The chemistry of the PPP can be tailored to suit the bond type. This is further
demonstrated based on the application of a silicalike primer layer, which was shown to enhance the adhesion of aluminum to an
epoxy adhesive (70). The primer layer was obtained from an HMDSO precursor deposited under oxidizing conditions. It was
suggested that the enhanced adhesion was due to the ability of this primer to chemically react with the epoxy adhesive. When
a different precursor trimethyl silane (TMS) was evaluated in the same study, however, poor adhesion was observed. This poor
adhesion was attributed to the lack of a chemical reaction between TMS and the epoxy adhesive.
In another study, a plasma polymerized primer was deposited from a mixture of Tetraethyl orthosilicate (TEOS) and poly-
hydrogenmethyl siloxane liquid precursors using the PlasmaStream system (71). This mixture produced an Si–H rich siloxane
coating on a stainless steel substrate onto which a silicone elastomer was bonded. The surface roughness of the steel substrate was
found to increase substantially with the application of the siloxane coating. This was mainly found to be associated with the
presence of particulates embedded in the deposited coating. Based on peel tests, the silicone elastomer bonded to the plasma
polymerized coated steel exhibited a 15-fold increase in adhesive fracture energy. This enhanced adhesion was found to be asso-
ciated with the high Si–H ratio in the primer mixture and the high surface roughness, which could potentially enhance mechanical
interlocking with an adhesive. As demonstrated in Figure 15, the enhancement in peel strength achieved with the primer layer is
substantially higher than that achieved using a plasma surface activation only.
Biomaterial applications. Table 2 provides examples of the potential use of plasma processing technology in the biomaterials
industry (72).
APPs are increasingly being investigated for use in biomedical applications. These include surface modification of biomaterials
to enhance implant integration, the development of targeted drug delivery systems for more effective localized treatment of diseases,
as well as therapeutic applications such as wound healing and sterilization.
When an implant material is placed within the body, there are a number of interactions that occur. These interactions take place
at the interface between the material surface and the biological environment. As a result, low-temperature atmospheric plasma
modification offers a potentially excellent route to alter the surface properties of an implant material to enhance integration, while
retaining the operational functionality provided by the bulk material.
Atmospheric plasmas have also been used for the application of biofunctional coating to reduce inflammation, which may result
in the formation of biofilms and bacteria and consequently cause the rejection of implant materials. To enhance the
180
160
Adhesive fracture energy, J m–2
140
120
100
80
60
40
20
0
Untreated Activated PPP
Figure 15 Peel test equipment (left). Enhanced adhesion between stainless steel and a siloxane elastomer achieved with plasma polymerized
primer (PPP) (right). Note the substantial increase in adhesion with the primer compared with the plasma-only treatment (activated). Reproduced from
Nwankire, C. E.; Dowling, D. P. Influence of nm Thick Atmospheric Plasma Deposited Coatings on the Adhesion of Silicone Elastomer to Stainless
Steel. J. Adhes. Sci. Technol. 2010, 24 (7), 1291–1302.
www.iran-mavad.com
Table 2 Common research areas and application of plasma treatments in biomaterials engineering
Area Application
Blood-compatible surface Vascular grafts, catheters, stents, heart-valves, membranes (hemodialysis), filters (blood cell separation),
biomolecules immobilized on surfaces
Nonfouling surfaces Intraoculars (IOLs) contact lenses, wound healing, catheters, biosensors
Tissue engineering and cell culture Cell growth, antibody production, essays, vascular grafts
Sterilization of surgical tools and devices Cutting tools of surgeons, tweezers
Biosensors Biomolecules immobilized on surfaces
Barrier coatings Controlled drug-release, gas-exchange membranes, device protection, corrosion protection,
reduction of leaches (e.g., additives, catalysts, plasticizers, etc.)
Reproduced from Favia, P.; d’Agostino, R. Plasma Treatments and Plasma Deposition of Polymers for Biomedical Applications. Surf. Coat. Technol. 1998, 98, 1102–1106.
hemocompatibility of blood contacting biomaterials, it is often beneficially to reduce the attachment of serum proteins, which can
lead to the formation of thrombin, inflammation, and implant rejection. It is widely regarded that proteins tend to adsorb more
favorably onto surfaces with hydrophobic properties. Thus by deposition of plasma polymer films with specific chemical func-
tionality, a reduction in the attachment of proteins, which can lead to biofilm formation, can be achieved. Through the deposition
of siloxane films with varying water contact angle, it has been shown that cell attachment can be controlled as illustrated in
Figure 16 (73). For the siloxane coatings investigated in this study, the optimum MG63 (osteoblast) cell adhesion was observed at
a water contact angle of approximately 64 . Surfaces, which were more hydrophilic or hydrophobic, led to a progressive reduction
in the level of cell adhesion. For fluorinated siloxane coatings, the level of cell adhesion was found to be directly dependent on the
level of fluorination. For example, a 13-fold decrease in cell adhesion was observed for the surface with a water contact angle of
155 compared with that obtained at 110 . It is concluded from this and other studies that adhesion was significantly influenced by
cell type, and that compared with the surface roughness, the surface chemistry was found to exhibit a greater influence on cell
adhesion.
The level of surface control possible through deposition of plasma polymer films makes them ideal for use in drug delivery
systems. Highly cross-linked, pinhole-free coatings exhibiting retention of original precursor monomer molecules offer a controlled
environment in which certain drugs may be encapsulated. Amorosi et al. (74) used an APP to deposit films from methacrylic acid
and ethylene glycol dimethacrylate. The physical and chemical properties of these films showed the progressive release of acet-
aminophen by dipping the film into a deionized water solution.
A number of studies have highlighted the use of atmospheric plasma to deposit antithrombogenic coatings. For example, Osaka
et al. (75) indicated that the TMCTS coated surfaces prevented blood plasma leakage, while also providing an improved antith-
rombogenic surface. Clarotti et al. (76) used a low-temperature plasma system to deposit fluorocarbon coatings on polymer
membrane material to improve material biocompatibility and hemocompatibility. The thrombogenicity of the treated membranes
was shown to reduce after fluorocarbon coating, while the filtering properties of the membranes remained unaffected.
In addition to organosilicon precursors, one of the most extensively researched precursor chemistries is poly(ethylene) glycol
(PEG), often referred to as poly(ethylene) oxide (PEO), depending on its molecular weight. These surfaces have been documented
to show a significant reduction in the adsorption of biological proteins (77). Many methods have been employed for the formation
of PEO surfaces, such as surface grafting, radiation-induced cross-linking, and self-assembled monolayer techniques. There has,
however, been increasing interest in the use of nonthermal plasma techniques for the formation of such surfaces. Nisol et al. (78),
for example, used an atmospheric pressure rf plasma system to deposit plasma-polymerized PEG films using a liquid or gaseous
tetraglyme as a precursor. Lopez et al. (79) used a glow discharge plasma for the deposition of tetraethylene glycol dimethyl ether to
create surfaces that were found to inhibit the adsorption of blood plasma proteins and cells. These studies indicated that there is
a significant reduction in both the initial adsorption and retention of biological proteins on plasma modified surfaces.
Figure 16 Influence of the plasma polymerized siloxane coated polystyrene water contact angle on osteoblast cell (MG63) cell adhesion. Reproduced
from Dowling, D.; Miller, I. S.; Ardhaoui, M.; Gallagher, W. M. Effect of Surface Wettability and Topography on the Adhesion of Osteosarcoma Cells
on Plasma-Modified Polystyrene. J. Biomater. Appl. 2011, 26 (3), 327–347.
www.iran-mavad.com
In summary, APPs offer a potentially excellent route for postmodification of material surfaces. Both surface activation and the
deposition of functional coatings provide an efficient route to enhance material surface properties while retaining the operational
functionality of the bulk material. These advantages have seen the use of APPs in the improvement of surface bonding, the
development of antifouling coatings, and for tailoring interactions in biological systems.
4.07.7 Plasma Medicine
The use of atmospheric plasmas to deposit functional coatings on medical devices is closely related to the rapidly developing area
known as plasma medicine. In this area, the focus is on the use of plasmas in therapeutic applications, such as in wound care and
dentistry. Cold atmospheric plasmas can be applied on living tissue due to their nonequilibrium property. As described earlier, the
energetic electrons generate reactive radicals and ions through collisions with neutral gas molecules, while the gas temperature
remains near room temperature. As a result, the plasma can be controlled so as not to cause thermal damage to heat-sensitive
biological systems such as cells and living tissue. Some of the plasma sources that are applied to the human body utilize DBD
in which the body itself can serve as one of the two electrodes (80). Both reactive oxygen and nitrogen species present in the plasma
have been shown to have an important role in plasma medicine (81). For example, Deng et al. (82) presented a quantitative analysis
of the potential mechanisms of bacterial inactivation by plasmas. Reactive oxygen species were shown to play a dominant role in the
inactivation process with heat, UV photons, electric field, and other charged particles playing only minor roles. It has been
established that the applicator design, the electrical drive frequency, the gas type, and the gap distance (between the plasma tip and
the surface of samples) enable considerable control over the plasma dynamics and chemistry, which in turn affects the time and the
level of response in biological samples.
With exposure of cells to plasmas at high intensities or with longer exposure times, either cell apoptosis (programmed cell death)
or necrosis (accidental cell death) can occur. Thus the discharge can be used to create sterile surfaces (83). Atmospheric plasmas have
also shown potential in the treatment of chronic foot and leg ulcers by sterilizing the surface of the wound (84). Following plasma
treatment sessions, this antibacterial effect has led to the inactivation of all the organisms tested that are susceptible to being present
in skin ulcers, such as Staphylococcus epidermidis, Escherichia coli, Streptococcus pyogenes, Bacillus cereus, and Pseudomonas aeruginosa.
After exposure to low intensity plasmas for a short period, live cell detachment can occur. These cells can then reattach to the
substrate and proliferate after a short incubation time. Inhibition of cell migration and temporary cell-membrane permeabilization
were caused by a short-term exposure to APPs (85). This and the decrease in bacteria in treated wounds may prove to accelerate
wound healing (86).
The ability of plasmas to effect in vivo blood coagulation has also been reported using DBD plasma on live SKH1 mice (87). It
should be noted, however, that the use of plasmas in surgery was first developed in the 1960s, when the plasma afterglow jet of an
inert gas was first applied in the sectioning of tissue. Associated with this treatment, blood coagulation is also obtained and the
technique is now referred to as the ‘plasma scalpel’ (88). Sladek et al. (89) tested a nonthermal atmospheric plasma for its anti-
microbial activity against biofilm cultures of a key carcinogenic bacterium Streptococcus mutants. This study showed a suppression of
the reformation of bacteria following single plasma treatments of 1 min duration.
In a study carried out by Jean-Michel Pouvesle et al., the potential antitumor effect of an in vivo plasma treatment on a U87-luc
glioma tumor was investigated (90). A pulsed DBD with ms pulses at moderate power was used for this treatment. It was concluded
that while the plasma treatment was safe for mice, long time plasma treatment (20 min, 3 consecutive days) produces a superficial
burn. An increase of subcutaneous temperature and a reduction in the cutaneous skin pH were observed. After 5 days of plasma
treatment, a dramatic reduction in U87 bioluminescence was observed that was associated with a reduction in tumor volume in U87
glioma-bearing mice. The exact mechanism of action of cold atmospheric treatments for reduction in cancer is yet to be fully
understood (91). The three important parameters highlighted were the presence of an electric field, the possible diffusion of species
through the skin in connection with a subcutaneous pH change, and to a lesser extent, the slight increase of temperature (90).
Plasma medicine has quickly moved from a basic research phase to clinical trials, and an increasing number of plasma systems
are undergoing clinical trials, with some having already obtained regulatory approval, such as the kINPen system APPJ developed
by Neoplas (Figure 17) (91). Isbary et al. (92) conducted a clinical trial including 150 patients with chronic infected wounds using
Figure 17 An example of tissue-tolerant atmospheric plasma jet (left), kINPen® system APPJ developed by Neoplas (right).
www.iran-mavad.com
low-temperature argon plasma treatment (MicroPlaster system) once daily over 2–5 min. This study showed a highly significant
reduction of bacterial load in wounds treated with plasma in comparison to the control wounds with a high level of patient
tolerance and no side effects observed.
4.07.8 Summary
This chapter has provided an introduction to the applications of cold atmospheric plasma in both materials processing and plasma
medicine. As a processing field it is still in its infancy, with large-scale plasma processing systems only recently being developed.
Factors such as the growing shortage of helium are likely to limit the use of plasma processes based on this gas to high value added
coating chemistries or its use in plasma medicine. The use of air plasmas for cleaning, polymer activation, and coating deposition,
however, are likely to see widespread adoption in manufacturing industry due to the ease of process scale-up and environmental
benefits such as the elimination of organic solvents in substrate cleaning. Plasma Medicine has seen an enormous growth in interest
in recent years for sterilizing surfaces, but a fundamental understanding of the mode of action of the plasma in wound healing or in
cancer treatment, for example, is still unclear due to the complexity of both plasma and biological systems.
Acknowledgments
The author would like to thank the members of the UCD Surface Engineering Group for their valuable assistance in the preparation
of this chapter. This work is supported in part by Science Foundation Ireland Grant 08/SRCI1411.
References
1. Thomson Reuters. Web of Knowledge. [Online] 2012. Available at: http://apps.webofknowledge.com/UA_GeneralSearch_input.do?product¼UA&search_

mode¼GeneralSearch&SID¼Y1okAi4CkFgMbb4KpIJ&preferencesSaved (accessed Sep 4, 2012).
2. Heberlein, J. New Approaches in Thermal Plasma Technology. Pure Appl. Chem. 2002, 74, 327–336.
3. Fridman, A.; Chirokov, A.; Gutsol, A. Non-Thermal Atmospheric Pressure Discharges. J. Phys. D Appl. Phys. 2005, 38, 1–24.
4. Roth, J.; Rahel, J.; Dai, X.; Sherman, D. The Physics and Phenomenology of One Atmosphere Uniform Glow Discharge Plasma (OAUGDP) Reactors for Surface Treatment
Applications. J. Phys. D Appl. Phys. 2005, 38, 555–567.
5. Wargo, S. L. Biological Applications of Atmospheric Pressure Plasma Liquid Deposition. Swanson School of Engineering, Pittsburgh: University of Pittsburgh, 149, 2009.
6. Ramamoorthy, A.; Rahman, M.; Mooney, D. A.; DonMacElroy, J. M.; Dowling, D. P. Thermal Stability Studies of Atmospheric Plasma Deposited Siloxane Films Deposited on
Vycor TM glass. Surf. Coat. Technol. 2008, 202 (17), 4130–4136.
7. Madani, M.; Ajmera, P. Characterization of Silicon Oxide Films Grown at Room Temperature by Point-to-Plane Corona Discharge. J. Electron. Mater. 1993, 22, 1147–1152.
8. Carreno, F.; Bernabeu, E. On Wire-to-Plane Positive Corona Discharge. J. Phys. D Appl. Phys. 1994, 27, 2136–2144.
9. Selwyn, G.; Herrmann, H.; Park, J.; Henins, I. Materials Processing Using an Atmospheric Pressure, RF-Generated Plasma Source. Contrib. Plasma Phys. 2001, 41, 610–619.
10. Tynan, J.; Law, V.; Ward, P.; Hynes, A.; Cullen, J.; Byrne, G.; Daniels, S.; Dowling, D. Comparison of Pilot and Industrial Scale Atmospheric Pressure Glow Discharge Systems
Including a Novel Electro-Acoustic Technique for Process Monitoring. Plasma Sources Sci. Technol. 2010, 19, 5015.
11. Nwankire, C. E.; Law, V. J.; Nindrayog, A.; Twomey, B.; Niemi, K.; Graham, W. G.; Dowling, D. P. Electrical, Thermal and Optical Diagnostics of the PlasmaStream™
Atmospheric Plasma Jet System. Plasma Chem. Plasma Process. 2010, 30, 537–552.
12. Toshifuji, J.; Katsumata, T.; Takikawa, H.; Sakakibara, T.; Shimizu, I. Cold Arc-Plasma Jet under Atmospheric Pressure for Surface Modification. Surf. Coat. Technol. 2002,
171, 302–306.
13. Bae, Y.; Lee, W.; Ko, K.; Lee, Y.; Namkung, W.; Cho, M. Characteristics of a Microwave Plasma Torch with a Coaxial Field-Structure at Atmospheric Pressure. J. Korean Phys.
Soc. 2006, 48, 67.
14. Roth, J. R. Industrial Plasma Engineering: Applications to Nonthermal Plasma Processing; Inst of Physics Pub Inc: Bristol, UK, 2001; Vol. 2.
15. Massines, F.; Gherardi, N.; Fornelli, A.; Martin, S. Atmospheric Pressure Plasma Deposition of Thin Films by Townsend Dielectric Barrier Discharge. Surf. Coat. Technol. 2005,
200, 1855–1861.
16. www.surfxtechnologies.com.
17. Lommatzsch, U.; Pasedag, D.; Baalmann, A.; Ellinghorst, G.; Wagner H-, E. Atmospheric Pressure Plasma Jet Treatment of Polyethylene Surfaces for Adhesion Improvement.
Plasma Process. Polym. 2007, 4, 1041–1045.
18. Baltazar-y-Jimenez, A.; Bistritz, M.; Schulz, E.; Bismarck, A. Atmospheric Air Pressure Plasma Treatment of Lignocellulosic Fibres: Impact on Mechanical Properties and
Adhesion to Cellulose Acetate Butyrate. Compos. Sci. Technol. 2008, 68, 215–227.
19. Scopece, P.; Viaro, A.; Sulcis, R.; Kulyk, I.; Patelli, A.; Guglielmi, M. SiOx Based Gas Barrier Coatings for Polymer Substrates by Atmospheric Pressure Plasma Jet Deposition.
20. Lommatzsch, U.; Ihde, J. Plasma Polymerization of HMDSO with an Atmospheric Pressure Plasma Jet for Corrosion Protection of Aluminum and Low-Adhesion Surfaces. Plasma
Process. Polym. 2009, 6, 642–648.
21. Bringmann, P.; Rohr, O.; Gammel, F. J.; Jansen, I. Atmospheric Pressure Plasma Deposition of Adhesion Promotion Layers on Aluminium. Plasma Process. Polym. 2009, 6,
496–502.
22. http://www.plasmatreat.com.
23. Dowling, D. P.; Tynan, J.; Ward, P.; Hynes, A. M.; Cullen, J.; Byrne, G. Atmospheric Pressure Plasma Treatment of Amorphous Polyethylene Terephthalate for Enhanced
Heatsealing Properties. Int. J. Adhes. Adhes. 2012, 35, 1–8.
24. Law, V. J.; Ramamoorthy, A.; Dowling, D. P. Real-Time Process Monitoring during the Plasma Treatment of Carbon Weave Composite Materials. J. Mater. Sci. Eng. 2011, 1 (2),
164–169.
25. Dowling, D. P.; O’Neill, F. T.; Langlais, S. J.; Law, V. J. Influence of DC Pulsed Atmospheric Pressure Plasma Jet Processing Conditions on Polymer Activation. Plasma Process.
Polym. 2011, 8 (8), 718–727.
26. Noeske, M.; Degenhardt, J.; Strudthoff, S.; Lommatzsch, U. Plasma Treatment of Five Polymers at Atmospheric Pressure: Surface Modifications and the Relevance for
Adhesion. Int. J. Adhes. Adhes. 2004, 24, 171–177.
www.iran-mavad.com
27. O’Neill, L.; Shephard, N.; Leadley, S. R.; O’Hare, L. A. Atmospheric Pressure Plasma Polymerised Primer to Promote Adhesion of Silicones. J. Adhes. 2008, 84, 562–577.
28. www.vitoplasma.com/en/vision.
29. www.tritonsys.com.
30. Paital, S. R.; Dahotre, N. B. Calcium Phosphate Coatings for Bio-Implant Applications: Materials, Performance Factors, and Methodologies. Mater. Sci. Eng. R 2009, 66, 1–70,
http://dx.doi.org/10.1016/j.mser.2009.05.001.
31. Pirzada, S. Plasma Treatment of Polymer Films. In Forty-second Annual Technical Conference Proceedings; Chicago, 1999, 301–306.
32. Schutze, A.; Jeong, J.; Babayan, S.; Park, J.; Selwyn, G.; Hicks, R. The Atmospheric-Pressure Plasma Jet: A Review and Comparison to Other Plasma Sources. IEEE Trans.
Plasma Sci. 1998, 26, 1685–1694.
33. Malik, M.; Ghaffar, A.; Malik, S. Water Purification by Electrical Discharges. Plasma Sources Sci. Technol. 2001, 10, 82–91.
34. Oda, T. Non-Thermal Plasma Processing for Environmental Protection: Decomposition of Dilute VOCs in Air. J. Electrost. 2003, 57, 293–311.
35. Dai, L.; Griesser, H.; Mau, A. Surface Modification by Plasma Etching and Plasma Patterning. J. Phys. Chem. B 1997, 101, 9548–9554.
36. Bogaerts, A.; Neyts, E.; Gijbels, R.; Van der Mullen, J. Gas Discharge Plasmas and Their Applications. Spectrochim. Acta Part B At. Spectrosc. 2002, 57, 609–658.
37. Shenton, M. J.; Lovell-Hoare, M. C.; Stevens, G. C. J. Phys. D Appl. Phys. 2001, 34, 2754.
38. Kruse, A.; Kruger, G.; Baalmann, A.; Hennemann, O. D. J. Adhes. Sci. Technol. 1995, 9, 1611.
39. Strobel, M.; Sullivan, N.; Branch, M.; Park, J.; Ulsh, M.; Kapaun, R.; Leys, B. J. Adhes. Sci. Technol. 2000, 10, 1243.
40. Briggs, D.; Rance, D. G.; Kendall, C. R.; Blathe, A. R. Polymer 1980, 21, 895.
41. Kaplan, S. L. Plasma Processes in the Plastics Industry. In 35th Ann. Tech. Conf; Soc. of Vacuum Coaters, 1992; Vol. 273.
42. Dowling, D. P.; Twomey, B.; Byrne, G. Deposition of Functional Coatings Using an Inline Atmospheric Pressure Plasma Apparatus. In 48th Ann. Conf; Society of Vacuum
Coaters, 2005; Vol. 214.
43. Gerenser, L. J. Surface Chemistry of Plasma Treated Polymers (chapter E3.1). In Handbook of Thin Film Process Technology; Glocker, D. B., Shah, S. L., Eds.; Institute of
Physics, Publishing: UK, 1995.
44. Arefi-Khonsari, F.; Tatoulian, M.; Bretagnol, F.; Bouloussa, O.; Rondelez, F. Processing of Polymers by Plasma Technologies. Surf. Coat. Technol. 2005, 200 (1–4), 14–20.
45. Shanahan, M. E. R.; Bourges-Monnier, C. Effects of Plasma Treatment on the Adhesion of an Epoxy Composite. Int. J. Adhes. Adhes. 1995, 16 (2), 129–135.
46. Biederman, H.; Osada, Y. Plasma Chemistry of Polymers. In Polymer Physics; Springer: Berlin/Heidelberg, 1990; pp 57–109.
47. Yasuda, H. Glow Discharge Polymerization. J. Polym. Sci. Macromol. Rev. 1981, 16, 199–293.
48. Schneider, J.; Kiesler, D.; Leins, M.; Schulz, A.; Walker, M.; Schumacher, U. Development of Plasma Polymerised SiOx Barriers on Polymer Films for Food Packaging
Applications. Plasma Process. Polym. 2007, 4, 155–159.
49. Lin, Y.; Yasuda, H. Effect of Plasma Polymer Deposition Methods on Copper Corrosion Protection. J. Appl. Polym. Sci. 1996, 60, 543–555.
50. Janocha, B.; Hegemann, D.; Oehr, C.; Brunner, H.; Rupp, F.; Geis-Gerstorfer, J. Adsorption of Protein on Plasma-Polysiloxane Layers of Different Surface Energies. Surf. Coat.
Technol. 2001, 142–144, 1051–1055.
51. Biederman, H. Plasma Polymer Films, Imperial College Press: London, 2004, 386.
52. Meeks, R. G. The Dow Corning Siloxane Research Program: An Overview and Update. Med. Dev. Diagn. Ind. 1999, 21, 112.
53. Smith, J. Analysis of Organo-Silicon Compounds, with Special Reference to Silanes and Siloxanes. A review. Analyst 1960, 85, 465–474.
54. Stallard, C. P.; McDonnell, K. A.; Onayemi, O. D.; O’Gara, J. P.; Dowling, D. P. Evaluation of Protein Adsorption on Atmospheric Plasma Deposited Coatings Exhibiting
Superhydrophilic to Superhydrophobic Properties. Biointerphases 2012, 7 (1–4), 31.
55. McCann, M. T. P.; Mooney, D. A.; Dowling, D. P.; MacElroy, J. M. D. Silica Nanofilms Deposited by Atmospheric Pressure Plasma Liquid Deposition. Thin Solid Films 2012,
520 (7), 2619–2626.
56. Howling, A. A.; Dorier, J. L.; Hollenstein, C. Negative Ion Mass Spectra and Particulate Formation in Radio Frequency Silane Plasma Deposition Experiments. Appl. Phys. Lett.
1993, 62, 1341–1343.
57. De Bleecker, K.; Bogaerts, A.; Goedheer, W.; Gijbels, R. Investigation of Growth Mechanisms of Clusters in a Silane Discharge with the Use of a Fluid Model. IEEE Trans. Plasma
Sci. 2004, 32, 691–698.
58. Favia, P. Plasma Deposition of Fluoropolymer Films in Different Glow Discharge Regimes. In Plasma Polymer Films; Vol. 25; Biederman, H., Ed.; Imperial College: London,
2004; Vol. 25.
59. Clarotti, G.; Schue, F.; Sledz, J.; Aoumar, A. A. B.; Geckeler, K. E.; Orsetti, A. Modification of the Biocompatible and Haemocompatible Properties of Polymer Substrates by
Plasma-Deposited Fluorocarbon Coatings. Biomaterials 1992, 13, 832–840.
60. Kumar, V.; Pulpytel, J.; Rauscher, H.; Mannelli, I.; Rossi, F.; Arefi-Khonsari, F. Fluorocarbon Coatings via Plasma Enhanced Chemical Vapor Deposition of 1H,1H,2H,2H-
Perfluorodecyl Acrylate–2, Morphology, Wettability and Antifouling Characterization. Plasma Process. Polym. 2010, 7 (11), 926–938.
61. Favia, P.; Perez-Luna, V.; Boland, T.; Castner, D.; Ratner, B. Surface Chemical Composition and Fibrinogen Adsorption-Retention of Fluoropolymer Films Deposited from an RF
Glow Discharge. Plasmas Polym. 1996, 1, 299–326.
62. Marabotti, I.; Morelli, A.; Orsini, L. M.; Martinelli, E.; Galli, G.; Chiellini, E. Fluorinated/Siloxane Copolymer Blends for Fouling Release: Chemical Characterisation and Biological
Evaluation with Algae and Barnacles. Biofoul. J. Bioadhesion Biofilm Res. 2009, 25, 481–493.
63. Nwankire, C. E.; Favaro, G.; Duong, Q.-H.; Dowling, D. P. Enhancing the Mechanical Properties of Superhydrophobic Atmospheric Pressure Plasma Deposited Siloxane
Coatings. Plasma Process. Polym. 2011, 8 (4), 305–315.
64. Ji, Y-Y.; Kim, S-S.; Kwon, O. P.; Lee, S-H. Easy Fabrication of Large-Size Superhydrophobic Surfaces by Atmospheric Pressure Plasma Polymerization with Non-Polar Aromatic
Hydrocarbon in an In-Line Process. Appl. Surf. Sci. 2009, 255, 4575–4578.
65. Kim, S. H.; Kim, J.-H.; Kang, B.-K.; Uhm, H. S. Superhydrophobic CFx Coating via In-Line Atmospheric RF Plasma of He-CF4-H2. Langmuir 2005, 21, 12213–12217.
66. Scardino, A.; De Nys, R.; Ison, O.; O’Connor, W.; Steinberg, P. Microtopography and Antifouling Properties of the Shell Surface of the Bivalve Molluscs Mytilus galloprovincialis
and Pinctada imbricata. Biofoul. J. Bioadhesion Biofilm Res. 2003, 19, 221–230.
67. Li, X.; Reinhoudt, D.; Crego-Calama, M. What Do We Need for a Superhydrophobic Surface? A Review on the Recent Progress in the Preparation of Superhydrophobic Surfaces.
Chem. Soc. Rev. 2007, 36, 1350–1368.
68. Wu, Y.; Bekke, M.; Inoue, Y.; Sugimura, H.; Kitaguchi, H.; Liu, C.; Takai, O. Mechanical Durability of Ultra-Water-Repellent Thin Film by Microwave Plasma-Enhanced CVD. Thin
Solid Films 2004, 457, 122–127.
69. DowCorning. A Guide to Silane Solutions from DowCorning, 30, 2008.
70. Twomey, B.; Rahman, M.; Byrne, G.; Hynes, A.; O’Hare, L.-A.; O’Neill, L.; Dowling, D. P. Effect of Plasma Exposure on the Chemistry and Morphology of Aerosol-Assisted,
Plasma-Deposited Coatings. Plasma Process. Polym. 2008, 5, 737–744.
71. Nwankire, C. E.; Dowling, D. P. Influence of nm Thick Atmospheric Plasma Deposited Coatings on the Adhesion of Silicone Elastomer to Stainless Steel. J. Adhes. Sci. Technol.
2010, 24 (7), 1291–1302.
72. Favia, P.; d’Agostino, R. Plasma Treatments and Plasma Deposition of Polymers for Biomedical Applications. Surf. Coat. Technol. 1998, 98, 1102–1106.
73. Dowling, D.; Miller, I. S.; Ardhaoui, M.; Gallagher, W. M. Effect of Surface Wettability and Topography on the Adhesion of Osteosarcoma Cells on Plasma-Modified Polystyrene.
J. Biomater. Appl. 2011, 26 (3), 327–347.
74. Amorosi, C.; Ball, V.; Bour, J.; Bertani, P.; Toniazzo, V.; Ruch, D. One Step Preparation of Plasma Based Polymer Films for Drug Release. Mater. Sci. Eng. 2012, 32,
2103–2108.
75. Osaki, S. Preparation of a Plasma Polymerized Tetramethylhydrocyclotetrasiloxane Membrane on Microporous Hollow Fibers. Surf. Modif. Polym. Biomater. 1997, 61.
www.iran-mavad.com
76. Walsh, J. L.; Kong, M. G. Contrasting Characteristics of Linear-Field and Cross-Field Atmospheric Plasma Jets. Appl. Phys. Lett. 2008, 93, 111501.
77. Harris, J.; Zalipsky, S. Poly (Ethylene Glycol): Chemistry and Biological Applications; An American Chemical Society Publication: Washington, DC, 1997; Vol. 680.
78. Nisol, B.; Poleunis, C.; Bertrand, P.; Reniers, F. Poly (Ethylene Glycol) Films Deposited by Atmospheric Pressure Plasma Liquid Deposition and Atmospheric Pressure Plasma-
Enhanced Chemical Vapour Deposition: Process, Chemical Composition Analysis and Biocompatibility. Plasma Process. Polym. 2010, 7 (8), 715–725.
79. Löpez, G. P.; Ratner, B. D.; Tidwell, C. D.; Haycox, C. L.; Rapoza, R. J.; Horbett, T. A. Glow Discharge Plasma Deposition of Tetraethylene Glycol Dimethyl Ether for Fouling-
Resistant Biomaterial Surfaces. J. Biomed. Mater. Res. 1992, 26 (4), 415–439.
80. Kuchenbecker, M.; Bibinov, N.; Kaemlimg, A.; Wandke, D.; Awakowicz, P.; Viol, W. Characterization of DBD Plasma Source for Biomedical Applications. J. Phys. D Appl. Phys.
2009, 42, 045212–045222.
81. Friedman, A. Plasma Chemistry. In Plasma Assisted Wound Healing; Cambridge University Press, 2008. ISBN: 13, 978-0-0521-84735-3, (chapter 12.4).
82. Deng, X.; Shi, J. J.; Kong, M. G. Physical Mechanisms of Inactivation of Bacillus subtilis Spores Using Cold Atmospheric Plasmas. Plasma Sci. 2006, 34 (4), 1310–1316.
83. Kong, M. G. A Complementary Sterilisation Strategy Using Cold Atmospheric Plasmas. Med. Dev. Technol. 2006, 17 (3), 26–28.
84. Pompl, R. Efficiency and Medical Compatibility of Low-Temperature Plasma Sterilization. In 6th Int. Conf. Reactive Plasmas Matsushima, Japan, 2006.
85. Park, G. Y.; Park, S. J.; Choi, M. Y.; Koo, I. G.; Byun, J. H.; Hong, J. W.; Sim, J. Y.; Collins, G. J.; Lee, J. K. Atmospheric-Pressure Plasma Sources for Biomedical Applications.
Plasma Sources Sci. Technol. 2012, 21 (4), 043001.
86. Kong, M. G.; Kroesen, G.; Morfill, G.; Nosenko, T.; Shimizu, T.; van Dijk, J.; Zimmermann, J. L. Plasma Medicine: An Introductory Review. New J. Phys. 2009, 11, 115012.
87. Fridman, G.; Shereshevsky, A.; Peddinghaus, A.; Gutsol, A.; Vasilets, V.; Brooks, A.; Balasubramanian, M.; Friedman, G.; Fridman, A. Bio-Medical Applications of Non-Thermal
Atmospheric Pressure Plasma. 37th AIAA, Plasmadynamics and Lasers Conference, San Francisco, California, 2006.
88. Glover, J. L.; Bendick, P. J.; Link, W. J.; Plunkett, R. J. The Plasma Scalpel: A New Thermal Knife. Lasers Surg. Med. 1982, 2 (1), 101–106.
89. Sladek, R. E. J.; Filoche, S. K.; Sissons, C. H.; Stoffels, E. Treatment of Streptococcus mutans Biofilms with a Nonthermal Atmospheric Plasma. Lett. Appl. Microbiol. 2007, 45,
318–323.
90. Vandamme, M.; Robert, E.; Pesnel, S.; Barbosa, S.; Dozias, S.; Sobilo, J.; Lerondel, S.; Le Pape, A.; Pouvesle, J. M. Antitumor Effect of Plasma Treatment on U87 Glioma
Xenografts: Preliminary Results. Plasma Process. Polym. 2010, 7 (3–4), 264–273.
91. www.neoplas.eu (accessed Sep 14, 2012).
92. Isabary, G.; Morfill, G.; Schmidt, H. U.; Georgi, M.; Ramrath, K.; Heinlin, J.; Karrer, S.; Landthaler, M.; Shimizu, T.; Steffes, B.; Bunk, W.; Monetti, R.; Zimmermann, J. L.;
Pompl, R.; Stolz, W. A 1st Operative Randomized Controlled Trial to Decrease Bacterial Load Using Cold Atmospheric Argon Plasma on Chronic Wounds in Patients. Br. J.
Dermatol. 2010, 163 (1), 78–82.
www.iran-mavad.com
www.iran-mavad.com
4.08 Ion Beam Deposition: Recent Developments
N Toyoda and S Matsui, University of Hyogo, Hyogo, Japan

4.08.2 Irradiation Effects of Ion Beam during Thin Film Depositions 187
4.08.3 Ion Beam Assisted Deposition Systems 189
4.08.4 Gas Cluster IBAD 190
4.08.5 Focused Ion Beam Assisted Nanostructure Formation 195
4.08.6 Conclusion 199
References 199
4.08.1 Introduction
Ion beam assisted deposition (IBAD) is one of the physical vapor depositions, which uses energetic particles during thin film
deposition in order to improve properties or microstructures. The additional energy deposited on thin films by ion bombard-
ment induces atomic displacements, surface migration, or transformation of growth mode. In addition, ion mixing between the
substrate and the grown film results in the improvement of adhesion. Ion irradiation during evaporation have been attracted
many attention in order to form nanostructures (1). Also, low-energy ion beam irradiation during deposition can be used for
biaxially aligned films (2).
In 1960s, it was shown that energetic ions within plasma have an important influence on film properties (3). The role of ions
during thin film deposition was studied in 1970s and 1980s in order to improve the density, adhesion, stress, morphology,
composition, nucleation density, and low-temperature deposition (4–6). The motivation of the use of simultaneous irradiation of
ion beam and physical vapor deposition is increased control of directionality and ion energies as compared to plasma-based
processes.
However, it is necessary to develop new ion beam techniques to realize precise structure or films on the order of nanometers.
Recently, emerging IBAD techniques have been developed. One is the gas cluster ion beam (GCIB) assisted deposition (7,8), and
another is the focused ion beam (FIB) assisted deposition for nanostructure formation (9). GCIBs are aggregates of several thou-
sands of gas atoms or molecules. Therefore, the energy per each atoms or molecules can be easily reduced to several eV. Besides,
GCIB irradiation gives a dense energy deposition on a target surface, which induces local and transient heating (10). Therefore, the
target surface reaches a very high temperature and high pressure without substrate heating, which leads to improvement of thin film
properties at low temperatures. In terms of unique characteristics of GCIB irradiation, GCIB shows surface smoothing of thin films
owing to lateral sputtering effects (11). Therefore, high-density thin films with quite smooth interfaces can be deposited with GCIB-
assisted deposition.
As to the FIB-assisted deposition, Ga-FIB induces dissociation of the phenanthrene molecule, which results in the amorphous
carbon formation. Since the bombardment area with FIB is limited within a several tens of nanometers, a three-dimensional (3-D)
structure made of amorphous carbon can be formed.
Since there are various review articles describing details of IBAD (12–16), general information regarding fundamental charac-
teristics of IBAD will not be discussed in large part; however, the recent progress of IBAD such as GCIB-assisted deposition and FIB
deposition for 3-D nanostructure formation are mainly reported.
4.08.2 Irradiation Effects of Ion Beam during Thin Film Depositions
A typical configuration of IBAD equipment is shown in Figure 1 (12). Materials are delivered from an evaporation source (electron
beam evaporation or thermal evaporation). The ions are extracted from a broad-beam grid ion source at low energy (0.2–2 keV)
(12). The bombardment of energetic ions or atoms on a thin film surface produces various effects that are beneficial to thin film
depositions. Since the energy of assisted ions can be changed in the range from 1 to 10 000 eV, nonequilibrium processes that
cannot be attained with thermal evaporations are expected. Comparing electron or photon irradiation, the effects of ion
bombardment on a thin film differ substantially because the mass of ions is of the same order as that of thin films. In general, effects
of ion irradiation during thin film depositions are summarized as (1) atomic displacement, (2) densification, (3) film stress, (4) ion
mixing, and (5) texture formations. As shown in Figure 2 (12), ion irradiation originates collision cascades in coating and substrate.
They create a mixed zone between the original substrate surface and films produced by ion beam mixing. This mixed zone helps
improve adhesion.
In thin film depositions, there are three types of film growth, known as (1) Frank–van der Merwe type (layer-by-layer film growth
with complete coverage), (2) Volmer–Weber growth (islands formation and growth before coalescing), and (3) Stranski–Krastanov
growth (monolayer formation followed by island growth). When there are ion bombardments during thin film depositions, the

www.iran-mavad.com
188 Ion Beam Deposition: Recent Developments
Figure 1 Typical configuration of an IBAD system. Hirvonen, J. K. Mater. Sci. Rep. 1991, 6, 215..
Figure 2 Schematic of the IBAD process, indicating the interaction of coating atoms in the collision cascade of the ions and mixed interface zone.
Hirvonen, J. K. Mater. Sci. Rep. 1991, 6, 215.
growth mode changes dramatically. Trushin et al. reported the kinetic model of the atomic assembly during ion beam assisted
growth of GaN films (17). They modeled GaN growth with molecular beam epitaxy (MBE) and additional irradiation of nitrogen
ions, and claimed that the additional bombardment with low-energy ions leads to a transformation of the growth mode from island
growth to layer-by-layer growth. Figure 3 shows the physical model explaining the differences in growth modes induced by nitrogen
ion irradiation during epitaxy, which is based on the kinetic theory of nucleation and growth of thin crystalline films (18–20).
It takes into account the fundamental processes of Ga and N adatom deposition, desorption, and diffusion; GaN cluster formation
and growth; and ion impact-induced height reduction of clusters. The hyperthermal ion bombardment increases the detachment of
atoms from the tops of the growing islands. Therefore, the growth mode changes from 3-D growth to two-dimensional (2-D)
growth with ion irradiation. This model shows a good agreement with the experimental results (21).
With energetic ion irradiation during evaporation, a microstructure of a thin film changes and a densification of the film is
usually observed. Muller modeled the evolution of a 2-D lattice under ion bombardment using molecular dynamics simulations
(22). It was found that the bombardment of energetic ion removes overhanging atoms, induces movement of atoms, produces local
heating, and collapses voids. In the case of optical thin films, this densification effect increases the density of the film, and leads to
the increase of refractive index of an optical film.
Thin film stress can be changed by ion irradiation. At low energies, the films are porous with a large void and tensile intrinsic
stress (23). With increasing energy, there is a transition to compressive stress due to densification by impinging ions. These stresses
can be changed by ion energy and a current density. Relatively high-energy ion irradiation also causes ion mixing of substrate and
grown film, which leads to increase of adhesion of the film. The adhesion of thin films on substrates depends on a variety of factors
such as interface chemistry, film stress, difference of thermal expansion, contaminant at the interface, and surface morphology (24).
Recently, texture or self-organized pattern formations with ion beam irradiation during thin film deposition have attracted much
attention in order to create nanostructures. Groves et al. reported biaxial texture development in the IBAD of MgO film (2), which is
www.iran-mavad.com
Ion Beam Deposition: Recent Developments 189
Figure 3 Schematic representation of the fundamental differences between (a) MBE and (b) IBA–MBE growth (without adsorption, desorption, and
diffusion). Trushin, Y. V.; Kulikov, D. V.; Safonov, K. L.; Gerlach, J. W.; Hoche, Th.; Rauschenbach, B. J. Appl. Phys. 2008, 103, 114904.
used as templates for high-temperature superconductor wires. They modeled the texturing as the MgO nucleation from an
amorphized MgO layer and subsequent development of texture by the film growth from these seed crystals.
Besides texture formation with ion beam irradiation during deposition, oblique incidence irradiation of ion beams on thin film
surface results in a nanostructure formation. In general, these nanostructures are formed by a self-organization process. They are
created by the competition of a sputtering and a surface relaxation by atomic migration, which is modeled by Bradley and Harper
(25). Without sample rotation, ripples with period of several tens of nanometers are formed. On the other hand, nanodots are
formed when there is a sample rotation during ion beam irradiation. These nanostructure formations are useful for periodic pattern
formation such as photonic devices (26), patterned media (27), antireflection coating (28), and so on.
4.08.3 Ion Beam Assisted Deposition Systems
As shown in Figure 1, a typical IBAD system consists of two main parts: the source for the low-energy particles, which usually
constitute the major part of the film to be deposited, and the source for simultaneous irradiation with highly energetic particles,
that is, the ion source. For the low-energy species, mostly vapor sources are used where the vapor is generated by resistive heating
or electron beam heating of the material. Electron beam evaporation is widely used because of its controllability, versatility, and
high deposition rates even for materials with low vapor pressure. Another possibility for generating the required atom flux is
sputtering (29).
The average energies of the species emitted from the above mentioned vapor sources differ from each other considerably. Atoms
from a crucible possess the lowest energies in the thermal regime. Puttered particles and ionized species from a cluster source possess
much higher energies. These energies are, however, still several orders of magnitude lower than those of energetic ions from a source.
The Kaufman ion source is most commonly used ion source; it is named after its inventor. Figure 4 shows a schematic view of the
broad-beam, multiaperture, Kaufman-type ion source (13). In this ion source, the gas is introduced into the chamber, which has
a hot cathode emitting electrons. There is a cylindrical anode surrounding the cathode and gases are ionized. A magnetic field is
applied transverse to the motion of electrons for electron confinement and increased ionization efficiency. Screen grids and
accelerator grids are used to extract ions. The ions produced by electron impacts pass through these grids and bombard a target.
The typical energy of assisted ions produced by these ion sources are several hundreds of eV; however, Mach et al. developed an
ultralow energy (30–200 eV) ion source for IBAD (30). Figure 5 shows a schematic diagram of the ion source. The source combines
www.iran-mavad.com
Figure 4 Schematic view of the broad-beam, multiaperture, Kaufman-type ion source. Mohan, S.; Krishna, M. G. Vacuum 1995, 46, 645.
Figure 5 Schematic of the ion–atomic beam source generating beams of thermal atoms and ultralow energy ions (30–200 eV). The potential setup
provides ions of 100 eV energy at the target. Mach, J.; Samuril, T.; Voborny, S.; Kolibal, M.; Zlamal, J.; Spousta, J.; Dittrichova, L.; Sikola, T.
Rev. Sci. Instrum. 2011, 82, 083302.
an effusion cell and an electron impact ion source and produces ion beams with ultralow energies. Decreasing ion beam energy to
hyperthermal values (w10 eV) without losing optimum ionization conditions has been achieved by the incorporation of an
ionization chamber with a grid transparent enough for electron and ion beams. The ion energy and the ion current density are about
10 eV and 10 nA cm2, respectively.
There has been much research using IBAD for various applications such as compound photovoltaic films (31), amorphous
carbon films for protective coating (32), thin films for spin valves used for hard disk heads (33), deposition for organic light-
emitting diodes (34), and thin films that are biocompatible (35).
4.08.4 Gas Cluster IBAD
There are various methods for deposition of high-quality thin film such as sputtering, IBADs, ion-plating deposition, and so on.
Recently, the GCIB process has been studied to replace the conventional ion beam in the low-energy region, such as surface
smoothing (36,37), reactive cluster ion etching (38), precise mold fabrication (39), and ultrashallow ion implantations (40).
Figure 6 shows various irradiation effects of GCIB. Gas clusters are huge aggregates containing several to thousands of atoms (41).
As each atom in a cluster shares total acceleration energy, an ultralow energy ion beam with several eV per atom can be easily
www.iran-mavad.com
Figure 6 Irradiation effects and applications of GCIB. Reproduced from Toyoda, N.; Hirota, T.; Yamada, I.; Yakushiji, H.; Hinoue, T.; Ono, T.; Matsumoto,
H. IEEE Trans. Magn. 2010, 46, 1599.
realized. Also, as the irradiated area with cluster ion realizes high-temperature and high-pressure conditions due to the dense energy
deposition in a local area, there are enhancements of chemical reactions near the surface, which realizes high-density films without
heating the substrate itself (42). The most interesting characteristic of cluster ion beams is that it exhibits strong surface-smoothing
effects because of its lateral sputtering effect (43,44). In addition, GCIB has been used for the ion beams for surface analysis such as
secondary ion mass spectrometry (45,46) and a depth profiling of X-ray photoelectron spectroscopy (47,48). The details of GCIB
system and fundamental irradiation effects are summarized in these review papers (49).
GCIB can be used as assist ions for very smooth and high-density thin film deposition at low temperatures. Figure 7 shows
a schematic diagram of a GCIB-assisted deposition system (42). There is an electron beam evaporator to evaporate materials such as
Ta2O5, Nb2O5, and SiO2. A GCIB source is mounted under the deposition chamber. In the case of oxygen cluster ion beams, O2
neutral clusters are formed by supersonic expansions of high-pressure (w1 MPa) O2–He mixture gas (30% of helium) through
a Laval nozzle into a vacuum chamber. Neutral clusters are formed by the cooling effect of supersonic expansion. From time of flight
mass spectrometry measurements, the average number of O2 molecules in a GCIB was approximately 1000–3000 molecules per
cluster. The neutral O2 jet formed by the nozzle is collimated by a skimmer and O2 neutral clusters are led into an ionizer.
Subsequently, they are ionized by electron bombardments and are accelerated up to 10 keV to a target. O2 cluster ion current density
was 1.0 mA cm2 at the target. The deposition rates of Ta2O5, Nb2O5, and SiO2 were 0.1 nm s–1. The vacuum pressure during
depositions was less than 1 104 Torr. The temperatures of the targets were kept at room temperature.
Figure 8 shows the average surface roughness of Ta2O5 films deposited without O2-GCIB, with neutral O2 cluster beam irra-
diation, and with O2-GCIB irradiation at 7 keV. The O2 neutral cluster irradiation means that ionization by electron bombardments
Figure 7 Schematic diagram of a high-current GCIB-assisted deposition system. Reproduced from Shirai, K.; Fujiwara, Y.; Takahashi, R.; Toyoda, N.;
Matsui, S.; Mitamura, T.; Terasawa, M.; Yamada, I. Jpn. J. Appl. Phys. 2002, 41, 4291.
www.iran-mavad.com
Surface roughness [nm]

1.5
Ra = 1.71 nm
0.5
Ra = 0.37nm
0
Without
Not assisted With neutral
Neutral beam With 7 keV
O2-GCIB assisted
O2 cluster O2-GCIB
Figure 8 Average surface roughness of Ta2O5 films deposited without O2-GCIB, with neutral O2 cluster beam irradiation, and with O2-GCIB
irradiation at 7 keV.
was not performed although O2 cluster beam was introduced into the deposition chamber. In the case of Ta2O5 film without
O2-GCIB, the average surface roughness was 1.7 nm and there were many bumps due to columnar structure of the film. In the case of
the neutral O2 cluster beam irradiation, it did not show significant reduction of the average surface roughness. However, with
O2-GCIB irradiation at acceleration energy of 7 keV, the average surface roughness dramatically improved to 0.4 nm. As the O2
pressure in the deposition chamber was the same as that with O2 neutral beam irradiations, it indicates that energetic O2 cluster ions
are required to obtain smooth Ta2O5 surfaces. Compositions of tantalum oxide film deposited with or without O2-GCIB irradia-
tions were measured with Rutherford backscattering spectrometry (RBS). Without O2-GCIB irradiations, the composition was
Ta2O4.6; however, it was Ta2O4.98 with O2-GCIB irradiation at acceleration energy of 7 keV. It also confirms that energetic O2 cluster
ions were required to obtain stoichiometric tantalum oxide films.
Figure 9 shows ion current density dependence of refractive indexes and surface roughness of Ta2O5 films (46). The accel-
eration energy of O2-GCIB was fixed at 7 keV. Ion current density of 0 represents the result with neutral O2 cluster beams. In the
case of neutral O2 cluster beam assisted deposition, the refractive index of Ta2O5 was the same as that of an electron beam
deposition. However, with increasing the cluster ion current density, the refractive index gradually increased and showed
saturation value at 2.2 over ion current density of 0.8 mA cm2. In Figure 9, the surface roughness suddenly decreased from 1.6 to
0.6 nm around the ion current density of 0.5 mA cm2. In the case of SiO2, the drop of the surface roughness was about
0.7 mA cm2.
Figure 10(a) shows a cross-sectional image of Ta2O5/SiO2 multilayer films (50). The fourth and fifth layers from the
bottom were deposited without O2-GCIB irradiations, and the other layers were deposited with O2-GCIB. The acceleration
energy and the ion current density of O2-GCIB were 7 keV and 1.0 mA cm–2, respectively. Figure 10(b) shows an atomic force
microscope image of the top layer (Ta2O5) with O2-GCIB irradiation and Figure 10(c) shows that of the fifth layer (SiO2)
deposited without O2-GCIB irradiation. From Figure 10(a), it is clearly shown that the films became coarse and columnar
structures when there was no O2-GCIB irradiation as shown in the fourth and fifth layers. However, the film became very
uniform and high density when there were O2-GCIB irradiations. When there was not a O2-GCIB irradiation, the average
surface roughness of SiO2 was 1.5 nm as shown in Figure 10(c). However, when a Ta2O5 film was deposited with O2-GCIB
irradiation on this rough surface, the surface roughness was significantly improved as shown in Figure 10(b). The average
roughness of this Ta2O5 film was 0.7 nm, which was almost half of that of the layer underneath SiO2. In general, it is difficult
to improve the surface roughness with conventional deposition methods. Once a rough surface was formed, the surface
roughness increased with increased numbers of layers. Therefore, the surface roughness improvement is one of the advantages
of the GCIB-assisted deposition.
Low-energy GCIB can be used as assist ions for hard amorphous carbon film formation. Figure 11 shows the GCIB-assisted
diamondlike carbon (DLC) film coater, which uses Ar-GCIB irradiation during evaporation of C60 in order to form ultrahard DLC
thin films at room temperature (51). GCIB creates a locally very high-temperature and high-pressure zone by dense energy
deposition, which enhances to produce high-density amorphous carbon films with a large amount of sp3 bonds. Figure 12 shows
the density of various DLC films (radio frequency (RF) plasma, electron cyclotron resonance plasma, ion plating, filtered cathodic
vacuum arc (FCVA), and GCIB-assisted deposition). In the case of a conventional DLC deposition method, the film density is about
2.0 g cm3. However, GCIB-assisted deposition achieves a film density of 2.8 g cm3, and it is close to those of the films deposited
using the FCVA technique. The amount of sp3 bonds was determined by near-edge X-ray absorption fine structure (NEXAFS) using
www.iran-mavad.com
Figure 9 Cluster ion current density dependence of refractive index and surface roughness of Ta2O5 and SiO2 films. The acceleration energy of O2-GCIB
and the thickness of films were 7 keV and 200 nm, respectively.
Figure 10 (a) Cross-sectional scanning electron microscope image of Ta2O5/SiO2 multilayer and atomic force microscope images of the surfaces,
(b) Ta2O5 film with 7 keV O2-GCIB, and (c) SiO2 film without O2-GCIB. Reproduced from Toyoda, N.; Fujiwara, Y.; Yamada, I. Nucl. Instrum. Methods Phys.
Res. B 2003, 206, 875.
synchrotron radiation facility as shown in Figure 13. It was found that a high fraction of sp3 bonds was included in the amorphous
carbon film (30% higher amount of sp3 compared to the DLC film deposited using the RF plasma technique) (52). The hardness of
the DLC films formed with Ar-GCIB showed 50 GPa at the energy of 5 keV. Also, DLC films with hardness above 40 GPa were
deposited with 5–9 keV acceleration energy. On the other hand, DLC films deposited with ion plating methods were 25 GPa (51).
The surface preparation of a substrate by GCIB to form a thin-film multilayer can also improve device performance. GCIB
smoothing of glass substrates for extremely narrow band-pass filters used for optical telecommunications became the first
production application of GCIB. Polishing of certain types of glass substrates to the limits of available conventional polishing
www.iran-mavad.com
Figure 11 Schematic diagram of the apparatus for DLC film formation with various Ar-ion beam assisted deposition.
Figure 12 Density of various DLC films.
Figure 13 NEXAFS C K-edge spectra of various DLC films. The DLC film denoted as GCIB was formed with Ar-GCIB bombardment of 5 kV acceleration
voltage.
www.iran-mavad.com
techniques can leave the glass surfaces with high densities of extremely small local asperities that have dimensions of only a few
nanometers. Upon subsequent deposition of a complex optical film stack containing more than 100 film layers, with the
thickness of each layer precisely controlled, the substrate asperities become replicated and amplified so as to create scattering
centers that result in significant degradation of the filter performance and unacceptable production yield losses. GCIB smoothing
can be used to readily eliminate the surface asperities from glass. By employing GCIB smoothing of optical filter substrates prior
to film deposition, the formation of scattering centers within the optical film stack does not occur and production yields are
substantially increased.
4.08.5 Focused Ion Beam Assisted Nanostructure Formation
In 1979, Seliger et al. reported Ga ion beams down to 100 nm diameter with current density in the focal spot of 1.5 A cm2 (53).
In 1990, Wargner et al. demonstrated pillars and walls with high aspect ratios achieved using FIB chemical vapor deposition (FIB-
CVD) (54). In 2000, Matsui et al. demonstrated 3-D nanostructure fabrication by FIB-CVD (9).
Matsui et al. used two commercially available FIB systems (SMI9200, SMI2050, SII Nanotechnology Inc., Tokyo, Japan) with
a Gaþ ion beam operating at 30 keV. The FIB-CVD used a precursor of phenanthrene (C14H10) as the source material. The beam
diameter of SMI9200 was about 7 nm and that of SMI2050 was about 5 nm. The SMI9200 system was equipped with two gas
sources in order to increase the gas pressure. The nozzles faced each other and were directed at the beam point. The nozzles were set
a distance of 40 mm from each other and positioned about 300 mm above the substrate surface. The inside diameter of a nozzle was
0.3 mm. The phenanthrene (C14H10) gas pressure during pillar growth was typically 5 105 Pa in the specimen chamber, but the
local gas pressure at the beam point was expected to be much higher. The crucible of the source was heated to 85 C. The SMI2050
system, on the other hand, was equipped with a single gas nozzle. The FIB is scanned in order to be able to write the desired pattern
via computer control, and the ion dose is adjusted to deposit a film of the desired thickness. The experiments were carried out at
room temperature on a silicon substrate.
The deposited film was characterized by observing it with a transmission electron microscope (TEM). A thin film of carbon
(200 nm thick) was deposited on a silicon substrate by 30 keV Gaþ FIB using phenanthrene precursor gas. The cross sections
of the structures created and its electron diffraction patterns were observed using a 300 kV TEM. There were no crystal
structures in the TEM images and diffraction patterns. It was therefore concluded that the deposited film was amorphous
carbon (a-C).
Raman spectra of the a-C films were measured at room temperature with the 514.5 nm line of an argon ion laser. The
Raman spectra were recorded using a monochromator equipped with a charge-coupled device (multichannel detector. Raman
spectra were measured at 0.1–1.0 mW to avoid thermal decomposition of the samples. A relatively sharp Raman band at
1550 cm–1 and a broad-shouldered band at 1400 cm–1 were observed in the spectra excited by the 514.5 nm line. Two Raman
bands were plotted after Gaussian line shape analysis. These Raman bands, located at 1550–1400 cm–1 originate from the
trigonal (sp2) bonding structure of graphite and tetrahedral (sp3) bonding structure of diamond. This result suggests that the a-
C film deposited by FIB-CVD is DLC, which has attracted attention due to its hardness, chemical inertness, and optical
transparency.
The coordination of carbon atoms in the carbon-based material formed by CVD of phenanthrene assisted by a Ga-FIB was
investigated by the measurement of NEXAFS spectra of the carbon K-edge over the excitation energy range 275–320 eV (55). A novel
peak observed at 289.0 eV was assigned to the 1s to s* transition of carbon neighboring to the residue gallium. The material formed
by this method was found to be Ga-doped DLC, which consists of a high-sp3 hybridized carbon.
The atomic fraction of the FIB-CVD DLC film has been determined as C:Ga:H ¼ 87.4:3.6:9.0 at% using RBS and elastic recoil
detection analysis (56). Hydrogen content of the FIB-CVD DLC film was relatively lower than that of DLC films formed by other
CVD methods.
Beam-induced CVD is widely used in the electrical device industry in repair of chips and masks. This type of deposition is mainly
done on 2-D pattern features, but it can also be used to fabricate a 3-D object. Koops et al. demonstrated a nanoscale 3-D structure
construction (57) by applying electron-beam-induced amorphous carbon deposition to a microvacuum tube. However, FIB-
induced CVD seems to have many advantages for the fabrication of 3-D nanostructures (9). The key issue to realizing such 3-D
nanostructures is the short penetration depth of the ions (a few tens of nanometers) into the target material, where the penetration
depth of the ions is much shorter than that of electrons (several micrometers). This short penetration depth reduces the dispersion
area of the secondary electrons, and so the deposition area is restricted to roughly several tens of nanometers. A 3-D structure usually
contains overhang structures and hollows. Gradual position scanning of the ion beam during the CVD process causes the position of
the growth region around the beam point to shift. When the beam point reaches the edge of the wall, secondary electrons appear at
the side of the wall and just below the top surface. The DLC then starts to grow laterally; the width of the vertical growth is also about
80 nm. Therefore, combining the lateral growth mode with rotating beam scanning, it is possible to obtain 3-D structures with
rotational symmetry like a wineglass.
The process of fabricating 3-D structures by FIB-CVD is illustrated in Figure 14 (9). In FIB-CVD processes, the beam is scanned in
digital mode. First, a pillar is formed on the substrate by fixing the beam position (position 1). After that, the beam position is
moved to within a diameter of the pillar (position 2) and then fixed until the deposited terrace thickness exceeds the range of the
ions (a few tens of nanometers). This process is repeated to make 3-D structures. The key point to making 3-D structures is to adjust
www.iran-mavad.com
Primary ion
Ga+ (30 keV)
Phenanthrene
(C14H10) Dissociation
of adsorbed molecules
3
1 2
Secondary
electron
DLC
Si
Figure 14 Fabrication process for 3-D nanostructure by FIB-CVD.
the beam scan speed so that the ion beam remains within the deposited terrace, which means that the terrace thickness always
exceeds the range of the ions. The growth in the x and y directions are controlled by both beam deflectors. The growth in the
z direction is determined by the deposition rate; that is, the height of structure is proportional to an irradiation time when
a deposition rate is constant.
To demonstrate the possibilities of this field, a ‘micro-wineglass’ was created on a Si substrate and a human hair as a work of
microstructure plastic arts as shown in Figures 15(a) and 16, respectively. A micro-wineglass with an external diameter of 2.75 mm
and a height of 12 mm was formed. The fabrication time was 600 s at a beam current of 16 pA. This beautiful micro-wineglass shows
the potential of the field of microstructure plastic art. A micro-Colosseum and a micro Leaning Tower of Pisa were also fabricated on
a Si substrate as shown in Figures 15(c) and 17, respectively.
Various microsystem parts have been fabricated using FIB-CVD. Figure 15(b) shows a microcoil with a coil diameter of 0.6 mm,
a coil pitch of 0.7 mm, and a line width of 0.08 mm. The exposure time was 40 s at a beam current of 0.4 pA. It was formed by
reducing the diameter of the microcoil. The diameter, pitch, and height of the microcoil were 0.25, 0.20, and 3.8 mm, respectively.
The exposure time was 60 s at a beam current of 0.4 pA. The results show that FIB-CVD is a highly promising technique for realizing
parts of a microsystem, although their mechanical performances must be measured.
The 3-D structure is built up as a multilayer structure. In the first step of this 3-D pattern-generating system, a 3-D model of the
structure, designed using a 3-D CAD system (3-D DXF format), is needed. In this case, we realized a structure shaped like
a pendulum. The 3-D CAD model, which is a surface model, is cut into several slices, as shown in Figure 18. The thickness of the
slices depends upon the resolution in the z direction (the vertical direction). The x and y coordinates of the slices are then used to
create the scan data (voxel data). To fabricate the overhanging structure, the ion beam must irradiate the correct positions in the
correct order. If the ion beam irradiates a voxel located in midair without a support layer, the ions intended for the voxel will be
deposited on the substrate. Therefore, the sequence of irradiation is determined, as shown in Figure 17.
The scan data and blanking signal therefore include the scan sequence, the dwell time, the interval time, and the irradiation
pitch. These parameters are calculated from the beam diameter, x–y resolution, and z resolution of fabrication. The z resolution is
proportional to the dwell time and inversely proportional to the square of the irradiation pitch. The scan data are passed to the
Figure 15 (a) Micro-wineglass with an external diameter of 2.75 mm and a height of 12 mm. (b) Microcoil with a coil diameter of 0.6 mm, a coil-pitch of
0.7 mm, and a line width of 0.08 mm. (c) Micro-Colosseum.
www.iran-mavad.com
Figure 16 Micro-wineglass with an external diameter of 2.75 mm and a height of 12 mm on a human hair.
Figure 17 Micro Leaning Tower of Pisa.
www.iran-mavad.com
Figure 18 Data flow of 3-D pattern-generating system for FIB-CVD.
beam deflector of the FIB-CVD, as are the blanking data. The blanking signal controls the dwell time and interval time of the ion
beam.
Figure 19 shows a 3-D CAD model and a SIM image of the Starship Enterprise NCC-1701D (from the television series Star Trek),
which was fabricated by FIB-CVD at 10–20 pA (58). The nanospaceship is 8.8 mm long and was realized at about a 1:100 000 000
scale on silicon substrate. The dwell time, interval time, irradiation pitch, and total process time were 80 ms, 150 ms, 2.4 nm, and
2.5 h, respectively. The horizontal overhang structure was fabricated successfully. Figure 20 shows a ‘nano T4 bacteriophage,’ which
is an artificial version of the virus fabricated by FIB-CVD on silicon surface. The size of the artificial nano ‘ T4 bacteriophage’ is about
10 times as that of the real virus.
Figure 19 Micro Starship Enterprise NCC-1701D, 8.8 mm long.
www.iran-mavad.com
Figure 20 T4 bacteriophage.
4.08.6 Conclusion
This chapter is a brief introduction to IBAD. Also, emerging IBAD techniques such as the GCIB-assisted deposition and the FIB-
induced nanostructure formations were introduced. IBAD represents a superior method for thin film growth; however, in general,
IBAD is considered a mature technology and there is little space to improve it. However, there are still many opportunities to expand
it into various applications by using ultralow energy ions or ultrasmall ion beams.
References
1. Paredez, P.; Marchi, M. C.; Maia da Costa, M. E. H.; Figueroa, C. A.; Kleinke, M. U.; Ribeiro, C. T. M.; Sanchez-Lopez, J. C.; Rojas, T. C.; Alvarez, F. J. Non-Cryst. Solids 2006,
352, 1303.
2. Groves, J. R.; Hammond, R. H.; Matias, V.; DePaula, R. F.; Stan, L.; Clemens, B. M. Nucl. Instrum. Methods Phys. Res. B 2012, 272, 28.
3. Mattox, D. M. Film Deposition Using Accelerated Ions; Sandia Corporation, 1963. Report SC-DR-281–63.
4. Pranevicius, L. Thin Solid Films 1979, 63, 77.
5. Martin, P. J.; Netterfield, R. P.; Sainty, W. G. J. Appl. Phys. 1984, 55, 235.
6. Colligon, J. S.; Hill, A. E.; Kheyrandish, H. Vacuum 1984, 34, 843.
7. Akizuki, M.; Matsuo, J.; Shin, I.; Harada, M.; Ogasawara, S.; Yamada, I. Nucl. Instrum. Methods Phys. Res. B 1997, 121, 166.
8. Qin, W.; Howson, R. P.; Akizuki, M.; Matsuo, J.; Takaoka, G. H.; Yamada, I. Mater. Chem. Phys. 1998, 54, 258.
9. Matsui, S.; Kaito, T.; Fujita, J.; Komuro, M.; Kanda, K.; Haruyama, Y. J. Vac. Sci. Technol. B 2000, 18, 3181.
10. Insepov, Z.; Sosnowski, M.; Yamada, I. Trans. Mater. Res. Soc. Jpn. 1994, 17, 111.
11. Insepov, Z.; Yamada, I. Nucl. Instrum. Methods Phys. Res. B 1995, 99, 248.
12. Hirvonen, J. K. Mater. Sci. Rep. 1991, 6, 215.
13. Mohan, S.; Krishna, M. G. Vacuum 1995, 46, 645.
14. Harper, J. M. E.; Cuomo, J. J.; Gambino, R. J. Nucl. Instrum. Methods Phys. Res. B 1985, 7/8, 886.
15. Hubler, G. K.; Carosella, C. A.; Donovan, E. P.; Vanvechten, D.; Bassel, R. H.; Andreadis, T. D.; Rosen, M.; Mueller, G. P. Nucl. Instrum. Methods Phys. Res. B 1990, 46, 384.
16. Manova, D.; Gerlach, J. W.; Mandl, S. Materials 2010, 3, 4109.
17. Trushin, Y. V.; Kulikov, D. V.; Safonov, K. L.; Gerlach, J. W.; Hoche, Th.; Rauschenbach, B. J. Appl. Phys. 2008, 103, 114904.
18. Trushin, Y. V. J. Nucl. Mater. 1991, 185, 268.
19. Kharlamov, V. S.; Kulikov, D. V.; Trushin, T. V. Vacuum 1999, 52, 407.
20. Trushin, Y. V.; Safonov, K. L.; Ambacher, O.; Pezoldt, J. Tech. Phys. Lett. 2003, 29, 663.
21. Sienz, S.; Gerlach, J. W.; Hoche, T.; Frost, F.; Rauschenbach, B.; Benndorf, G. Appl. Phys. Lett. 2006, 88, 011902.
22. Muller, K. H. Phys. Rev. B 1987, 35, 7906.
23. McKenzie, D. R. J. Vac. Sci. Technol. B 1993, 11, 1928.
24. Baglin, J. In Handbook of Ion Beam Processing Technology; Cuomo, J., Rossnagel, S., Kaufman, H., Eds.; Noyes: Park Ridge, NJ, 1989.
25. Bradley, R. M.; Harper, J. M. E. J. Vac. Sci. Technol. A 1988, 6, 2390.
26. Patzig, C.; Rauschenbach, B.; Erfurth, W.; Milenin, A. J. Vac. Sci. Technol. B 2007, 25, 833.
27. Teichert, C.; de Miguel, J. J.; Bobek, T. J. Phys. Condens. Matter 2009, 21, 224025.
28. Xiao, J.; Xu, L.; Geng, L.; Tong, L.; Yang, F.; Xu, J.; Su, W.; Liu, D.; Yu, Y.; Ma, Z.; Chen, K. Appl. Surf. Sci. 2011, 257, 10893.
29. Ensinger, W. Rev. Sci. Instrum. 1992, 63, 5217.
30. Mach, J.; Samuril, T.; Voborny, S.; Kolibal, M.; Zlamal, J.; Spousta, J.; Dittrichova, L.; Sikola, T. Rev. Sci. Instrum. 2011, 82, 083302.
31. Zachmann, H.; Puttnins, S.; Daume, F.; Rahm, A.; Otte, K. Thin Solid Films 2011, 519, 7317.
www.iran-mavad.com
32. Nabot, J.-P. H.; Andre, B.; Paidassi, S. Surf. Coat. Technol. 1990, 43/44, 71.
33. Quan, J. J.; Wolf, S. A.; Wadley, H. N. J. Appl. Phys. 2007, 101, 074302.
34. Olzon-Dionysio, D.; Chubaci, J. F. D.; Matsuoka, M.; Faria, R. M.; Guimarães, F. E. G. Surf. Coat. Technol. 2010, 204, 3096.
35. Song, J. S.; Lee, S.; Cha, G. C.; Jung, S. H.; Choi, S. Y.; Kim, K. H.; Mun, M. S. J. Appl. Polym. Sci. 2005, 96, 1095.
36. Matsuo, J.; Toyoda, N.; Yamada, I. J. Vac. Sci. Technol. B 1996, 14, 3951.
37. Toyoda, N.; Hagiwara, N.; Matsuo, J.; Yamada, I. Nucl. Instrum. Methods Phys. Res. B 1999, 148, 639.
38. Toyoda, N.; Kitani, H.; Matsuo, J.; Yamada, I. Nucl. Instrum. Methods Phys. Res. B 1997, 121, 484.
39. Toyoda, N.; Houzumi, S.; Mashita, T.; Mitamura, T.; Mochiji, K.; Yamada, I. Surf. Coat. Technol. 2007, 201, 8624.
40. Goto, K.; Matsuo, J.; Tada, Y.; Tanaka, T.; Momiyama, Y.; Sugii, T.; Yamada, I. IEDM Tech. Dig. 1997, 471.
41. Yamada, I.; Matsuo, J.; Toyoda, N.; Kirkpatrick, A. Mater. Sci. Eng. 2001, R34, 231.
42. Shirai, K.; Fujiwara, Y.; Takahashi, R.; Toyoda, N.; Matsui, S.; Mitamura, T.; Terasawa, M.; Yamada, I. Jpn. J. Appl. Phys. 2002, 41, 4291.
43. Chu, W. K.; Li, Y. P.; Liu, J. R.; Wu, J. Z.; Tidrow, S. C.; Toyoda, N.; Matsuo, J.; Yamada, I. Appl. Phys. Lett. 1998, 72, 246.
44. Toyoda, N.; Hirota, T.; Yamada, I.; Yakushiji, H.; Hinoue, T.; Ono, T.; Matsumoto, H. IEEE Trans. Magn. 2010, 46, 1599.
45. Toyoda, N.; Matsuo, J.; Aoki, T.; Yamada, I.; Fenner, D. B. Appl. Surf. Sci. 2003, 203–204, 214.
46. Matsuo, J.; Ninomiya, S.; Nakata, Y.; Honda, Y.; Ichiki, K.; Seki, T.; Aoki, T. Appl. Surf. Sci. 2008, 255, 1235.
47. Hada, M.; Ibuki, S.; Ninomiya, S.; Seki, T.; Aoki, T.; Matsuo, J. Jpn. J. Appl. Phys. 2010, 49, 036503.
48. Miyayama, T.; Sanada, N.; Suzuki, M.; Hammond, J. S.; Si, S. Q. D.; Takahara, A. J. Vac. Sci. Technol. A 2010, A28 (L1), 1116.
49. Toyoda, N.; Yamada, I. IEEE Trans. Plasma Sci. 2008, 36, 1471.
50. Toyoda, N.; Fujiwara, Y.; Yamada, I. Nucl. Instrum. Methods Phys. Res. B 2003, 206, 875.
51. Kitagawa, T.; Yamada, I.; Toyoda, N.; Tsubakino, H.; Matsuo, J.; Takaoka, G. H.; Kirkpatrick, A. Nucl. Instrum. Methods Phys. Res. B 2003, 201, 405.
52. Kanda, K.; Kitagawa, T.; Shimizugawa, Y.; Haruyama, Y.; Matsui, S.; Terasawa, M.; Tsubakino, H.; Yamada, I.; Gejo, T.; Kamada, M. Jpn. J. Appl. Phys. 2002, 41, 4295.
53. Seliger, R. L.; Kubena, R. L.; Olney, R. D.; Ward, J. W.; Wang, W. J. Vac. Sci. Technol. 1979, 16, 1610.
54. Wargner, A.; Levin, J. P.; Mauer, J. L.; Blauner, P. G.; Kirch, S. J.; Long, P. J. Vac. Sci. Technol. B 1990, 8, 1557.
55. Kanda, K.; Igaki, J.; Kato, Y.; Kometani, R.; Saikubo, A.; Matsui, S. Radiat. Phys. Chem. 2006, 75, 1850.
56. Igaki, J.; Saikubo, A.; Kometani, R., et al. Jpn. J. Appl. Phys. 2007, 46, 8003.
57. Koops, H. W. Jpn. J. Appl. Phys. 1994, 33, 7099.
58. Hoshino, T.; Watanabe, K.; Kometani, R., et al. J. Vac. Sci. Technol. B 2003, 21, 2732.
www.iran-mavad.com
4.09 Plasma Polymer Deposition and Coatings on Polymers
D Hegemann, Empa, St. Gallen, Switzerland
4.09.1 Plasma Polymer Deposition 201

4.09.1.1 Introduction to Plasma Polymerization 201
4.09.1.1.1 Quasi-Arrhenius Approach 204
4.09.1.1.2 Energy Density in the Gas Phase 205
4.09.1.1.3 Energy Density at the Surface 207
4.09.1.1.4 Influence of Pressure 209
4.09.1.2 Plasma Polymer Deposition from Different Monomers 209
4.09.1.2.1 Hexamethyldisiloxane 209
4.09.1.2.2 O2/HMDSO 212
4.09.1.2.3 Hydrocarbons 212
4.09.1.2.4 Hydrocarbons with Reactive Gases 215
4.09.1.2.5 Acrylic Acid 215
4.09.1.2.6 Further Monomers 216
4.09.1.3 Conclusion 216
4.09.2 Coatings on Polymers 217
4.09.2.1 General Aspects for the Coating on Polymers 217
4.09.2.1.1 Influence of Polymers on the Plasma Conditions 219
4.09.2.1.2 Aging of Plasma Polymer Depositions 219
4.09.2.1.3 Adhesion of Plasma Polymer Depositions 220
4.09.2.1.4 Adhesion of Metallizations 222
4.09.2.2 Industrial Aspects for the Coating on Polymers 222
4.09.2.2.1 Plasma Sources 223
4.09.2.2.2 Atmospheric Pressure Plasmas 223
4.09.2.3 Conclusion 225
References 225
4.09.1 Plasma Polymer Deposition
Deposition of plasma polymers occurs via plasma activation of a precursor (typically a monomer-containing gas) creating reactive
intermediates that yield macromolecule formation. Primary processes are activated by electron impact, since the electrons take most
of the delivered energy leaving the plasma ‘cold’ (nonequilibrium plasma). Even at a low mean electron temperature Te (e.g., 1 eV),
some electrons gain energies up to tens of eV due to the tail of the electron energy distribution function (eedf ). Hence, all chemical
bonds can be broken in a plasma. The intermediates adsorb at surfaces and stick together depending on their reactivity. Generally
starting with islandlike growth (and thus inherently comprising a pinhole structure), plasma polymers subsequently cover surfaces
in a layer-by-layer growth (1). The deposits are highly branched and cross-linked, different than conventional long-chain polymers.
The notice of deposits in electric discharges goes back to the 1870s. These early plasma polymers, however, were thought of as
undesirable byproducts associated with electric discharges and little attention was given to their properties (2). Only from 1960 on,
plasma polymer deposition and their film properties were studied with more care (3). Today, plasma polymers are important in
diverse fields such as microelectronics, tools, machine parts, glasses, medical devices, solar cells, packaging, filtration, sensors, tissue
engineering, composites, textiles, and fibers.
The process to deposit plasma polymers is often named plasma polymerization, which is also be used throughout this chapter.
Other common names are plasma chemical vapor deposition (CVD), plasma-enhanced CVD, plasma-assisted CVD, glow discharge
polymerization, or luminous CVD. Furthermore, plasma polymers can be deposited by sputtering of polymer targets (4). Plasma
polymer deposition is a dry and resource-saving technology, since only low amounts of precursors are required.
4.09.1.1 Introduction to Plasma Polymerization

Plasma polymerization enables the deposition of a broad range of plasma polymers covering properties from soft and swellable
up to hard and cross-linked coatings (Figure 1). As swellable coatings, mainly PEO-like films are investigated as potential non-
fouling surfaces (5). These coatings require a low energy input (low fragmentation and cross-linking) in order to maintain
functional C–O–C groups. Thus, they show a limited stability. Increasing cross-linking yields more permanent coatings, while still
functional groups such as –COOH, –OH, –NH2, and –CH3 can be retained. Hydrophilic or hydrophobic surfaces can be obtained
and the penetration of, e.g., gases through the plasma polymer films can be controlled. Finally, high cross-linking avoids the
Comprehensive Materials Processing, Volume 4 http://dx.doi.org/10.1016/B978-0-08-096532-1.00426-X 201

www.iran-mavad.com
202 Plasma Polymer Deposition and Coatings on Polymers
Figure 1 Plasma polymerization covers a broad range of film properties, depending on the energy input into the plasma process responsible for
fragmentation and densification.
incorporation of terminal groups resulting in dense, hard coatings. Such plasma polymers can be quartzlike or diamondlike; i.e.,
they are no longer ‘polymeric.’
Hence, the properties of plasma polymers mainly depend on the energy which is consumed within the deposition process. But
what exactly is meant by the term ‘energy density’?
It is known since 1920 that the plasma chemical decomposition of gases is determined by the energy invested per particle within
the active plasma zone, depending on the power input per plasma volume multiplied by the residence time, also named Becker’s
formula (6). This reaction parameter was originally used in gas synthesis (e.g., ozone formation). A dimensionless reaction
parameter S can then be written as
Wsplasma T
S¼ [1]
pVplasma T0
with the power input W, the residence time in the plasma zone splasma, the pressure p, the volume of the plasma zone Vplasma, and
the temperature T (related to the standard temperature T0 ¼ 273 K). Using the ideal gas law, it can easily be seen that this parameter
describes the energy invested per particle in the plasma zone. Furthermore, considering that
splasma 1 pT0
¼ [2]
Vplasma F p0 T
with the gas flow rate F at standard conditions (given in standard cubic centimeters (sccm) per minute) and the standard pressure p0
(¼101325 Pa), an energy density 3plasma can be defined as the energy delivered per plasma volume (specific energy) by the external
parameters power input and gas (monomer) flow rate, W/F:

W
3plasma ¼ [3]
F plasma
This energy density 3plasma, which is independent of pressure, delivers the energy per molecule in the gas phase (plasma), which
is required, e.g., for dissociation processes. Note that 1 eV per molecule corresponds to 4.3 J cm–3 (or 42 in the dimensionless
eqn [1]) using the ideal gas law.
In the 1970s, Yasuda started to examine mass deposition rates in plasma polymerization processes depending on the energy
input into the plasma (2). Empirically, he found that the mass deposition rate Rm increases almost linearly with power input W at
low energy input (energy-deficient regime). At high energy input, on the other hand, Rm appeared to be mainly proportional to the
monomer flow rate F (monomer-deficient regime) as depicted in Figure 2. Note that Yasuda actually introduced the reaction
parameter W/FM with the molecular gas of the monomer M. Although widely accepted, the theoretical basis of this approach and its
practical value are still under discussion. Most of all, this approach neglects surface processes, which are obviously important in film
deposition.
Therefore, different models were developed such as the competitive ablation and polymerization principle (CAP), the acti-
vated growth model (AGM), as well as the subplantation model considering film growth, densification, and ablation processes at
the surface (7–9). Further plasma polymerization models are, e.g., discussed in a recent review by Friedrich (10). For the
deposition of hard coatings, the energy of incident ions at the growing film surface, which strongly correlates with bias voltage in
asymmetric reactors, is often considered a crucial factor in plasma polymerization. As an example, SiOx films were deposited from
O2/hexamethyldisiloxane (HMDSO) (15:1) radio frequency (RF) discharges at fixed plasma conditions (pressure of 7 Pa, power
of 100 W, and gas flow rate of 60 and 4 sccm), but varying bias voltage. While the mass deposition rate (i.e., the flux of film-
forming species depending on the energy density in the gas phase) was maintained in all cases, the film density was increased
with bias voltage as shown in Figure 3. Hence, both gas phase and surface processes have to be considered for plasma
polymerization.
www.iran-mavad.com
Plasma Polymer Deposition and Coatings on Polymers 203
Figure 2 Empirical finding of the dependence of mass deposition rate per gas flow rate, Rm/F, on energy input, W/F, as observed by Yasuda.
Figure 3 Variation of the bias voltage in an O2/HMDSO discharge at otherwise fixed plasma deposition conditions. Higher ion energies result in
increasing film density at maintained ion and particle flux (constant deposition rate).
However, films deposited at comparable surface conditions (bias voltage of 600 V at a pressure 1.4–1.5 Pa), but different
power input into the gas phase (110 W vs. 50 W) reveal different film properties (e.g., hardness of 15 GPa vs. 6 GPa) as shown for
tetramethylsilane (TMS)-fed discharges (11). A lower power density in the plasma yields reduced electron densities and thus
reduced fragmentation. Since the ion flux to the substrate is influenced by ion density (equal to electron density in electropositive
discharges) and Bohm velocity (lower for higher ion masses), a lower power input results in a reduced energy flux to the substrate,
even when the ion energies are comparable. For surface processes, the energy density during film growth has to be considered, which
is given by the energy flux per deposition rate
Gi Emean
3surf ¼ [4a]
R
with the incident ion flux Gi, the mean ion energy Emean, and the deposition rate R. This energy density gives the energy that is
dissipated in the growing film volume and can easily be converted to energy per deposited atom (in eV per atom) regarding the
atom deposition rate (12). The number of atoms of a particular element per film volume is given by
NA catom Matom
natom ¼ r [4b]
Matom catom Matom þ ci Mi
with the film density r, the Avogadro number NA, the molecular mass of the considered atom Matom, and the concentration of the
atom in the film catom (where the subscript i denotes further elements in the film). Amorphous hydrocarbon layers, a-C:H, for
example, contain w67 carbon atoms per nm3 at 1.4 g cm–3, w82 C atoms per nm3 at 1.7 g cm–3, and w98 C atoms per nm3 at
2.0 g cm–3. Note that higher electron densities can enhance both the ion flux and the deposition rate in a similar way, which explains
the often-observed good agreement of film properties with bias voltage.
Plasma polymerization can thus be investigated on the basis of the energy density both in the gas phase (e.g., in eV per monomer
molecule yielding dissociation) and at the surface (e.g., in eV per deposited atom yielding cross-linking) (13). Surface processes
www.iran-mavad.com
Figure 4 Influence of the energy density in the plasma, W/F, and at the surface, 3surf, on the nature of the resulting plasma polymer depositions.
might have the stronger impact as schematically shown in Figure 4, since larger fragments or adsorbed, unreacted monomers can
still be densified by energetic particle bombardment as well as by etching of weaker phases (14).
Excess energy conditions as they are obtained in plasma polymerization processes actually allow all kinds of chemical (endo-
thermic) reactions to take place.
4.09.1.1.1 Quasi-Arrhenius Approach

For a chemical reaction to take place between two colliding particles, normally, an activation barrier has to be overcome. The
reaction rate constant k then depends on the temperature T as

Ea
k ¼ n exp [5]
RT
where n is the collision frequency and Ea the activation energy (with the gas constant R). Since the collision frequency only weakly
depends on temperature, it can be considered as being constant yielding in the well-known Arrhenius equation (at least around Ea
where the increase with T is strong).
In thermal CVD processes, where the chemical reactions depend on thermally activated dissociation and surface diffusion
processes, the kinetics of the surface reactions can thus be examined using the Arrhenius approach, i.e., a logarithmic plot of the
mass deposition rate Rm over the inverse temperature (Figure 5) (15). Typically, two regimes can be obtained, if one dominant
reaction pathway occurs: (1) reaction rate limited regime (the deposition rate is limited by the surface reactions), and (2) mass
transport limited regime (the deposition rate is limited by diffusion of the reactants to the surface).
With increasing temperature, different reaction channels such as reverse reactions might contribute to the net deposition rate
showing deviations from the simple Arrhenius approach. However, for a certain temperature range, i.e., the activation regime
around the corresponding activation energy Ea, the observed activation energy by the Arrhenius plot might be related to a disso-
ciation step providing radical-generating species required to initiate the polymerization process (16). Recently, it was shown that the
dissociation of a weak peroxide bond (O–O) in an initiator molecule (tert-butyl peroxide) can also be achieved by a low-energy
plasma where the film formation rate follows the Arrhenius law (17).
Figure 5 Arrhenius plot as typically obtained in thermally activated CVD processes.
www.iran-mavad.com
Figure 6 Increase in deposition rate, depending on the energy input for plasma polymer depositions with an activation barrier.
At the end of the 1980s, Park and coworkers were assuming similar activation reactions in plasma polymerization processes,
where the role of the thermal energy RT is now replaced by the specific energy W/F, which delivers the energy for dissociation
processes in the plasma (18). They could show for an RF-driven capacitively coupled discharge with hexamethyldisilazane that the
mass deposition rate per monomer gas flow rate follows the quasi-Arrhenius relation
0 1
Rm B Ea C
¼ C exp@ . A [6]
F W F
over a broad parameter range. This quasi-Arrhenius regime is determined by the apparent activation energy Ea and a constant C
comprising the decomposition rate of the monomer which strongly depends on the eedf of the plasma (17,19). Figure 6 shows
a graph of eqn [6]. Compared to Figure 2, the apparent activation energy Ea now separates the two regimes (energy deficient and
monomer deficient) as obtained by Yasuda, yet also comprising the transition regime (20).
The conversion of monomer into film growth is given by
Rm RT0
xdep ¼ A [7]
F dep Mr p0
with the deposition area Adep and the molecular mass of the monomer Mr. The constant C in eqn [6] thus gives the maximum
conversion within the Arrhenius regime (when the exponential term approaches 1) (21).
Regarding the work of Rutscher and Wagner, this approach should be adequate if a dominant plasma–activated reaction
pathway exists; i.e., the active plasma zone is the source of the film-forming species (22). In this case, the mass deposition rate is
proportional to the flux of film-forming species incident on the substrate surface enabling a macroscopic description of plasma
polymerization (13,23). In practice, different plasma chemical reaction channels (with different activation energies) might exist.
One example is the formation of ozone in a cold oxygen discharge. At low energy input into the plasma, ozone synthesis proceeds
mainly via excited states (2O2(1Dg) þ O2 / 2O3) requiring less than 1 eV per molecule as activation energy, while electron
dissociation reactions prevail at enhanced energy input with a threshold of 6.0 and 8.4 eV (O2 þ e / 2O þ e and
O þ O2 þ M / O3 þ M) (22,24). Note that the electron dissociation channel typically requires higher energies than the dissociation
energy (e.g., 5.4 eV for O]O), since it also proceeds via excited states.
The film growth processes are also influenced by the energy flux incident on the substrate surface according to eqn [4], which can
support, limit, or reverse the incorporation of film-forming species. The net deposition rate might thus look as depicted in Figure 7,
where also adsorption of unreacted monomers (at low energy input) and ion-induced etching processes (at high energy input)
might contribute to the deposited mass.
Nevertheless, the macroscopic approach based on a proper selection of experimental conditions – also considering the plasma
reactor geometry – can help to get insights into plasma polymerization processes, while a microscopic approach based on eedf and
rate constants may still be too complex when all possible reactions have to be considered.
4.09.1.1.2 Energy Density in the Gas Phase

To enhance the benefit of the quasi-Arrhenius approach by using the external parameters power input and gas flow rate, well-
defined plasma conditions are required. Power losses and secondary plasmas should be minimized, which can be controlled using
V/I probes to record excitation voltage V0, discharge current I0, and phase angle 4. Power absorption in the plasma, Wabs, can thus be
calculated and followed over the examined range of nominal power, W. For example, V0 (and also Vbias) often shows a square root
dependence on power input following an ohmic relation (25). Increasing power losses with increasing nominal power, which often
occur, can then be detected by deviations (26). Moreover, the plasma volume expands with increasing power input in a nonconfined
www.iran-mavad.com
Figure 7 Typical Arrhenius-type plot for plasma polymer depositions showing different activation barriers and additional influences on the
deposition rate.
geometry (25). The electron density ne thus scales linearly with W/Vplasma (14). Varying pressure also influences the plasma volume.
While the energy density 3plasma according to eqn [3] is independent of pressure, since it gives the energy per molecule, different
plasma volumes might be considered involving the influence of pressure (21).
Furthermore, a part of the gas flow might bypass the active plasma region, thus not participating in the deposition process. Those
effects can be regarded by a factor rdep comprising the geometry of the plasma reactor set-up used for plasma polymer deposition:

W W
3plasma ¼ ¼ rdep [8a]
F plasma F
with
Wabs dact Vgas
rdep ¼ [8b]
W dgas Vdis
where dact represents the width of the active plasma zone, dgas the way of the gas traveling through the plasma to the substrate, Vgas
the volume consumed by the total gas flow and Vdis the volume consumed by the total gas discharge (Figure 8(a)) (21,27). This
geometrical factor helps in finding the real energy delivered to the plasma volume with respect to the location of the plasma
deposition, which gives the energy per molecule in the gas phase and supports the comparison of different experimental set-ups. The
situation actually becomes more defined using a confined, symmetric plasma reactor set-up with a uniform, vertical gas inlet into
the plasma zone by a gas shower head, where Vgas ¼ Vdis and dact/dgas ¼ const (e.g., plasma length per electrode gap in a plane
parallel electrode set-up, as shown in Figure 8(b)) (23).
When using gaseous mixtures, i.e., inert carrier gases and reactive gases mixed with the monomer, another correction can be
carried out in order to obtain the energy delivered per monomer molecule. Assuming that a part of the energy is consumed by the
‘nonpolymerizable’ gases, which also contributes to the plasma chemical reaction pathway, this fraction can be considered by
a factor rgas introduced for the total gas flow rate
F ¼ Fm þ rgas Fc [9]
with the monomer flow rate Fm and the carrier (or reactive) gas flow rate Fc (20). Note that by evaluation of the deposited mass out
of the plasma volume according to eqn [6], inherently, only processes contributing to film growth are covered. Thus, rgas gives the
Figure 8 Plasma reactor geometry and plasma expansion in an asymmetric RF discharge (a) and a symmetric, confined discharge (b).
www.iran-mavad.com
Figure 9 Correction factors introduced in order to obtain the energy invested per monomer molecule in the plasma.
reactivity of the added gas with respect to the formation of the film-forming species. The quasi-Arrhenius approach can nevertheless
be always performed starting with the external parameter W/F (neglecting rdep and rgas), since mainly the slope of the Arrhenius fit is
shifted by the corrections (Figure 9).
4.09.1.1.3 Energy Density at the Surface

In order to estimate the energy density at the surface, the energy flux incident on the substrate should be known besides the
deposition rate according to eqn [4]. Therefore, V/I probes, Langmuir, heat flux, and ion flux probes might be used. The mean energy
Emean of the ions impinging on the substrate (in (eV)) depends on the sheath voltage Vsh (potential drop between plasma potential
and substrate potential, e.g., bias or floating potential) according to a study by Manenschijn and Goedheer (28):
Emean ¼ eVsh cos 4 [10a]
The sheath voltage in an RF discharge can be obtained by averaging over an RF cycle depending on excitation voltage and
(negative) bias voltage as Vs ¼ 0.39 V0 0.73 Vbias. The factor cos 4 regards collisions in the plasma sheath and depends on the
sheath length s, the ion mean free path li, and the type of collisions (where c z 0.5 for elastic collisions and somewhat higher – but
below 1 – for charge exchange):
1
cos 4 ¼ 1 þ c s=l [10b]
i
Different expressions can be found for cos 4 in the literature; this simple expression, however, matches already well with
experimental and modeling data (Figure 10(a)) (12,28). In the case that a substrate is considered at floating potential (e.g., placed
within the active plasma zone), the smaller floating potential Vfl (about five times the electron temperature Te) replaces Vsh in eqn
[10a] and the sheath length s is only a few Debye lengths (thus typically collision free) (14).
The ion flux Gi incident on the substrate (in (nm–2 s–1)) can be estimated using a model proposed by Lieberman et al. (29):
Gi ¼ n0 vB cos q [11a]
with n0 being the plasma density in the plasma bulk (which equals the electron density ne for electropositive gases) and vB, the
Bohm velocity of the ions entering the plasma sheath. The reduction of the number of ions fulfilling the Bohm criterion due to
Figure 10 Reduction of mean ion energy (a) and ion flux (b) incident on a substrate surface due to collisions in the gas phase. As examples, Ar and C2H2
discharges are shown.
www.iran-mavad.com
collisions in the plasma bulk (presheath) is considered by the factor cos q, which in turn depends on the plasma length l and the ion
mean free path li

l 0:5
cos q ¼ 0:86 3 þ [11b]
2li
This expression holds over a broad range of pressure (Figure 10(b)) (29).
The Bohm velocity vB is given by
sffiffiffiffiffiffiffi
kTe
vB ¼ [11c]
mi
with the electron temperature Te, the mean ion mass mi, and the Boltzmann constant k. vB gives the velocity of the ions entering the
plasma sheath, which they gained by the small (Bohm) potential kTe/2 in the presheath (Bohm criterion). The presheath expands
(at least in symmetric discharges) over half of the plasma length. In the case of a floating substrate, the induced plasma perturbation
expands only over a few Debye lengths and can thus be considered as collision free; i.e., cos q z 0.5 (14,29). Note that in many
textbooks a factor of 0.61 is given using a simple derivation (which might be an adequate approximation for asymmetric
discharges). Electron temperatures are typically independent of power input, while electron density scales with power input per
plasma volume. High electron temperatures as, for example, in hydrogen-rich plasmas can thus increase the ion flux. The mean ion
mass can be reduced with increasing power, i.e., in highly fragmented monomer-containing discharges (30). The energy flux can
thus be calculated from eqns [10] and [11].
Film-forming species adsorb at the exposed surface depending on their reactivity (sticking coefficient). While physical
adsorption consumes an energy of less than 1 eV per atom, higher energies are required for chemical reactions to take place. The
surface binding energy for this chemisorption process is typically 10–50% higher than the covalent binding energy to create
a covalent bond. This energy is provided by the energy flux from the plasma incident on the surface. Hence, the energy density
during film growth enables the creation of nucleation sites and cross-linking reactions, but also ion-assisted etching.
Since mean ion energy roughly scales with the square root of power (Vsh w W0.5) and ion flux is roughly proportional to power
(ne w W), the energy flux increases with W1.5 (12). According to eqn [6], the deposition rate, on the other hand, increases strongly
around Ea, but less at enhanced power input (for a fixed flow rate). Overall, the energy density at the surface, i.e., the dissipated
energy during film growth, thus dominates for the high-power range and increases further as soon as the flux of film-forming species
is reduced due to, e.g., a change in the plasma-chemical reaction pathway (Figure 11).
Windischmann has shown that the induced stress in plasma coatings actually scales with the square root of energy (of the
incident particles) multiplied by their flux (number per area and time) following the knock-on linear cascade theory of forward
sputtering (31). The formation of intrinsic (compressive) stress during film growth is thus momentum rather than energy driven.
Davis further advanced this approach by relating the ‘momentum flux’ (E0.5 mean$Gi) to the deposition rate R, i.e., the momentum
transfer per deposited atom (32). His work investigated the build-up and relaxation of stress based on knock-on implantation of
deposited atoms. To initiate these processes, the incoming ion should bring an energy exceeding some threshold energy. Since the
induced volumetric strain is also linked to film density, the densification during plasma polymer growth should be determined by
the momentum density
Gi E0:5 3surf
psurf ¼ mean
¼ 0:5 ; [12]
R Emean
i.e., collision cascades within the growing film (33). Note that an additional factor of 2(2mi)0.5 might be considered by relating the
mean energy to the mean momentum.
Figure 11 Typically, the energy dissipated during film growth increases with power input above the activation barrier. A different behavior is obtained
around Ea due to strongly increasing deposition rates.
www.iran-mavad.com
4.09.1.1.4 Influence of Pressure

Several authors already noticed that plasma polymer deposition behaves quite similar independent of pressure, i.e., when
comparing low pressure and atmospheric pressure plasma processes, especially for siloxanes such as HMDSO and tetraethyl
orthosilicate (TEOS) (34–36). This agrees well with the concept that the energy input into the plasma, W/F, which governs the
plasma chemical reaction pathway, is actually independent of pressure, since it gives the energy invested per molecule in the active
plasma zone. At increasing pressure the reactor volume is filled up with an increasing amount of carrier gas in addition to the
monomer. At low pressure the energy is mainly transferred by direct electron impact dissociation of the molecules, whereas at higher
pressure the carrier gas is excited which then transfers energy to the monomer molecules. Therefore, helium works very well as carrier
gas due to its high mobility and its abundant excited states (metastables) up to the ionization energy of 24.6 eV. Hence, the
formation of film-forming species might be considered as being independent of pressure.
The actual film growth conditions, however, depend on the energy dissipated during film growth as well as on the transport of
the film-forming species to the substrate.
With increasing pressure the plasma zones (plasma bulk and plasma sheaths) tend to become smaller due to the decreasing
mean free path length yielding local phenomena and stronger gradients. While mean ion energies are low (e.g., in the range of
1–3 eV at atmospheric pressure), the ion flux to the substrate can still be high at elevated pressure due to high electron densities and
small plasma volume (high energy density). A high energy density at the growing film surface can thus be reached by adjusting
energy flux and deposition rate according to eqn [4]. In addition, excited molecules can transfer energy.
Increased collision frequencies also result in convection and gas heating. In order to overcome those phenomena, high carrier gas
flow rates are thus required at elevated pressures. At the low pressure side, on the other hand, plasma polymerization is limited by
the residence time and the achievable gas flow rate of the monomer according to eqn [2]. At pressures around 0.1 Pa, the ion flux
incident on the substrate was found to become comparable to the deposition rate, while at higher pressures primarily neutral species
contribute to the plasma polymer deposition (37). For these reasons, most plasma polymerization processes are carried out in the
pressure range of 0.1–1000 Pa with, e.g., hard a-C:H coatings at the low pressure and microcrystalline Si:H films at the high pressure
end. Increasing effort is nevertheless spent for atmospheric pressure processes due to the expected easier handling of substrates,
which will be further discussed in the Section 4.09.2.
4.09.1.2 Plasma Polymer Deposition from Different Monomers

This section considers different monomers as an example for plasma polymerization processes yielding functional coatings.
Monomer chemistry, gas phase, and surface processes determine the film growth conditions. However, the detailed chemical
reaction pathway is generally only partly understood. Addition of reactive gases further influences the plasma chemistry, which is
also considered.
4.09.1.2.1 Hexamethyldisiloxane
HMDSO, (CH3)3–Si–O–Si–(CH3)3, is a frequently used organosilicon monomer. It is a liquid precursor with a rather high vapor
pressure and safe to handle. It might thus serve as a model plasma process in order to achieve a better understanding of plasma
polymerization and to compare different plasma reactors. The deposition of HMDSO-fed discharges is one example where the
plasma chemical reaction pathway and the film-forming species seem to be known. First of all, direct plasma activation of the
monomer is required (in the absence of a reactive gas), i.e., electron impact dissociation (38). By electron impact in the plasma,
mainly hydrogen (C–H bond energy of 3.5 eV) or methyl groups (Si–C bond energy of 4.6 eV) are abstracted (39). Note that
electron impact dissociation requires activation energies exceeding the bond energies by a few eV due to molecule excitation (40).
Therefore, many of these fragments (and reaction products) are found in the gas phase (41,42). However, species that are abun-
dantly detected in the gas phase likely have a low reactivity at surfaces with respect to film formation. Dissociation of the Si–O bond
(8.3 eV), on the other hand, yields the radical (CH3)3–Si–O, which has a high affinity at surfaces to be incorporated into the
growing film. Conversion rates of the HMDSO molecule into film growth (eqn [7]) of more than 30% can be obtained (e.g.,
70 nm min–1 for 6 sccm HMDSO and Adep of 1500 cm2) indicating a sticking coefficient close to 1. Further abstraction of a methyl
group (CH3) – most likely at the surface – enables the formation of a strong Si–O backbone, which contributes to the high sticking
coefficient (reactivity) of the film-forming species. For this process, the surface binding energy has to be overcome (6.7 eV for C on Si
surfaces for, e.g., abstraction of –CH3) (43). Under conditions that supply sufficient energy in the gas phase and at the surface,
a predominant chemical reaction pathway can thus be observed. Note that very low deposition rates (w1 nm min–1) have been
observed in inert gas-based HMDSO plasmas at atmospheric pressure where the energy flux to the substrate was actually small
(36,44). Also in downstream or remote plasma operation, a reduction in deposition rate occurs, accompanied by reduced energy
fluxes (45,46). In a low-power, low-pressure HMDSO plasma, the sticking probability was estimated to be as low as w0.1% at
insufficient energy density at the growing film surface (according to eqn [4]) (37). The deposition rate is thus limited by the
formation of nucleation sites and cross-linking. Enhanced deposition rates, on the other hand, can be obtained at atmospheric
pressure plasma when the specific energy consumptions are increased (47).
As mentioned before, many different species are created in the plasma which even require lower dissociation energies (compared
to the Si–O bond). However, those species show low sticking coefficients or form stable volatile compounds. Although the second
radical formed by Si–O dissociation, (CH3)3–Si, enables film growth as known from TMS discharges, it seems not to contribute in
HMDSO plasma polymerization. First, it shows a lower reactivity as evidenced by strongly reduced deposition rates obtained by
www.iran-mavad.com
TMS compared to HMDSO (48). Second, the radicals might be displaced by (CH3)3–Si–O at the growing film surface due to its
stronger bond formation. These findings also agree with the observation of TMS in the gas phase of HMDSO discharges (39,42).
If we now assume that the plasma chemical reaction pathway to form siliconelike plasma polymers from HMDSO discharges
proceeds predominantly via the (CH3)3–Si–O radical, a rather high activation barrier should be obtained for the required
dissociation step. The measurement of the mass deposition rate of pp-HMDSO films, which is proportional to the flux of film-
forming species (in this case, (CH3)3–Si–O), under well-defined deposition conditions, is thus suitable to prove this assumption.
Therefore, symmetric, capacitively coupled RF plasma reactors were used, where the plasma volume as well as the power
consumption is well known. Using two different reactors – one with an electrode area of 700 cm2 and an electrode gap of 5 cm
(symmetric small) and the other with 1150 cm2 and 8 cm gap (symmetric large), where the gas inlet was opposite the substrates and
the pumping system – the mass deposition rate was measured over a broad range of energy input, W/F, into the plasma zone.
Both sets of experiments reveal straight lines in the Arrhenius-type plot (Figure 12) that are in parallel. The (negative) slope yields
the apparent activation energy Ea giving 51 2 J cm–3 (12 eV per HMDSO molecule). This rather high activation barrier is sufficient
to meet the requirements for the Si–O bond opening in HMDSO, while a reaction pathway via, e.g., methyl group abstraction, would
have resulted in a lower activation energy. At low energy input the slope might indeed be less steep, which points toward the
contribution of larger fragments within this low energy range. Close to the reaction barrier, however, a predominant reaction
pathway becomes apparent (Arrhenius regime), while at higher energy input more fragmentation and abstraction of –CH3 occurs.
The situation is more complex using asymmetric plasma reactors which is, however, the usual situation. The (fade) luminous
discharge of a pure HMDSO plasma is concentrated in front of the smaller, driven RF electrode and expands with increasing power
input. This situation can be described by the geometrical factor as defined in eqn [8]. Using proper factors within different
asymmetric set-ups with 150 cm2 (small), 180 cm2 (medium) and 1500 cm2 (large) electrode area, the Arrhenius regime could be
obtained with parallel slopes, again showing deviations at low energy input (Figure 13) (49).
Although the obtained activation energies are the same for the different used plasma reactors, the net deposition rates
(normalized to the monomer gas flow) are found to be different. The prefactor in the Arrhenius-like equation, which gives the
maximum conversion of monomer within the Arrhenius regime, is thus also influenced by the actual conditions in the plasma
Figure 12 Arrhenius-type plot of the deposited mass from HMDSO discharges using two symmetric RF reactors of different size.
Figure 13 Arrhenius-type plot of the deposited mass from HMDSO discharges using three asymmetric RF reactors of different size.
www.iran-mavad.com
reactor: (1) the reaction rate for the electron impact dissociation is influenced by the eedf (giving the decomposition rate of the
monomer), (2) the transport of the film-forming species to the substrate depends on the gas flow direction (drift) and the diffusion
length as well as the residence time of the species at the surface, and (3) the incorporation of the species into film growth is affected
by nucleation sites and deposition vs. ablation conditions. For the latter, the energy density during film growth is thus decisive (see
below). The eedf, on the other side, depends strongly on the reactor geometry (wall to plasma volume ratio) and the plasma source.
The influence of gas drift in the reactor can be examined using the same plasma reactor with two different set-ups for the
pumping unit. While gas inlet opposite to the pumping system (and to the substrate) enables a vertical flow through the plasma
zone toward the substrate (directed drift), pumping from the same side as gas inlet forces a deflected gas flow and omits the directed
drift towards the substrate leaving only diffusion. Figure 14 reveals that the net deposition rate is reduced in the set-up with the
deflected gas flow.
Note that due to these additional factors influencing the net deposition rate (beside undefined expansion of plasma volume,
secondary discharges, or undefined power losses), the Arrhenius behavior might not be observed by measuring deposition rates
when inappropriate conditions were chosen. The Arrhenius-like approach, however, can always be used to examine the specific
conditions in a given plasma reactor set-up (50). It was found that an energy input into the plasma slightly higher than the apparent
activation energy is required in order to obtain hydrophobic, polydimethylsiloxane-like (PDMS) coatings (51).
In order to obtain permanent hydrophobic coatings using HMDSO discharges, surface processes should be taken into account as
well. The energy density during film growth (energy flux per deposition rate) enables nucleation sides and cross-linking. It can be
seen from Figure 15 that the film density increases with energy density showing an almost logarithmic dependence, while the
carbon content of the SiOC:H films is reduced. It can be assumed that increasing abstraction of hydrogen and methyl groups (–CH3)
occurs accompanied by cross-linking via –CH2– groups (51). Contact angles against water are thus also reduced (Figure 15).
Chemical composition, film density, and contact angles indicate that the coatings deposited at low energy density
(w100 eV nm3) indeed resemble PDMS. Well cross-linked and stable plasma polymers, on the other hand, are typically obtained
at energy densities higher than 200 eV nm–3. Note that the increase in film density might be stronger correlated to the momentum
Figure 14 Arrhenius-type plot of the deposited mass from HMDSO discharges using the same symmetric RF reactor with opposite location of the
connection to the pumping system.
Figure 15 Increase of film density of plasma polymer depositions obtained from different HMDSO discharges and related carbon content depending on
the energy density during film growth. The residual hydrocarbon content also defines the advancing and receding contact angles (as given in parentheses).
www.iran-mavad.com
transferred per deposited atom (according to eqn [12]), while a threshold energy seems to be required to avoid a void-containing
plasma polymer structure.
4.09.1.2.2 O2/HMDSO
Oxygen is typically added to HMDSO discharges in order to obtain inorganic SiOx coatings. Following chemical thermodynamics,
a O2/HMDSO ratio of 12 is required to fulfill the chemical reaction (CH3)3–Si–O–Si–(CH3)3 þ 12O2 / 2SiO2 þ 6CO2 þ 9H2O.
Using optical emission spectroscopy (OES), it can indeed be shown that atomic oxygen can be detected only in the gas phase for
O2/HMDSO ratios > 12, i.e., when more oxygen is available as can be consumed (52).
Using the quasi-Arrhenius approach, it was found that the obtained activation energies increase linearly with O2/HMDSO ratio
(53). As depicted in Figure 9 and given by eqn [9], the factor rgas giving the contribution of the reactive gas in the plasma poly-
merization process can thus be determined from the different Arrhenius regimes for each O2/HMDSO ratio (Figure 16).
Regarding the total gas flow rate F (¼ Fm þ rgasFc) for the reaction parameter W/F, the same activation energy is obtained as for
pure HMDSO discharges with rgas ¼ 0.6. Hence, it can well be assumed that the same plasma chemical reaction pathway (via
(CH3)3–Si–O radicals by opening of one Si–O bond) initiates the plasma polymerization process, even if oxygen is added to the
discharge. The observed rather high value of 0.6 for rgas indicates that more than one-third of the energy input is consumed for the
excitation (dissociation) of oxygen in order to abstract hydrocarbons – most likely at the growing film surface – and to form an
inorganic Si–O network (20,53). Increasing deposition rates (per monomer flow rate) also point toward a direct participation of
oxygen during film growth.
By varying the O2/HMDSO ratio, the residual hydrocarbon content in the growing film can be influenced, which also determines
the wetting properties (from hydrophobic to hydrophilic) (53). The related film density (and thus, e.g., barrier properties), however,
is also governed by the energy dissipated within the growing film as induced by ion bombardment (see Figure 3). Increasing the
O2/HMDSO ratio by reducing the monomer flow rate at maintained energy flux (fixed power input and pressure), an increase in
film density can be recognized (Figure 17). This increase is caused by ion-induced etching of hydrocarbons, which might both be
due to increasing atomic oxygen content and be due to the reduced deposition rate (and thus increased 3surf).
In this way, dense SiOx films with good barrier properties could be achieved for high O2/HMDSO ratios (54). Too high energy
densities (or more precisely, momentum densities), however, induce high internal stresses that cause cracks and reduced barrier
properties (53). In order to examine influences by plasma chemistry, care has to be taken in order to compare experiments at similar
energetic conditions during film growth.
4.09.1.2.3 Hydrocarbons
Hydrocarbon monomers can be used to deposit a broad range of plasma polymer films from polymerlike to hard coatings. Due to
their branched, cross-linked structure with substantial hydrogen incorporation, they are commonly named amorphous hydro-
carbon films (a-C:H). The activation of hydrocarbon monomers in the plasma can be considered as a rather unselective process due
to the relative similarity of most bond strengths in the starting molecules and intermediates with respect to the electron energy
distribution function. Plasma polymer deposition can thus be obtained with any hydrocarbon precursor, even those that cannot be
polymerized in the conventional way, such as, e.g., saturated hydrocarbons. However, stabilizing processes such as plasma polymer
growth might be rather selective due to different activation barriers, reactivities, and sticking coefficients of the intermediates (55).
The formation of neutral species in the plasma was found to depend uniquely on the energy dissipated per source gas molecule, W/F
(56). The overall reaction pathway is generally difficult to clarify.
Figure 16 Arrhenius-type plot of the deposited mass from O2/HMDSO discharges with different gas ratios.
www.iran-mavad.com
Figure 17 Increase of film density of plasma polymer depositions from O2/HMDSO discharges by reducing the monomer flow rate (at fixed O2 flow rate).
A reduction in deposition rate is observed.
Measuring the deposited mass is a method that inherently takes only the contribution of film-forming species into account.
Evaluation of the deposition rates in methane (CH4) discharges, for example, revealed a rather low activation energy Ea of around
5 eV, pointing toward abstraction of hydrogen (with a bond strength of 4.5 eV) as predominant step in the chemical reaction
pathway (Figure 18) (14,24). In order to incorporate the resulting CH3 radicals into film growth, the surface binding energy of
a-C:H films of 7.4 eV has yet to be overcome (57). This energy results from the energetic particle bombardment during film growth
according to eqn [4]. Nucleation sites are thus generated by hydrogen abstraction and the films become densified by cross-linking.
The importance of atomic hydrogen for the abstraction of hydrogen at the film surface was described by von Keudell yielding
polymer-like a-C:H films (with an H/C ratio around 1) (58). This mechanism can be enhanced by ion bombardment and was found
to hold over a broad range of energy input (9,12,14). Only at very high energy densities, the deposition mechanism might change as
also indicated by a drop in deposition range. Temperature effects, chemical sputtering, and different film-forming species and
growth zones are thus discussed (9,59,60).
Using higher hydrocarbons (CnHm, with n 2), the deposition rate is found to be increased compared to methane. The
comparison of the deposition rates with ethane (C2H6) using an Arrhenius-type plot reveals an interesting behavior. For low energy
input, the apparent activation energy is close to the one as obtained for methane too. A similar chemical reaction pathway might be
assumed involving the abstraction of hydrogen (with a bond strength of 4.3 eV). The potential film-forming radical, C2H5, then
yields roughly double the deposition rate as CH3 for a comparable sticking coefficient as obtained in Figure 18. In this energy
regime, mainly polymer-like coatings are formed. Increasing energy input, on the other hand, shows strongly increasing deposition
rates with a noticeable higher apparent activation energy (20 eV per C2H6 molecule). This rather high activation energy indicates
that multistep reactions might be involved, yielding highly reactive film-forming species.
The unsaturated hydrocarbon ethylene, C2H4, shows a further increased deposition rate at a lower activation energy of 12 eV per
molecule in the plasma. Note that acetylene, C2H2, was found to exhibit an apparent activation energy of 9 eV (23). Since electron
Figure 18 Arrhenius-type plot of the deposited mass from different hydrocarbon discharges.
www.iran-mavad.com
Figure 19 Increase of film density of plasma polymer depositions obtained from different methane and ethylene discharges depending on the energy
density. A steady increase is obtained up to about 2000 eV nm–3 (w25 eV per deposited C atom).
impact processes yielding breaking of the ChC bond require more than 17 eV, it can be assumed that H abstraction (bond energy
5.5 eV) and opening of the triple bond (bond energy 1.9 eV) contribute to the chemical reaction pathway instead (61). Likewise,
opening of the double bond in C2H4 (scission of the p-bond with an energy of 2.6 eV) beside H abstraction (bond energy 4.8 eV)
might yield the film-forming species in ethylene discharges. Hence, a different chemical reaction pathway can be assumed for
saturated and unsaturated hydrocarbons.
Film density was found to steadily increase with energy density during film growth, 3surf, over a broad parameter range
(Figure 19). The observed saturation in film density (at around 1.7–1.8 g cm3 for methane discharges) is caused by relaxation
processes at increasing energy density. Additional energy is converted to thermal energy allowing the bonding to relax, e.g., toward
more graphite-like structures in a-C:H films (62).
Plotting the film density of the a-C:H coatings against the momentum density psurf (instead of 3surf), a linear relationship
becomes apparent at medium values of the momentum transfer (Figure 20). The densification is thus indeed momentum rather
than energy derived within this range.
The reduced film density at the low momentum side (as examined for C2H4 discharges) might be indicative of a threshold energy
for the momentum-induced densification. Calculating the energy per deposited C atom, the observed threshold is approximately
5 eV (which might be related to the abstraction of hydrogen atoms). At energies below this threshold a void structure of the
deposited plasma polymer can be assumed similar as observed in metal growth under low energy ion bombardment (63). Higher
hydrogen contents in the a-C:H films also point into this direction.
At the high momentum side, on the other hand, a threshold energy can be assumed for the relaxation of the compressed network
by releasing atoms from their metastable positions to the surface of the film (32). Film densities become thus saturated at an energy
level of around 40 eV per deposited C atom. Highest hardness values for a-C:H coatings were indeed obtained around 50 eV per C
Figure 20 Increase of film density of plasma polymer depositions as obtained in Figure 19 depending on the momentum density. A linear increase is
obtained in a central range.
www.iran-mavad.com
atom, slightly depending on the type of hydrocarbon monomer (CH4 vs. C2H2) (12). The actual flux of atomic hydrogen to the
substrates might be decisive, limiting the film density and hardness that can be achieved (64). Thus, methane-derived a-C:H films
exhibit a lower film density and hardness than ethylene or acetylene discharges (61).
Generally, hydrogen plays an important role in plasma polymerization, since an increasing hydrogen fraction in the plasma
yields more H atoms, lower electron densities, and higher electron temperatures (H-rich plasmas). Net deposition rates are thus also
influenced besides hydrogen incorporation. Lowered deposition rates were found to yield higher hardness values by increasing the
H2/C2H2 ratio (12).
The deposition of hydrocarbon plasma polymer films can thus be controlled with respect to the energy and momentum density
within the plasma process in order to adjust their properties, e.g., for wettability, friction reduction, permeability, and hardness.
4.09.1.2.4 Hydrocarbons with Reactive Gases

While pure hydrocarbon discharges are mainly examined for the deposition of hard coatings, the admixture of reactive gases such as
H2O, CO2, N2, and NH3 enables the incorporation of (oxygen or nitrogen) functional groups during film growth. The different
reactive gases show a different impact on the chemical reaction pathway, as was already extensively discussed by Yasuda (2). Hence,
different curves are obtained in an Arrhenius-type plot (Figure 21).
In water vapor–containing discharges, OH radicals are formed, which act as quenching species. Thus dangling bonds (trapped
free radical sites) within the growing hydrocarbon films are suppressed and some oxygen-functional groups are incorporated. The
observed deposition rates are found to be unaffected as long as the H2O admixture and the energy input remain moderate (65).
Most likely, the occurrence of atomic oxygen in the plasma at enhanced energy input and/or H2O/hydrocarbon ratio changes the
growth mechanism by inducing etching effects as observed by strongly reduced deposition rates. Addition of carbon dioxide results
in reduced deposition rates already at lower admixtures and energy input by consumption of the formed atomic oxygen (scavenger
effects) (60). The incorporated oxygen content scales with the gas ratio. With increasing energy input, a further drop in deposition
rate is obtained which can be attributed to chemical etching (13). Reduced oxygen incorporation is observed with this change in
growth mechanism. A similar behavior is found by addition of nitrogen reactive gases, which also reduce the deposition rates
compared to pure hydrocarbon discharges and yield incorporation of nitrogen-functional groups scaling with the nitrogen/
hydrocarbon ratio (66). The observed drop in deposition rate with increasing energy density, however, was identified to be caused
by ion-induced etching starting at around 1000 eV nm–3 (w15 eV per C atom) (13,67). In this regime, hard carbonitride coatings
can be obtained. By varying the corresponding gas ratios of reactive gas to hydrocarbon monomer (see eqn [9]), a gas reactivity
factor rgas was determined to be <0.05 for H2O, w0.15 for CO2, w0.3 for N2, and w0.5 for NH3. This plasma sensitivity series
shows the reactivity of the reactive gases within a hydrocarbon plasma environment. Thus the quasi-Arrhenius approach offers
deeper insights into plasma polymerization processes compared to, e.g., Yasuda’s in-out rule of thumb that nitrogen has the higher
reactivity to be incorporated than oxygen (7). Permanent functional plasma polymers can be achieved within the Arrhenius regime
at energy densities not yet causing a drop in deposition rate. The control of both the energy input into the gas phase and the energy
density during film growth can thus be used to optimize the functionality and stability of oxygen and nitrogen functional plasma
polymers. Again, film densities were found to scale with energy respectively momentum density (13).
4.09.1.2.5 Acrylic Acid

The monomer acrylic acid, CH2]CHCOOH, is extensively investigated for the deposition of carboxylic group–containing plasma
polymers. The plasma chemical reaction pathway, however, is still under discussion (26,68,69). Plasma polymer deposition can
already be obtained at low energy density, since plasma-induced chain-growth polymerization of acrylic acid is initiated with a low
activation energy of w1 eV by opening of the C]C double bond (scission of the p-bond) (10). High retention of carboxylic groups
within poorly cross-linked plasma polymers results. An important aspect, however, is the stability of such plasma polymer films in
aqueous environments, since they are, for example, examined as biointerfaces. Therefore, enhanced power input into the plasma or
Figure 21 Representative Arrhenius-type plot of the deposited mass from different admixtures of reactive gases to an ethylene discharge.
www.iran-mavad.com
Figure 22 Arrhenius-type plot of the deposited mass from a pure acrylic acid discharge and with admixture of hydrogen (with a ratio of 3:1). Two
regimes with a different activation barrier become apparent.
admixture of hydrogen is investigated (70,71). A thorough study of both parameters using a well-defined symmetric plasma reactor
(RF excitation at a pressure of 10 Pa) reveals different activation energies with increasing energy input into the plasma. Moreover,
a reduction in deposition rate can be observed by admixture of H2 as shown in Figure 22.
While the smaller activation energy (Ea(1) of about 7 eV per molecule) at lower energy input in the pure acrylic acid discharge
might be indicative of hydrogen abstraction beside activation via the vinyl group, the higher activation energy (Ea(2) of about 23 eV
per molecule) involves stronger fragmentation. Indeed, this energy regime was found to be comparable to a CO2/C2H4 discharge,
which introduces the same chemical elements as acrylic acid into the plasma (68). Admixture of hydrogen maintains these different
activation regimes, however, at enhanced energy input. A higher activation energy is thus required. For this reason as well as due to
the reduced deposition rates in the H-rich plasma, the energy density during film growth is increased, yielding ion-induced effects
that blur the Arrhenius behavior at enhanced energies. Similar effects could be observed with different reactor geometries (26,68).
Stable, carboxylic group–containing plasma polymers can thus be obtained with increasing cross-linking, which in return reduces
the amount of carboxylic functionalities.
Monomers comprising a vinyl group, H2C]CHR, such as acrylic acid or an allyl group, H2C]CH–CH2R, such as allylamine and
allyl alcohol, require lower activation energies and show higher deposition rates as was also demonstrated for C2H4 compared to
C2H6. Thus, they are extensively examined for the deposition of functional plasma polymers (2,10,72). For a comparison with the
saturated counterpart as for example propionic acid, CH3CH2COOH, care should be taken, since the deposition rate was found to
be about 7–8 times lower than for acrylic acid (73). With comparable energy flux, the energy dissipated during film growth is thus
also 7–8 times higher for propionic acid according to eqn [4]. The plasma chemical reaction pathway is thus strongly superimposed
by surface processes due to the corresponding high energy density at film growth. Nevertheless, a different plasma chemical reaction
pathway might be expected with respect to the results of the comparison of C2H4 and C2H6.
4.09.1.2.6 Further Monomers

The applicability of the quasi-Arrhenius approach for plasma polymerization processes has been shown for many further mono-
mers comprising the elements C, Si, H, O, N, and B (23). Although not yet thoroughly investigated, different silico-organic
molecules as well as silanes (including SiH4, TMS, and vinyltrimethoxysilane), higher hydrocarbons, allylamine, and allyl alcohol
among others are expected to follow the quasi-Arrhenius approach. Fluorocarbons, however, show a different behavior, since the
surface itself can be the source for film-forming species such as CF2 (74). This effect becomes stronger with increasing energy density
during film growth. In any case, surface processes might be dominant for a broad range of plasma conditions, thus blurring the
plasma chemical reaction pathway.
4.09.1.3 Conclusion
Plasma polymer deposition is investigated on the basis of the energy density in the active plasma zone for the generation of film-
forming species and the energy density at the surface required for chemisorption of such species.
A reaction cycle for plasma polymerization can be drawn as shown in Figure 23. The gases introduced in the reactor become to
a large part activated in the active plasma zone. A smaller part of these activated species contribute to film growth after traveling
toward a surface due to their different sticking coefficients (cycle 1 for film-forming species). Also nonpolymerizing species can
contribute to film growth either by incorporation (of, e.g., oxygen or nitrogen) or etching processes (cycle 2 for reactive species).
Surface reaction processes might also be a source of film-forming species (cycle 3 as, e.g., CF2 in fluorocarbon deposition) yielding
rather complex reaction pathways that strongly depend on the energy density in the gas phase as well as at the surface.
www.iran-mavad.com
Figure 23 Schematic cycle of the particle flow in a plasma polymer deposition process. Cycle 1 denotes deposition from film-forming species
arising from the active plasma zone, cycle 2 shows the contribution of reactive species to deposition, and cycle 3 shows the surface as source of film-
forming species.
The conversion of the source gases into film-forming species (decomposition rate) depends on rate constants and the eedf, while
activation barriers for specific reactions determine the chemical reaction pathway to a large extent. The incorporation of potentially
film-forming species into the plasma polymer depends, in addition, on the transport to the substrate, the residence time at the
surface, available nucleation sites, and ablation processes. The plasma chemical reaction pathway yielding film-forming species can
thus more or less be superimposed by surface processes. For surface processes, energy thresholds were obtained for cross-linking,
ion-induced etching, and relaxation processes, while the densification of the plasma polymer depositions scales with the
momentum transferred to the deposited atoms via collision cascades induced by ion bombardment (and energetic neutral species).
Considering those relevant factors, plasma polymer deposition enables surface functionalization for a broad range of appli-
cations, allowing highly economic and ecologically friendly industrial processes.
4.09.2 Coatings on Polymers
This section mainly considers general aspects of coatings on polymers. While polymers possess a broad range of excellent bulk
characteristics (such as light weight, flexibility, machinability, and costs), some missing properties (such as reactive functional
groups, stability, barrier effect, and electrical conductivity) might be added by plasma coatings. For this purpose, understanding of
both the behavior of polymers in plasmas and plasma polymer formation is required (75). Actually, all polymers (even the same
polymer type from different manufacturers) behave differently when exposed to a plasma (due to different additives, etc.). Plasma
polymer depositions, in particular hard coatings, exhibit different mechanical properties than conventional polymers due to their
branched, cross-linked structure. Therefore, the plasma deposition conditions should be individually adapted for each polymer
substrate, especially for the initial phase of film growth (76,77).
4.09.2.1 General Aspects for the Coating on Polymers

Polymers are distinguished by their long polymer chains, i.e., their molar masses, which constitute their mechanical properties.
Since more chain ends can be observed toward the surface of a polymer (Figure 24), the glass transition temperature is found to be
reduced within the near-surface layer. For any (cross-linked) coating on an unmodified polymer surface, a weaker boundary layer is
thus obtained between the coating and the bulk polymer. Moreover, polymer surfaces typically lack in functional groups and show
contaminations at their surface (e.g., by adsorption processes in ambient air). Therefore, plasma treatment is an attractive method to
modify a polymer surface and to prepare it for a subsequent coating process.
Figure 24 Schematic drawing of a polymer surface showing different phases, crystallinity, fillers, additives, chain ends toward the surface, roughness,
adsorption layers, contamination, surface charging, and wetting.
www.iran-mavad.com
The plasma interaction with a polymer surface transfers sufficient energy to open C–C, C–H, C–O, C–N, etc. bonds that are
present in the polymer. Since those bonds show similar dissociation energies, the plasma interaction is highly unselective. Following
the inevitable chain scission (producing radical sites), recombination can occur, leading to cross-linking, which is the desired process
during plasma treatment in order to stabilize the polymer surface. Mainly gas composition and energetic conditions during plasma
interaction (such as ion-assisted radical-based interactions) affect the chain scission/cross-linking mechanism. Nevertheless, it
depends, first of all, on the type of polymer itself, which can be described by the polymer’s susceptibility to plasma etching. Higher
rates of weight loss (etch rate) in an inert gas plasma point toward predominant chain scission reactions. Figure 25 gives a qualitative
order of plasma susceptibility for different polymers using inert gas plasmas (Ar or He) at moderate energetic conditions (77,78).
The huge differences in plasma susceptibility indicate that not only ion bombardment (energy flux) but also different factors
contribute to the plasma etching of polymers: (1) vacuum ultraviolet radiation (VUV, l < 200 nm) induces chain scission with
a higher penetration depth (5–50 nm) than ions (1–3 nm), (2) radical-induced deterioration or hindrance of this mechanism by
shielding through neighboring groups, (3) rate of radical recombination (cross-linking), and (4) abstraction of reactive elements,
e.g., oxygen species from O-containing polymers, which enhances etching. All these factors strongly depend on the type of polymer.
Both high and low plasma susceptibility hinders the plasma coating of the respective polymer. While the degradation of pol-
yoxymethylene (POM) and polymethyl methacrylate (PMMA) is enhanced by VUV radiation (photo degradation), fluorocarbon
polymers such as polytetrafluoroethylene show a high stability of the generated polymer free radicals (through fluorine abstraction)
impeding further etching, but also limiting chemical bond formation. Silicone polymers, on the other hand, form strong, yet brittle
inorganic oxides close to the surface, which strongly reduces the etching process. Plasma coatings on silicones are thus subject to
a mechanical mismatch. Low etching rates also imply a high thermal load in the near-surface layer, since the dissipated energy is
high according to eqn [4] regarding the etching rate Retch (instead of the deposition rate) (79).
Cross-linking reactions are supported in polymers containing at least one hydrogen atom bonded to adjoining carbon atoms
such as polyethylene (PE), polypropylene (PP), polystyrene (PS), and polyamide. Among the polymers which seem to be suitable
for plasma coatings, there are still differences due to the ability to form chemical bonds, which is, e.g., higher in polyethylene
terephthalate (PET) and polycarbonate (PC) compared to the polyolefins (PE and PP) (80). Therefore, the plasma pretreatment
should be optimized for each polymer type. Further polymers that are suitable for plasma coating comprise technical polymers such
as polyether ether ketone, cyclic olefin copolymer, and highly crystalline polymers (e.g., ultra-high-molecular-weight polyethylene,
aramids, polyphenylene benzobisoxazole, liquid crystal polymers).
Predominant deterioration of a polymer when exposed to a plasma leads to the formation of low molecular weight fragments
(oligomers), causing a weak boundary layer. Coatings are thus subject to film failure. Short plasma exposure times (around 1 s)
might avoid degradation effects and are already sufficient for the generation of free radicals at the surface (via H abstraction), while
cross-linking requires longer treatment times (several seconds to minutes) (75). PC substrates, for example, showed good adhesion
of SiOx coatings after 5 min N2 plasma pretreatment (20 Pa, 0.3 W cm–2), while PP sheets were found to be more cross-linked after
2 min He plasma (20 Pa, 0.3 W cm–2), which is mainly an effect due to VUV radiation in N2 or He plasmas (77). The occurrence of
low-molecular weight oxidized materials, however, is also a common phenomenon by treating polyolefins or polyesters in an
oxygen-containing plasma indicating overtreatment (81,82). The plasma exposure time and the power input, i.e., the energy input
that also determines the VUV radiation, should thus be adapted to the polymer.
Beside the modification of the polymer chains by plasma interaction, metastable radicals are generated which are (partly)
quenched by functional groups (either during the plasma interaction by using reactive gases or by postplasma reactions with air).
The polymer surface thus becomes more polar, supporting chemical bond formation for coatings. Even if different gases and
different plasma conditions are used, the polymer functionalization is always ‘polysort’; i.e., different polar groups will be generated
Figure 25 Relative susceptibility to plasma etching of different polymers in inert gas discharges.
www.iran-mavad.com
(75). Some tendencies can be observed showing; e.g., more –OH groups for water vapor (H2O) plasmas compared to O2 and more
–NHx groups for ammonia (NH3) compared to N2 plasmas.
4.09.2.1.1 Influence of Polymers on the Plasma Conditions

Polymers introduced into the gas discharge can influence the discharge characteristics in two ways: (1) outgassing or decomposition
of the polymer sample might yield a change in gas composition, and (2) secondary electron emission from the polymer sample
might yield a change in the composition of species present in the discharge (e.g., ion and electron densities).
In many cases, however, the gas-phase processes are not much affected when polymer samples are introduced. Hence, the
formation of film-forming species (plasma-chemical reaction pathway) is not altered. Measurement of excitation voltage, discharge
current, and electron densities should thus show only minor changes. The actual surface conditions, however, should be taken into
account. As long as the coatings are not yet closed during the initial stages of film growth, the substrate actually influences the film
growth (1). Hence, the properties of nanoscaled coatings are affected, as well as adhesion and roughness (83). The initial deposition
rate might be reduced due to a lower sticking coefficient on the specific substrate. This in turn increases the energy density during
deposition favoring etching processes. Hence, a mixed zone of (modified) substrate surface and growing film results, which is called
interphase.
When the deposition is examined on metal surfaces (i.e., also small, thin substrates on a metal electrode), the secondary electron
emission is not much affected by the actual surface state of the metal (cleaned vs. dirty) for typical plasma polymerization
conditions (84). Dielectrics such as polymers, however, have much higher secondary electron emission yields, which acts in
reducing the energy flux to the substrate, since the sheath voltage is reduced (as, e.g., indicated by a reduced bias voltage) (85). This
effect should thus be considered when the polymer samples cover a substantial part of the electrode (Figure 26). Also the height of
the polymer samples influences the voltage drop across the plasma sheath due to a reduced electric field by a combination of bound
and induced charges, when placed at the electrode (86). Different deposition rates along the samples and edge effects were observed
in this case (87). An insulating, growing plasma coating on polymers can also lead to charging effects that further reduce the sheath
voltage (88). Although the energy density in the gas phase might still be unchanged, yielding the same flux of film-forming species
incident on the polymer substrate, the energy density during film growth can be strongly affected. Thus, weaker plasma polymer
films result due to reduced cross-linking, densification, and etching. A noticeably lowered energy flux might even reduce the
observed deposition rates in order to overcome the surface binding energy and to create sufficient nucleation sites.
Some of these obstacles might be avoided by coating polymers at floating potential, where the energy flux to the substrate is well
defined (14). Here, the transport of film-forming species to the sample surface is crucial in order to obtain conformal coatings –
similar to thermal CVD processes. In addition, edge effects have to be considered, since the higher electric field intensity close to
the edges of a substrate at floating potential can enhance gas-phase reactions and increases ion bombardment (89). Polymeric fabrics
and fibers, for example, can well be plasma coated by continuously transferring them directly through the active plasma zone.
4.09.2.1.2 Aging of Plasma Polymer Depositions

Polymers exposed to a plasma are modified within a thin surface layer, typically below 100 nm. While their bulk properties can thus
be maintained, the surface properties are altered. Energy dissipated in the near-surface layer generates radicals, which are partly
quenched by functional (polar) groups. The accompanying hydrophilization, however, is prone to aging effects. Diffusion,
migration, reorganization, chemical reaction, and adsorption processes contribute to the so-called hydrophobic recovery, since the
polymer structure remains rather mobile (76). Most of all, fluorocarbon polymers show a strong tendency of their terminal CFx
groups to reorientate back to the surface, since the near-surface layer cannot be efficiently cross-linked in the plasma environment.
In order to improve the permanency of functional polymer surfaces, plasma polymer depositions are an attractive method. Due
to small film-forming units, the plasma polymers are highly branched and cross-linked. Polar groups chemically bond within the
plasma polymer network have a much lower degree of freedom to reorientate compared to the direct functionalization of the
polymer surface (77). Nevertheless, plasma-produced metastable radicals are trapped during the growth of the plasma polymer film,
causing postplasma reactions (2). Therefore, plasma polymer films also undergo aging effects, which are related to the number of
trapped free radicals and depend thus also on the storage conditions (76). The number of trapped radicals during plasma polymer
film growth of, e.g., hydrocarbons, can be reduced by adding reactive gases such as O2, CO2, H2O, N2O, N2, and NH3 (plasma
copolymerization), which act in quenching of the radicals (2). By this means, functional groups are formed within the plasma
polymer network. Using suitable gaseous mixtures and plasma conditions, aging can be strongly reduced (Figure 27).
Most of all, corona (atmospheric pressure plasma) treatments in air reveal strong aging effects, which can be reduced by using,
e.g., O2 (5–20%)/Ar mixtures at low pressure. The observed short-term aging effects are mainly due to diffusion of surface oligomers
Figure 26 Polymer substrates of different size placed at different positions within a plasma reactor. (a) Small, thin sample placed at a metal surface
(e.g., electrode), (b) large, thin sample covering the metal surface, (c) sample height influencing the plasma, and (d) sample at floating conditions.
www.iran-mavad.com
Figure 27 Aging of different plasma treatments on PET substrates followed over one month. Plasma polymer depositions show reduced aging effects
compared to low pressure plasma or air corona activation.
and reorientation of polar groups. Cross-linked a-C:H:O coatings show moderate wettability, but stable, functional surfaces (90).
Such coatings are thus used, e.g., for blood filtration. SiOx layers as deposited, e.g., by O2/HMDSO discharges, are also known as
stable, hydrophilic surfaces (comprising silanol groups, –Si–OH, at the surface). Surface topography, i.e., the specific surface area,
can also contribute to the permanency of functionalized surfaces. A nanoporous plasma polymer with accessible amine-functional
groups was found to show noticeably reduced hydrophobic recovery (66,91). Stable plasma polymers showing a void structure
through terminal functional group incorporation can be obtained under moderate energy density in the gas phase and low-to-
moderate energy density during film growth. Nevertheless, it should be regarded that the functional –NHx group itself is prone to
aging, which, however, results in oxidation (92). Such plasma polymers can be stabilized by subsequent wet-chemical attachment of
molecules (e.g., for bio applications).
In order to combine functionality with sufficient stability (e.g., for applications in aqueous environments), moderate energy
input (mainly at the surface) is required for the plasma deposition process (see Figure 4). Similar to the direct activation of the
polymer surfaces, the functional plasma polymer layers comprise different, ‘poly-sort’ functional groups. More specific function-
alizations (monofunctional surfaces) can be obtained at lower energy density, most of all by using pulsed-plasma conditions (93).
Numerous examples in the literature deals with this approach, which is not further discussed here, since sufficient stability
(permanency) for potential applications remains an issue due to the formation of a loose, swellable plasma polymer network at low
energy input conditions (68,70,72). Nevertheless, such coatings might be interesting as, e.g., chemical or bio sensors, since
subsequent micro/nanostructuring is enabled by local cross-linking (94).
4.09.2.1.3 Adhesion of Plasma Polymer Depositions

By applying coatings to polymer substrates in order to tailor their surfaces for particular applications, first of all, adhesion of the film
has to be regarded. Adhesion of coatings is a complex phenomenon depending on surface energy, chemical bonds, surface
roughness, cleanness, and mechanical strength. Therefore, excellent adhesion should be expected from functionalized, roughened,
cleaned, and cross-linked polymer surfaces, enabling covalent bonding of the coating. If a coating finally adheres within certain
conditions (and is thus suitable for its application) depends on the adhesion strength (bonding strength), which is the amount of
energy required to separate the coating from the substrate. Adhesion strength can thus be considered as the maximum force per
width of the sample that the film/substrate system can withstand without showing film failure.
The actual force is made up of the internal (intrinsic) film stress sint and the external load sext on the system, which mainly acts
on the interfacial bonds between the coating and the polymer substrate. Internal stress arises during film growth due to the energy
density 3surf, which induces mechanical stresses (mismatch of the elastic modulus of film and substrate) and thermal stresses
(mismatch in the coefficient of thermal expansion) (95). For laterally (in the xy plane) homogeneous plasma coatings, the internal
force acting on the interface is then given by
Z
Fint ¼ sint ðzÞw dz [13]
dfilm
with the film thickness dfilm (in z-direction) and the width of the sample w. For a uniform internal stress, which is the case for
homogeneous plasma polymer deposition conditions, the force is thus proportional to the film thickness. A plasma-coated polymer
shows sufficient adhesion if the adhesion strength is higher than the combined (internal and external) forces acting on the interface:
Fext ðsext Þ
Sadh > sint dfilm þ [14]
w
www.iran-mavad.com
Especially for the deposition of hard coatings, high internal stresses can be observed due to the densification during film growth,
yielding compressive stresses (Figure 28). The resulting internal forces induce bending of the substrate and lead to film delami-
nation when exceeding the adhesion strength. For typical (good) adhesion strengths of 10 N cm–1, an intrinsic stress of 2.5 GPa
causes film delamination for a film thickness greater than 400 nm (without considering external stress).
The saturation of the compressive stress level (and even reduction at higher energy input) is caused by relaxation processes.
Additional energy is converted to thermal energy, allowing compressed atoms to relax by leaving their position. The measured stress
values can well be explained by the Davis model: volumetric strain induces stress proportional to the momentum density, whereas
relaxation is enabled above a certain excitation energy (32,62).
Since generally film failure should be avoided, the coating on polymers requires adapted plasma polymer films. Low energy
input yields swellable, water-soluble coatings, whereas high energy input results in hard coatings showing crack formation by elastic
and plastic deformation of the soft substrate (eggshell effect). Both effects might provoke film failure. In case the achievable
adhesion strengths of plasma polymer films on polymer substrates are not sufficient, the internal stress acting on the interface
should be reduced. This might be attained by thinner plasma coatings or through a gradient coating increasing the film density over
a broader interphase. The adhesion and permanence of hard SiOx coatings, for example, can be strongly increased by an adapted
gradient during the initial film growth (e.g., by steadily increasing the O2/HMDSO ratio) (50,77). The mechanical properties
(mismatch in elastic modulus) of the coating are thus adjusted to the substrate properties (Figure 29).
Most of all, the initial plasma polymerization conditions should be adapted to the type of polymer, since the energy dissipated
during initial etching/deposition conditions generates a highly mobile polymer surface due to chain scission. At suitable conditions,
a mixed interphase of polymer/film is generated, supporting adhesion. Inappropriate conditions, on the other hand, favor
Figure 28 Compressive stress in a-C:H plasma polymer films derived from a methane discharge as induced by the momentum density.
Figure 29 SiOx plasma coating deposited on polycarbonate substrate by making use of interphase formation (mixed phase of polymer and film) and
gradual adaption of the intrinsic stress (from 0.03 to 0.12 GPa) resulting in enhanced adhesion strength.
www.iran-mavad.com
formation of a weak boundary layer, which also affects the plasma polymer film growth, resulting in poor adhesion and pinhole
formation (96).
Plasma polymer depositions on practical substrates are actually nonpinhole free (as mentioned earlier), since the substrate
always induces defects. For example, the barrier effect of plasma coatings strongly depends on the surface roughness of the substrate
and is thus always limited (97). The number of pinholes, however, can be reduced by smooth, planarized surfaces. Polymer
substrates might also be planarized using atomic layer deposition or initiated CVD.
4.09.2.1.4 Adhesion of Metallizations

Although this chapter mainly focuses on plasma polymer depositions, the metallization of polymers is of high interest, e.g., for
electrical conductivity, optical reflectivity, and catalysis, and as antimicrobial treatment (mainly with Ag). Among other methods
such as electroless plating, physical vapor deposition is an industrially relevant process also fitting the scope of this chapter (98).
Evaporation of metals offers high deposition rates at low energy flux, thus at low energy density during film growth. High-quality
metal coatings with excellent adhesion, yet at lower deposition rates, can be realized by sputtering processes. Sufficient energy per
target atom has to be transferred from the plasma to release target atoms via collision cascades (Sigmund’s theory) (99). Hence, the
momentum of the incident particles, i.e., their masses, also contributes to the sputtering process. In order to obtain a high energy
flux incident on the target, low-pressure plasmas (around 1 Pa) with high plasma densities (e.g., by using a magnetron source) are
thus required. The sputtered atoms typically carry energies of 1–3 eV, which are low, referring to the typical surface binding energies
of metals (3–5 eV) (100). Adsorbed atoms at a polymer surface thus form islandlike growth, which proceeds at low energy density
during film growth. A needlelike structure containing voids is thus formed which exhibits low adhesion due to high stress at the
small interfacial area (Figure 30).
To improve adhesion, energetic conditions are required that enable surface diffusion of the adsorbed metal atoms (15). This
energy can be delivered by the plasma used for sputtering (with the substrate at floating potential) or by an additional bias voltage
applied to the substrate. Considering the energy flux incident on the polymer substrate and the deposition rate, which is connected
to the sputtering yield of the used metal, suitable energetic conditions can be obtained, yielding a dense film structure (101).
Momentum transfer in knock-on processes requiring an energy density higher than the surface binding energy was found to be
necessary for dense metal films (63). The films typically become continuous (closed) after about 20 nm of film growth (for
a polymer surface roughness of a few nanometers). Due to the microstructural evolution of metallizations during film growth, the
intrinsic stress is not homogeneous over the film thickness (102). Typically, there is a maximum in stress just when the film becomes
continuous. Therefore, further film growth enhances adhesion by lowered stress and increasing cohesion (for void-free
metallizations).
Beside these film deposition requirements, the type of polymer and its pretreatment are crucial in order to obtain a high peel
strength (adhesion and cohesion strength) of the metal layer. First, the polymer surface should be cleaned, and second, it should be
functionalized to obtain chemical bonding between the metal atoms and the polymer (103). Oxygen- or nitrogen-containing
plasma pretreatments were found to noticeably improve adhesion of metal layers (80). Cross-linking of the polymer surface and
intermixing during the initial phase of film growth (interphase formation via rivaling ablation and deposition processes) help in
adjusting the mechanical properties. The formation of a weak boundary layer as induced by an ‘overtreatment’ of the polymer
surface, however, should be avoided (Figure 30).
For the deposition of metals that form insufficient chemical bonds with the polymer substrate (mainly the noble metals), metallic
interlayers such as Cr or Ti that form both carbides and oxides can be used. Adhesion between the metal layers is then mainly based
on interdiffusion (100). Furthermore, plasma polymer interlayers might be applied, such as an adapted SiOx coating (104).
4.09.2.2 Industrial Aspects for the Coating on Polymers

Material selection is generally based on bulk properties, surface characteristics, machinability, availability, and costs. Ninety percent
of today’s global production of polymers comprises the polymers PE, PP, polyvinylchloride, PS, polyurethane, and PET. It is thus
Figure 30 Initial phase of metal film growth on a plasma-treated polymer surface. Low surface diffusion yields needlelike growth, while more energetic
deposition conditions result in dense metal films, but might also provoke the formation of a weak boundary layer.
www.iran-mavad.com
attractive to adapt certain properties of low-cost polymers by surface modification/coating to the needs of particular applications. In
a cost-driven market, the coated materials should be lower in price than special polymers that might also fulfill the requirements
(105). Hence, the coating process has to become fast and cost efficient. Plasma-based modifications of polymers should thus enable
a high throughput and reliable process conditions. The investment costs for the plasma apparatus appear to be secondary compared
to the running costs, which are discussed next.
4.09.2.2.1 Plasma Sources

The selection of an appropriate plasma source is important for the throughput and the reliability of the plasma process. For simple
processes such as plasma activation with nonpolymerizable gases, alternating current generators that do not require a matching
network are typically used. Examples are the hydrophilization of polymer parts in preparation for finishings, lacquering, lamina-
tion, and gluing (enhancement of adhesion strength). Huge pumping systems allow high throughput of the vacuum-based plasma
treatment.
For more sophisticated applications, plasma coatings are applied, which require a higher process control with respect to the
energy input into the gas phase and at the surface as extensively discussed within this chapter. For laboratory-scale reactors, mainly
capacitively coupled RF-driven plasma sources (CCP) that enable uniform discharges over large areas, high power consumption,
and excellent control over plasma polymerization processes are used. Transfer to industry is enabled based on the scaling parameters
power input per gas flow rate (3plasma) and energy flux per deposition rate (3surf) (67). Experts are, however, needed – not only for
process development but also for industrial processes. RF-driven plasma processes are limited by their deposition rate, which can be
further enhanced by an increase in plasma density. More electrons enable the formation of a higher number of film-forming species
and thus higher deposition rates (increased prefactor C in eqn [6]) as long as the particular film growth conditions at the surface are
considered.
Using RF inductively coupled plasma (ICP) at 13.56 MHz or microwave-excited plasma (MW) at 2.45 GHz, the deposition rates
can be strongly increased compared to RF CCP, since the electron densities are roughly one order of magnitude higher (e.g.,
1011 cm3 compared to 1010 cm3) (29,106–108). Therefore, the production rate of film-forming species can be enhanced within
the same range of energy input into the plasma, increasing the overall conversion of monomer into film growth. The ion flux is also
increased by electron density (see eqn [11]). The energy flux incident on the substrate can thus be enhanced, even when the sheath
voltage might be lowered. Note that CCP enables high potential differences by the bias voltage, while deposition at floating
potential yields comparable ion energies. If the deposition rate is increased in the same way as the energy flux, the energy density
during film growth can be maintained according to eqn [4]. For example, comparable functional plasma polymer depositions can
be performed both in ICP and in CCP discharges – with about six times higher deposition rates using ICP (27).
In order to control both the plasma density and the ion energies, an additional bias voltage can be applied by capacitive coupling
to the substrate, while the plasma is mainly excited by ICP or MW (29). Dual-frequency discharges (operated at, e.g., 2 and 27 MHz)
are investigated for the same purpose (109). Such plasma sources thus gain increasing interest for industrial applications.
4.09.2.2.2 Atmospheric Pressure Plasmas

Vacuum-based processing is a highly developed, well-controlled technology that, however, requires evacuation of the samples for
treatment. Aside from this, plasma processes working without vacuum equipment have gained increasing interest from both
academia and industry. Some reasons are obvious, such as sample handling, treatment of substrates that are not suitable for vacuum
processing, local treatment, and interaction during running processes. Therefore, atmospheric pressure plasmas are mainly inves-
tigated in two directions: plasma pens for local treatment (e.g., local activation of polymers for adhesion enhancement or for wound
treatment) and dielectric barrier discharges (DBDs) for two-dimensional treatment of thin substrates (e.g., foils and textiles) (110–
113). The reasons are the reduced plasma expansion (in the range of 0.1 mm) at atmospheric pressure, which limits the treatment of
3D-shaped parts. Further arguments often stated in many papers considering plasma polymer deposition at atmospheric pressure
include (1) reduced treatment times, (2) reduced costs, and (3) simplified technological transfer (112,114). This reasoning,
however, is not as obvious as it might seem.
As mentioned earlier, the quasi-Arrhenius approach is independent of pressure. Thus, the same concept of energy input into the
plasma can be used at low pressure (LP) and atmospheric pressure (AP). The deposition rate then depends on (1) the energy
invested per monomer with respect to the activation energy (exponential term in eqn [6]), (2) the conversion of the monomer
driven by the eedf (prefactor in eqn [6]), (3) the transport of film-forming species to the substrate, and (4) the degree of incor-
poration. While at LP mainly direct electron impact is responsible for the formation of the film-forming species, excitation of the
carrier gas transferring its energy to the monomer molecules is the predominant mechanism for AP. The latter works best in helium
due to its many metastable states and its high mobility, but also in argon or nitrogen. For the transport of the species to the
substrate, AP requires high carrier gas flow rates (several standard liter per minute (slm)) in order to overcome convection, to
maintain a rather low temperature, and to reduce impurities (from the environment). For comparable energy densities in the gas
phase, similar deposition rates can then be expected, resulting in at best comparable treatment times at AP. Note that the conversion
in LP almost reaches the maximum for optimized conditions as discussed for HMDSO discharges.
Sawada et al. mentioned already in 1995 that comparable deposition rates at LP and AP are observed for plasma polymer
depositions from siloxane discharges such as HMDSO and TEOS by using similar monomer flow rates (34). In order to compare the
deposition rates, both the monomer flow rate Fm and the deposition area Adep have to be considered according to eqns [6] and [7].
The deposition rate is thus proportional to Fm/Adep, which should be considered for a comparison of LP to AP. In Figure 31, the
www.iran-mavad.com
Figure 31 Comparison of deposition rates within HMDSO-fed plasmas at low pressure (LP) and atmospheric pressure (AP). Two different deposition
areas (corresponding to different reactor sizes) were considered. In order to adjust the energy invested per HMDSO molecule in the gas phase, the
correction factors rdep ¼ 0.1 and rgas ¼ 0.001 were used for AP.
deposition rates of HMDSO discharges within two different LP reactors (with a deposition area of 180 and 1500 cm2) using
a monomer gas flow of 4 and 28 sccm are depicted. The deposition rates clearly scale with Fm/Adep, revealing some loss in
conversion for increasing reactor size and flow rate. This is caused by different eedfs and incorporation rates as discussed earlier
(14,67). Therefore, literature data of AP HMDSO plasmas meeting the same range of monomer flow rate and deposition area were
searched. AP conditions found comprise DBDs with 1–4 sccm HMDSO (in 10 slm He), 15 cm in diameter, and with 1–24 sccm
HMDSO (in 2 slm N2) using a 23 cm long rod and a rotating ground electrode (34,114). In order to display the data in the same
Arrhenius-like plot, the factors rdep and rgas (w0.1 and w0.001, respectively) were fitted to meet the same range of energy invested
per HMDSO molecule; i.e., only a small part of the activated carrier gas contributes directly to deposition at AP conditions
(Figure 31). For comparable monomer flow rates and deposition areas, indeed similar deposition rates can then be obtained.
Therefore also the processing times are comparable, i.e., not reduced at AP.
This finding enables a comparison of a production process at LP and AP, e.g., for the continuous coating of membranes (w20 cm
in width) in order to render them hydrophobic by a 50-nm-thick SiOC:H plasma polymer film (reel-to-reel process). Considering
a deposition area of 70 21 cm2, a deposition rate of 360 nm min–1 was obtained for 100 sccm HMDSO in LP. For the same
conditions at AP, about 40 slm of an inert carrier gas (e.g., He or N2) are required.
Comparable conditions mean the same output (here, 5 km per day at 5 m min–1 production speed) and the same fixed costs
(maintenance, personal, infrastructure, etc.), which are reasonable assumptions (115). An investment of 500 kV for the LP appa-
ratus and 250 kV for the AP equipment is further assumed. The cumulative costs of LP and AP processes can thus be calculated and
compared. As can be seen in Figure 32, the breakeven of the LP plasma polymer deposition is already reached after 8 months
when helium is taken as carrier gas in AP and after 14 months when a cheaper carrier gas such as nitrogen or argon can be used.
The reasons are the high running costs due to the high carrier gas flow rates required.
Figure 32 Cost analysis of HMDSO discharges in low pressure (LP), atmospheric pressure in helium (AP – He), and in nitrogen (AP – N2). While
assuming comparable deposition rates using 0.1 slm HMDSO, the high gas flow rates required for AP processes (here 40 slm) cause higher running costs.
www.iran-mavad.com
Figure 33 Different forms of plasma polymer depositions on a polymer substrate. Thick coatings induce stress on the film/substrate interface, yielding
crack formation and delamination. Functional surfaces can be obtained by thin films, vertical gradients, multilayers, nanocomposites (with adlayer),
nanoporous coatings, domainlike coatings, or lateral gradients.
A reduction of the gas flow rates would result in higher thermal load and lower deposition rates (116). Therefore, AP plasma
processes based on air as carrier gas are of high interest. However, this is not an option for hydrophobic coatings, but, e.g., for the
deposition of SiOx coatings (117,118). Nevertheless, air-based AP plasma polymerization remains a challenge for the scientific
community as well as for the development of reliable industrial processes.
Furthermore, the hydrophobic properties of AP plasma–coated polymer substrates and the washing fastness of AP plasma–
coated textiles are still found to be inferior compared with the corresponding HMDSO-derived coatings in low pressure – most
likely due to an insufficient energy density during film growth (114,116,119). Reduced deposition rates at AP helps in improving the
film properties regarding eqn [4] (117). Within an industrial low-pressure plasma reactor, on the other hand, a washing fastness of
30 times (industrial washing at 60 C) could be realized, which is superior to conventional wet-chemical treatments (120). Hence,
some arguments as used for AP processes could not be verified.
4.09.2.3 Conclusion
Polymers are difficult to coat by their nature. Long-chain molecules with little cross-linking implicate different mechanical prop-
erties than most coatings. Reduced mechanical stability of a near-surface layer and lack in functional groups add to the situation. For
the coating of plasma polymers (or metal layers) the polymer surface has to be conditioned. Besides cleaning, stabilization (cross-
linking) of the polymer surface (avoiding the formation of a weak boundary layer) and creation of functional groups are aimed for.
Due to their different plasma susceptibility, some polymers are more suitable for plasma coating than others, which should already
be considered in the material’s selection. Nevertheless, plasma polymer deposition is an attractive method to tailor the properties of
polymer surfaces. Layer-by-layer deposition and control of the energetic conditions during film growth enable adaption of the
plasma polymer properties to the polymer substrate. A broadened film/substrate interphase can be obtained during the initial film
growth, which is mainly responsible for the excellent adhesion that can be observed for plasma polymers on suitable polymer
substrates. Overtreatment and formation of weak boundary layers, however, should be avoided.
Plasma polymer depositions on polymers further offer a wide range of different film structures. While uniform, rather thick
coatings might induce a mechanical mismatch with the substrate, thin layers, gradient (lateral and vertical) films, multilayers,
nanoporous, or nanoparticle-containing coatings enable highly functional and stable surfaces (Figure 33).
Although widely used for industrial applications, further understanding of plasma polymer deposition as well as of plasma/
polymer interaction will open up additional fields of application, since surface characteristics can be controlled over a broad
range. For a particular application, plasma polymer depositions, however, compete with other emerging technologies. The
strengths of plasma coatings are their versatility and flexibility, adaptability to polymer substrates, and ability to combine with
other techniques. Finally, plasma polymer deposition presents an environmentally friendly technology that is economically
feasible.
References
1. Vasilev, K.; Michelmore, A.; Martinek, P.; Chan, J.; Sah, V.; Griesser, H. J.; Short, R. D. Early Stages of Growth of Plasma Polymer Coatings Deposited from Nitrogen- and
Oxygen-Containing Monomers. Plasma Process. Polym. 2010, 7, 824–835.
2. Yasuda, H. Plasma Polymerization; Academic Press: Orlando, Florida, 1985.
3. Goodman, J. The Formation of Thin Polymer Films in the Gas Discharge. J. Polym. Sci. 1960, 44, 551–552.
4. Biederman, H. Plasma Polymer Films; Imperial College Press: London, UK, 2004.
5. Choukourov, A.; Gordeev, I.; Arzhakov, D.; Artemenko, A.; Kousal, J.; Kylian, O.; Slavinska, D.; Biederman, H. Does Cross-Link Density of PEO-Like Plasma Polymers Influence
Their Resistance to Adsorption of Fibrinogen? Plasma Process. Polym. 2012, 9, 48–58.
6. Wagner, H. E. In Low Temperature Plasma Physics; Hippler, Ed.; Wiley-VCH: Weinheim, Germany, 2001; p 305.
www.iran-mavad.com
7. Yasuda, H.; Yasuda, T. The Competitive Ablation and Polymerization (CAP) Principle and the Plasma Sensitivity of Elements in Plasma Polymerization and Treatment. J. Polym.
Sci. Polym. Chem. 2000, 38, 943–953.
8. d’Agostino, R.; Favia, P.; Fracassi, F. The Effect of Power on the Plasma-Assisted Deposition of Fluorinated Monomers. J. Polym. Sci. Polym. Chem. 1990, 28, 3387–3402.
9. Robertson, J. The Deposition Mechanism of Diamond-Like a-C and a-C:H. Diamond Relat. Mater. 1994, 2, 361–368.
10. Friedrich, J. Mechanisms of Plasma Polymerization – Reviewed from a Chemical Point of View. Plasma Process. Polym. 2011, 8, 783–802.
11. Klages, C. P.; Dietz, A.; Höing, T.; Thyen, R.; Weber, A.; Willich, P. Deposition and Properties of Carbon-Based Amorphous Protective Coatings. Surf. Coat. Technol. 1996, 80,
121–128.
12. Peter, S.; Graupner, K.; Grambole, D.; Richter, F. Comparative Experimental Analysis of the a-C: H Deposition Processes Using CH4 and C2H2 as Precursors. J. Appl. Phys.
2007, 102, 053304 (18 pp).
13. Hegemann, D.; Schütz, U.; Körner, E. Macroscopic Approach to Plasma Polymerization Using the Concept of Energy Density. Plasma Process. Polym. 2011, 8, 689–694.
14. Trieschmann, J.; Hegemann, D. Plasma Polymerization at Different Positions in an Asymmetric Ethylene Discharge. J. Phys. D Appl. Phys. 2011, 44, 475201 (9 pp).
15. Ohring, M. Materials Science of Thin Films, 2nd ed.; Academic Press: San Diego, California, 2002.
16. Ozaydin-Ince, G.; Gleason, K. K. Transition between Kinetic and Mass Transfer Regimes in the Initiated Chemical Vapor Deposition from Ethylene Glycol Diacrylate. J. Vac. Sci.
Technol. A 2009, 27, 1135–1143.
17. Coclite, A. M.; Gleason, K. K. Initiated PECVD of Organosilicon Coatings: A New Strategy to Monomer Structure Retention. Plasma Process. Polym. 2012, 9, 425–434.
18. Park, S. Y.; Kim, N.; Kim, U. Y.; Hong, S. I.; Sasabe, H. Plasma Polymerization of Hexamethyldisilazane. Polym. J. 1990, 22 (3), 242–249.
19. Hegemann, D. Macroscopic Investigation of Reaction Rates Yielding Plasma Polymer Deposition. J. Phys. D: Appl. Phys. 2013, 46, 205204 (8 pp).
20. Hegemann, D.; Hossain, M. M. Influence of Non-Polymerizable Gases Added during Plasma Polymerization. Plasma Process. Polym. 2005, 2, 554–562.
21. Hegemann, D. Influence of Pressure of an Asymmetric, Radio Frequency Discharge with Methane. Thin Solid Films 2006, 515, 2173–2178.
22. Rutscher, A.; Wagner, H. E. The Model of Macroscopic Kinetics in Non-Equilibrium Plasma Chemical Reactions. Contr. Plasma Phys. 1985, 25 (4), 315–328.
23. Hegemann, D.; Hossain, M. M.; Körner, E.; Balazs, D. J. Macroscopic Description of Plasma Polymerization. Plasma Process. Polym. 2007, 4, 229–238.
24. Eliasson, B.; Hirth, M.; Kogelschatz, U. Ozone Synthesis from Oxygen in Dielectric Barrier Discharges. J. Phys. D Appl. Phys. 1987, 20, 1421–1437.
25. Hegemann, D.; Schütz, U. Activation Energy of an Asymmetrical, Radio Frequency Discharge with Methane. Thin Solid Films 2005, 491, 96–103.
26. Hegemann, D.; Körner, E.; Guimond, S. (Reply to): Testing the Hypothesis: Comments on Plasma Polymerization of Acrylic Acid Revisited. Plasma Process. Polym. 2010, 7,
371–375.
27. Hegemann, D.; Körner, E.; Chen, S.; Benedikt, J.; von Keudell, A. Functional Plasma Polymers Deposited in Capacitively and Inductively Coupled Plasmas. Appl. Phys. Lett.
2012, 100, 051601 (3 pp).
28. Manenschijn, A.; Goedheer, W. J. Angular Ion and Neutral Energy Distribution in a Collisional rf Sheath. J. Appl. Phys. 1991, 69, 2923–2930.
29. Lieberman, M. A.; Lichtenberg, A. J. Principles of Plasma Discharges and Materials Processing, 2nd ed.; Wiley: Hoboken, NJ, 2005.
30. Dhayal, M.; Bradley, J. W. Using Heated Probes in Plasma Polymerising Discharges. Surf. Coat. Technol. 2004, 184, 116–122.
31. Windischmann, A. An Intrinsic Stress Scaling Law for Polycrystalline Thin Films Prepared by Ion Beam Sputtering. J. Appl. Phys. 1987, 62 (5), 1800–1807.
32. Davis, C. A. A Simple Model for the Formation of Compressive Stress in Thin Films by Ion Bombardment. Thin Solid Films 1993, 226, 30–34.
33. Hegemann, D.; Körner, E.; Blanchard, N.; Drabik, M.; Guimond, S. Densification of Functional Plasma Polymers by Momentum Transfer during Film Growth. Appl. Phys. Lett.
2012, 101, 211603 (4 pp).
34. Sawada, Y.; Ogawa, S.; Kogoma, M. Synthesis of Plasma-Polymerized Tetraethoxysilane and Hexamethyldisiloxane Films Prepared by Atmospheric Pressure Glow Discharge.
J. Phys. D Appl. Phys. 1995, 28, 1661–1669.
35. Guo, Y. B.; Hong, F. C. N. Radio-Frequency Microdischarge Arrays for Large-Area Cold Atmospheric Plasma Generation. Appl. Phys. Lett. 2003, 82 (3), 337–339.
36. Babayan, S. E.; Jeong, J. Y.; Schutze, A.; Tu, V. J.; Moravej, M.; Selwyn, G. S.; Hicks, R. F. Deposition of Silicon Dioxide Films with a Non-Equilibrium Atmospheric-Pressure
Plasma Jet. Plasma Sources Sci. Technol. 2001, 10, 573–578.
37. Michelmore, A.; Bryant, P. M.; Steele, D. A.; Vasilev, K.; Bradley, J. W.; Short, R. D. Role of Positive Ions in Determining the Deposition Rate and Film Chemistry of Continuous
Wave Hexamethyl Disiloxane Plasmas. Langmuir 2011, 27, 11943–11950.
38. Rau, C.; Kulisch, W. Mechanism of Plasma Polymerization of Various Silico-Organic Monomers. Thin Solid Films 1994, 249, 28–37.
39. Basner, R.; Foest, R.; Schmidt, M.; Becker, K.; Deutsch, H. Absolute Total and Partial Electron Impact Ionization Cross Sections of Hexamethyldisiloxane. Int. J. Mass
Spectrom. 1998, 167, 245–252.
40. Rusanov, V. D.; Fridman, A. A.; Sholin, G. V. The Physics of a Chemically Active Plasma with Nonequilibrium Vibrational Excitation of Molecules. Usp. Fiz. Nauk 1981, 134,
185–235.
41. Wang, Y.; Zhang, J.; Shen, X. Y. Surface Structures Tailoring of Hexamethyldisiloxane Films by Pulse RF Plasma Polymerization. Mater. Chem. Phys. 2006, 96,
498–505.
42. Sonnenfeld, A.; Tun, T. M.; Zajickova, L.; Kozlov, K. V.; Wagner, H. E.; Behnke, J. F.; Hippler, R. Deposition Process Based on Organosilicon Precursors in Dielectric Barrier
Discharges at Atmospheric Pressures – A Comparison. Plasmas Polym. 2001, 6, 237–266.
43. Kudriavtsev, Y.; Villegas, A.; Godines, A.; Asomoza, R. Calculation of the Surface Binding Energy for Ion Sputtered Particles. Appl. Surf. Sci. 2005, 239, 273–278.
44. Zhu, X.; Arefi-Khonsari, F.; Petit-Etienne, C.; Tatoulian, M. Open Air Deposition of SiO2 Films by an Atmospheric Pressure Line-Shaped Plasma. Plasma Process. Polym. 2005,
2, 407–413.
45. Dennler, G.; Houdayer, A.; Raynaud, P.; Seguy, I.; Segui, Y.; Wertheimer, M. R. Growth Modes of SiOx Films Deposited by Evaporation and Plasma-Enhanced Chemical Vapor
Deposition on Polymeric Substrates. Plasmas Polym. 2003, 8, 43–59.
46. Theirich, D.; Soll, C.; Leu, F.; Engemann, J. Intermediate Gas Phase Precursors during Plasma CVD of HMDSO. Vacuum 2003, 71, 349–359.
47. Foest, R.; Adler, F.; Sigeneger, F.; Schmidt, M. Study of an Atmospheric Pressure Glow Discharge (APG) for Thin Film Deposition. Surf. Coat. Technol. 2003, 163–164,
323–330.
48. Corbella, C.; Bialuch, I.; Kleinschmidt, M.; Bewilogua, K. Modified DLC Coatings Prepared in a Large-Scale Reactor by Dual Microwave/Pulsed-DC Plasma-Activated Chemical
Vapour Deposition. Thin Solid Films 2008, 517, 1125–1130.
49. Hegemann, D.; Schütz, U.; Oehr, C. RF Plasma Deposition of SiOx and a-C: H as Barrier Coatings on Polymers. In Plasma Processes and Polymers; d’Agostino, R., Favia, P.,
Oehr, C., Wertheimer, M. R., Eds.; Wiley-VCH: Weinheim, Germany, 2005; pp 23–37.
50. Hegemann, D.; Oehr, C.; Fischer, A. Design of Functional Coatings. J. Vac. Sci. Technol. A 2005, 23 (1), 5–11.
51. Hegemann, D.; Brunner, H.; Oehr, C. Plasma Treatment of Polymers to Generate Stable, Hydrophobic Surfaces. Plasma Polym. 2001, 6 (4), 221–235.
52. Goujon, M.; Belmonte, T.; Henrion, G. OES and FTIR Diagnostics of HMDSO/O2 Gas Mixtures for SiOx Deposition Assisted by RF Plasma. Surf. Coat. Technol. 2004, 188–189,
756–761.
53. Hegemann, D.; Brunner, H.; Oehr, C. Evaluation of Deposition Conditions to Design Plasma Coatings like SiOx and a-C:H on Polymers. Surf. Coat. Technol. 2003, 174–175,
253–260.
54. Schneider, J.; Kiesler, D.; Leins, M.; Schulz, A.; Walker, M.; Schumacher, U.; Stroth, U. Development of Plasma Polymerised SiOx Barriers on Polymer Films for Food
Packaging Applications. Plasma Process. Polym. 2007, 4, S155–S159.
55. Suhr, H. Application of Nonequilibrium Plasmas in Organic Chemistry. Plasma Chem. Plasma Process. 1983, 3 (1), 1–61.
56. Bauer, M.; Schwarz-Selinger, T.; Kang, H.; von Keudell, A. Control of the Plasma Chemistry of a Pulsed Inductively Coupled Methane Plasma. Plasma Sources Sci. Technol.
2005, 14, 543–548.
www.iran-mavad.com
57. Kleber, R.; Weiler, M.; Krüger, A.; Sattel, S.; Kunz, G.; Jung, K.; Ehrhardt, H. Influence of Ion Energy and Flux Composition on the Properties of Plasma-Deposited
Amorphous Carbon and Amorphous Hydrogenated Carbon Films. Diamond Relat. Mater. 1993, 2, 246–250.
58. von Keudell, A. Formation of Polymer-Like Hydrocarbon Films from Radical Beams of Methyl and Atomic Hydrogen. Thin Solid Films 2002, 402, 1–37.
59. Keudell, A.; Jacob, W. Elementary Processes in Plasma-Surface Interaction: H-Atom and Ion-Induced Chemisorption of Methyl on Hydrocarbon Film Surfaces. Prog. Surf.
Sci. 2004, 76, 21–54.
60. Hegemann, D.; Körner, E.; Albrecht, K.; Schütz, U.; Guimond, S. Growth Mechanism of Oxygen-Containing Functional Plasma Polymers. Plasma Process. Polym. 2010, 7,
889–898.
61. Weiler, M.; Sattel, S.; Giessen, T.; Jung, K.; Ehrhardt, H.; Veerasamy, V. S.; Robertson, J. Preparation and Properties of Highly Tetrahedral Hydrogenated Amorphous
Carbon. Phys. Rev. B 1996, 53 (3), 1594–1608.
62. Maitre, N.; Girardeau, T.; Camelio, S.; Barranco, A.; Vouagner, D.; Breelle, E. Effects of Negative Self-Bias on Hydrogenated Amorphous Carbon Films Deposited by PECVD
Technique. Diamond Relat. Mater. 2003, 12, 988–992.
63. Yehoda, J. E.; Yang, B.; Vedam, K.; Messier, R. Investigation of the Void Structure in Amorphous Germanium Thin Films as a Function of Low-Energy Ion Bombardment.
J. Vac. Sci. Technol. 1988, A 6 (3), 1631–1635.
64. Neyts, E.; Bogaerts, A.; van de Sanden, M. C. M. Effect of Hydrogen on the Growth of Thin Hydrogenated Amorphous Carbon Films from Thermal Energy Radicals. Appl.
Phys. Lett. 2006, 88, 141922 (3 pp).
65. Hegemann, D. Plasma Polymerization and Its Application in Textiles. Indian J. Fibre Textile Res. 2006, 31, 99–115.
66. Hossain, M. M.; Hegemann, D.; Fortunato, G.; Herrmann, A. S.; Heuberger, M. Plasma Deposition of Permanent Superhydrophilic a-C:H:N Films on Textiles. Plasma
Process. Polym. 2007, 4, 471–481.
67. Guimond, S.; Schütz, U.; Hanselmann, B.; Körner, E.; Hegemann, D. Influence of Gas Phase and Surface Reactions on Plasma Polymerization. Surf. Coat. Technol. 2011,
205, S447–S450.
68. Hegemann, D.; Körner, E.; Guimond, S. Plasma Polymerization of Acrylic Acid Revisited. Plasma Process. Polym. 2009, 6, 246–254.
69. Short, R. D.; Steele, D. A. Testing the Hypothesis: Comments on Plasma Polymerization of Acrylic Acid Revisited. Plasma Process. Polym. 2010, 7, 366–370.
70. Detomaso, L.; Gristina, R.; Senesi, G. S.; d’Agostino, R.; Favia, P. Stable Plasma-Deposited Acrylic Acid Surfaces for Cell Culture Applications. Biomaterials 2005, 26,
3831–3841.
71. Sciarratta, V.; Vohrer, U.; Hegemann, D.; Müller, M.; Oehr, C. Plasma Functionalization of Polypropylene with Acrylic Acid. Surf. Coat. Technol. 2003, 174–175, 805–810.
72. Siow, K. S.; Britcher, L.; Kumar, S.; Griesser, H. J. Plasma Methods for the Generation of Chemically Reactive Surfaces for Biomolecule Immobilization and Cell Colonization
– A Review. Plasma Process. Polym. 2006, 3, 392–418.
73. Barton, D.; Shard, A. G.; Short, R. D.; Bradley, J. W. The Effect of Positive Ion Energy on Plasma Polymerization: A Comparison between Acrylic and Propionic Acids.
J. Phys. Chem. B 2005, 109, 3207–3211.
74. Hegemann, D. Macroscopic Control of Plasma Polymerization Processes. Pure Appl. Chem. 2008, 80 (9), 1893–1900.
75. Friedrich, J. The Plasma Chemistry of Polymer Surfaces; Wiley-VCH: Weinheim, Germany, 2012.
76. Liston, E. M.; Martinu, L.; Wertheimer, M. R. Plasma Surface Modification of Polymers for Improved Adhesion: A Critical Review. J. Adhesion Sci. Technol. 1993, 7,
1091–1127.
77. Hegemann, D.; Brunner, H.; Oehr, C. Plasma Treatment of Polymers for Surface and Adhesion Improvement. Nucl. Instr. Meth. Phys. Res. B 2003, 208, 281–286.
78. Yasuda, H.; Lamaze, C. E.; Sakaoku, K. Effect of Electrodeless Glow Discharge on Polymers. J. Appl. Polym. Sci. 1973, 17, 137–152.
79. Engelmann, S.; Bruce, R. L.; Weilnboeck, F.; Oehrlein, G. S.; Nest, D.; Graves, D. B.; Andes, C.; Hudson, E. A. Dependence of Polymer Surface Roughening Rate on
Deposited Energy Density during Plasma Processing. Plasma Process. Polym. 2009, 6, 484–489.
80. Grace, J. M.; Gerenser, L. J. Plasma Treatment of Polymers. J. Dispersion Sci. Technol. 2003, 24, 305–341.
81. Strobel, M.; Lyons, C. S.; Strobel, J. M.; Kapaun, R. S. Analysis of Air-Corona-Treated Polypropylene and Poly(Ethylene Terephthalate) Films by Contact-Angle
Measurements and X-Ray Photoelectron Spectroscopy. J. Adhesion Sci. Technol. 1992, 6, 429–443.
82. Guimond, S.; Wertheimer, M. R. Surface Degradation and Hydrophobic Recovery of Polyolefins Treated by Air Corona and Nitrogen Atmospheric Pressure Glow Discharge.
J. Appl. Polym. Sci. 2004, 94 (3), 1291–1303.
83. Barz, J.; Haupt, M.; Vohrer, U.; Hilgers, H.; Oehr, C. Ultrathin Carbon-Fluorine Film Processing. Surf. Coat. Technol. 2005, 200, 453–457.
84. Phelps, A. V.; Petrovic, Z. L. Cold-Cathode Discharges and Breakdown in Argon: Surface and Gas Phase Production of Secondary Electrons. Plasma Sources Sci. Technol.
1999, 8, R21–R44.
85. Ahedo, E. Presheath/Sheath Model with Secondary Electron Emission from Two Parallel Walls. Phys. Plasmas 2002, 9 (10), 4340–4347.
86. Barnat, E. V.; Hebner, G. A. Electric fields in a sheath near a metal-dielectric interface. Appl. Phys. Lett. 2004, 85 (16), 3393–3395.
87. Hegemann, D.; Brunner, H.; Oehr, C. Deposition Rate and Three-Dimensional Uniformity of RF Plasma Deposited SiOx Films. Surf. Coat. Technol. 2001, 142–144,
849–855.
88. Yi, W.; Jeong, T.; Yu, S. G.; Lee, J.; Jin, S.; Heo, J.; Kim, J. M. Study of the Secondary-Electron Emission from Thermally Grown SiO2 Films on Si. Thin Solid Films 2001,
397, 170–175.
89. Amanatides, E.; Mataras, D. Growth Kinetics of Plasma Deposited Microcrystalline Silicon Thin Films. Surf. Coat. Technol. 2011, 205, S178–S181.
90. Körner, E.; Guimond, S.; Hegemann, D. Influence of RF Plasma Reactor Setup on Carboxylated Hydrocarbon Coatings. Plasma Process. Polym. 2009, 6, 119–125.
91. Hegemann, D.; Hossain, M. M.; Balazs, D. J. Nanostructured Plasma Coatings to Obtain Multifunctional Textile Surfaces. Progr. Organic Coat. 2007, 58, 237–240.
92. Ruiz, J. C.; St-Georges-Robillard, A.; Theresy, C.; Lerouge, S.; Wertheimer, M. R. Fabrication and Characterization of Amine-Rich Organic Thin Films: Focus on Stability.
93. Oehr, C.; Müller, M.; Elkin, B.; Hegemann, D.; Vohrer, U. Plasma Grafting – A Method to Obtain Monofunctional Surfaces. Surf. Coat. Technol. 1999, 116–119, 25–35.
94. Perez-Roldan, M.; Colpo, P.; Gilliland, D.; Ceccone, G.; Rossi, F. Surface Characterization of PEO-Like Microstructures by Means of ToF-SIMS, XPS and SPR. Surf. Interface
Anal. 2013, 45, 240–243.
95. Tamulevicius, S. Stress and Strain in the Vacuum Deposited Thin Films. Vacuum 1998, 51, 127–139.
96. Yun, J.; Lee, S.; Bae, T. S.; Yun, Y.; Lee, S.; Kwon, J. D.; Lee, G. H. Adhesive and Structural Failures of Oxide Coatings on Plasma-Treated Polymers. Plasma Process.
Polym. 2011, 8, 815–831.
97. Johansson, K. S. Gas Barrier Properties of Plasma-Deposited Coatings – Substrate Effects. In Polymer Surface Modification: Relevance to Adhesion, Vol. 2; Mittal, K. L.,
Ed.; VSP: Utrecht, The Netherlands, 2000; pp 575–603.
98. Hegemann, D.; Amberg, M.; Ritter, A.; Heuberger, M. Recent Developments in Ag Metallised Textiles Using Plasma Sputtering. Mater. Technol. 2009, 24, 41–45.
99. Franz, G. Low Pressure Plasmas and Microstructuring Technology; Springer: Berlin, Germany, 2009.
100. Westwood, W. D. Sputter DepositionIn: AVS Education Committee Book Series, Vol. 2; AVS: New York, 2003.
101. Müller, K. H. Monte Carlo Calculation for Structural Modifications in Ion-Assisted Thin Film Deposition due to Thermal Spikes. J. Vac. Sci. Technol. 1986, A 4 (2), 184–188.
102. Floro, J. A.; Chason, E.; Cammarata, R. C.; Srolovitz, D. J. Physical Origins of Intrinsic Stresses in Vollmer-Weber Thin Films. MRS Bull. 2002, 1, 19–25.
103. Amberg, M.; Kasdallah, C.; Ritter, A.; Hegemann, D. Influence of Residual Oils on the Adhesion of Metal Coatings to Textiles. J. Adhesion Sci. Technol. 2010, 24,
123–134.
104. Tiller, H. J.; Göbel, R.; Magnus, B.; Garschke, A.; Musil, R. A New Concept of Metal-Resin Adhesion Using an Intermediate Layer of SiOx–C. Thin Solid Films 1989, 169,
159–168.
www.iran-mavad.com
105. Lange, J.; Wyser, Y. Recent Innovations in Barrier Technologies for Plastic Packaging – A Review. Packag. Technol. Sci. 2003, 16, 149–158.
106. Hopwood, J. Review of Inductively Coupled Plasmas for Plasma Processing. Plasma Sources Sci. Technol. 1992, 1, 109–116.
107. Wertheimer, M. R. Comparison of Microwave and Lower Frequency Plasmas for Thin Film Deposition and Etching. J. Vac. Sci. Technol. 1985, A3 (6), 2643–2649.
108. Kaiser, M.; Baumgärtner, K. M.; Schulz, A.; Walker, M.; Räuchle, E. Linearly Extended Plasma Source for Large-Scale Applications. Surf. Coat. Technol. 1999, 116–119,
552–557.
109. Schulze, J.; Gans, T.; O’Connell, D.; Czarnetzki, U.; Ellingboe, A. R.; Turner, M. M. Space and Phase Resolved Plasma Parameters in an Industrial Dual-Frequency
Capacitively Coupled Radio-Frequency Discharge. J. Phys. D Appl. Phys. 2007, 40, 7008–7018.
110. Noeske, M.; Degenhardt, J.; Strudthoff, S.; Lommatzsch, U. Plasma Jet Treatment of Five Polymers at Atmospheric Pressure: Surface Modifications and the Relevance for
Adhesion. Int. J. Adhes. Adhes. 2004, 24 (2), 171–177.
111. Fridman, G.; Friedman, G.; Gutsol, A.; Shekhter, A. B.; Vasilets, V. N.; Fridman, A. Applied Plasma Medicine. Plasma Process. Polym. 2008, 5, 503–533.
112. Pappas, D. Status and Potential of Atmospheric Plasma Processing of Materials. J. Vac. Sci. Technol. 2011, 29 (2), 020801 (17 pp).
113. Radic, N.; Obradovic, B. M.; Kostic, M.; Dojcinovic, B.; Kuraica, M. M.; Cernak, M. Deposition of Silver Ions onto DBD and DCSBD Plasma Treated Polypropylene Nonwoven.
Surf. Coat. Technol., 2012 206 (23), 5006–5011.
114. Siliprandi, R. A.; Zanini, S.; Grimoldi, E.; Fumagalli, F. S.; Barni, R.; Riccardi, C. Atmospheric Pressure Plasma Discharge for Polysiloxane Thin Films Deposition and
Comparison with Low Pressure Process. Plasma Chem. Plasma Process. 2011, 31, 353–372.
115. Hegemann, D. Deposition of Functional Plasma Polymers by Control of Film Growth Conditions. In Proc. 54rd SVC Annual Technical Conf., Chicago, IL, 2011, p 5.
116. Kale, K. H.; Palaskar, S. Atmospheric Pressure Plasma Polymerization of Hexamethyldisiloxane for Imparting Water Repellency to Cotton Fabric. Textile Res. J. 2010, 81 (6),
608–620.
117. Trunec, D.; Zajickova, L.; Bursikova, V.; Studnicka, F.; Stahel, P.; Prysiazhnyi, V.; Perina, V.; Houdkova, J.; Navratil, Z.; Franta, D. Deposition of Hard Thin Films from
HMDSO in Atmospheric Pressure Dielectric Barrier Discharges. J. Phys. D Appl. Phys. 2010, 43 (22), 225403 (18 pp).
118. Starostin, S. A.; Premkumar, P. A.; Creatore, M.; de Vries, H.; Paffen, R. M. J.; van de Sanden, M. C. M. High Current Diffuse Dielectric Barrier Discharge in Atmospheric
Pressure Air for the Deposition of Thin Silica-Like Films. Appl. Phys. Lett. 2010, 96, 061502 (3 pp).
119. Hegemann, D.; Körner, E.; Guimond, S. Macroscopic Approach to Investigate Plasma Polymer Growth. In Proc. 53rd SVC Annual Technical Conf., Orlando, FL, 2010,
pp 409–413.
120. Hossain, M. M.; Grabher, G. Applications of Low-Pressure Plasma in Textile Finishing. Melliand Int. 2011, 17, 149–151.
www.iran-mavad.com
4.10 Thermal Spray Coating Processes
P Vuoristo, Tampere University of Technology, Tampere, Finland

4.10.2 Spray Materials 231
4.10.2.1 Powders 231
4.10.2.2 Wires and Rods 233
4.10.2.3 Suspensions and Solutions 235
4.10.2.4 Substrate Materials 236
4.10.3 Thermal Spray Processes 236
4.10.3.1 Classification of Processes 236
4.10.3.2 Flame Spraying 238
4.10.3.2.1 Powder Flame Spraying 239
4.10.3.2.2 Wire Flame Spraying 239
4.10.3.2.3 Spray and Fuse Process 240
4.10.3.3 Electric Arc Spraying 240
4.10.3.4 Plasma Spraying 241
4.10.3.4.1 Atmospheric Plasma Spraying 241
4.10.3.4.2 Low-Pressure Plasma Spraying 242
4.10.3.4.3 High-Power and Special Plasma Spray Processes 243
4.10.3.5 High-Velocity Flame Spraying 244
4.10.3.5.1 HVOF Spraying 246
4.10.3.5.2 Warm Spraying 246
4.10.3.5.3 HVAF Spraying 246
4.10.3.5.4 High-Velocity Combustion Arc Spraying 247
4.10.3.5.5 Detonation Gun Spraying 247
4.10.3.6 Cold Spraying 248
4.10.3.6.1 High-Pressure Cold Spraying 249
4.10.3.6.2 Low-Pressure Cold Spraying 249
4.10.3.7 Suspension and Solution Spraying 250
4.10.4 Formation of Sprayed Coatings 251
4.10.4.1 Coatings from Molten Droplets 251
4.10.4.2 Formation of Sprayed and Fused Coatings 255
4.10.4.3 Solid-State Formation in Cold Spraying 255
4.10.4.4 Posttreatments 257
4.10.5 Coating Materials and Properties 257
4.10.5.1 Coating Materials 258
4.10.5.1.1 Pure Metals 258
4.10.5.1.2 Metal Alloys 258
4.10.5.1.3 Hard Metals 258
4.10.5.1.4 Oxide Ceramics 259
4.10.5.1.5 Polymers 259
4.10.5.1.6 Composites and Other Materials 259
4.10.5.2 Coating Properties 259
4.10.5.2.1 Microstructure, Phase Structure, and Density 260
4.10.5.2.2 Mechanical Properties 260
4.10.5.2.3 Corrosion Properties 262
4.10.5.2.4 Wear Properties 262
4.10.5.2.5 Friction Properties 264
4.10.5.2.6 High-Temperature Properties 265
4.10.5.2.7 Biomedical Properties 265
4.10.6 Industrial Applications 265
4.10.6.1 Mechanical Engineering 268
4.10.6.2 Corrosion Protection in Infrastructures and Steel Structures 268
4.10.6.3 Paper and Pulp 268
4.10.6.4 Gas Turbines 270
4.10.6.5 Textile Industry 270

www.iran-mavad.com
230 Thermal Spray Coating Processes
4.10.6.6 Printing Industry 270

4.10.6.7 Electrical Industry 272
4.10.6.8 Steel Industry 272
4.10.6.9 Biomedical Industry 272
4.10.6.10 Power Generation Plants 273
4.10.6.11 Miscellaneous Applications 273
4.10.7 Development Trends 275
4.10.8 Concluding Remarks 275
References 275
4.10.1 Introduction
Thermal spraying includes a group of coating processes in which metallic and nonmetallic materials are spray deposited as fine
particles in a molten or semimolten condition or even in fully solid state to form a coating. Figure 1 presents the principle of
thermal spray coating process. In the process, a heat source, which can be a combustion flame, an electric arc, or an ionized gas, i.e.,
plasma, is used to melt the feedstock material (powder, wire, stick, suspension/solution) and accelerate the particles toward the
solid surface of the substrate, which is the material or component to be coated. Other thermal or kinetic energy sources used today
are, for instance, laser beams in laser coating and inert gas flow jets in the cold kinetic spray process. The sprayed coatings are formed
from flattened and solidified particles.
Thermal spraying originated a long time ago when Dr M. U. Schoop from Switzerland built his first metal spray gun in about
1910 (1). History says that he got the idea either from the trace that a lead bullet left when impacting on the surface of a stone wall or
from a method to prepare lead powder. Dr Schoop was working for some time in the laboratory of Thomas A. Edison in New Jersey,
where he developed a new manufacturing method for lead powder pigment used in paints. After returning to Switzerland, Schoop
filed a patent application for a method in which molten lead was poured into a fast flowing compressed air jet. After several trials
and experiments, he finished his first commercialized wire spray guns – thermal spraying was actually called during the first years the
Schoop spray method. Just prior to the First World War, he sold the technique to a German company, Metallizator. This equipment
was sold and rented in Europe and in the United States in the early 1920s. A British company, Metallisation Ltd., was established
during these years. The first thermal spray guns in the United States were manufactured by the Metallizing Company of America in
Los Angeles during the great depression, starting in 1929. The company still exists with the name Mogul. In 1931, Schoop’s patent
expired, which was followed by the establishment of other manufacturing companies and coating producers. However, the thermal
spray method was at that time not yet convincing enough, making commercial activities still difficult. One of the new companies
resulting from Mogul in the mid-1930s was Metallizing Engineering Company Inc., later known as Metco. In the beginning of
the1940s, during the Second World War, the US Army successfully started several workshops for repair of heavy military equipment
by thermal spraying in China, Burma, India, and Italy. From those years forward thermal spraying has become a noteworthy coating
technology among the many other processes, although it has been still fairly unknown by lay people. During the past 20 years,
thermal spraying has developed much more than during the whole history of the process before.
Coating materials in thermal spraying include pure metals, metal alloys, hard metals (carbides), oxide ceramics, plastics,
cermets, composites, and blended materials (2). Materials suitable for thermal spraying include practically all materials, except
compounds that decompose irreversibly during melting and the ones that do not have a stable melt state and vaporize excessively in
the spray process. Some materials, e.g., silicon ceramics and magnesia, are difficult to spray thermally. The deposition rate, i.e.,
coating production rate, in thermal spraying is high compared with many other deposition technologies. Depending on the
applications, thermally sprayed coatings have thickness ranging from some micrometers up to several millimeters. Due to the
versatility of thermal spraying regarding coating materials, thermal spray processes, coatings properties and in many case also cost
efficiency, thermally sprayed coatings are industrially used very widely, e.g., in mechanical engineering applications, aviation and
Figure 1 Principle of the thermal spray coating process. Source: Tampere University of Technology.
www.iran-mavad.com
Thermal Spray Coating Processes 231
energy technologies, paper and printing industries, infrastructure applications, biomedical applications, electronics, etc. The flex-
ibility of thermal spraying is probably the highest among all different deposition technologies. This is due to several reasons,
particularly because of the following features of the thermal spray technology:
l Very large amounts of different types of spray materials and compositions are available, including pure metals, metal alloys, hard
metals (carbides/cermets), ceramics, polymers, and combinations of these materials.
l Coating properties, for instance, wear resistance, hardness, and adhesion strength are clearly adequate for many industrial
applications.
l Various properties and functionalities can be obtained by thermally sprayed coatings.
l The production rate of thermal spraying is remarkably high and is suitable also for large-area surfaces.
l The heat input to the components to be coated by thermal spraying is noticeably low; no metallurgical changes or distortion of
components is present. An obvious exception is the spray and fuse process for self-fluxing alloys.
l Coating manufacturing costs are acceptable for most industrial applications.
l Thermal spraying is regarded as green technology; particularly in comparison with hard chromium plating.
l Novel variations of thermal spray technologies have been developed, e.g., cold kinetic spraying and suspension and solution
spraying, allowing new possibilities in totally new application areas.
In this chapter, thermal spray technology is reviewed. The presentation includes a review of spray materials, thermal spray
processes, the basics of coating formation, types and properties of coating materials, and industrial applications and a summary of
recent developments in thermal spray technology. For more detailed information about the topic, several good books are also
available for further reading (2–7).
Thermal spray coating- and processing-related chapters are published in such refereed journals as Journal of Thermal Spray
Technology (ASM International), Surface and Coatings Technology (Elsevier), Surface Engineering (Maney), and several other journals,
and in the proceedings of some scientific conferences such as International Thermal Spray Conference (ITSC), RIPT conference,
International Conference on Metallurgical Coatings and Thin Films. Societies such as ASM Thermal Spray Society (ASM/TSS) and
European Thermal Spray Association (ETSA) are both active in promoting thermal spray science and technology worldwide.
4.10.2 Spray Materials
Thermal spray coating processes use powder, wire, or rod-shaped feedstock materials for coating manufacturing. All these types are
used and depend highly on the particular thermal spray process equipment. Some of the spray processes, such as plasma, high-
velocity oxygen-fuel (HVOF), and powder flame spray systems, use only powder materials, and some processes, e.g., electric
wire arc and wire flame spraying, are based on wire materials as feedstock. Some exceptions to this may exist; also under research
and in the early stages of use are liquid-type feedstock materials, e.g., suspensions, which are liquids containing nanometer- or
micrometer-sized solid particles, and fully liquid materials containing specific chemicals and metallic salts for coating formation.
It is notable that the number of different spray materials is extremely high; several hundreds or even up to the level of one thousand
different materials in different forms of feedstock are available. Thermal spray technology is clearly the most versatile coating
manufacturing technology among all the deposition technologies when the comparison is done according to the availability of
different coating materials and structures and the flexibility of using many of these coating materials in the same particular spray
equipment. Both powder and wire materials show advantages and disadvantages in materials selection for the coating manufac-
turer, as well as the end user and the final application. The main advantage of wires is simple handling, but material variety is
somewhat limited and significantly less than that available as spray powders. Selection of the form of the coating material is usually
determined by the availability of appropriate spray equipment and by the required coating structures and coating properties. Wires
used in thermal spray processes, particularly in electric arc spraying, are usually either thinner (1.6 and 2.5 mm) solid or cored wires
similar to those in arc welding or wires with higher diameter (1.4–4.8 mm) used particularly in the flame spray process. Moreover,
relatively thick, flexible powder-containing filled wires (flexicords) with polymer cover and thick solid ceramic rods or sticks are
used in flame spraying. Particularly for thermal spray powders, powder characteristics and coating properties obtained are
remarkably determined by the powder-production methods. Based on weight, the annual consumption of wires is notably higher
than that of spray powders. However, coatings prepared from powder materials typically result in improved coating properties and
are also used in more demanding industrial applications, e.g., in aerospace, energy, and mechanical engineering applications.
Besides powders and wires, suspensions and even liquids without solid particles are used in thermal spray processes. So far,
suspension and liquid spraying are not yet widely used; instead these types of feedstock materials are in the research and devel-
opment stage in several laboratories. Some evaluations and attempts for industrial use are seen for these types of novel materials.
4.10.2.1 Powders
Thermal spray powders are highly developed products with tailored compositions and carefully adjusted particle sizes and size
distributions. The most important spray powder types and their manufacturing processes are (2,3,7):
l gas and water atomized powders,
l fused and crushed powders,
www.iran-mavad.com
l sintered and crushed powders,

l clad/composite powders and dense/porous coated powders,
l agglomerated and sintered/spray-dried powders, as well as
l special powders, e.g., sol–gel and mechanically milled/alloyed powders.
For metallic powders, gas and water atomizing process has the highest importance. Inert gas atomizing is more popular for the
production of metal and alloy powders for flame spray, plasma spray, and high-velocity flame spray processes. In the gas atomizing
process, the metal is induction melted and the molten metal is injected as a stream to a vacuum chamber, where the molten material
is spheroidized and collected by separators and cyclones as fine-sized spherical powders. In the water atomizing process, a water jet
is used for the spheroidization. In atomizing processes, it is notable that melting of the raw material can occur under a protective gas
or in air, depending on the tolerated oxygen content of the spray powder produced. The atomizing medium, either inert gas (argon,
nitrogen, etc.) or water, has a significant influence on the gas content (oxygen) as well as the particle shape of the spray powder. Inert
gas-atomized powders are typically highly spherical and have significantly lower gas contents than water-atomized powders.
Figure 2 presents some state-of-the-art inert gas-atomized powders for flame spray and HVOF thermal spray processes.
Fusing or sintering with subsequent crushing has been used traditionally for production of several different type of materials,
including hard metals, metals, and oxide ceramic powders. These powder-production technologies are still in use, particularly for
the production of ceramic powders for plasma spraying. Other powder manufacturing techniques, e.g., spray drying and sintering,
are gathering more use, particularly in the production of multicomponent composite powders based on hard metals and mixed
oxides. Figures 3 and 4 illustrate examples of sintered and crushed WC–10%Co–4%Cr powder for HVOF and spraying and fused
and crushed ceramic oxide powders for different plasma spray processes, respectively. Sintered or fused and subsequently
mechanically crushed powders show dense microstructures and blocky morphologies.
Agglomeration by the spray drying process with a subsequent sintering step is one of the most modern production methods for
spray powders, particularly for hard metal and oxide ceramic spray powders. The production process is usually followed by sintering
in furnaces or spheroidizing by plasma treatment in order to compact and densify the powder. The spray drying process readily
results in spherical particles and the sintering step then prevents the agglomerated powder from being destroyed during powder
transport from the feeder to the spray gun, or in the combustion flame or plasma used to melt the material. The main benefits of
spray dried and sintered powders are the spherical particle morphology ensuring excellent powder flowability from the powder
feeder to the thermal spray gun and good homogeneity of various elemental components in the powder particles. In hard metal
Figure 2 Morphologies of gas-atomized powders for different thermal spray processes: (a) coarse NiCrBSi powder for flame spray and fusing
process and (b) fine-sized stainless steel AISI 316 powder for HVOF spray process. Source: Tampere University of Technology.
Figure 3 Morphology (a) and cross-sectional microstructure (b) of sintered and crushed WC–10%Co–4%Cr powder for HVOF spraying.
Source: Tampere University of Technology.
www.iran-mavad.com
Figure 4 Morphologies of fused- and crushed-type ceramic spray powders for plasma spraying: (a) ZrO2-based thermal barrier coating (TBC)
powder, (b) very coarse Cr2O3 powder for water-stabilized plasma spray process. Source: Tampere University of Technology.
powders, such as WC–Co(Cr) and Cr3C2–NiCr, good control of carbon content allows production of microstructures in which small
micrometer- or even nanometer-sized hard carbides (WC) are embedded uniformly in the soft metallic binder phase. Figure 5 shows
the morphology and cross-sectional microstructure of an agglomerated and sintered WC–10%Co–4%Cr powder for HVOF spraying.
Spray drying and sintering is also a widely used powder manufacturing method for micrometer- and nanometer-sized multicom-
ponent ceramic oxide powders, such as various Cr2O3–25%TiO2, ZrO2–8%Y2O3, and Al2O3–(3–40)%TiO2 compositions.
There exist also other useful powder manufacturing technologies, which are widely used due to their capabilities of producing
specific powder compositions and properties. Clad powder is one important example. In clad powders, coating of the so-called
primary grains with fine secondary grains and an additional organic binder is the method for producing such composite powders.
Also, relatively rough primary grains can be electroplated with nickel or cobalt. Nickel/graphite is typical for this group of materials,
which are used as abradable coatings in gas turbine applications. In such powders, the core of the powder particle is graphite, which
is then protected from burning during spraying by a nickel cladding layer. Powder production by mixing two or more different
powder types and compositions is one more possibility for widening the technical applications of thermal spray coatings; such types
of powder blends are still widely used and some examples of compositions are, for instance, WC/NiCrBSi and Mo/NiCrBSi blends.
When selecting spray materials, attention should be paid to the fact that the properties of sprayed coatings can be completely
different from the properties of cast or wrought corresponding materials. Not only density and porosity but also internal stresses
and oxide content of sprayed coatings differ significantly from bulk materials. Tables 1–3 summarize the importance of powder
properties on sprayability and coating properties, characterization techniques for thermal spray powders, and typical appropriate
particle sizes for different thermal spray processes, respectively. It is obvious that the size distribution and shape of powders
determines such important requirements as good flowability and melting behavior in specific spray processes. A wide range of
compositions of thermal spray powders is standardized and presented, e.g., in EN 1274 and other standards. For highly demanding
industrial applications, significantly more strict powder specifications are prepared by powder manufacturers and advanced thermal
spray coating suppliers and manufacturers.
4.10.2.2 Wires and Rods

The majority of thermal spraying feedstock materials on the basis of weight are used as wire materials. This is because practically all
corrosion protection materials based on cathodic protection are zinc and aluminum. Wire flame spraying and electric arc spraying
Figure 5 Morphology (a) and cross-sectional microstructure (b) of agglomerated (spray dried) and sintered WC–10%Co–4%Cr powder for HVOF
spraying. Source: Tampere University of Technology.
www.iran-mavad.com
Table 1 General requirements for thermal spray powders
Characteristic Consequence
Particle size l Suitability to particular spray process

l Reproducible coating properties
Spherical shape of powder particles l Improved flowability
l Uniform powder feeding and clogging-free spraying
l Uniform melting of powder particles
Narrow particle size range l Uniform melting of powder particles
l Reproducible coating properties
l High deposition efficiency
Homogeneous alloying l Homogeneous microstructure of coating
l Homogeneous coating properties
Low interparticle porosity l Uniform melting characteristics of powder
l Dense coating structure, particularly in low-temperature processes
Manufacturing by mass production processes l Appropriate powder price for industrial use
Table 2 Characterization techniques for thermal spray powders
Property/characteristic Analysis/research method Example: WC/Co 88/12
Chemical composition All relevant analytical methods, including spectroscopic 12%Co, 3.8%C, max. 2%Fe, bal. W, also light
techniques and X-ray fluorescence elements and impurities
Particle size and size distribution Sieve analysis; techniques based on X-ray 90 þ 45 mm
absorption and laser scattering
Production method/particle shape SEM or optical micrographs Fused, crushed/blocky
Apparent density Scott method, EN 23923-2 6.8 g cm3
Hall flow behavior Hall test, EN ISO 4490 14 s 50 g1
Porosity Mercury porosimeter. Nitrogen absorption –
Microstructure Cross-sectional specimen from powder particles SEM images
Phase structure and composition X-ray diffraction analysis, microprobe Phase structure: WC, W2C, Co–W–C mixed phases
Melting behavior (powder for Differential thermal analysis –
spray and fuse process)
SEM, scanning electron microscope.
Table 3 Typical particle sizes of powders used in different thermal spray processes
Spray process Nominal particle sizes (mm) Comments
Flame spraying 30–125 –

Powder welding 20–90 –
Plasma spraying 10–70 Different size ranges used
HVOF spraying 10–60 Typical 15–45 mm
HVAF spraying 10–45 Typical 10–30 mm
Cold spraying 10–30 Fine powders in LPCS; coarser in HPCS
PTA weld surfacing 50–150 Coarse powders
Laser spraying/cladding 50–150 Typically PTA grade powders used
LPCS, low-pressure cold spraying; HPSC, high-pressure cold spraying; PTA, plasma transferred arc; HVAF, high-velocity air-fuel.
are the processes that use only wire materials. The wires are either solid wires or tubular wires containing various additives
inside, called cored wires, see Figure 6. Solid wires are manufactured by drawing from materials that will enable such a wire
manufacturing process. The use of cored wires has become more common in recent years. These wires can be made of highly
alloyed iron-, nickel-, or cobalt-based alloys. The principle is that a thin and ductile metal, e.g., stainless-steel strip, is used to
prepare the tubular wire and this metallic tube is packed with alloying elements and/or with hard particles, such as WC or other
carbides, and the strip is then processed to a cored wire. Table 4 shows a summary of the different types of thermal spraying wires
for wear protection. Wire-type feedstock materials are standardized, for example, in EN 14919. Wires for thermal spray processes,
such as wire flame spray and electric arc spraying, are basically based on the following material groups:
l zinc and its alloys (pure Zn, Zn–15%Al);
l aluminum and its alloys (Al 99.0%, Al–5%Mg, Al–12%Si);
l lead alloys (Pb–12.5%Sb–9.2%Sn–0.5%Cu);
www.iran-mavad.com
Figure 6 Different types of thermal spray wires and production method of cored wires. Source: Tampere University of Technology.
Table 4 Wire feedstock materials for wear and tribological applications
Material class Example of composition Coating hardness
Chromium steels Fe þ 13%Cr þ 0.35%C 360–400 HB

Stainless steels Fe þ 17%Cr þ 12%Ni þ 2.5%Mo 140–200 HB
Carbon steels Fe þ 0.1–1.1%C 150–550 HBa
Nickel-based alloys Ni þ 20%Al þ 20%Fe þ 4%Cr 300–350 HB
Molybdenum Mo 99.9% 37–60 HRC
Aluminum bronze Cu þ 9%Al þ 1%Fe 180–200 HB
Brass Cu þ 40%Zn 60–100 HB
Babbitt metals Sn þ 7.5%Sb þ 3.5%Cu þ 0.25%Pb –
a
Depends on carbon content and other alloying.
l tin and its alloys (Sn 99.9%, Sn–7.5%Sb–3.5%Cu–0.25%Pb);

l copper and its alloys (Cu 99.8%, Cu–9%Al–1%Fe, Cu–40%Zn);
l iron and its alloys (Fe–13%Cr–0.35%C, Fe–18%Cr–%8Ni, Fe 29%Cr–3.7%B–1.25%Si);
l nickel and its alloys (Ni–20%Al, Ni–30%Cu);
l molybdenum (Mo 99.9%);
l other metals (Ag–38%Cu–32%Zn); and
l oxide ceramics (rigid stick and flexicord wires) (Cr2O3, Al2O3–TiO2, and other compositions).
Spray wire based on filled wires, i.e., the cord wires, are also widely available. By adding alloying elements and/or hard particles
into the cord wires, coatings can be produced containing high alloying levels in metals. For example, most of the cobalt-based
hardfacing wires, the so-called Stellite compositions, particularly grades other than Stellite 21, are available only as cord wires for
electric arc spraying. Cord wires may also contain hard carbide particles, which are known to produce wear-resistant coatings.
4.10.2.3 Suspensions and Solutions

The use of suspensions and liquids/solutions as feedstock materials is a relatively new approach in thermal spray technology and
has been under more active research and development by various researchers only during the past 10–15 years (8–10). The
motivation for using suspensions or solutions in thermal spraying is the possibility to:
l prepare thin and nanostructured coatings (8–10);
l prepare new materials for specific applications, e.g., for SOFC and photocatalytic applications (10–12); and
l simplify and shorten the feedstock material production chain for thermal spray coating manufacturing by use of solution
precursor spraying (12).
www.iran-mavad.com
Table 5 Suitability of various substrate materials in thermal spray coating processing
Substrate material group Suitability to thermal spraying; remarks
Mild steels, low alloyed, Excellent materials to be coated, prevent oxidation by avoiding heating above 150 C
and steel alloys
Hardened steels Can be coated, high substrate hardness lowers efficiency of grit blasting, for steels <50 HRC
Stainless steels, nickel-based alloys Can be thermally spray coated, high thermal expansion coefficient must be noted, preheat to max. 100 C
Copper and Cu alloys Can be coated, high thermal expansion coefficient must be noted, no preheating due to oxidation sensitivity
Aluminum and Al alloys Can be coated, high thermal expansion coefficient must be noted, no preheating due to oxidation sensitivity
Titanium and magnesium Can be coated, no preheating
Polymers and polymer composites Can be coated but requires low heat input, low temperature resistance, low bond strength,
buffer/intermediate layers recommended
The suspensions are liquids including fine solid particles, which may be nanosized or up to a few micrometers. Both water- and
alcohol-based suspensions are used. Suspensions contain also additives such as plasticizer, as well as dispersing agents. Commonly
used suspension-sprayed coating materials are, for example, Al2O3, TiO2, ZrO2, Cr2O3, and other ceramic materials (8). Also,
WC–12%Co nanostructured coatings have been prepared from suspensions (13). So far, suspension and liquid spraying are not
widely used industrially; instead these types of feedstock materials are undergoing much research and development in several
laboratories. Some evaluations and attempts for industrial use are seen for these types of novel materials.
4.10.2.4 Substrate Materials

Since thermal spraying is in most cases a deposition method that does not increase much the temperature of the substrate during
processing, it allows the use of many different substrate materials other than metals and alloys, even polymers and their composites.
Table 5 summarizes the use of different materials as substrates in thermal spray coating.
4.10.3 Thermal Spray Processes
Thermal spray technology involves several different deposition processes that all use concentrated heat sources to melt the feedstock
materials and to project the melted particles at different levels of kinetic energies toward the surface to be coated. Thermal spray
processes can be based on thermal energies obtained from burning of combustible gases, such as hydrocarbons or hydrogen, or
liquid fuels or can be obtained from electric discharges. Besides actual thermal spray processes, such as flame spray with powder or
wire material, electric arc spraying, plasma arc spraying, and HVOF spraying, processes entirely based on deposition by kinetic
energies are also classified under the general term thermal spray processes. In the following sections, each of the thermal spray
processes will be presented in more detail after a more general classification and comparison of these processes. Thermal spray-
related processes such as plasma transferred arc (PTA) surfacing, laser spraying or laser cladding, and melt spraying are excluded
from this chapter and readers are recommended to read about these techniques from elsewhere, e.g., Refs. (14–16).
4.10.3.1 Classification of Processes

Thermal spray processes are usually classified according to the type of energy source used to melt the feedstock material, as is
presented in Figure 7. The most typical energy sources in thermal spraying are thermal (and/or kinetic) energy obtained from
combustion of gases, typically hydrocarbon or hydrogen, or liquids; from thermal energy obtained from electric discharges such as
electric arcs or ionized plasma gases; or from purely kinetic energy sources in cold kinetic spray process. In addition to these, energy
from molten liquids or high-power laser beams may be used in thermal spray-like deposition of materials. Depending on the type of
energy source, thermal spray processes can be further classified according to the spray gun principle or design, type of feedstock
material used in the process, type of fuel (gas or liquid), type of deposition atmosphere (atmospheric, low/high pressure, inert gas,
under water, etc.), type of oxidizer in combustion, etc. It is notable that in spite of the availability of several more or less differing
techniques in thermal spray processes, they all have their own important niches and only minor overlaps in performance,
investment and operation cost levels, and coating material processability and coating properties exist.
Table 6 is a very general comparison of different thermal spray processes and gives some most important but very generalized
numerical data and information for each one. It should be noted that all values and numbers presented in the table are based on
general interpretation of data from several different sources and partly based on the experience of the author in the topic. The values
and numbers may be highly material, parameter, and process/equipment sensitive and act therefore only as comparative data for
general comparison of the different thermal spray processes. It is obvious that processability of different coatings to obtain high
quality plays a large role in selection of an appropriate thermal spray process. Selection and use of an optimal thermal spray process
for particular material and application may look highly complex and demanding. This is partly true, but understanding the
technological and economical advantages and limitations of different processes allows successful use of these in varying technical
applications. A highly simplified expression of the information from Table 6 allows us to draw conclusions from its contents.
www.iran-mavad.com
Figure 7 Classification of thermal spray coating processes according to the type of energy source used in the process. Source: Tampere University
of Technology.
Table 6 Comparison of characteristics of coatings deposited by different thermal spray processes
Spray Temperature Particle velocity Adhesion Oxide content Porosity Spray rate Relative cost d, Typical coating
method ( C)a (m s1) (MPa)b (%)c (%) (kg h1) low ¼ 1 high ¼ 5 thickness (mm)
Flame 3000 40 8 10–15 10–15 2–6 1 0.1–15

Electric arc 4000 100 12 10–20 10 10–25 2 0.1–15
HVOF 3000 600–800 >70 1–5 1–2 2–8 3 0.1–2
HVAF 2000–3000 600–1200e >70 ***f 0–0.2 Hard metals: 2–30; 2 0.1–12
metals: 2–23g
Detonation 4000 800–1200 >70 1–5 1–2 0.5–2 4 0.05–0.3
APS 12 000 200–400 10–70 1–3 1–5 2–10 4 0.1–1
LPPS/VPS 12 000 400–600 >70 0 <0.5 2–6 5 0.1–1
LPCS 200–650 300–500 5–30 0 <0.5 0.5–3 1 0.2–2
HPCS 500–1000 400–800 10–40 0 <0.5 4–12 4 0.3–4
a
Temperature of the heat source.
b
Depends on deposit material.
c
Oxide content in metallic deposits.
d
Investment cost of process.
e
Depends on equipment type; e.g., M2 AC-HVAF, 600–700 m s1; M3 SAP, 800–1200 m s1.
f
Oxide content is approximately 1.5–2 times the oxide content of the feedstock.
g
Depends on gun type; M2 AC-HVAF spray gun; M3 supersonic spray gun.
Flame – flame spraying, wire and powder method; electric arc – electric arc wire spraying; detonation – detonation gun spraying; HVOF – high-velocity oxygen-fuel spraying;
HVAF – high-velocity air-fuel spraying; LPPS/VPS – low-pressure/vacuum plasma spraying; LPCS – low-pressure cold spraying; HPCS – high-pressure cold spraying/kinetic
spraying; APS – atmospheric plasma spraying.
The temperature of the heat source is important; refractory materials, i.e., high-melting-point ceramics and refractory metals can be
melted only by plasma spray-based processes. Easily oxidizing metals may require spray atmospheres in which oxygen has been
eliminated, i.e., inert gas shroud or low-pressure plasma spray (LPPS) or vacuum plasma spray (VPS) processes. Cold spraying may
be a clear exception, because in this process the powder material does not melt at all, and therefore can be processed to coatings
even in air atmosphere. Particle velocities in different spray processes will clearly have an impact on the coating properties, such as
adhesion to the substrate (bond strength), coating density, and other technical properties including hardness, wear, and corrosion
resistance. The residual porosity in thermal spray coatings is usually lowest in the high-velocity spray processes. Such features as
www.iran-mavad.com
Figure 8 Overview of particle temperatures and velocities for different spray processes; process names from TopGun to K2 refer to various HVOF
systems. Source: H.C. Starck GmbH.
spray rate may act as economic factors influencing the cost of the process and coated part. Investment and operation costs of
different processes definitely have a large influence on the possibility of using such a process in the specific application; other surface
treatments and coating methods may in such cases be more economical. Some industrial applications may allow the use of only
very conventional spray processes, such as flame spraying, due to cost level limitations. Maximum coating thickness available in
different processes may also recommend some techniques instead of others. Some thermal spray processes are more ideal for spray-
specific coating materials. The high-velocity flame spray processes (HVOF, high-velocity air-fuel spraying (HVAF), and detonation
spraying) are more suitable for spraying hard metal powders, such as WC–Co(Cr) and Cr3C2–NiCr, due to their high particle
velocities and more suitable lower flame temperatures. Different thermal spray processes can also be located to specific characteristic
places in the process temperature vs particle velocity charts or maps, as is done in Figure 8. The arrow indicates development trends
in which process temperatures are continuously decreased, while particle velocities show simultaneous increase is a general trend in
the development of novel thermal spray processes, particularly for manufacturing dense and/or hard coatings with improved
mechanical and corrosion properties. The following chapters describing different thermal spray processes in more detail will give
further information about their applicability. The formation of thermally sprayed coatings, examples of coating materials, their
microstructures and properties, and areas of applications are presented in the following chapters.
4.10.3.2 Flame Spraying

Conventional flame spray was the first thermal spray process developed. Both powder and wire feedstock materials are used in flame
spraying. In spite of their long history, these techniques are still widely used particularly due to the low cost of equipment and
processing. However, the importance of flame spray technologies in more demanding applications has significantly reduced as more
advanced processes such as plasma spraying and HVOF spraying have become more predominant. Modern flame spray torches
have changed fairly little since the 1950s. Flame spraying process and other technologies such as HVOF use the chemical energy
of combusting fuel gases to generate heat and consequently accelerate the molten particles toward the substrate. Oxygen acetylene
torches are the most common, using acetylene (C2H2) as the main fuel in combination with oxygen to generate the highest
combustion temperature of approximately 3160 C. Other gases in use are propane (C3H8), propylene (C3H6), hydrogen (H2), and
ethane (C2H4) (14). Powders, wires, or rods are introduced axially through the rear of the nozzle into the flame at the nozzle exit.
The feedstock materials are melted and the particles/droplets accelerated toward the substrate surface by the expanding gas flow and
in some cases also by air jets. Externally combusted, or open-flame, jet temperatures are generally above 2600 C and are controlled
by the combustion temperatures of the fuel/oxygen mixtures and the mixing patterns of the combustion gases with the surrounding
air. The temperature curves for fuel gases are parabolic and centered roughly around the stoichiometric ratio with oxygen. Adjusting
the fuel/oxygen ratio, either side of stoichiometry, will cool the flame. The reason behind the adjustment is to make the flame either
oxidizing or reducing as the feedstock material and application dictate. Flames are set to be reducing (fuel-rich) for most metal
spraying in order to minimize oxidation. Generally, flame spray torches are designed to scavenge part of the oxygen needed for
combustion from the atomizing air or cooling air jets. This design further helps produce metal coatings with lower oxides. Flame
spray is capable of depositing a wide range of materials, ranging from polymers to ceramics and refractory metals.
The different components of typical flame spray systems are (2):
l gas supplies (acetylene and oxygen);
l air supply (compressed air);
l gas hoses;
l gas regulators for oxygen, fuel, and air (if used);
l rotameters for gas flow controls;
l flashback arrestors at the gun and regulators;
l flame spray torch comprising a torch body, nozzle, and atomizing air cap; and
l feedstock delivery system comprising a separate powder feeder, powder hopper, and air turbine drive for wires and rods.
www.iran-mavad.com
Powder is fed into flame spray torches in one of two ways: by carrier gas or by gravity. Gravity-fed devices have powder canisters
or bottles mounted directly to, and on top of, the torch. Powder feed rate is controlled by a pinch valve that meters powder into the
body of the torch, where it is aspirated by the gases flowing through the torch. Carrier-gas-fed units use externally mounted powder
feeders. External powder feeders use a carrier gas (typically nitrogen, air is also used) stream to transport the powder from the feeder
through a hose to the spray torch. Wire- and rod-fed devices use air turbines built into the torch that power the drive rolls, which pull
feedstock from the source and push it through the nozzle.
4.10.3.2.1 Powder Flame Spraying

Figure 9 illustrates schematically the structure of a powder flame spray torch. In flame spray torches, a mixture of acetylene and
oxygen is combusted at the front of the torch, external to the nozzle. In some designs, a constricting flow of air is fed through
a concentric set of holes at the nozzle exit and is used to pinch the expanding gases, generating a higher temperature combustion
zone. The flame spray torch consists of gas distribution passages through the nozzle and mixing orifices at the nozzle end. Nozzles
are used to shape the spray pattern. Powder flame spray processes are used widely due to the low equipment and processing cost.
Typical materials sprayed are iron, nickel, cobalt, copper, and some other pure metals and their alloys. Due to relatively low particle
velocities, the coatings show only moderate properties, which are definitely adequate for repair and restoration and also for various
new components. In addition, ceramic powders are sprayed by powder flame spraying, but the resulting coating properties again do
not reach those of plasma-sprayed coatings. Powder flame spraying is the most used process to spray nickel- and cobalt-based self-
fluxing alloy powders, which are not used in the as-sprayed state, but instead posttreated by a fusing/sintering process. These alloys
are typically NiCrBSi with varying amounts of chromium and carbon, and the self-fluxing additives boron and silicon. The methods
used in posttreatment are fusing by oxyacetylene torch, furnace sintering, induction fusion, or other techniques such as laser
treatment.
4.10.3.2.2 Wire Flame Spraying

Figure 10 shows schematically the structure of a wire flame spray torch. Basically the gun is similar to a powder flame spray torch, but
in wire flame spraying, special air caps are added to these torches, producing a concentrated air jet to atomize the tip of the molten
wire or rod. The wires and the rods are mechanically fed through the center of the nozzle into the heating zone. Oxide and metal
rods and wires can be used as coating feedstock. One significant advantage of wires and rods over powders is that the degree of
melting is significantly higher than in powder flame spraying, producing denser coatings. All material that is emitted from the hot
wire tip is fully molten. In powder flame spraying process, some powder particles, particularly the largest ones, do not necessarily
melt fully and may cause porosity formation in the coating and may also lower the deposition efficiency (D.E). D.E. in thermal
spraying is the ratio of coating deposited divided by the amount of feedstock material fed to the spray torch (excluding off-sprayed
coating material).
Figure 9 Schematic presentation of a flame spray gun using powder feedstock material. Source: Tampere University of Technology.
Figure 10 Schematic presentation of a flame spray gun using wire as feedstock material. Source: Tampere University of Technology.
www.iran-mavad.com
In the wire flame spraying process, the atomizing air can produce finer droplets, which in turn may produce finer and smoother
coatings. In flame spray processes, fuel/oxygen ratio and total gas flow rates are adjusted to produce the desired thermal output
needed to melt the specific feedstock material. Optional air jets, downstream of the combustion zone, may also further adjust the
thermal profile of the flame. Spray gas speeds typically are below 100 m s1, generating particle speeds up to approximately
80 m s1 before impact.
4.10.3.2.3 Spray and Fuse Process

Spraying with a postheat treatment by sintering/fusing can be carried out to obtain dense coatings with metallurgical bonding
(diffusion bonds). This thermal spray processing consisting of two separate stages – flame spraying and postfusion – is clearly
different from other thermal spray processes, which are one-step processes and in which the coating adherence to the substrate
material is mainly of the type of mechanical anchoring and bonding to the substrate surface asperities created by grit blasting of the
surface prior to thermal spray coating. The fusing is done with an oxyacetylene flame, inductive heating, or in a furnace (gas furnaces
or electric furnaces with protective gas atmosphere). The fusing occurs at a temperature where only deposited material can melt to
close interparticle gaps and where fluxing, i.e., removing of the oxides occurs. The fluxing elements, particularly boron and silicon,
have direct influence on the wetting behavior of the practically molten coating material on the substrate. When the fusing stage is
complete, the surface has a highly reflective, glassy look. These types of coatings are homogeneous with very good bonding strength,
350–500 MPa. Because of the high fusing temperature (approximately 1050 C), there is a risk of deformation and structural
transformation of the substrate. One type of fused layer is made by simultaneous fusing, so-called spray welding. The bond strength
is high and comparable to brazing. Figure 11 presents the fusing stage of a flame spray-coated sleeve with oxyacetylene torch.
4.10.3.3 Electric Arc Spraying

Figure 12 presents an electric arc two-wire spray gun schematically. Arc spraying is used to apply only electrically conductive
materials including different metals, metal alloys, and metal–metal oxide or metal–carbide mixtures (cord wires). In arc spraying,
an electric arc between two wires is used to melt the coating feedstock material. Compressed gas, usually air, is used to atomize and
propel the molten material to the substrate. Sometimes air is replaced with nitrogen to lower the degree of oxidation of the metallic
coating material. An electric arc spray system consists of a high-power direct current (DC) power supply, two wire coils (which can
Figure 11 Postfusing operation for flame-sprayed NiCrBSi self-fluxing alloy coating with oxyacetylene torch. Source: Kuopion Konepaja Oy.
Figure 12 Schematic presentation of an electric arc wire spray gun using two electrically conducting wires as feedstock material. Source: Tampere
University of Technology.
www.iran-mavad.com
be different material), wire feeding mechanisms, wire feeding hoses, electrical conducting cables, and the arc spray gun. The two
wires are continuously fed to the gun at a uniform speed. A low-voltage (18–40 V) DC power supply is used, with one wire serving as
the cathode and the other as the anode. Coating quality and properties can be controlled by varying the atomization pressure, air
nozzle shape, power, wire feed rate, traverse speed, and spray distance (standoff distance). Arc spraying is generally the most
economical thermal spray process for applying corrosion-resistant metal coatings, including zinc, aluminum, and their alloys.
Energy costs are lower and production rates are significantly higher than they are in competing methods such as wire flame spray.
4.10.3.4 Plasma Spraying

Plasma spraying process is a widely used method of manufacturing very different types of coatings. The process due to its high-
temperature heat source, a thermal plasma, makes it possible to melt practically all possible material including all ceramic materials
and refractory metals. A thermal plasma, often called the fourth state of matter, normally consists of neutral atoms, positive ions,
and free electrons. Plasma is produced by transferring energy into a gas until the energy level is sufficient to ionize the gas, allowing
the electrons and ions to act independent of one another. The plasma state is achieved when, under an electric field, currents can be
sustained as the free electrons move through the ionized gas. Once the energy input is removed, the electrons and ions recombine,
releasing heat and light energy. Figure 13 shows the enthalpy vs temperature behavior of common plasma-forming gases in thermal
spraying. Monoatomic gases such as argon (Ar) and helium (He) ionize directly from atomic state, whereas diatomic gases such as
nitrogen (N2) and hydrogen (H2) first dissociate and then become ionized. Argon and nitrogen are used in plasma spraying as the
primary plasma gases and hydrogen and helium as secondary gases. The combination of argon/hydrogen is common; argon ionizes
easily and hydrogen significantly increases the enthalpy enabling efficient melting capacity of the plasma torch. Hydrogen addition,
typically only up to 25% to prevent electrode and nozzle wear, also increases the thermal conductivity of the gas mixture and allows
increase of the arc voltage for higher plasma power levels.
4.10.3.4.1 Atmospheric Plasma Spraying

Plasma spray torches are the most flexible of the thermal spray processes with respect to materials that can be sprayed, because of the
inert gases, high gas velocity, and extremely high temperatures. Plasma spray processes operating in ambient atmospheres are called
atmospheric (or air) plasma spraying (APS) processes. A schematic structure of a typical DC plasma gun is illustrated in Figure 14.
Such plasma torches consist of a circular anode/nozzle made of copper and a cathode made of thoriated tungsten. Powder feeding is
Figure 13 Enthalpy vs temperature behavior of common plasma-forming gases in plasma. Source: Tampere University of Technology.
Figure 14 Schematic presentation of a plasma spray gun using DC power for plasma forming. The powder is injected radially to the plasma jet.
www.iran-mavad.com
Figure 15 Mixing of the surrounding air with the plasma gas as function of distance from plasma gun nozzle exit. Reproduced from Matejka, D.;
Benko, B. Plasma Spraying of Metallic and Ceramic Materials; John Wiley & Sons, 1989; p 280.
usually realized by pneumatically injecting fine spray powder radially to the plasma jet. The electric arc discharge, supported by
a powerful DC generator (80 kW and above) through the connectors heats up by ionization the plasma forming gases. When the
ionized gas returns back to atoms/molecules by recombining, a high amount of energy is released. The gas expands in the atmo-
sphere and forms an extremely hot and high-velocity gas jet. The particles of the powder after being melted and accelerated in the jet
impact the substrate and form the coating. DC plasma guns are intensely water cooled to prevent melting and to minimize cathode
and nozzle wear. The plasma forming gases are introduced through the rear of the gun. Typically, these gases enter the arc chamber
through a gas ring (gas injector) that imparts a spin or vortex flow to the gas. The vortex stabilizes the arc at the cathode tip in the
low-pressure region of the vortex and rotates the arc attachment or arc root at the anode. This is called vortex stabilization. The
plasma forming gases are heated as they pass around and through the arc. The heated gases expand radially and axially, accelerating
as they expand and exit through the nozzle. Depending on the pressure ratio between the upstream arc region and the downstream
nozzle exit of the gun, the gases (plasma jet) either expand subsonically or supersonically. Many gun designs use converging/
diverging de Laval nozzles to achieve supersonic expansion, which increases significantly the particle velocity, thus improving the
coating properties.
Most plasma spray equipment operates in atmospheric conditions, i.e., some air is entrapped into the hot plasma jet causing
some interaction between the fast-flying hot spray powder particles and the surrounding gas atmosphere. Figure 15 shows this
behavior schematically. Therefore, it is apparent that sprayed coatings also tend to react with oxygen from the entrapped air in the
otherwise inert gas atmosphere of plasma spraying. In order to totally prevent the interaction of the molten spray particles and
oxygen from air, special plasma spray processes called vacuum or low-pressure plasma spraying has been developed.
Atmospheric plasma spraying has some separate modifications from the conventional 40-kW level plasma spray processes, e.g.,
high-power plasma spray (HPPS), axial plasma spray, and water-stabilized plasma spray (WSP) processes. These technologies will
be reviewed later in this chapter.
4.10.3.4.2 Low-Pressure Plasma Spraying

LPPS/VPS processes were developed in 1974 by E. Muehlberger. This new process demonstrated that very clean coatings with
practically no oxide inclusions could be produced. The main difference of LPPS/VPS processes compared with APS is that these
processes are carried out inside a vacuum chamber, which fully eliminates the interaction of molten spray particles with oxygen.
Main uses of LPPS/VPS are in manufacturing oxidation-sensitive coating materials such as MCrAlYs (M ¼ Ni, Co, NiCo, CoNi;
sometimes Fe) and reactive metals such as titanium (Ti) and tantalum (Ta) coating. Such coatings when sprayed in chambers have
porosities significantly below 1% and are free of oxide inclusions, with clean substrate/coating interfaces. The LPPS/VPS spray
equipment consists of a conventional plasma spray gun equipped with a nozzle modified for higher pressure expansion ratios.
These high ratios are achievable because the plasma jet exhausts into a chamber at below atmosphere pressure, typically at the level
of 100 mbar. The closed environment of the chambers necessitates remote manipulation of the spray guns, for example, with an
industrial robot, and a door in the chamber or load-lock systems to transfer parts into the chamber. Other unique aspects of
chamber spraying include vacuum pumping, exhaust cooling, filtration, and chamber wall cooling.
LPPS results in the following process and coating characteristics relative to conventional APS (2):
l broader and longer spray jets;
l cleaner interfaces, made possible by reverse transferred arc sputter cleaning;
l oxide-free coatings;
l high coating densities, close to theoretical;
www.iran-mavad.com
l low residual stresses;

l increased deposition thickness capability (>25 mm); and
l high substrate/deposit preheating capability: greater than 1000 C prior to and during coating.
4.10.3.4.3 High-Power and Special Plasma Spray Processes

Besides APS and LPPS/VPS plasma-based processes, there exist some variants and developments. In order to increase the powder
feeding rates for more efficient and productive plasma spraying, technologies such as HPPS (17), axial plasma spraying (2), and
WSP (18) have been developed during the past years. The main driving forces in developing such new processes have been high-rate
spraying obtainable by high powder throughputs (high feeding rates), high D.E. to reach high powder usage, and high process
stability for uniform melting of the spray material. The solutions have been primarily the use of high power levels up to 200 kW
(normal APS has 40-kW gun power), developing improved stability of the plasma arc by new gun designs, and getting optimal
powder utilization by new powder injection principles, e.g., axial powder feeding by special multielectrode gun designs.
Figure 16 is a schematic presentation of an HPPS torch, in which nitrogen is the primary gas (argon can not be used in this
process) and hydrogen is the secondary plasma gas; the gas flow rates are high (300–400 l min1; in conventional APS approxi-
mately 50 l min1), the DC power level is high (165–220 kW), and the powder spray rates are increased from those of conventional
APS (100–200 g min1 vs 40 g min1).
In a very specific plasma spray process, the WSP using water as plasma-forming media, the cathode is made of graphite and the
anode is an internally water-cooled rotating copper disk located in front of the gun nozzle. The main differences of WSP compared
with APS are the plasma-forming media used (gas vs water), significantly different torch design, plasma temperature, power level in
process (160 kW vs 40 kW), and powder feeding rates. A comparison of these processes is presented in Table 7 (19).
There exist basically three different variants of conventional plasma spray torch, where the plasma guns are based on using
more than one electrode; typically three electrodes (cathodes) are used in plasma spray torches such as Axial III and Triplex. The
first one is based on using three pairs of separate cathodes and anodes, which are located symmetrically around the center powder
feeding nozzle, allowing powder feeding axially directly to the plasma plume. Deposition efficiencies of the order of 75–90%
(conventional APS has typically 40–60%) combined with high powder feeding rates makes such technologies attractive for
industrial plasma spraying (Table 8). It is also evident that in all multielectrode plasma torches (Axial, Triplex) the stability of the
plasma arc discharge is much improved from conventional single-electrode plasma torches enhancing more uniform powder
melting in the process; the single-electrode plasma torch may show significant instabilities of the plasma arc, particularly when the
Figure 16 Schematic presentation of a high-power plasma spray torch (Plazjet), which consists of an extended nozzle/anode. Source: Praxair
Surface Technologies Inc.
Table 7 Comparison of conventional gas-stabilized and water-stabilized

plasma spray processes
WSP Conventional plasma
Plasma forming media Water Gas (Ar, H2, N2, He)

Arc current (A) 500 380–400
Arc voltage (V) 300–320 68–75
Plasma power (kW) 160 40
Plasma temperature (K) 30 000 10 000–15 000
Spray distance (mm) 350 100
Injection distance (mm) 30 6
Powder feed rate (kg h1) 20–50 3–6
Modified from Sampath, S.; Gansert, R.; Heramn, H. Research Summary: Plasma-spray
Forming Ceramics and Layered Composites. JOM (USA) 1995, 47 (10), 30–33.
www.iran-mavad.com
Table 8 Deposition efficiencies and powder feeding rates obtainable with axial powder feeding plasma torch
Material Feed rate (kg h1) Feed rate (g min1) Deposition efficiency, D.E. (%)
Hard metals (cermets)

WC–17%Co 9–13.5 150–225 75–90
WC–12%Co 9–13.5 150–225 65–80
WC–Cr3C2–Ni 7.2–11.4 120–190 75–90
Cr3C2–NiCr 3.6–6.9 60–115 65–90
Ceramics (oxides)
Al2O3 1.8–4.5 30–75 75–90
Al2O3–13%TiO2 2.7–5.4 45–90 75–90
ZrO2–8%Y2O3 2.1–3.6 35–60 60–85
Cr2O3 1.8–5.4 30–90 45–75
Abradables
AlSi–Polyester 1.8–3.6 30–60 70–90
Al–Si–Graphite 1.8–2.7 30–45 45–50
Metals and alloys
AISI 316L stainless steel 4.5–11.4 75–190 70–90
Titanium (c.p.) 1.8–4.5 30–75 75–90
NiAlMo 8.1–9.9 135–165 75–90
NiCoCrAlY 4.5–9 75–150 75–90
Copper 9–10.8 150–180 75–90
Molybdenum 9–18 150–300 75–90
Source: Northwest Mettech Corp.
electrode pairs are already worn by arc erosion, which may lead to nonuniform melting of the spray powder. This will lead to
lowered D.E. and changes in the coating quality.
The conventional technique to ionize gases for plasma spraying is the use of DC plasma configurations. However, induction coils
powered with high frequency can also be used to ionize gases for plasma spraying and other material processing. Besides the type of
power supply, some other significant deviations from DC plasma torches exist in inductively coupled plasma torches. The gas
velocity in induction plasma is significantly lower than that in DC plasma torches; larger powder particles have to be used in
induction plasma processes. The torch design is quite different. Induction plasma spraying systems are not used to such an extent as
DC plasma torches for coating production, due to bulky equipment and the lower coating properties achievable. Examples of
applications are powder spheroidization and production of coatings for highly specific applications.
4.10.3.5 High-Velocity Flame Spraying

High-velocity flame spraying (HVFS), also more frequently called HVOF spraying involves nowadays several different combustion
flame spray processes in which a gaseous or liquid fuel is combusted in a combustion chamber of the spray gun with oxygen or air
as oxidizer. Particularly typical of these processes is the very high gas velocity leading to high particle velocities of the order of
500–800 m s1, depending on the type of spray gun and moderate particle temperatures. In such spray processes, the thermal energy of
the process is partly replaced with the high kinetic energy, i.e., increased particle velocities. This means that the particle temperatures are
not excessively high as in many other spray processes but are kept as low as possible at a level in which deposition still occurs with
acceptable D.E. The particle velocities in HVOF processes are markedly higher than in other conventional spray processes, e.g., in plasma
spraying. High particle velocities combined with moderate particle temperatures result in coatings that are dense, well-adhered to the
substrate, and show a low amount of oxidation in the case of metallic and hard metal coatings. Since HVFS processes were primarily
developed to spray tungsten carbide-based cermet (WC–Co), another obvious benefit was the high amount of retained carbides (WC)
in the structure resulting in markedly improved mechanical and wear properties versus, for example, plasma sprayed coatings.
HVFS involves techniques such as high-velocity oxygen-fuel (HVOF) spraying, high-pressure high-velocity oxygen-fuel (HP-HVOF)
spraying, HVAF spraying, high-velocity impact-fusion spraying, and other novel techniques, e.g., the warm spraying process.
Basically all techniques are similar, but they differ from each other in the gun design, type of fuel (gas or liquid), type of oxidizer
(oxygen or air), particle temperature and velocities, and in other less important characteristics. HVFS plays today very important role
in the thermal spray process palette and the most important spray techniques are HVOF, HP-HVOF, and HVAF. In the next section,
the HVOF (includes here also HP-HVOF), HVAF, and warm spraying are described in more detail due to such factors as the
importance and the novelty of the processes. Also, a hybrid method composed of HVFS and electric arc spraying is described
separately in its another chapter. The detonation gun spray process (DGS or D-Gun), a traditional high-velocity flame spray process
known already for a long time, is also described due to its similarity with the later developed spray techniques.
Table 9 presents some important properties, such as maximum temperature when combusted with oxygen and the corre-
sponding heat of combustion for different fuel gases and a liquid fuel frequently used in HVOF and similar high-velocity flame
www.iran-mavad.com
Table 9 Properties of different fuel gases and a liquid fuel used in HVOF and other similar processes
Maximum flame Heat of combustion Oxygen/fuel ratio Oxygen/fuel ratio Oxygen/fuel ratio
Fuel temperature ( C) (MJ m3) for max. flame temperature for neutral flame for HVOF applications
Propane 2828 93.2 4.5 5.0 3.0–8.0

Propylene 2896 87.6 3.7 4.5 3.5–7.0
Hydrogen 2856 10.8 0.42 0.5 0.3–0.6
Ethylene 2924 59.5 2.4 3.0 2.0–5.0
Acetylene 3160 56.4 1.5 2.5 1.3–4.0
Kerosene ca. 2900 37.3 MJ l1 2.9 3.4a 2.8–4.8a
a
Standardized mass ratio oxygen/kerosene.
Reproduced from Kreye, H.; Gärtner, F.; Kirsten, A.; Schwetzke, R. High-velocity Oxy-fuel Flame Spraying State of the Art, Prospects and Alternatives. In 5. Colloquium on High
Velocity Flame Spraying; GTS e.V.: Erding, Nov. 16–17, 2000.
Figure 17 Influence of oxygen/fuel ratio l on flame temperature (l ¼ 1 for stoichiometric combustion). Reproduced from Kreye, H.; Gärtner, F.;
Kirsten, A.; Schwetzke, R. High-velocity Oxy-fuel Flame Spraying State of the Art, Prospects and Alternatives. In 5. Colloquium on High Velocity Flame
Spraying; GTS e.V.: Erding, Nov. 16–17, 2000.
spray processes. Figure 17 shows the influence of normalized oxygen–fuel ratio on the temperature of the flame for gaseous and
liquid fuels in HVOF spraying (20). It is evident that the choice of fuel type has a marked influence not only on the obtained
maximum temperature but also on the cost of HVOF spraying. It should be mentioned that acetylene is seldom used in HVOF
spraying, but in turn is the main fuel gas in powder and wire flame spraying processes. Flame temperatures in HVOF processes can
be easily adjusted by changing the ratio of the fuel gas and oxygen flow rates. Combustion of liquid fuel, and other hydrocarbon fuel
gases, with compressed air in the HVAF process lowers significantly the flame temperature and may result in significant
improvements in coating quality. It is known that hard metal coatings, e.g., WC–Co(Cr), and metallic coatings show in HVAF-
sprayed coatings better properties, such as higher amount of retained carbide phases for improved wear resistance and ductility and
lower level of oxidation for improved corrosion characteristics.
www.iran-mavad.com
Figure 18 Schematic presentation of an HVOF spray torch. Source: Tampere University of Technology.
Figure 19 Influence of type of fuel gas and spray distance from the HVOF gun exit on measured powder particle temperatures and their average
velocities. Reproduced from Nestler, M. C.; Benary, R.; Rusch, W. Innovations in HVOF Coating Solutions. In 5. Colloquium on High Velocity
Flame Spraying; GTS e.V.: Erding, Nov. 16–17, 2000.
4.10.3.5.1 HVOF Spraying

Figure 18 shows an HVOF spray gun schematically. HVOF spray guns consist of a fuel gas/oxygen mixer, a combustion chamber,
and an expansion nozzle, which is usually in modern HVOF guns a converging/diverging de Laval-type nozzle. Such nozzles lead
to high gas velocities, which in turn accelerate the powder particles to very high velocities of the order of 600–800 m s1
(Figure 19) (21). The HVOF guns may also be either water or sometimes only air cooled. Full HVOF spraying systems consist of
different units such as (2):
l gas supplies (fuel gas/liquid fuel, oxygen);
l air supply (compressed air);
l gas hoses;
l gas regulators for oxygen, fuel, and air (if used);
l rotameters or mass flow controllers for gas flow controls;
l flashback arrestors at the gun and regulators;
l HVOF spray torch composed of a torch body, gas mixer, combustion chamber, de Laval nozzle, and gun cooling;
l powder feeding system; and
l spray gun manipulator/robot (unless manually operated).
4.10.3.5.2 Warm Spraying

Warm spray is a novel method that aims to compete with cold spray and HVAF in depositing dense coatings without changing the
phase or chemical compositions of the feedstock. Warm spray is very similar to HVOF with the difference of an inert gas (nitrogen)
being introduced to the combustion jet before the powder in a mixing chamber, thus allowing more control over the temperature
and velocity of the supersonic jet. The temperature of the jet is lower than in corresponding HVOF processes and the velocities are
same as in HVOF processes. The melting temperature of most materials is approximately 800–1500 K, and the warm spray system
enables the control of the jet temperature very specifically depending on the material. Warm spraying has been primarily used to
spray WC–Co(Cr) and some metallic powders. When spraying carbides, it has been observed that warm spray does induce some
decarburization of WC, although significantly less than the HVOF process (22–25).
4.10.3.5.3 HVAF Spraying

HVAF spray process can be regarded as a modification of HVOF (26,27). The HVAF uses a mixture of air and a fuel gas in the
combustion chamber and a second mixture of air and fuel in the nozzle to propel the powder feedstock from the gun to high
www.iran-mavad.com
Figure 20 An illustration of a third-generation HVAF spray torch (M3 type) using hydrocarbon gas and air as process gases. Source: Uniquecoat
Technologies Inc.
velocities. Since the HVAF uses air instead of oxygen for the combustion, the operating costs are significantly lower due to safer and
more controllable combustion, which also lowers expenses on security, and to the notably lower cost of air compared with oxygen.
Additionally, the powder feeding rates are markedly higher than in HVOF systems. The main advantages regarding coating qualities
arise from lower flame temperatures, which minimizes the amount of dissolved carbides in WC–Co(Cr) coatings and produces less
oxidized metallic coatings. The feedstock used in HVAF guns is powder. In activated combustion high-velocity air-spraying
(AC-HVAF), a hot catalyst is used to stabilize the combustion and make it even. Recent types of HVAF torches such as the third-
generation equipment that utilize supersonic air-fuel technology can have particle velocities well above those of the previous HVAF
torches and other HVOF torches that are still in use. The recognized advantages of this method are high coating density, due to high
velocities, and low oxidation, due to lower temperature, of the coating. An illustration of a third-generation HVAF torch is presented
in Figure 20.
4.10.3.5.4 High-Velocity Combustion Arc Spraying

High-velocity combustion arc spray (HV-Arc) process is a hybrid thermal spray process, which consists of a special electric arc spray
torch integrated in a high-velocity air-fuel combustion chamber and spray nozzle, as is schematically presented in Figure 21. In
comparison with the traditional electric arc spray process, the HV-Arc process has higher particle velocities, the molten material is
atomized by the fast gas jet into finer droplets, and the coatings consist of densely packed fine splats producing better coating
properties. Propane is a typical fuel gas in the HV-Arc spray process (28).
4.10.3.5.5 Detonation Gun Spraying

Detonation gun spray equipment is basically a long water-cooled barrel with inlet valves for gases and powder (Figure 22). Oxygen
and fuel gas, acetylene being the most common one, is fed into the barrel along with a charge of powder. A spark plug is used to
ignite the gas mixture and the resulting detonation heats and accelerates the powder to supersonic velocity down the barrel. A pulse
Figure 21 Schematic structure of a high-velocity arc spray torch. Source: Uniquecoat Technologies Inc.
www.iran-mavad.com
Figure 22 Schematic structure of a detonation gun spray equipment. Source: Tampere University of Technology.
Table 10 Temperatures of the detonation products with various

gas mixtures. Amounts of components in mixture, %
C2H2 O2 N2 Td.p ( C)
38 62 0 2900
33 55 12 2550
28 47 25 2200
24 39 37 1900
20 31 49 1600
Reproduced from Astakhov, E. A.; Klimenko, V. S.; Skadin, V. G.; Zverev, A. T.
Sov. Powder Metall. Met. Ceram. 1982, 21, 307–310.
of nitrogen or air is used to purge the barrel after each detonation. This process is repeated many times a second. The high kinetic
energy of the hot powder particles on impact with the substrate results in the buildup of a very dense and strong coating. The firing
(explosion) frequency of various detonation gun systems can be in the range of 1–50 Hz (shots per second), but is typically in the
range of 5–10 Hz (29,30).
In detonation gun spraying, the flame temperature is easily controlled by varying the oxygen/fuel gas ratio and/or the amount of
diluent gas (nitrogen or air) in the gas mixture. Table 10 shows the influence of a diluent gas (nitrogen) on the flame temperature of
an acetylene/oxygen mixture.
The development of HVOF technology and particularly the availability of HVOF spray equipment has lowered markedly the
industrial importance of DGS. The main drawbacks of the detonation spray technology has been its poor availability, particularly in
the 1980s and the 1990s, and low production rates (low powder feed rates) typical to this process.
4.10.3.6 Cold Spraying

Cold spraying was developed in the former Soviet Union in the 1980s. Figure 23 is a schematic presentation of the cold kinetic spray
equipment. Cold spraying is a thermal spray process that enables production of metallic and metallic–ceramic coatings with dense
(very low porosity level) and pure (low oxygen content) structures. The process differs entirely from all other thermal spray
processes, because the powder feedstock material is always unmelted; the process is a solid-state spray method.
Cold-sprayed coatings have some specific characteristics compared to other thermal spray processes. Due to solid-state pro-
cessing combined with high particle velocities, the metallic coatings are pure, i.e., free of oxides, and noticeably dense. In cold
spraying, the heat input not only to the spray material but also to the substrate is significantly low, not changing the substrate
Figure 23 Schematic presentation of cold kinetic spray process. Source: Linde AG.
www.iran-mavad.com
Table 11 Comparison of the characteristics of high-pressure and low-pressure cold

spray processes
Parameter HPCS LPCS
Process gas N2, He, or mixtures Air (or N2, He)

Pressure 10–50 bar 6–10 bar
Gas preheat temperature RT: 800–1000 C RT: 650 C
Gas flow rate 0.85–2.5 m3 min1 N2 0.3–0.4 m3 min1
Max 4.2 m3 min1 He
Powder feed rate 4.5–13.5 kg h1 0.3–3 kg h1
Spraying distance 10–50 mm 5–15 mm
Electric power 17–47 kW Max. 3.3 kW
Particle size 1–50 mm 5–30 mm
properties. The temperature of the gas is well below the melting point of all sprayed material and thus particles are not melted in the
gas flow. Some preheating of the spray powder particles is sometimes beneficial in order to soften the particles for improved
deformation and flattening. Furthermore, residual stresses are usually compressive due to the shot peening effect of impinging solid
particles.
Cold spraying consists of two differing processes, which are high-pressure cold spraying (HPCS) and low-pressure cold spraying
(LPCS) divided by the pressure level used in the processes (40 bar vs max. 10 bar). Generally, cold spraying is based on higher
particle velocities and lower process temperatures than in other thermal spray processes. The coating is formed in a solid state when
powder particles impact on a sprayed surface with high kinetic energy, deform, and adhere to the substrate or to other particles.
Therefore, a high level of plastic deformation and adiabatic shear instability are required for a tight bonding between powder
particles and for the formation of dense microstructures. Moreover, in cold spraying, many factors, e.g., powder characteristics and
compositions, spraying parameters, and posttreatments affect the formation and properties of the coating. Table 11 compares the
two types of cold spray processes and shows their main parameters and characteristics.
Cold spray processes and coatings have been reviewed in several books, journals, and conference papers. The reader is rec-
ommended to search for more detailed information about the topic, e.g., from Refs. (31–38).
4.10.3.6.1 High-Pressure Cold Spraying

Cold spraying is divided into two different processes based on the pressure level used: HPCS and LPCS processes. The HPCS process
enables the use of pressure as high as 40 bar and preheating temperature of max. 800 C; some new equipment reach the level of
1000 C. Typically, pure metal and metal alloys are used as powder materials in the HPCS process. Furthermore, the selection of
powder materials is larger with the HPCS process compared to the LPCS process. Typically, metals and metal alloys, e.g., Cu, Al, Ta,
Ni, Cu–Sn, NiCr, and Ni–Al are used as powder materials. Figure 24 shows a schematic presentation of the HPCS gun. Basically, in
the HPCS process, highly pressurized gas (N2, He, or mixture of these gases) is heated and flows to the nozzle to which powder is fed
with a carrier gas. Then, the gas–powder flow is fed through the nozzle in which high velocity is generated. After that, particle–gas
flow exits the nozzle with high speed and impacts on the sprayed surface. The HPCS process is controlled with a control unit.
In HPCS, a converging–diverging nozzle is used to produce supersonic gas flow in which particles achieve supersonic velocity.
Supersonic velocity is created in the nozzle throat, in the diverging part of the nozzle. Powder is fed with a carrier gas in the axial
direction and then mixed with the main gas forming a particle–gas flow before the nozzle. A de Laval-type converge/diverge nozzle
is the most commonly used. The schematic presentation of one cold spray gun with de Laval nozzle is shown in Figure 24. The
colors indicate temperatures in different regions of the gun, from warmer (red) input to colder (blue) output.
4.10.3.6.2 Low-Pressure Cold Spraying

In the LPCS process, preheating temperatures of the process gas are between room temperature (RT) and 650 C, and pressures are
between 5 and 9 bar. Usually, compressed air is used as process gas to spray powder mixtures. Typically, the LPCS process is
a method to spray metallic powders (e.g., Cu, Al, Ni, Zn) with an addition of ceramic powder (Al2O3, SiC) as powder mixtures.
Figure 24 Schematic presentation of cold spray gun. Colors indicate temperatures; in the red areas temperature is the highest and in the blue areas
temperature is the lowest. Source: Linde AG.
www.iran-mavad.com
Figure 25 Schematic presentation of spraying gun in LPCS process. Powder is fed radially to the diverging part of the converging–diverging-type
nozzle. Some LPCS equipment have nearly barrel-type nozzles. Source: Obnisk Centre for Powder Spraying.
The main functions of ceramic powder addition are to activate the sprayed surface and to keep the nozzle of the gun clean. In
addition to these, ceramic particles affect the coating by mechanical hammering of the substrate/sprayed layers or by the so-called
shot peening via particle impacts. The ceramic particle addition in the powder mixtures has also a compacting effect during the
impacts, indicating improved coating properties and D.E. of metallic coatings. The hard phase can also be used for reinforcing and
strengthening the metallic matrix in the LPCS process.
The main difference between the spray guns used in the HPCS and LPCS processes is that in the LPCS process, powder is fed to
the gas flow perpendicularly (radial injection) to the diverging part of the nozzle, as presented in Figure 25. In addition to that,
Papyrin et al. (32) have presented the possibility of injecting two powders, e.g., Cu þ Al, and Cu þ SiC, separately to the nozzle (two
powder feeders at the same time) to form composite coatings. Additionally, another difference between the LPCS and HPCS
processes is that heating of gas is done only in the spraying gun in the LPCS process, whereas in the HPCS process, the gas is
preheated in a separate heating unit and also, in addition to that, in the spraying gun. Although powder feed rates are lower in the
LPCS process, particle concentration in gas flow can be higher than in the HPCS process.
4.10.3.7 Suspension and Solution Spraying

In addition to cold spray processes, the use of liquids, i.e., suspensions or solutions, is one of the latest thermal spray technologies
under active research and evaluation for future industrial use (8). Suspension and liquid spraying is usually done by using a plasma,
HVOF, or flame spray torch as the heat source for the process. However, in order to be able to spray suspensions, specific hardware
components are needed. These are a suspension feeder and a suspension injector. Nowadays suspension and solution spraying is
done with only minor modification of the thermal spray torch itself; i.e., guns initially designed for this purpose do not generally
exist. An exception maybe a liquid flame spraying process developed for clearly other purposes than traditional thermal spray
coating. Figures 26 and 27 illustrates two different suspension spray processes; one based on plasma spraying and one on HVOF
spraying techniques, respectively (8–10).
Coatings sprayed using liquids or suspensions have much finer surfaces than those sprayed using powders. They are also thinner
and need more combustion or electrical energy to be sprayed. This results from the fact that a considerable part of the energy is used
to evaporate the liquid. The most important area of application of solution or suspension thermal spraying are nanostructured
coatings that have improved mechanical or electrical properties. An important advantage is the possible cost saving. The smooth
surfaces of suspension-sprayed coatings may render it unnecessary to use postspray mechanical finishing. Suspension spraying may
be the last step in powder spraying as the spray torch with the sprayed material remains the same. The resulting coatings would have
a gradient of crystal grain size.
Figure 26 Schematic presentation of suspension thermal spray torch-based plasma spraying with radial liquid feeding. Source: Technical University
of Stuttgart
www.iran-mavad.com
Figure 27 Schematic presentation of suspension thermal spray torch-based HVOF spraying with axial liquid feeding. Source: Technical University of
Stuttgart.
4.10.4 Formation of Sprayed Coatings
The formation of thermally sprayed coatings is fairly unique compared with many other deposition technologies. In thermal
spraying, the coatings are built up from individual spray material particles that impact on the substrate surface, spread, and solidify
rapidly. The particles can be in fully or partly molten state at the moment of impact in all spray processes where the feedstock
material (powder or wire) is aimed to be melted or in solid state in all cold spray processes. The thermal history that the spray
particles experience when traveling in-flight from the spray torch to the surface of the substrate highly influences the structure and
therefore also the properties of the coatings. Such factors as the type of the spray materials and the spray process, parameters of the
spray torch, kinematic parameters, and coating procedures all have their marked influences on the resulting coating structures and
properties. In the next section, the formation of thermally sprayed coatings for material melting type of processes, solid-state spray
processes, and spray processes involved with posttreatments is presented. Figure 28 shows the variety of process parameters in
thermal spraying influencing the structure and properties of the coatings. It is evident that thermal spraying process consists of
a large number of influencing parameters and factors, which are also in many cases interrelated.
4.10.4.1 Coatings from Molten Droplets

Most thermal pray processes involve melting of the feedstock material and spraying the particles in molten state. Cold spraying, and
similar processes, which aim to process the material in solid state, are a clear exception. In thermal spraying with molten particles,
the coatings are built up from individual spray material particles that impact on the substrate surface, spread, and solidify rapidly.
The particles are primarily in a fully molten state at the moment of impact, thus allowing easy spreading of the particles on the
substrate surface or on the previously deposited coating layer. The thermal history the spray particles experience from the spray torch
to the surface of the substrate influence is schematically presented in Figure 29. It is evident that the material meets a highly specific
thermal history starting from heating of the particles by the energy from the spray torch, melting and acceleration of the particles by
the high-temperature high-velocity gas jet in millisecond-scale periods, spreading and solidification of the particles associated with
rapid quenching in microseconds, and cooling of the sprayed coating layer in the next several minutes. Figure 30 illustrates the
Figure 28 Summary of various parameters influencing thermal (plasma) spray process. Source: Tampere University of Technology.
www.iran-mavad.com
Figure 29 Thermal history of powder particles in a thermal spray process. Reproduced from Pawlowski, L. The Science and Engineering of Thermal
Spray Coatings; John Wiley & Sons Ltd., 2008.
Figure 30 Principle of coating formation in thermal spraying processes from molten droplets. Source: Tampere University of Technology.
formation of the coating from molten droplets. The coatings are formed by this type in most of the thermal spraying processes,
including powder and wire flame spraying, electric arc spraying, plasma spraying, and in high-velocity methods such as in HVOF
and detonation spraying. It is the case even also in suspension spraying process, but, however, in somewhat different size scale.
In thermal spraying, it is important to carefully optimize the spray parameters so that adequate melting would occur for high
D.E. without excessive evaporation of the material. In the process, solid particles would easily rebound or remain weakly connected
to the rest of the coating as unmelted ingredients. At the beginning of the coating buildup, particles impact directly onto the
substrate. The phenomena occurring at this stage determine the adhesion of the coating to the substrate. The molten particles
deform, become lamellae, and solidify into columnar or fine-grained equiaxial crystals. The spraying torch moves over the substrate
and the first layer usually composed of 5–15 lamellae, depending on the processing parameters, such as powder feed rate, spray
distance, particles size, and linear velocity of the spray torch, is formed.
Molten particles on impact with a substrate (or a previously deposited coating) transform into lamellae. The transformation is
associated with the processes of deformation and solidification, in turn depending on such factors as (2):
l particle velocity, size, and phase content (totally liquid, partly liquid, etc.) on impact;
l particle material properties in the liquid state (viscosity, surface tension, etc.);
l ability of wetting the substrate by liquid particles;
l temperature of substrate; as well as
l substrate roughness.
The solidification and deformation processes occur more or less simultaneously. The temperature at the particle’s interface with
the substrate at impact, called the contact temperature, influences adhesion of the lamellae and, consequently, adhesion of the
coating to the substrate. The mechanisms of adhesion can be, most probably, mechanical anchorage and, less probably, metal-
lurgical bonding. The temperature in the particle bodies, which can be at present measured by using high-speed and high-sensitivity
sensors, strongly influences the microstructure of the coatings. First, its evolution with time determines material solidification and
cooling rates, which, in turn, influences the formation of crystal phases and crystal size.
www.iran-mavad.com
The phenomena occurring inside the generated layer determine the crystal structure and size, porosity formation, and formation
of microcracks. The torch returns after a little time, depending on the work piece size, to the same spot. Meanwhile, the surface of the
layer is subjected to the action of the environment, i.e., oxidation (for metals or alloys) and/or cooling. The cooling leads to
generation of residual stresses, which may, eventually, cause the coating to crack. During spraying of one layer of coating, the torch
also heats up the previously deposited material. Additional heat fluxes result from solidification of the particles and their cooling
down to the temperature of equilibrium. The final coating thickness is reached in a few tens to a few hundred passes of the torch
over the substrate. Thereafter, the coating is cooled down to RT. This last stage of buildup is often crucial because of the generation of
thermal stresses. The latter in worst case causes detachment of the entire coating from its substrate.
Spraying conditions and the particular spray material predominantly influence the properties of spray coatings. Coating structure
and the configuration of sprayed coatings determine the characteristics of the coating/substrate system. Spray coatings grow when
individual spray particles impact the substrate surface, deform or splatter, and solidify due to heat transfer into the workpiece.
Thermal spray coatings usually have lamellar and, depending on the spray technique and material, more or less porous, micro-
cracked, heterogeneous, and anisotropic structure. Furthermore, they contain partially molten particles or particles that solidified
before impacting the surface or have reacted with gas from the atmosphere. Figure 31 presents schematically possible features in
thermally sprayed coatings. Figure 32 in turn shows a cross-sectional fracture and surface view of plasma-sprayed ceramic coatings.
The coating consists of flattened lamellar particles layered on each other. A typical microstructure of an electric arc-sprayed metallic
coating is shown in Figure 33. Typical thermally sprayed coating features such as individual flattened particles, interparticle regions,
oxides, and some residual porosity is present in the coating.
Novel thermal spray technologies, in which the particle velocities are high, result in coating structures largely without such voids
and other features, illustrated in Figure 31. This is obtained by replacement of part of the thermal energy with high level of kinetic
energy. Figure 34 presents an HVAF-sprayed very dense hard metal coating of good quality.
Suspension and solution spraying has been researched actively during the past years. In suspension spraying, the coating
formation is based on several stages involving evaporation of the liquid carrier, in-flight sintering of some fine solid particles,
melting of the solids and agglomerates, potential evaporation from the melt, and finally, impact of the molten particles to form
the coating. Figure 35 illustrates these stages in suspension thermal spraying process (39). The micrographs in Figure 36 show
the structure of suspension HVOF-sprayed alumina (Al2O3) coating on a steel substrate. Figure 36(b) is a high-magnification
Figure 31 Schematic illustration of the structure of thermally sprayed coating. Source: Tampere University of Technology.
Figure 32 Microstructure of plasma-sprayed ceramic oxide coating. (a) Cross-sectional fracture structure of plasma-sprayed Al2O3 coating and (b)
surface view of plasma-sprayed Al2O3 coating. Source: Tampere University of Technology.
www.iran-mavad.com
Figure 33 Cross-sectional microstructure of arc-sprayed NiCrMo (Inconel 625) coating showing individual flattened particles, interparticle regions,
oxides, and some residual porosity. Source: Tampere University of Technology.
Figure 34 Cross-sectional microstructure of good-quality HVAF-sprayed WC–10%Co–4%Cr coating showing no voids in the structure. Source:
Tampere University of Technology.
Figure 35 Coating formation stages in suspension thermal spraying process. Reproduced from Pawlowski, L. Suspension and Solution Thermal
Spray Coatings. Surf. Coat. Technol. June 25, 2009, 203 (19), 2807–2829.
Figure 36 Microstructure of suspension HVOF-sprayed alumina coating on steel substrate as studied by FE-SEM. (a) General view of the coating
cross-section and (b) detailed microstructure of the coating. Source: Tampere University of Technology.
www.iran-mavad.com
Figure 37 Microstructure of solution precursor HVOF-sprayed MnCo oxide coating. Reproduced from Puranen, J.; Laakso, J.; Kylmälahti, M.;
Vuoristo, P. J. Therm. Spray Technol. 2013, 22, 622–630.
field-emission-scanning electron microscopic (FE-SEM) image of the structure of the coating showing areas formed from flattened
molten particles with some unmelted or partially melted particles.
Besides suspensions, solutions without solid particles can be used to prepare coatings by thermal spraying. Now the coating
formation mechanisms may involve chemical reactions during spray processing in the gas phase forming reaction products and
chemical vapor deposition or spray pyrolysis-type coating formations on the substrate surface. Figure 37 illustrates the micro-
structure of high-velocity solution precursor flame-sprayed manganese–cobalt oxide coating.
4.10.4.2 Formation of Sprayed and Fused Coatings

Self-fluxing alloy powders are widely sprayed by flame spray process. These coatings are usually used to a smaller extent in the as-
sprayed state, but are posttreated by fusing the coating at high temperature of the order to 1050–1150 C, depending on the
composition. These coating materials are applied by powder flame spraying or welding using commonly available oxygen–
acetylene equipment. During thermal spraying or welding, the molten alloy forms a metallurgical bond with the substrate
surface. Coatings thus formed are very dense and show bond strength values significantly higher than those obtained with other
spray techniques.
Both coating and substrate are treated by torch or furnace heating. Self-fluxing alloys contain elements that react with oxygen or
oxides to form low-density oxides that float to the surface, thus improving density, bonding, and mechanical properties. In general,
these are nickel- or cobalt-based alloys that contain boron or silicon, either singly or in combination, in excess of 1.5%, as well as
carbon and iron. Chromium is added to many alloys for increased wear and corrosion resistance. Boron and silicon also promote
wetting by reducing oxides of nickel, cobalt, chromium, and iron and control surface tension and fluidity. During fusing, a boro-
silicate glass is formed that is believed to further reduce oxidation. Boron and silicon are fluxing agents, hence the term self-fluxing.
It is evident that the coatings prepared by flame spraying with subsequent fusing are dense and the interface also confirms the
metallurgical joining to the substrate material.
4.10.4.3 Solid-State Formation in Cold Spraying

Formation of a cold kinetic-sprayed coating is definitely very different from conventional thermally sprayed coatings. The method is
applicable to several pure metals and alloys. Other materials such as ceramic are not typically deposited by cold spraying. Cold
spraying is a solid-state process and therefore the coating is formed without melting of the material. However, there are some
similarities with other thermal spray processes. In both types of coating formation, the powder particles are accelerated to high
velocities by fast gas flow, the particles impact the surface, and they spread; in thermal spraying with molten particles, the melt flows,
but in cold spraying, the solid material flows due to high kinetic energy and the particles deform on impact and stick to the previous
layers by explosive welding-type mechanism. In cold spraying, the powder particles need a certain critical velocity (vcrit) for
successful coating formation. If the particle velocity is lower than the material-dependent critical velocity, the particles tend to
rebound and do not form a coating. In turn, when the particle velocities are above the critical velocity, the particles adhere to the
surface and build up the coating. Particles are stuck to the surface by plastic deformation or they just penetrate or are embedded into
the sprayed surface. Formation of the cold-sprayed coating is schematically presented in Figure 38.
The first particles cause the buildup of the coating (adhered particles) and substrate cratering (rebounding particles). Formation
of the cold-sprayed coating is based on two steps: buildup of the first layer (particle–substrate interactions) and deposit of the
coating (particle–particle interactions). In the first step, the adhesion strength between coating and substrate is defined, whereas step
2 defines the cohesive strength of the coating. In the first stage, particle–substrate interactions occur and are affected by substrate and
coating materials as well as surface preparation. After that, the next incoming particles adhere to the previous particles, deform, and
realign; now the coating is built up. In the second stage, particles interact with each other and form the coating structure. During
the spraying, metallic bonding and void reduction occur and after that, the coating structure is densified and work hardened.
www.iran-mavad.com
Figure 38 Coating formation in cold kinetic spraying process. Reproduced from Van Steenkiste, T.; Smith, J.; Teets, R.; Moleski, J.; Gorkiewicz, D.;
Tison, R.; Marantz, D.; Kowalsky, K.; Riggs, W.; Zajchowski, P.; Pilsner, B.; McCune, R.; Barnett, K. Surf. Coat. Technol. 1999, 111, 62–71.
It is notable that cold kinetic-sprayed metallic coatings are highly deformed and thus work hardened. The hardness of cold-sprayed
coatings is noticeably higher than that of the corresponding wrought metal. The elongation is low and the coating may show brittle
behavior. Figure 39 presents the microstructure of a fully dense and oxide-free cold-sprayed metallic tantalum coating. Figure 40 is
a higher magnification SEM image of high-pressure cold-sprayed copper coating showing the highly deformed copper particles
(etched structure).
Figure 39 Structure of cold-sprayed fully dense and oxide-free tantalum metal coating. Source: Tampere University of Technology.
Figure 40 Structure of high-pressure cold-sprayed copper coating; etched microstructure. Source: Tampere University of Technology.
www.iran-mavad.com
Table 12 Properties of several sealants for thermally sprayed coatings
Wetting angle on Viscosity, Curing shrinkage Curing shrinkage by

Sealant Base polymer Curing parameters g-alumina ( ) h (mPa s1) by weight, % volume, %
A Inorganic aluminum phosphate 200 C/7 h 97 505 48 –

B Methacrylatea 60 C/1 h 15 6.5 3.8 17.6
C Methacrylatea 60 C/1 h 15 8.6 2.6 5.3
D Phenola RT 10 1.5 79.6 8
E Epoxy 80 C/2 h 39 162 0.1 0.2
F Methacrylatea UV 26 28 0.5 <2
G Epoxy 60 C/2 h 42 179 1 0.1
H Furan 60 C/1 h – RT 67 626 2 2
I Vinyl ester 60 C/1 h – RT 83 319 0.6 2.4
a
Commercial sealant.
Reproduced from Knuuttila, J.; Sorsa, P.; Mäntylä, T. J. Therm. Spray Technol. 1999, 8, 249.
Figure 41 Laser fusion of thermally sprayed coating. Source: Force Technology.
4.10.4.4 Posttreatments
Thermally sprayed coatings are sometimes posttreated in order to modify the structure of the coating and to improve coating
properties. Such treatments include impregnating the coatings with organic or inorganic sealers, fusing the self-fluxing alloy
coatings, or treating the surfaces by laser beams, induction melting, etc. Machining and grinding of thermal spray coatings is also
a conventional and widely used process, but will not be discussed here. Thermally sprayed coatings fairly often contain a noticeable
amount of residual porosity, which particularly in plasma-sprayed ceramic coatings is also through porosity. Sealants are frequently
used to remove the open porosity of the sprayed coatings and to provide better corrosion protection to the coating. Table 12 gives
some characteristics of several polymeric sealants and one inorganic sealer for plasma-sprayed ceramic coatings (40). Important
properties of the sealants are, for instance, their chemical resistance, capability to fill the pores in the coating, low curing shrinkage,
wetting behavior, and thermal treatment required.
As was mentioned above, there are also other techniques such as laser beam melting of the sprayed coating. Figure 41 presents such
treatment operation for thermally sprayed NiCrBSi coating. In laser surface remelting, it is important to avoid excessive melting of the
substrate material, which would in such case be transferred to the molten coating material and interfere with its corrosion properties by
dilution with iron from the substrate. The coatings can be only partially melted, i.e., only the surface layer of the coating is treated with
the laser beam. The main disadvantages related to laser beam remelting of thermally sprayed coatings are the need of a two-step
process, quality of adjacent treated joints, tendency of crack formation in coatings, particularly in hard coatings, and cost of processing.
4.10.5 Coating Materials and Properties
As has been already mentioned, thermal spray technology seems to be clearly the most versatile coating manufacturing technology
among all deposition technologies. This claim is particularly relevant when the availability of different coating materials and
structures and the flexibility of using many of these coating materials in the same particular spray equipment are compared. Nearly all
materials can be processed by thermal spray technologies into sprayed coatings; this includes all pure metals and alloys, all hard
www.iran-mavad.com
Table 13 Main groups of coating materials and their typical properties when sprayed by different common techniques
Property Material Flame HVOF Arc Plasma
Gas temperature ( C) 3000 2600–3000 4000 (Arc) 12 000–16 000

Spray rate (kg h1) 2–6 1–9 10–25 2–10
Particle velocity (m s1) Up to 50 Up to 700 Approx. 150 Up to 450
Bond strength (MPa) Ferrous alloys 14–21 48–62 28–41 21–34
Nonferrous alloys 7–34 48–62 14–48 14–48
Self-fluxing alloys 83þ (fused) 70–80 15–50 –
Ceramics 14–34 – – 21–41
Carbides 34–48 83þ – 55–69
Coating thickness (mm) Ferrous alloys 0.05–2.0 0.05–2.5 0.1–2.5 0.4–2.5
Nonferrous alloys 0.05–5.0 0.05–2.5 0.1–5.0 0.05–5.0
Self-fluxing alloys 0.15–2.5 0.05–2.5 – –
Ceramics 0.25–2.0 – – 0.1–2.0
Carbides 0.15–0.8 0.05–5.0 – 0.15–0.8
Hardness (HRC) Ferrous alloys 35 45 40 40
Nonferrous alloys 20 55 35 50
Self-fluxing alloys 30–60 30–60 – 30–60
Ceramics 40–65 – – 45–65
Carbides 45–55 55–72 – 50–65
Porosity (%) Ferrous alloys 3–10 <2 3–10 2–5
Nonferrous alloys 3–10 <2 3–10 2–5
Self-fluxing alloys <2 (fused) <2 – –
Ceramics 5–15 – – 1–2
Carbides 5–15 <1 – 2–3
Source: Sulzer Metco AG.
metal-like materials also named as cermets or carbides (e.g., WC–Co), as well as all ceramic materials with a few exceptions (magnesia
and silicon nitride and carbide as pure compounds), all thermoplastics, and mixtures of these. Due to the very large variation of
coating materials, it is not reasonable to deal with all these here, but to give an overview of typical thermal spray coating materials.
4.10.5.1 Coating Materials

Table 13 summarizes the main groups of coating materials and their typical properties when sprayed by different common tech-
niques. Typical process temperatures, feedstock feeding rates, particles velocities, coating bond strength numbers, typical coating
thickness, hardness, and porosities are presented.
4.10.5.1.1 Pure Metals

Pure metals are available as powders and wire. The most important pure metals in thermal spraying are aluminum (Al), zinc (Zn),
copper (Cu), nickel (Ni), molybdenum (Mo), titanium (Ti), tantalum (Ta), and iron in steels. Such materials as titanium and
tantalum show high reactivity with oxygen when sprayed in air atmosphere, resulting in highly oxidized coatings with unsatisfactory
properties and poor sprayability. These materials require spraying in vacuum or inert gas, by VPS/LPPS techniques or shrouded
plasma spraying, respectively. However, the development of cold spray technology has offered possibility to these materials to spray
them in atmospheric conditions without any practical oxidation of the coating. All other pure metallic coating materials mentioned
above can be processed by all normal thermal spray technologies.
4.10.5.1.2 Metal Alloys

This group of coating materials in thermal spraying is so large that it is not completely discussed here. Metallic alloys used as
powders and wires are largely based on alloys of iron, nickel, cobalt, aluminum, zinc, molybdenum, and copper, i.e., metallic alloys
commonly used as construction metals, and some more other alloys, which are used primarily only as sprayed coatings, e.g., zinc
alloys (Zn–15%Al). Table 14 summarizes some common metal alloys used as coatings in thermal spray processing.
4.10.5.1.3 Hard Metals

Hard metal-like coatings, also referred to as carbides or cermets, are common coating materials for wear prevention (Table 15).
These materials are composite materials and consist of a very high volume fraction of hard carbide phase, typically tungsten car-
bide (WC), chromium carbide (Cr3C2), or titanium carbide (TiC), and a metallic binder phase, typically cobalt (Co) or nickel (Ni).
For improved corrosion properties for the metallic binder phase, alloying with chromium (Cr) is commonly done. The spray
processes used for these materials are plasma, detonation, HVOF, and HVAF, with coating properties improving also in the same
order. The compositions all show excellent wear properties. WC-based compositions can be used in operation temperatures up to
540 C in air, but higher temperatures lead to catastrophic oxidation of WC to WO3. Cr3C2-based hard metals in turn can be used at
significantly higher temperatures up to 900 C in air. WC–Ni coatings can be used in mild corrosive applications and in nuclear
www.iran-mavad.com
Table 14 Examples of metal alloys for thermal spraying
Group Compositions Application
Ferrous Fe–18%Cr–8%Ni, Fe–13%Cr, FeCrAl Buildup, wear, corrosion

Nickel NiCrMo alloys (Inconel 625, Hastelloy, Monel 400, Ni–(5–20)%Al, NiCrAlY, etc.) Corrosion, oxidation, bond layers, etc.
Cobalt Stellites, triballoys, CoNiCrAlY, etc. Wear, corrosion, oxidation
Aluminum Al–Mg, Al–Zn, Al–Si, etc. Corrosion, etc.
Zinc Zn–15%Al Corrosion
Copper Cu–Al, Cu–Ni, etc. Corrosion, sliding wear, buildup, etc.
Molybdenum Mo/NiCrBSi, etc. (note: blended powders) Wear, friction, bond layer
Table 15 Examples of hard metal-like coatings and typical compositions
Group Compositions Application
WC–Co WC–6%Co, WC–12%Co, WC–17%Co Wear

WC–CoCr WC–10%Co–4%Cr, WC–6%Co–8%Cr Wear, corrosion
WC–Ni WC–12%Ni Wear, corrosion, nuclear
WC–NiCr WC–20%Cr–7%Ni Wear, corrosion
WC–Alloy binder WC–15%Hastelloy, WC–15%FeCrAl Wear, corrosion
Cr3C2–NiCr Cr3C2–(20–50)%NiCr, e.g., Cr3C2–25%NiCr Wear, corrosion, oxidation
TiC–Ni/Co TiC–(20–40)%Ni or Co Wear and corrosion
applications where Co is not allowed to be used. Alloying of the metallic binders with Cr leads to marked improvements in the
corrosion resistances in aqueous solutions.
4.10.5.1.4 Oxide Ceramics

Typical ceramic coating materials in thermal spraying are oxides of aluminum, titanium, chromium, and zirconium. Other ceramic
materials may also exist as thermally sprayed coatings, such as borides, carbides without metallic binder, and nitrides, but these are
not commonly used. Typical oxides for wear and corrosion applications are:
l alumina, Al2O3, as pure oxide or alloyed with other oxides, e.g., titania (2.5–40%TiO2), silica (SiO2), zirconia (20–40%ZrO2),
or chromia (Cr2O3);
l chromia, Cr2O3, as pure oxide or alloyed with TiO2, SiO2, or Al2O3; as well as
l titania as pure oxide.
Another noticeable group of ceramic oxide coating materials used in thermal spraying are the TBCs based on zirconia ZrO2.
Zirconia is not used as a pure oxide due to its poor properties under thermal changes, but is instead stabilized partially or fully with
such other oxides as magnesia (MgO), yttria (Y2O3), calcia (CaO), or ceria (CeO2). Other novel TBC compositions also exist.
4.10.5.1.5 Polymers
Polymeric materials such as thermoplastics can also be processed by thermal spraying. In spraying of polymers it is essential not to
degrade the polymer structure by excessive heating, i.e., burning of the material. Powder flame spraying, particularly using propane
and air as process gases, is the most suitable spray technique for polymers. Such polymers as polyolefines (polyethylene (PE),
polypropylene (PP), copolymers), other low temperature thermoplastics (ethylene vinyl acetate, ethylene-methacrylic acid,
polyamide), high-temperature thermoplastics (polyether ether ketone, polyphenylene sulfide), and some fluoropolymers (ethylene
chlorotrifluoroethylene, fluorinated ethylene propylene, polyvinylidene fluoride) have been reported to be processable by thermal
spraying. Particularly polyolefins such as PE and PP, and polyamides (nylon) can produce dense and corrosion-resistant coatings.
Special care should be taken when spraying polymers that emit harmful species (e.g., fluoropolymers).
4.10.5.1.6 Composites and Other Materials

The category of composites is an extremely broad definition of thermal spray coating materials. All mechanical powder mixtures/
blends can be defined as composites in thermal spraying. Also, there exist a number of real composite materials, e.g., in the group of
abradable coatings. These may be, for instance, nickel–graphite composites, which consist of graphite core particle clad with a layer
of nickel. More information about various composite materials in thermal spraying can be found elsewhere in the literature (2,3,7).
4.10.5.2 Coating Properties

The structure and consequently the technical properties of thermally sprayed coatings are highly dependent on several factors, such
as material composition, class, and feedstock material characteristics; particular spray process and process parameters used; coating
www.iran-mavad.com
Figure 42 Factors influencing the required properties of thermally sprayed coatings. Source: Tampere University of Technology.
formation, and posttreatment procedures. Figure 42 schematically presents all main factors influencing the needed coating
properties of thermally spray materials planned to fill the specific requirements of the intended technical application.
Several factors are known to have influence on the structure and properties of thermal sprayed coatings:
l feedstock material type and composition;
l powder manufacturing methods;
l powder properties such as particle size, size distribution, presence of small particles, presence of excessive large particles, carbide
size, e.g., in WC–Co;
l wire type and composition;
l thermal spray method;
l spray torch parameters, e.g., gas types and flow rates; electrical parameters such as voltage and current (in plasma); state of
electrodes/nozzles;
l kinematic parameters, such as surface velocity, distance between passes/increment, spray distance, degree of overlapping, etc.;
l substrate temperature and cooling media and configuration;
l heat load to substrate and temperature cycling during spraying;
l powder/wire feeding rate and layer thickness per pass;
l evaporated species from the process;
l external dust removal and presence air knives and air cooling; as well as
l stability of power suppliers, powder/wire feeding, and gas flows.
The presence of such a large number of thermal spraying parameters gives huge possibilities for tailoring coating properties and
producing excellent coatings with specific properties. However, this also makes the process to some extent sensitive to variations in
spray conditions, which may result in deviation of coating quality. Nevertheless, thermal spray technology can be a highly
reproducible deposition process if all important parameters are monitored and controlled during spraying.
4.10.5.2.1 Microstructure, Phase Structure, and Density

As was presented already in connection with the coating formation mechanisms, thermally sprayed coatings may contain several
different features, which can be:
l Residual porosity, often through porosity, the amount of which, however, can vary highly depending on the spray process and
type of coating material; all variations from highly porous microstructures to fully dense and oxide-free coatings can exist.
l Oxidation in metallic coatings, particularly at the splat interfaces.
l Microcracks inside individual lamellae, primarily in ceramic coatings; such microcracks with porosity may improve resistance
against thermal shocks.
l Lack of bonding between overlapping lamellae/splats.
l Presence of stable crystalline phase and metastable or even amorphous phases due to rapid solidification phenomena during
coating formation.
However, is spite of such possible, and to some extent harmful, features, thermally sprayed coatings can show excellent coating
structures and properties, which allow their use in very many demanding and specific applications.
4.10.5.2.2 Mechanical Properties

Several thermally sprayed coatings are widely used in many applications that require good mechanical properties, such as high
hardness, low friction, high ductility, etc. For instance coating hardness may vary significantly depending on the resulting coating
microstructure, porosity, phase structure, and level of residual stress in the coating. Figures 43 and 44 and Table 16 present
www.iran-mavad.com
Figure 43 Comparison of microhardness of hard metal-like coatings deposited by HVOF, D-Gun, APS, and VPS processes. Reproduced from
Pawlowski, L. The Science and Engineering of Thermal Spray Coatings; John Wiley & Sons Ltd., 2008.
Figure 44 Comparison of hardness of different ceramic oxide coatings. Reproduced from Gärtner, F.; Voyer, J.; Qi, X.; Kreye, H.; Richter H. J.;
Krömmer, W. New Challenges for Wire and Rod Flame Spraying. In Proceedings of the 6th HVOF Colloquium, Nov. 27–28, 2003, Erding.
Table 16 Microhardness of some APS and VPS ceramic oxide coatings
Coating Atmospheric plasma sprayed Vacuum plasma sprayed
Al2O3 720 1310

Al2O3–40%TiO2 690 810
Cr2O3 930 1500
TiO2 730 750
Reproduced from Takeuchi, J.; Nakahira, H.; Nagai, J. In Proc. 2nd Plasma Technik Symposium, Vol. 2,
Switzerland, 1991.
www.iran-mavad.com
some useful and comparative microhardness data for different hard metal coatings, ceramic oxide coatings, and for APS- and
VPS-processed ceramic oxide coatings, respectively (41). Highest hardness values are usually obtained with high-velocity coating
processes, which simultaneously produce dense coating structures and, with hard metal coatings the highest amount of retained
carbides in the coating.
4.10.5.2.3 Corrosion Properties

Several factors are known to affect the corrosion properties of thermally sprayed coatings. Primarily the coating material
composition is responsible for this, but it is well accepted that in most cases the corrosion properties of thermally sprayed
coatings are not equal to those of the corresponding bulk material. This is mainly due to such reason as differences in coating
microstructure vs bulk material. Sprayed coatings often contain porosity, which may lead to corrosion of the substrate material
or at least corrosion of the coating material itself. The inhomogeneous composition and phase structure together with crevice
corrosion tendency will often lead to selective corrosion inside the coating. In particular, the heterogeneous microstructure
leads to selective corrosion of specific phases, sensitization, galvanic corrosion, etc. Coating porosity, often through porosity,
and the presence of cracks will lead to such internal crevice corrosion, concentrated chlorine ions, low pH, substrate/interface
corrosion, and finally delamination of the whole coating from the substrate. Sealing of the open porosity can prevent the
electrolyte penetration into the coating structure and substrate/coating interface, thus possibly solving corrosion attacks.
Ceramic coatings on mild steel or cast iron substrates are often equipped with corrosion-resistant bond layers, e.g., Ni–20Cr.
A thorough presentation of corrosion properties in this context is not possible and therefore only some examples are presented
in the next section.
Figure 45 compares the extent of corrosion of some HVOF-sprayed hard metal coatings. According to the results, it is evident that
addition of Cr to the WC–Co system clearly improves the corrosion resistance of HVOF-sprayed wear-resistant hard metal coatings.
Such coatings are widely used in wear applications, in which corrosion attack may also be an issue, e.g., in the paper manufacturing
industry.
Such thermal spray technologies, which can produce impermeable, i.e., fully dense and unoxidized and homogeneous
metallic coatings, can offer interesting possibilities for using the coatings in demanding corrosive applications. Figure 46
presents electrochemical results, which show that e.g., cold-sprayed tantalum coatings can be fully protective against corrosion
by, in this case, 3.5% NaCl solutions. Such coatings are also known to be resistant to other very aggressive strongly acidic and
basic electrolytes (43,44).
As was discussed earlier, thermally sprayed zinc and aluminum coatings can give cathodic protection to steel structures and be
used e.g., in applications exposed to the effect of sea water.
4.10.5.2.4 Wear Properties

Wear of materials includes several different types of wear phenomena and mechanisms. Typical wear types are abrasive wear,
adhesive wear, erosion wear, fretting, corrosion wear, cavitation wear, etc. Due to the high versatility of thermal spray technology
and the wide range of coating materials available, this technology is one of the main candidates for tackling wear problems
economically and effectively. Several different coatings with different properties and characteristics exist. Wear-resistant coatings can
be e.g.,
l hard coatings, e.g., hard metal-like coatings and oxides;

l soft solid lubricating coatings, e.g., bronze, lead alloys, and tin alloys; and
l hard and lubricious metallic coatings, e.g., molybdenum.
No. Powder Type System/fuel

1 WC–17Co Aggl./sint. Jetkote/propane
2 WC–17Co Aggl./sint. TopGun/ethylene
3 WC–17Co Aggl./sint. DJH-2600/hydrogen
4 WC–12Co Aggl./sint. DJH-2700/propane
5 WC–12Co Sintered DJH-2700/propane
6 WC–12Co Fused DJH-2700/propane
7 WC–10%Co–4%Cr Sint. crush. DJH-2700/propane
8 WC–10%Co–4%Cr Sintered DJH-2700/propane
9 WC–6Co–8Cr Sintered DJH-2700/propane
Figure 45 Comparison of corrosion behavior of some HVOF-sprayed hard metal coatings prepared from different spray powder types. Corrosion
test: neutral salt spray test and corrosion test in salt spray with SO2 atmosphere. Reproduced from Schwetzke, R.; Kreye, H. J. Therm. Spray Technol.
1999, 8, 433–439.
www.iran-mavad.com
Figure 46 Electrochemical corrosion behavior of fully dense cold-sprayed tantalum coating showing excellent corrosion resistance and protection
characteristics of the coating; results for a porous tantalum coating are also included for comparison. Reproduced from Koivuluoto, H.; Näkki, J.;
Vuoristo, P. J. Therm. Spray Technol. 2009, 18, 75–82.
Figure 47 Abrasion wear of thermally sprayed hard metal coatings compared with wear of hard chromium plating and mild steel. Sand abrasion
test: abrasive 0.1–0.6 mm quartz sand; Fe 52, mild steel; HCr, hard chromium plating; P, plasma sprayed; H, HVOF sprayed; D, detonation sprayed;
CrC, Cr3C2–25NiCr; WC, WC–12Co. Source: Tampere University of Technology.
Figure 47 compares the abrasion wear of thermally sprayed hard metal coatings with wear of hard chromium plating and mild
steel. It is obvious that the high-velocity spray methods produce excellent coating properties against abrasion wear by quartz sand.
Sprayed coatings show generally better properties than hard chromium plating, making thermal spraying a cost-effective alternative
to hazardous electrochemical processing. Table 17 is a comparison of wear, hardness, and corrosion properties of some hard metal
coatings manufactured by HVOF spraying. Figure 48 adds further information about the abrasive wear properties of several
WC–10%Co–4%Cr coatings prepared from five different powders of the same composition and by different thermal spray
processes, including plasma spraying, detonation, and HVOF spraying technologies. Comparative wear results of various ceramic
oxide coatings are in turn presented in Figure 49.
Table 17 Comparison of properties some hard metal coatings manufactured by HVOF spraying
Coating Hardness HV0.3 Weight loss (mg) Corrosion behavior
WC–12%Co 1171 2 Corrosion at 24 h

WC–17%Co 1112 4 Corrosion at 24 h
WC–20%Cr–7%Ni 1129 2 Uncorroded at 360 h
WC–10%Co–4%Cr 1064 2 Uncorroded at 120 h
WC/TiC–17%Ni 913 9 Uncorroded at 120 h
Mild steel – 55 Severe corrosion
Abrasion wear test: JIS H8615, abrasive is a cylinder coated with SiC emery paper.
Corrosion test: JIS Z2371 salt spray test (5% NaCl, temperature 35 C).
Source: Tocalo Co. Ltd.
www.iran-mavad.com
Figure 48 Comparison of several thermally sprayed WC–10%Co–4%Cr coatings prepared by APS (Ar/H2 and Ar/He gases), two HVOF processes
(CDS and DJH-2700), and detonation spraying. The five columns refer to five different powders or similar composition. Source: Tampere University
of Technology.
Figure 49 Dry abrasion wear results for different ceramic coatings. Reproduced from Gärtner, F.; Voyer, J.; Qi, X.; Kreye, H.; Richter H. J.;
Krömmer, W. New Challenges for Wire and Rod Flame Spraying. In Proceedings of the 6th HVOF Colloquium, Nov. 27–28, 2003, Erding.
Figure 50 Dry abrasion wear results of several different thermal spray coating materials as a function of wear distance. Reproduced from
Houdkova, S.; Zahalka, F.; Kasparova, M.; Berger, L. -M. Tribol. Lett. 2011, 43, 139–154.
Figures 50 and 51 summarize and compare abrasive wear properties of several different thermally sprayed coatings with some
bulk materials and other coatings (45). It is obvious that thermally sprayed coatings can provide excellent wear properties to surfaces
and the results and behavior depends a lot on the specific coating material used.
4.10.5.2.5 Friction Properties

Thermally sprayed coatings can be used not only as low-friction surfaces but also to increase friction. Table 18 presents some results
obtained by using a ball-on-disk tribometer, in which the mating ball was either alumina or hardened steel.
www.iran-mavad.com
Figure 51 Dry abrasion wear results of several different thermal spray coating materials as a function of wear distance. Reproduced from
Houdkova, S.; Zahalka, F.; Kasparova, M.; Berger, L. -M. Tribol. Lett. 2011, 43, 139–154.
Table 18 Comparison of friction properties of several coatings in a ball-on-disk tribometer
Coeff. of wear resistance coefficient of friction

Coating Kabr (mm3 Nm1) 106 (COF) for Al2O3 ball COF for steel ball
WC–Co – 0.369 0.008 0.780 0.011

WC–CoCr – 0.398 0.009 0.757 0.033
WC–Hastelloy – 0.382 0.011 0.811 0.008
Cr3C2–NiCr 5.5 0.549 0.008 0.717 0.093
(Ti)(Mo)(C,N)–NiCo 0.35 0.653 0.016 0.529 0.011
NiCrSiB 65 0.645 0.026 0.622 0.024
AISI 316L 182 0.644 0.021 0.638 0.030
Cr2O3 – 0.350 0.032 0.567 0.015
Hardened steel 6.2 0.568 0.008 –
Hard chrome 277 0.704 0.023 –
Reproduced from Houdkova, S.; Zahalka, F.; Kasparova, M.; Berger, L. -M. Tribol. Lett. 2011, 43, 139–154.
4.10.5.2.6 High-Temperature Properties

Coatings with high-temperature properties are frequently refractory metals, ceramic materials, or metallic alloys containing various
oxidation resistance-improving elements such as chromium, aluminum, and yttrium. The oxidation behavior of thermally sprayed
coatings is much dependent on the composition of the coating and the microstructure.
4.10.5.2.7 Biomedical Properties

Some thermal spray coatings materials such as titanium and hydroxyapatite (HAp), are bioinert and have bioactivity properties,
respectively. Titanium coatings are usually prepared so that the coating deposits as porous thus allowing the organic tissue to
penetrate into the surface of various implants. Similar behavior but with bioactivity occurs with HAp coatings also applied by
thermal spray techniques, e.g., plasma spraying.
4.10.6 Industrial Applications
Thermal spray coatings were originally used in the past only for repair and maintenance purposes; the technology indeed was called
‘metal spraying.’ However, soon the technology was adopted to several other application fields and today thermal spray coatings are
used also much in new production, i.e., components and equipment are designed to have thermal spray coatings for high
performance and durability. Applications in which a thermal spray process is one part of the full manufacturing chain are known.
At present, thermal spray coatings are used in very many different industrial applications and in industrial sectors, for instance in:
l refurbishment of worn or incorrectly machined parts;
l wear protection against abrasion, erosion wear, adhesive wear, etc.;
l corrosion protection against atmospheric and wet corrosion;
l protection against oxidation and high-temperature corrosion;
l thermal insulation at high temperatures, e.g., TBCs;
www.iran-mavad.com
l electrical insulation;
l improvement of thermal conduction, e.g., copper coatings;
l providing electrically conducting layers;
l modification of surface properties, e.g., lowering of increasing friction properties;
l abradable coatings to provide tolerance-control coatings;
l bioinert and bioactive surfaces in human implants in medical technology; as well as
l decorative coatings.
Table 19 summarizes with examples industrial applications of thermal spray coatings. It is evident that thermal spray technology
is the most versatile and most widely used among all coating technologies available. Thermally sprayed coatings can be manu-
factured by manual spraying, which finds much use due to its versatility and the importance of operator’s skill on quality. In
particular, conventional thermal spray techniques such as wire and powder flame and arc spraying may often be operated manually.
Figure 52 shows manual spraying in refurbishment of a worn component. Automated spraying using various levels of mechani-
zation and robotics is surely preferred over manual spraying in order to increase coating quality, for reproducibility, and to
eliminate potential health and environmental risks. Figure 53 shows an example of an automated mass production system for grit
blasting and coating of synchronizer rings by wire flame spraying. The capacity of such installation is approximately 500 rings per
hour in three shifts. Large components over 10 m in length, such as rolls and cylinders for paper machines, are frequently coated in
large spray booths such as the one presented in Figure 54. The cylinders are loaded into the spray booth through opened roofs.
Figure 55 illustrates HVOF spraying of a large paper machine cylinder by HVOF spray process. The spraying is done in a separate
spray booth and the spraying procedure is controlled and monitored remotely from the operator’s spray control panel. In the next
Table 19 Industrial applications of thermal spray coatings
Application Material
Corrosion protection of steel structures, steel parts in bridges, ship hulls, exhaust tubes, Zinc, Zn–15%Al alloys; aluminum, Al–5%Mg alloys
offshore oil drilling platforms, etc.
Corrosion protection of parts, rolls and cylinders, pump shafts, pump housings, propeller Stainless steels, Inconel 625, Hastelloy C 276
shaft, etc.
Corrosion and wear protection, sealing surfaces of pumps, sleeves, wear parts, mechanical Chromium oxide
seals, feeding screws, cylinders, rolls, etc. Chromium steels
Stainless steels
Self-fluxing alloys (NiCrBSi)
Thermal barrier coatings; combustion chambers, transition ducts, blades and vanes, piston Zirconium oxide (partly and fully stabilized)
crowns, exhaust valves, etc.
Oxidation, hot corrosion, erosion; heat exchangers surfaces in boilers, superheaters, water FeCrAl alloys
wall tubes, etc. Chromium steels
Self-fluxing alloys
Fretting wear in gas turbines Tungsten carbide (<500 C)
Chromium carbide (<815 C)
Cu–Ni–In
Abrasion wear resistance; guider rolls, shafts, piston rings, piston rods, feeding screws, Chromium steels
drawing drums Tungsten carbide
Adhesive wear resistance; piston rings, gear guides, synchronizing rings, drawing drums, Molybdenum
drawing tools, etc. Aluminum bronze
Tungsten carbides
Babbitt metals
Erosion and cavitation resistance; blower blades, wear plates, pump housings, etc. Chromium steels
Aluminum bronze
Tungsten carbide
Chromium carbide
Nonslip surfaces, friction surfaces, floor plates, rolls, paper feeders, etc. Aluminum
Chromium steels
Tungsten carbide
Electrical conducting layers, heaters, condensators, electrical contacts, etc. Silver, tin, aluminum, copper, nickel–chromium
Electrically insulating coatings, alternator parts, heaters, corona rolls, welding nozzles, etc. Aluminum oxide
Electromagnetic shielding, housings for electrical devices Zinc, aluminum, tin
Clearance surfaces; abradable coatings in gas turbines and turbine compressors Nickel–graphite, nickel aluminum, aluminum–
polyester, nickel–bentonite
Repair of mismachined components, shafts, pump parts, drums, bearing surfaces, etc. Chromium steels, stainless steels, special steel
grades, aluminum bronze, babbitt metals, self-
fluxing alloys
www.iran-mavad.com
Figure 52 Manual powder flame spraying in refurbishment. Source: Kuopion Konepaja Oy.
Figure 53 Automated production system for grit blasting and coating of synchronizer rings by wire flame spraying. Source: GTV mbH.
Figure 54 Thermal spraying booth for coating large paper machine rolls and cylinders. Source: Vaahto Group Oy.
www.iran-mavad.com
Figure 55 Spraying of a large paper machine cylinder by HVOF spray process. Source: Metso Paper Inc.
section, some industrial examples of thermally sprayed coatings are reviewed. The aim is not to present all possible applications, but
instead to show the high versatility of thermal spraying in very diverse industrial sectors.
4.10.6.1 Mechanical Engineering

In mechanical engineering sectors, thermally sprayed coatings are widely used in restoration and in providing specific surface
properties such are wear and corrosion protection, friction properties, and heat resistance. The application area is extremely wide.
Wear resistance of surfaces is a complex issue depending on the whole system, which includes the wear mode present, material
properties, and properties of the tribosystem. Besides the material combinations of the mating surfaces, this is influenced by many
factors including relative motion of the mating surfaces, strain properties, roughness of surfaces, atmosphere, temperature, etc. The
main wear mechanisms that can also occur simultaneously in certain applications are adhesive wear, abrasive wear, erosion wear,
surface fatigue, fretting, and corrosion wear. To avoid adhesive wear, metallurgical interactions between friction partners by micro-
welding have to be prevented. In engine applications, this can be realized by producing thermal spray coatings with a porous structure
that picks up and holds the lubricating oil; plasma-sprayed coating based on a mixture of molybdenum and self-fluxing alloy NiCrBSi
is one good choice. In abrasive wear applications, the surfaces require thermally sprayed coatings with high and uniform hardness. For
this, homogeneous and dense coatings are needed. HVOF spraying and its variants, e.g., HVAF, are optimal for producing such wear-
resistant coatings. Typical materials are WC–Co and NiCrBSi þ WC–Co mixtures. Coatings used in applications where surface fatigue
is a problem have to be uniform and free from pores, cracks, or delaminations. Therefore, very dense coatings produced under
protective gas atmosphere in vacuum, by HVOF, or by HVAF are best suited. The coating materials can be hard metallic alloys or
carbide-containing coatings. In corrosive wear applications, iron-, cobalt-, or nickel-based alloys or sealed chromium oxide and
aluminum oxide ceramic coatings have often be used. Typical industrial applications are, e.g., sealing surfaces of pumps, valves,
sleeves, wear parts, mechanical seals, feeding screws, shafts etc. Figure 56 shows HVOF-coated ball valve parts after spraying.
4.10.6.2 Corrosion Protection in Infrastructures and Steel Structures

Thermal spray coatings used for corrosion protection are either coatings providing cathodic protection or coatings that are more noble
than the substrate material. In the latter case, the coating needs to be impermeable, either directly in the as-sprayed state or after
postsealing treatment. In case of cathodic corrosion protection, the component (substrate) is as a cathode, whereas the thermally
sprayed zinc or aluminum coating is used as a sacrificial anode in the electrochemical circuit. Using more noble materials as a coating
involves opposite charge of coating and substrate, and therefore formation of an electrochemical circuit must be prevented. The only
possibility now is to deposit dense, corrosion-resistant coatings or sealed ceramic coatings. Ceramic coatings definitely often require
additional dense Ni–20%Cr or Ni–5%Al bond layers in order to protect the steel or cast iron substrate from corrosion and to guarantee
good bond strength for the ceramic coating. Applications in infrastructure are mainly based on using active coating materials such as
zinc and aluminum alloys, which provide cathodic corrosion protection to steel structures. Examples of applications are, e.g.,
corrosion protection of steel structures, steel parts in bridges, ship hulls, exhaust tubes, off-shore oil drilling platforms, etc.
4.10.6.3 Paper and Pulp

Applications of thermally sprayed coatings in the paper and pulp industry are numerous. The surface areas of paper machine rolls
are large, which imposes a number of special requirements on the reliability of thermal spray processes. The use of spray materials in
www.iran-mavad.com
Figure 56 HVOF-sprayed WC–10Co4Cr coatings on valve parts prior to surface grinding and polishing. Source: Tampere University of Technology.
these applications is also large. Examples of applications in the pulp and paper industry are pulp digesters, blow tanks, suction rolls,
calendar rolls, center press rolls, dryer rolls, yankee dryer cylinders, coater blades, and creping blades. Several other applications exist
including water removal elements etc. Figures 57–59 presents HVOF spraying of a large center press roll with HVOF, ceramic-coated
center press roll, and on-site coating of a large yankee dryer cylinder, respectively.
Figure 57 HVOF spraying of a center press roll with corrosion-resistant metal alloy coating for bond layer of ceramic top coating. Source: Metso
Paper Inc.
Figure 58 Paper machine center press roll with a thermally sprayed ceramic coating. Source: Metso Paper Inc.
www.iran-mavad.com
Figure 59 Thermal spray coating of a yankee dryer cylinder. With permission of Tocalo Co. Ltd.
4.10.6.4 Gas Turbines

Both aeroengine and industrial gas turbine use various different thermally sprayed coatings. The types of applications are oxidation- and
corrosion-resistant coatings on turbine blades and vanes; TBCs in combustion chambers, transition ducts, and blades and vanes;
abradable coatings as seals in the compressor and turbine sector; wear- and erosion-resistant coatings; and fretting-resistant coatings.
Corrosion protection in elevated and high-temperature applications with metal alloys is only possible by using relatively complex metal
alloys that are able to form Al2O3, SiO2, or Cr2O3 surface oxide layers at high-temperature exposure. Often the content of these alloying
elements in the coating has to be sufficiently high in order to form the protective oxidic layer several times. Coating materials called
MCrAlY alloys are thermal spray coating materials developed specifically for oxidation and hot corrosion-resistant coatings on gas
turbine blades and vanes. The basis for these alloys is a nickel, cobalt, or sometimes iron alloy that is adjusted for particular applications
by adding a large number of additional alloying elements. The ‘M’ represents one or more of the elements iron, nickel, or cobalt. In the
terminology of the business, these then become FeCrAlY, NiCrAlY, CoCrAlY, NiCoCrAlY, etc. These alloys form complex oxides from
the aluminium, chromium, and yttrium, which serve as oxygen and corrosion barriers to protect the less-resistant substrate. These
coatings are often combined with ceramic top coats, the TBCs. MCrAlY coatings are traditionally applied by VPS/LPPS processes, in
which oxidation of the alloying elements such as Cr, Al, and Y is totally avoided. Coated blades and blades are normally heat-treated to
create the creep properties of the base material; such heat treatment consists of a solution treatment at high temperatures well above
1100 C followed by a lower temperature precipitation treatment. The treatment procedures depend on the substrate material.
Nickel–graphite abradable coatings have been used successfully in turbine and engine applications for many years. After wear-in
during start-up of the engine, the coatings prevent performance loss due to friction of expanding components against the housing of
the turbine at the operating temperature.
The use of TBCs is expanding with increasing use in power generation. Yttria-stabilized zirconia (YSZ) is used as a TBC on
components in the hot section, i.e., turbine section, of aircraft engines and industrial gas turbines. The low thermal conductivity of
TBC materials facilitates increased performance by lowering component temperatures in the presence of high-temperature
combustion. YSZ has a relatively high coefficient of thermal expansion (CTE) and is near that of the nickel- and cobalt-based
superalloys used for turbine components. This fortuitous CTE match minimizes stress induced by differential expansion
between the coating and its substrate. To a lesser extent, other materials are used as stabilizers, including calcia (CaO), ceria (CeO2),
and magnesia (MgO). Stabilization is necessary to eliminate the volumetric change associated with the phase transformation of
zirconia from monoclinic to tetragonal during thermal excursions, which occurs as the turbine ramps from ambient to operational
temperatures. Figures 60–62 illustrate some thermal spray coating applications in gas turbines.
4.10.6.5 Textile Industry

Various thermally sprayed coatings, particularly plasma-sprayed wear-resistant oxides, are used in machines in textile industry.
These coatings are characterized by a precise definition of the morphology produced using various handling methods and the
topology of the surface. The surface texture is of particular importance for production components in contact with textile thread. In
order to maximize fiber production, coatings of ceramic oxides are used, usually with a nickel-based bond coat that also provides
corrosion protection. Some coated parts for textile industry are represented in Figure 63.
4.10.6.6 Printing Industry

Printing and graphic industries have also been users of thermally sprayed coatings. Plasma-sprayed electrically insulating ceramic
coatings based on alumina are used in corona discharge rolls to activate the surface of PE films. Corona rolls are typically
approximately 5 m long and the diameter is between 250 and 500 mm.
www.iran-mavad.com
Figure 60 Section of gas turbine with abradable coating for clearance control. Source: Patria Aviation Oy.
Figure 61 Thermal barrier-coated surface in a military aeroengine afterburner. With permission of Patria Aviation Oy.
Figure 62 Robotized plasma spraying of a transition duct of an industrial land-based gas turbine power plant. Source: Patria Aviation.
www.iran-mavad.com
Figure 63 Wearing parts with plasma-sprayed coatings for textile industry applications. Source: Advanced Coating S.A.
Another application is coating of the anilox rolls. Such roll is a cylinder usually constructed of a steel or aluminum core that is
coated by plasma-sprayed chromia and laser engraved to contain millions of very fine small cavities known as cells. The function of
the anilox roll is to transfer ink in the printing process.
4.10.6.7 Electrical Industry

Thermally sprayed coatings can also be used in electrical and electronics industry. Zinc coating prepared by electric arc spraying is
used as metallization of capacitor ends. Plasma- and HVOF-sprayed coatings can be used for electrical insulation; an example of
industrial use is an electrical insulating coating in the alternators of cars. Other applications are the use of electrically conducting
coatings for various purposes, including electromagnetic shielding.
4.10.6.8 Steel Industry

Examples of use of thermally sprayed coatings in steel industry are in continuous annealing lines used for the production of sheets
and in continuous galvanizing lines in the galvanizing section, which includes sink rolls. The rolls in the high-temperature
annealing zone are coated by Cr2C3–NiCr by HVOF and detonation spraying. Cermet coatings including MCrAlY as the metallic
part (15–60 vol.%) and oxides or carbides as reinforcements are also used.
The sink rolls are made of stainless steel and are coated with WC–Co coatings, and also with oxide coatings such as Al2O3,
Al2O3–TiO2, and Al2O3–ZrO2. The preferred deposition techniques are HVOF and detonation spraying. Figures 64 and 65 present
examples of thermal spray coating applications in steel industry.
4.10.6.9 Biomedical Industry

Biomedical applications are a relatively novel use for thermal spray processes and coatings. Metallic coatings, particularly titanium
and its alloys, are sprayed onto metal prostheses in vacuum processes to provide nonoxidized titanium coatings. The coatings are
bioinert surfaces with sufficient roughness and porosity allowing mechanical interlocking of the bone tissue. Bioactive HAp or
fluorapatite coatings with bone-type phosphate structures have even better compatibility and performance. These materials allow
growth of bone tissue on the surface of the prosthesis. Different human spare parts, e.g., hip, knee, and tooth implants, are today
coated by vacuum and atmospheric plasma processes. Hip implants with thermally sprayed coatings are illustrated in Figure 66.
Novel approaches in this field of industry are the use of cold spray technology for titanium spraying.
Figure 64 Coated rolls for continuous annealing lines used for the production of sheets. Source: Advanced Coating S.A.
www.iran-mavad.com
Figure 65 Thermally sprayed sink rolls for continuous galvanizing of steel sheet metal. Source: Advanced Coating S.A.
Figure 66 Hip implants with thermally sprayed coatings. Source: Sulzer Metco Inc.
4.10.6.10 Power Generation Plants

Thermally sprayed coatings are also widely used in various power generation plants including fluidized bed combustors, soda
boilers, and waste incineration plants. The main functions of the coatings are to provide resistance against erosion wear,
oxidation, and high-temperature corrosion. Energy industry uses coatings that are applied by combustion wire, electric arc,
plasma, and HVOF spray processes. Wear- and corrosion-resistant coatings are used in surfaces of heat exchanger such as the
water walls, superheater tubes, etc. The coating materials used are usually wear- and corrosion-resistant materials such as iron-
based alloys, e.g., FeCrAl and FeCrTi; nickel-based alloys such as NiCr, NiCrMo (Inconel 625), and NiCrBSi self-fluxing alloys;
and carbide-containing materials such as Cr3C2–NiCr. Several other materials including oxides are also used. Electric arc
spraying and HVOF are the most common techniques, but other novel techniques such as HV-Arc and HVAF are also used.
Sprayed coatings need to be dense if such coatings are used in very difficult corrosive conditions such as waste boilers. Figure 67
presents manual arc spraying of boiler tubes and Figure 68 mechanized spray with wire flame spray process using multiple
torches.
4.10.6.11 Miscellaneous Applications

Further examples of special thermal spray coating applications are electrically insulating Al2O3, superconductive YBaCu oxide
with perovskite-type structure, oxygen ion conduction in fuel cells, molybdenum and tungsten coatings for high-vacuum
technology, and decorative coatings on household products, which are often Al2O3–TiO2 coatings. Figures 69 and 70
present some special applications of thermally sprayed coatings. Ball bearings in Figure 69 have electrically insulating
coatings to prevent degradation of the balls by stray current-induced sparking. Figure 70 in turn presents a plasma spray-
formed free-standing part for lambda sensors of car engines. The coating material is magnesium aluminate, i.e., spinel oxide
ceramic.
The current and future applications of thermally sprayed coatings are numerous. More detailed information about applications
can be found, e.g., from the books in the references.
www.iran-mavad.com
Figure 67 Manual arc spraying of boiler tubes on-site. Source: Telatek Oy.
Figure 68 Mechanized spraying of boiler tubes with three wire flame spray torches. Source: Tocalo Co. Ltd.
Figure 69 Ball bearings with electrically insulating coatings. Source: Tocalo Co. Ltd.
www.iran-mavad.com
Figure 70 Plasma spray-formed free-standing part for lambda sensor. Coating material is magnesium aluminate spinel ceramic. Source: Advanced
Coating S.A.
4.10.7 Development Trends
Thermal spray technology is continuously developing and new applications are seen for thermally sprayed coating materials and
structures. Main development trends are in the following technical areas:
l thermal spray business is smoothly growing in all sectors of use;
l thermal spray processes used are becoming more cost-effective, more robust, less operator dependent, and more productive;
l coatings are more reproducible and properties are more predictable;
l new functional properties are used in applications, e.g., electrical properties, photocatalytic properties, etc.;
l novel spray technologies have been recently taken into research stage and industrial use; examples are HVAF, cold spraying,
suspension plasma spraying, high-velocity suspension/liquid spraying, direct write process, etc.;
l thermal spray coatings are found in new applications, e.g., in fuel cells, as hard chromium alternatives, etc.;
l novel coatings materials and structures are studied and used; e.g., novel nano/submicron-sized carbides, coatings with nanosized
crystals, nanocomposite materials, cost-efficient spray materials, liquids as feedstock materials, suspensions, etc.;
l hybrid processes are studied and used; these include, for instance, combining thermal spray processes with laser novel laser
materials processing; and
l life cycle assessments, environmental considerations, and sustainable use of spray materials and processes.
4.10.8 Concluding Remarks
The present overview of thermal spray coating processing is a generic summary about the topic trying to comprehensively cover
many important related aspects and to give the reader a general knowledge about the technology, formation of the coatings, coating
materials with their properties, and industrial applications of the coatings. The numbers of available processes, coating materials
and properties, and applications for thermal spray coatings are practically endless and cannot be perfectly covered in the frame of
this chapter. Additional information is available from high-level journal papers such as Journal of Thermal Spray Technology (ASM
International) and in related conferences, e.g., ITSC and in its proceedings. Societies such as ASM/TSS, German Gesellschaft
für Termisches Spritzen, and ETSA promote the thermal spray technology and its science, technology, and application.
References
1. Siegmann, S.; Abert, C. 100 Years of Thermal Spray: About the Inventor Max Ulrich Schoop. Surf. Coat. Technol. 2013, 220, 3–13.
2. Davis, J. R., Ed. Handbook of Thermal Spray Technology; ASM International, 2013.
3. Pawlowski, L. The Science and Engineering of Thermal Spray Coatings; John Wiley & Sons Ltd: Hoboken, NJ, 2008.
4. Matejka, D.; Benko, B. Plasma Spraying of Metallic and Ceramic Materials; John Wiley & Sons: New York, 1989; p 280.
5. Heimann, R. B. Plasma-spray Coating – Principles and Applications; VCH Verlagsgesellschaft mbH: New York, 1996; p 339.
6. Schneider, K. E.; Belashchenko, V.; Dratwinski, M.; Siegmann, S.; Zagorski, A. Thermal Spraying for Power Generation Components; Wiley-VCH: Weinheim, 2006; p 271.
7. Beczkowiak, J. Spray Materials. In Modern Surface Technology; Bach, Fr.-W., Laarmann, A., Wenz, T., Eds.; Wiley-VCH: Weinheim, 2006; pp 137–144.
8. Pawlowski, L. Suspension and Solution Thermal Spray Coatings. Surf. Coat. Technol. 2009, 203, 2807–2829.
9. Toma, F.-L.; Berger, L.-M.; Naumann, T.; Langner, S. Microstructures of Nanostructured Ceramic Coatings Obtained by Suspension Thermal Spraying. Surf. Coat. Technol.
2008, 202, 4343–4348.
10. Bolelli, G.; Cannillo, V.; Gadow, R.; Killinger, A.; Lusvarghi, L.; Rauch, J. Properties of High Velocity Suspension Flame Sprayed (HVSFS) TiO2 Coatings. Surf. Coat. Technol.
2009, 203, 1722–1732.
www.iran-mavad.com
11. Monterrubio-Badillo, C.; Ageorges, H.; Chartier, T.; Coudert, J.-F.; Fauchais, P. Preparation of LaMnO3 Perovskite Thin Films by Suspension Plasma Spraying for SOFC
Cathodes. Surf. Coat. Technol. 2006, 200, 3743–3756.
12. Puranen, J.; Laakso, J.; Kylmälahti, M.; Vuoristo, P. J. Therm. Spray Technol. 2013, 22, 622–630.
13. Oberste, J.; Marple, B.; Moreau, C. Suspension Plasma Spraying Nanostructured WC–12Co Coatings. J. Therm. Spray Technol. 2006, 15, 676–681.
14. Great Britain. Wear Resistant Surfaces in Engineering – A Guide to Their Production, Properties and Selection; London Her Majesty’s Stationery Office, Department of Trade and
Industry: UK, 1985.
15. Burakowsky, T.; Wierzchon, T. Surface Engineering of Metals; CRC Press LLC: Boca Raton, FL, 1999.
16. Toyserkani, E.; Khajepour, A.; Corbin, S. Laser Cladding; CRC Press: Boca Raton, FL, 2005.
17. Quint, M.; Kopech, H. High Energy Plasma Ceramic Coating Optimization. In 23rd Annual Conference on Composites, Advanced Ceramics, Materials, and Structures; Cocoa
Beach, FL, USA; 1999; Ceramic Engineering and Science Proceedings. 1999 Ceramic Engineering and Science Proceedings. Vol. 20, Issue 4, pp 335–345.
18. Chráska, P.; Hrabovsky, M. An Overview of Water Stabilized Plasma Guns and Their Applications. In Proceedings of the International Thermal Spray Conference & Exposition;
ASM International: Orlando Florida, USA, 28.5.–5.6.1992; pp 81–85.
19. Sampath, S.; Gansert, R.; Heramn, H. Research Summary: Plasma-spray Forming Ceramics and Layered Composites. JOM (USA) 1995, 47 (10), 30–33.
20. Kreye, H.; Gärtner, F.; Kirsten, A.; Schwetzke, R. High-velocity Oxy-fuel Flame Spraying State of the Art, Prospects and Alternatives. In 5. Colloquium on High Velocity Flame
Spraying; GTS e.V.: Erding, Nov. 16–17, 2000.
21. Nestler, M. C.; Benary, R.; Rusch, W. Innovations in HVOF Coating Solutions. In 5. Colloquium on High Velocity Flame Spraying; GTS e.V.: Erding, Nov. 16–17, 2000.
22. Kawakita, J.; Katanoda, H.; Watanabe, M.; Yokoyama, K.; Kuroda, S. Warm Spraying: An Improved Spray Process to Deposit Novel Coatings. Surf. Coat. Technol. 2008, 202,
4369–4373.
23. Kuroda, Seiji; Watanabe, Makoto; Kim, KeeHyun; Katanoda, Hiroshi Current Status and Future Prospects of Warm Spray Technology. J. Therm. Spray Technol. June 2011,
20 (4), 653.
24. Kawakita, J.; Kuroda, S.; Fukushima, T.; Katanoda, H.; Matsuo, K.; Fukanuma, H. Dense Titanium Coatings by Modified HVOF Spraying. Surf. Coat. Technol. 2006, 201,
1250–1255.
25. Kuroda, S.; Kawakita, J.; Watanade, M.; Katanoda, H. Sci. Technol. Adv. Mater. 2008, 9, 17.
26. Verstak, A.; Baranovski, V. Deposition of Carbides by Activated Combustion HVAF Spraying. In Proceedings of the International Thermal Spray Conference, Osaka, Japan; 10
May 2004 through 12 May 2004; pp 551–555.
27. Verstak, A.; Baranovski, V. Activated Combustion HVAF Coatings for Protection against Wear and High Temperature Corrosion. In Marple, B. R., Moreau, C., Eds., Vol. 1; ASM
International: Orlando, Florida, USA, 5–8 May, 2003; pp 535–541.
28. Verstak, A.; Baranovski, V. HVAF Arc Spraying. Osaka, Japan. In Proc. Int. Thermal Spray Conference; 2004; pp 696–701.
29. Kharlamov, Y. A. Bonding of Detonation-Sprayed Coatings. Thin Solid Films 1978, 54, 271–278.
30. Astakhov, E. A.; Klimenko, V. S.; Skadin, V. G.; Zverev, A. T. Sov. Powder Metall. Met. Ceram. 1982, 21, 307–310.
31. Champagne, V., Ed. The Cold Spray Materials Deposition Process: Fundamentals and Applications; Woodhead Publishing Ltd.: Cambridge, England, 2007; p 362.
32. Papyrin, A.; Kosarev, V.; Klinkov, S.; Alkimov, A.; Fomin, V. Cold Spray Technology, 1st ed.; Elsevier: Netherlands, 2007; p 328.
33. Grujicic, M.; Zhao, C.; Tong, C.; DeRosset, W.; Helfritch, D. Analysis of the Impact Velocity of Powder Particles in the Cold-gas Dynamic-spray Process. Mater. Sci. Eng. A
2004, 368, 222–230.
34. Dykhuizen, R.; Smith, M. Gas Dynamic Principles of Cold Spray. J. Therm. Spray Technol. 1998, 7 (2), 205–212.
35. Borchers, C.; Gärtner, F.; Stoltenhoff, T.; Assadi, H.; Kreye, H. Microstructural and Macroscopic Properties of Cold Sprayed Copper Coatings. J. Appl. Phys. 2003, 93 (12),
10064–10070.
36. Maev, R.; Leshchynsky, V. J. Therm. Spray Technol. 2006, 15, 198–205.
37. Van Steenkiste, T.; Smith, J.; Teets, R.; Moleski, J.; Gorkiewicz, D.; Tison, R.; Marantz, D.; Kowalsky, K.; Riggs, W.; Zajchowski, P.; Pilsner, B.; McCune, R.; Barnett, K. Kinetic
Spray Coatings. Surf. Coat. Technol. 1999, 111, 62–71.
38. Klassen, T.; Kliemann, J. O.; Onizava, K.; Donnen, K.; Gutzmann, H.; Binder, K.; Schmidt, T.; Gärtner, F.; Kreye, H. Cold Spraying – New Developments and Application
Potential. In Proceedings of the 8th Colloquium of HVOF Spraying, Nov. 5–6, 2009, Erding.
39. Pawlowski, L. Suspension and Solution Thermal Spray Coatings. Surf. Coat. Technol. June 25, 2009, 203 (19), 2807–2829.
40. Knuuttila, J.; Sorsa, P.; Mäntylä, T. Sealing of Thermal Spray Coatings by Impregnation. J. Therm. Spray Technol. 1999, 8, 249.
41. Gärtner, F.; Voyer, J.; Qi, X.; Kreye, H.; Richter H. J.; Krömmer, W. New Challenges for Wire and Rod Flame Spraying. In Proceedings of the 6th HVOF Colloquium, Nov. 27–28,
2003, Erding.
42. Schwetzke, R.; Kreye, H. Microstructure and Properties of Tungsten Carbide Coatings Sprayed with Various High-Velocity Oxygen Fuel Spray Systems. J. Therm. Spray Technol.
1999, 8, 433–439.
43. Koivuluoto, H.; Näkki, J.; Vuoristo, P. Corrosion Properties of Cold-Sprayed Tantalum Coatings. J. Therm. Spray Technol. 2009, 18, 75–82.
44. Koivuluoto, H.; Bolelli, G.; Lusvarghi, L.; Casadei, F.; Vuoristo, P. Corrosion Resistance of Cold-Sprayed Ta Coatings in Very Aggressive Conditions. Surf. Coat. Technol. 2010,
205, 1103–1107.
45. Houdkova, S.; Zahalka, F.; Kasparova, M.; Berger, L.-M. Comparative Study of Thermally Sprayed Coatings Under Different Types of Wear Conditions for Hard Chromium
Replacement. Tribol. Lett. 2011, 43, 139–154.
www.iran-mavad.com
4.11 Electroplating and Electroless Plating
A Yli-Pentti, Metropolia University of Applied Sciences, Vantaa, Finland

4.11.2 Principal Elements of Plating 278
4.11.2.1 Principles of Electroplating and Electroless Plating 278
4.11.2.2 Substrates 280
4.11.2.3 Deposition Process 280
4.11.2.3.1 Brightening 283
4.11.2.3.2 Stress in Coating Films 284
4.11.2.4 Principles of Electroless Plating 284
4.11.2.5 Pretreatment 285
4.11.2.6 Equipment 285
4.11.3 Electrolytic Plating 286
4.11.3.1 Zinc Plating 286
4.11.3.1.1 Comparison of Different Zinc Plating Baths 287
4.11.3.1.2 Post-Treatment of Zinc Coating 288
4.11.3.2 Chromium Plating 288
4.11.3.2.1 Bath Chemistry 289
4.11.3.2.2 Decorative Chromium Plating 290
4.11.3.2.3 Hard Chromium Plating 291
4.11.3.2.4 Trends in Chromium Plating 291
4.11.3.3 Nickel Plating 291
4.11.3.3.1 Processes 291
4.11.3.3.2 Properties 292
4.11.3.3.3 Trends in Nickel Plating 292
4.11.3.4 Copper Plating 293
4.11.3.5 Silver Plating 294
4.11.3.6 Gold Plating 295
4.11.3.7 Tin Plating 296
4.11.3.7.1 Processes 296
4.11.3.8 Other Metals 297
4.11.4 Alloy Plating 298
4.11.4.1 Principles 298
4.11.4.2 Brass Plating 298
4.11.4.3 Zinc–Nickel Plating 299
4.11.4.4 Other Alloy Coatings 300
4.11.5 Aluminum Finishing 300
4.11.5.1 Anodizing 300
4.11.5.2 Conversion Coatings of Aluminum 301
4.11.6 Electroless Plating Solutions 302
4.11.6.1 Electroless Nickel 302
4.11.6.1.1 Bath Types 302
4.11.6.1.2 Properties 304
4.11.6.2 Electroless Copper 305
References 305
4.11.1 Introduction
Electroplating is a coating method whereby a component to be plated is set as a cathode into an electrolyte solution containing
metal salts. The material used for plating is set as an anode or, alternatively, an inert anode may be used if the source metal is added
as a metal salt into the electrolyte. The metallic coating, typically within thickness range 1–1000 mm, will be formed in electrolysis. It
is prerequisite that the component surface be electrically conductive either inherently or treated to be conductive before electro-
plating. Typically, the electrolyte solutions are aqueous, but in rare cases molten salt or solid electrolytes could be used.
The main difference between electroplating and electroless plating is that in electroplating an external current source is required
and the process is electrolysis. In electroless plating, the redox reactions happen without an external current source on metallic

www.iran-mavad.com
278 Electroplating and Electroless Plating
Figure 1 Applications of electroplating: Silver-plated jewelry. Pictures are the property of Lapponia Jewelry Oy.
surface, which catalyzes the reactions of the chemicals that are able to reduce the coating metal from its salts. Therefore, it is often
also called autocatalytic plating or chemical plating. The thickness range made by electroless plating is 1–120 mm. An application of
electroless plating requires more control and is more expensive than electroplating. It is applied only when unique properties of
electroless plating, for example, even surface thickness distribution over complex shaped components, are required.
Plating is always a multistep operation that includes cleaning of the surface to be plated, activation of the surface and other
pretreatments, plating steps, post-treatments, and rinsing steps after each solution.
Typical applications for plating are decorative coatings (see Figures 1 and 2), corrosion-resistant coatings, enhancement of
hardness and wear properties of a surface, improvement of conductivity or reflectivity, repair of worn dimensions (Figures 3), and
so on.
In the following text, chemical processing and applications will be discussed. More detailed discussion of bath chemistry and
typical processing and operation sequences can be found in various textbooks (1–5). Detailed operation sequences for plating
solutions are also given by commercial solution providers. The applications for different coatings will be discussed in more detail in
other chapters of this volume. The recent development of plating methods will be discussed.
4.11.2 Principal Elements of Plating

4.11.2.1 Principles of Electroplating and Electroless Plating
Electroplating and electroless plating have their basis in electrochemistry. Plating takes place from electrolytes containing metal salts
by redox reactions; see Figure 4.
The amount of current required in electrolysis follows Faraday’s law:
Q ¼ It ¼ nzF;
where
Q is the quantity of electricity,
I is total current,
t is time,
www.iran-mavad.com
Electroplating and Electroless Plating 279
n ¼ mM1 where n is the molar number, m is the mass, and M is the molar mass of the reduced/oxidized species,
z is the amount of electrons transferred in an electrode reaction, and
F is the Faraday’s constant 96485 Asmol1.
Often in tables the current is expressed as current per area (i ¼ IA1), for example, in unit Adm2.
The applied current is the total current flowing through the electrolysis cell. During plating operations, there will be side
reactions besides the precipitation reaction in target, typically, hydrogen evolution. In alloy plating, a total current by the Faraday
equation includes all the partial currents of depositing species. The Faraday equation describes the total current of all reactions,
including side reactions. The amount of current consumed by side reactions varies in great amounts, depending on the plating
process – for example, from nearly zero percent in silver plating to 70–85% in hard chromium plating. The maintenance of a plating
bath composition has an effect on the amount of side reactions. The Faraday equation does not give information about the
proportions of the electrolysis reactions; therefore, the current efficiency is often determined separately and is used to estimate how
much of the total current is used for the desired plating reaction. The average coating thickness can be calculated by recording the
amperes and time, calculating the mass of the coating using current efficiency and dividing the result by the density and surface area
of the plated part.
Figure 2 Applications of electroplating: Gold-plated jewelry. Pictures are the property of Lapponia Jewelry Oy.
Figure 3 Applications of electroplating: Hard chromium-plated areas on a wheel axle. Plated areas are ground to measure tolerance.
www.iran-mavad.com
Power source
Direction of
– + electron flow
Cathode = an
object to be Copper anode
plated
2+ 2– 2–
Cu SO4 SO4
– –
e e
2+ 2+
Cu 2–
Cu
– SO4 –
e e
Copper sulfate electrolyte
Figure 4 A schematic presentation of a copper electroplating cell.
4.11.2.2 Substrates
The most common substrate to be treated is low-alloy steel. It can be plated with almost any metallic coating, but zinc, nickel, and
chromium are the most prevalent ones. Typical applications are corrosion-resistant coatings with zinc or electroless nickel, wear
resistance coatings with chromium, sliding surface coating with copper or cobalt, and repair coating with nickel and chromium.
The maximum operating temperature has an effect on the selection of a coating. Stainless steel products can also be plated,
although it is not necessary in most applications. Wear resistance, sliding surface, and repair coatings can be done with similar
methods as for carbon steel products, provided that pretreatment of the metal includes activation steps that break down the
passivation layer inherent for stainless steels. Nickel and cobalt-based alloys can also be plated analogously to stainless steels.
Copper and copper alloys can be plated with other metals quite readily. Anodizing is the most common coating method for
aluminum and magnesium to protect the surface from corrosion and to have wear resistance. Aluminum can be plated with other
metallic coatings using a multistep process that provides nickel coating on aluminum, which can thereafter be plated with other
coatings. Titanium is not usually plated, but in rare cases it can be done to achieve better sliding properties and wear resistance or
repair coating.
Plastics and other nonconductors can also be plated. Plastics like ABS (acrylonitrile-butadiene-styrene) and its blends with
polycarbonate (PC) have been plated for a long time. Recently, polyamide and polyoxymethylene (POM) have gained more
ground. For plastics, the steps before plating are etching the surface and activating it with precious metal species and electroless
deposition of a conductive metal film. Other nonconducting surfaces than plastics can also be made conductive in various ways,
using, for example, metallic or graphite powder, metallic paint, or other metallizing techniques.
4.11.2.3 Deposition Process

When metallic material is dipped into the electrolyte solution in a plating cell, a potential difference will form between anode and
cathode. When no current is passed through the cell, the potential difference depends on the electrode potentials, temperature, and
the concentrations, or more correctly, the activities of the species involved in redox reactions. The activities of the ions may differ
from concentration due to the interatomic effects and especially due to the complex agents. The activity of complex ions is much
lower than simple or hydrated ions, which has an effect on electrode potentials.
When the external power source is switched on, the potential between electrodes will change to a new value. The electrodes
become polarized. At the cathode surface, the concentration of the metal ions will be reduced. A diffusion layer will form on the
surface of the electrode. Its thickness in a solution depends on the agitation of the solution, deposition rate of the metal ions, and
temperature. Typically, the thickness is some fractions of a millimeter. The applied current affects the deposition rate. The diffusion
does not provide enough mass transfer through the layer, and a potential difference across the diffusion film will be formed.
Eventually, when the potential at the external power source is raised, the concentration of the deposited ions will be zero at the
metal surface and the limiting current density has been reached. The current will grow further only by side reactions such as
hydrogen evolution. Using potentials higher than at limiting current density, the coating quality will suffer greatly, leading to ‘burnt’
or powdery deposits.
Other factors that polarize the electrodes are the energy required for activation of the redox reactions and the ohmic resistance of
the plating solution. Also, the power loss in external current circuit may be substantial, for example, due to the resistance across
contact surfaces of the plating fixtures and racks. Usually, the plating step is not controlled by voltage but by current or ampere-
hours, which determine the total amount of plated metal when current efficiency is known.
www.iran-mavad.com
At the surface of the metal, there is a layer that is a few atoms thick, where hydrated metal ions are adsorbed on oppositely
charged metal surfaces. It is called the Helmholtz double layer after the inventor of the model. The charge transfer occurs within this
region. When the charge transfer takes place, the neutral metal ions will diffuse on the surface toward the metal lattice and crystallize
to have an energetic favorable state (see Figure 5). Specifically adsorbed anions and surfactants may have an effect on crystal
structure.
The current distribution at the cathode surface has a great effect on the quality of the coating. The current distribution is
generally divided into primary and secondary current distribution. The origin of the primary distribution is in the geometry of
the surface and its effect on current flow. No polarization is taken into consideration, only the size, shape, and distances of the
plated object, anodes, and tank walls and surfaces, the conductivity of the object, electrolyte, and anodes. Edges and corners
will have higher current density, leading to a thicker coating thickness or even burned coating. On the contrary, hollows or
recesses will have a thinner coating than average or may even be left bare. Figure 6 shows that the edges of the test panel have
gotten more current than the middle area. The layer thickness of the edges is nearly double compared to the center, and the
coating is burned.
W
W W
Electrolyte + W
W
W
1.
Diffusion layer
W W
W + W
2.
W + W
3.
Double layer
5.
4.
Coating layer
Metal surface + Metal ion
W Water molecule
Figure 5 Processes during bright plating: (1) Convection of hydrated metal ion from bulk solution. (2) Diffusion toward cathode surface through
diffusion layer, where part of the hydrated water molecules will be lost. (3) Charge transfer in the double layer, metal ion will be electrically neutralized, and
more hydrated water will be lost. (4) Surface diffusion of metal atom. (5) Crystallization into coating metal lattice.
www.iran-mavad.com
Figure 6 Burnt deposit on the edges of a bright nickel-plated test panel.
The secondary current distribution is caused by the concentration polarization. It will appear after a short time when electrolysis
is performed. It will level the effects of the primary current distribution. When a peak at the surface gets higher current density due to
the primary current distribution, the concentration of the metal ions will be depleted faster near the peak than in the valley, which
will lead to higher polarization resistance due to both activation and concentration change, hence leveling the primary current
distribution.
The relative sizes of the activation, concentration, and ohmic polarization are determined by the size of the geometrical profiles.
The smaller the surface profile is, the bigger is the relative influence of the secondary current distribution. This is due to the thickness
of the diffusion layer on the surface. If the profile size is about the same size as the diffusion layer, the secondary distribution is
dominant, while in larger profiles the primary profile is prevailing.
The throwing power of a bath is used to express how uniform coating thickness on the plated part can be expected. Macro-
throwing power is used to describe a bath’s ability to produce an even coating thickness on the large surface of a part, when the
dimensions of the profile under consideration are much bigger than the diffusion layer on the surface. Macrothrowing power can be
expressed as a relation of the lowest and highest coating thickness, and it should be near 1. The baths that have good macrothrowing
power will form a thick diffusion layer. Terms that increase polarization, that is, lowering metal ion concentration, increasing
current density, and decreasing agitation, will enhance macrothrowing power.
Microthrowing power describes the bath’s ability to give an even layer on the microprofile of a component to be plated.
Microprofile dimensions are smaller than the thickness of the diffusion layer on the component. If the layer thickness is even, it is
considered as good microthrowing power, and if the microprofile grows deeper, the bath has poor microthrowing power. The bath
may also smooth the profile and is considered to have microleveling, or simply leveling abilities. In microprofile, the diffusion layer
thickness varies between profile peaks and recesses. Depending on the sizes of the concentration and activation polarization, the
amount of metal deposition is different at the peaks and recesses. For example, when current density is high, and the diffusion of
the metal ions through the diffusion layer controls the deposition rate, the microthrowing power will be poor as the peaks, where
the layer is thinner, will grow faster than the recesses.
The leveling ability of a bath is the ability to make microprofiles more even. Peaks will have thinner coatings than recesses.
Usually, this effect is achieved by additives, which adsorb on peaks and slow down the metal deposition. Sometimes the bath has
inherent leveling abilities when the kinetics of the reduction reaction is depending on the concentration and current density. If the
peak gets higher current density and/or the metal concentration gets lower than in the valley, the kinetics of the reaction may slow
down and lead to leveling.
The minimum current density for layer growth is called coverage and can be determined by the Hull cell. The coverage depends
on current density, base metal properties, and metal ion concentration. Sometimes it is necessary to start the plating operation with
higher current density to initiate the layer growth, after which the current may be lowered. For example, in chromium plating, higher
current densities at the start may be required, while in silver plating the coverage is usually very good even with small current
densities.
In most cases, direct current is used in electroplating, but pulse plating and reverse pulse plating can also be used. Between the
current pulses, during pause, the diffusion layer will vanish, and the microthrowing properties will be better. If a reverse pulse is
www.iran-mavad.com
used, some metal will be dissolved from the corners and peaks. Larger current densities can be used with the pulse plating and the
quality of the coating will be better.
The electrolyte as a plating bath has several functions: It holds the metal salt, it is the conductive path for electric current, it may
prevent the surface from passivation, and it may clean the surface, although most of the cleaning should be done during
pretreatment. Plating solutions usually have a rich metal salt content to prevent the rapid depletion of the metal ions. Anions are not
reduced, but they affect the activity of the salt and may take part in the film formation process by specific absorption on the cathode
surface and thus control the structure of the coating. Typically, chlorides, sulfates, sulfamates, sulfonates, and fluoroborates are used
in simple salt baths. Complex ion solutions have been used extensively, mainly alkaline cyanide baths for silver, gold, copper, zinc,
and cadmium. The use of cyanide is still wide, although a lot of effort has been made to replace it due to the potential occupational
hazard for plating shop workers and a potential environmental hazard. Other complex ions used often in plating are stannates,
thiosulfates, amines, tartrates, and citrates. Wetting agents are required to help the air and evolved hydrogen gas bubbles to detach
from the surface, thus preventing pitting. Conducting the current through the electrolyte requires high conductivity. Some of the
power is consumed due to the ohmic loss of the solution. Conducting salts are added to the electrolyte in order to prevent ohmic
losses. They often act as pH buffers, and they also keep the plating operation conditions stable. Carbonates in cyanide baths, boric
acid in nickel baths, and acetates in zinc sulfate baths are widely used as buffering agents. Finally, some ions are added to promote
anode corrosion. For example, in some formulations, chloride ions are added to the sulfamate nickel bath for this reason.
Electroplating baths can usually be maintained to last a very long time, provided the amount of accumulating impurities is low.
However, there will be environmental load caused by the rinse waters that must be purified by appropriate means.
4.11.2.3.1 Brightening
Metal surfaces appear bright when they reflect most of the light with the same angle as the angle of arrival. When the surface is rough
and the dimensions of roughness are greater than the wavelength of the light (0.4–0.8 nm), the ray of light will diffuse into different
angles and the surface will appear as matt, see Figure 7.
Bright plating baths contain additives that are called brighteners. There are a large number of different kinds of these
compounds. They enhance both the leveling and brightening abilities of a bath at wide current density range. Often, a brightener
system is used, which consists of primary and secondary brighteners. Primary brighteners are usually organic compounds that
modify the microstructure of the coating and reduce the internal stress of the coating, preventing it from cracking or peeling. Thus
primary brighteners are required for bright coating, but alone they would not be sufficient for coating thicknesses used in practice.
Bright coatings have usually a lamellar structure. Without brighteners, the coating will have columnar structure that is coarse and
makes the appearance typically matt. The brighteners adsorb on cathode, which will add the polarization and slow down the crystal
formation. Brighteners also enhance nucleation, which will lead to a more fine-grained crystal structure. The former effect somewhat
contradicts the desired fine-grained structure, but probably by slowing down the growth rate the crystals will grow faces parallel to
the substrate, leading to a bright appearance. Saccharin or sulfonamides are typical and traditional examples of primary brighteners
in bright nickel baths, while polyvinyl alcohol is used in cyanide zinc baths. In nickel baths, primary brighteners are usually
consumed slowly during operations and no frequent additions are necessary. Their concentration is typically 1–10 g l1, and the
additions are made based on used ampere-hours.
Secondary brighteners may be organic or inorganic compounds that further refine the grain structure and leveling to make the
coating brighter. Leveling is typically achieved by adsorption on the profile peaks, thus hindering coating growth on peaks and
enabling the growth on lower areas of a profile. They usually make the coating harder and create tensile stresses in coating.
Secondary brightener molecules usually contain double or triple carbon–carbon, oxygen, or nitrogen atom bonds, for example,
aromatic sulfones or sulfonates in nickel plating, or aromatic aldehydes in cyanide zinc plating (6). Their concentrations may vary
considerably, between 0.01 and 10 g l1, depending on brightener type. They will coprecipitate and decompose during plating
operations and require more frequent additions than primary brighteners. However, they should not be added in excess because it
(a) (b)
Figure 7 Surface brightness depends on its roughness: surface roughness (a) greater, (b) smaller than the wavelength of light.
www.iran-mavad.com
could lead to brittle or porous coating; in large excess they could precipitate out of the solution. Due to the decomposition, the
coating will contain remains of a brightener, that is, sulfur.
A third group of compounds that affects brightening is wetting agents, which prevent the evolved hydrogen bubbles to attach on
a surface. Otherwise, pitting by gas bubbles could deteriorate the brightness. Some additives may also affect the pH of the surface to
reduce hydrogen evolution, or they may react with hydrogen to prevent gas bubbling.
4.11.2.3.2 Stress in Coating Films

Internal stress in coating films is common for many plated metals, for example nickel, chromium, and gold. Usually the stress is on
the tensile side, which is harmful for coating and substrate. The coating may be cracked or detached, and the stress may cause crack
formation or propagation for the substrate as well. Thin-wall parts may suffer deformation, and the plated parts are more vulnerable
for hydrogen embrittlement. However, the dislocations in crystal structure make the coating harder and more brittle. Compressive
stresses in a coating are usually beneficial, since they add the parts tolerance for cracking. The intensity of compressive stresses is
usually smaller than tensile stresses, so peeling or deformation is not an issue.
The origin for internal stresses is in crystal imperfections, which result from codeposition of impurities or other dislocations
formed during electrocrystallization. The most common codeposited constituents are hydrogen and sulfur. In chromium plating,
the codeposited hydrogen will form chromium hydrides that will be depleted from the coating after plating, causing internal stress;
see Section 4.11.3.2.1.1. Many factors affect the magnitude and nature of internal stresses, for example, bath type and concentra-
tions, involvement of chlorides in baths, bath impurities, current density, temperature, and pH. The stresses can be relieved by
a mild heat treatment.
4.11.2.4 Principles of Electroless Plating

In electroless plating, a chemical reducing agent reduces metal ions to form a coating on the surface that catalyzes the reduction
reaction. No external power is required, but the solution must usually be heated. The coating metal must also catalyze the reaction
to keep the coating process going on. Metallic coatings that are typically made with the electroless method are nickel and copper, but
there are also applications for cobalt, silver, gold, and palladium.
The method is in wide use due to its advantageous properties: Deposit thickness is uniform even on complex shaped parts and
small cavities, no electrical contact is required, nonconductors like plastics can be plated, and the deposits will have low porosity
and unique chemical, mechanical, and magnetic properties, for example, corrosion and wear resistance.
Typically, sodium hypophosphite is used as a reducing agent. Also, formaldehyde, hydrazine, borohydride, amine boranes, and
their derivatives could be utilized (7). However, hydrazines are carcinogenic substances and have no practical significance. When
sodium hypophosphite is used, the coating will also contain phosphorus, while with the other reducing agents the coating will
contain either boron or they may be almost pure metal. The hypophosphite will react according to
2H2 PO2 þ 2H2 O/2HPO3 2 þ 4Hþ þ 2Hads þ 2e
The metal will be depleted by the reaction
Me2þ þ 2e /Me
The total reaction for metal deposition is thus
2H2 PO2 þ 2H2 O þ Me2þ /Me þ 2HPO3 2 þ 4Hþ þ H2
Only about 35% of the hypophosphite will react according to the previous reaction while codeposition of phosphorus and
hydrogen evolution will take place according to
Hads þ H2 PO2 /P þ OH þ H2 O
2H2 O þ 2e /H2 þ 2OH

The phosphorous content in electroless nickel coatings is about 3–15 wt.%, depending on the bath chemistry.
See Figure 8 for a schematic presentation of the above reactions.
When borohydride or boranes are used as reducing agents, the reactions are as follows (7):
BH4 þ 8OH /BðOHÞ4 þ 4H2 O þ 8e
BðOHÞ4 þ 3e /B þ 4OH
R 2 NH BH3 þ 3H2 O/R 2 NH2 þ þ H3 BO3 þ 5Hþ þ 6e

An electroless plating bath typically has several components: a source of metal, reducing agent, chelating agents, buffering
agents, accelerators, stabilizers, and surface active agents. A metal sulfate is a typical metal source. Chelating agents are required to
keep the metal in solution for reduction reaction. Buffering agents are required to keep the pH value of the bath in such limits that
harmful side reactions are prevented when hydrogen ions are produced in the reduction reactions, thus having a strong effect on
www.iran-mavad.com
H2
Ni
P
HPO2–
salt 3
Figure 8 Processes during electroless plating.
solution pH. Accelerators are required to raise the precipitation speed to a practical level, and at the same time stabilizators are
required to prevent the autocatalytic reaction rate running out of control. Reduction reaction products like orthophosphates will
stay in the solution and gradually increase the salt content of the bath. Finally, elevated salt content slows down the reaction rate and
deteriorates the coating quality, which leads to the replacement of the solution.
4.11.2.5 Pretreatment
The need and methods for treating the component surface before plating heavily depend on the substrate and its quality.
Typically, all loose impurities should be removed, and possible passivation layers should be broken to get good adhesion for the
plating. Also, nonconductors can be plated by etching and activating the surface with a catalyst, followed by electroless coating.
Mechanical cleaning with abrasive blast methods are used to remove loose particles and rust or oxide layers from the surface.
Organic dirt is removed with various methods: Degreasing by solvents is an efficient method, but due to environmental reasons it
has largely been replaced by aqueous cleaning methods. Emulsion cleaning can be used for heavy dirt. Alkaline cleaning is a good
method for steels, but high-alkaline solutions will etch aluminum surface. Mild alkaline or acidic cleaners should be used for
aluminum. Copper cleaning should be done by mild nonoxidizing acid. Zinc as an active metal should be cleaned in mild
alkaline cleaner. Mechanical or ultrasonic agitation is frequently used to enhance the cleaning process. In electrocleaning an
electric current is used to form bubbles of hydrogen and oxygen from water in alkaline cleaner to enhance agitation. When
organic impurities have been removed, the surface is usually activated with acid pickling to remove inorganics and to break the
oxide layer.
4.11.2.6 Equipment
Plating can be done either by dipping the parts in a tank or brushing the parts with a specific brush that is wetted with the
electrolyte. Brush plating is frequently used in the repair coating of large parts. Plating techniques are usually simple and
relatively cheap. The scale of the plant can be anything between a teacup size used in small-volume jewelry plating and
industrial-scale production plants for coating components used in machinery or constructions. In all scales, the basic elements
are open tanks, power supply equipment, and ventilation. Electroless plating does not need power supply for plating but may
need it for anodic protection of the equipment. Depending on the scale and the coating, a specific building, piping systems, and
wastewater treatment facilities, exhaust air cleaning systems, heating and filtration systems, cranes and automation systems are
typically installed to comply the obligations of occupational and environment safety, quality and production efficiency. See
Figures 9 and 10 for a schematic presentation of a plating line. Monitoring of the chemistry of the plating baths must be
arranged. In metal-coated steel sheet production the plating line can also be arranged to convey a continuous metal strip through
a plating station.
www.iran-mavad.com
Figure 9 A design for a large anodization plant. Picture is property of Oy Galvatek Ab.
Figure 10 A design for a large anodization plant. Picture is property of Oy Galvatek Ab.
4.11.3 Electrolytic Plating

4.11.3.1 Zinc Plating
Zinc electroplating is widely used because of its ability to protect steel from corrosion. The sacrificial nature of the protection and
tolerance of scratches makes its performance superior to paints alone. However, use on base metals other than steel or stainless steel
is rare. Zinc plating can be prepared to have an attractive appearance, and as the cost is relatively low it is a very popular coating both
for small parts like bolts, nuts, rivets, washers, nails, hinges, and hooks, and for automotive parts, interior components, and so on. It
also works also as an effective undercoat for paints.
Typical minimum thickness ranges are specified between 5 and 25 mm (8). The corrosion resistance of a coating depends on the
layer thickness and applied post-treatment.
Three main processes are used for zinc plating: cyanide, alkaline noncyanide, and weakly acidic chloride solutions. The choice of
the solution depends on the effluent treatment abilities of a plant and the type and material of components to be plated.
The cyanide process has a good macrothrowing power, and it thus gives quite even coating thickness regardless of the
geometrical shape of the component. The bath is easy to maintain, and it tolerates rather large variations in concentrations. As the
bath is alkaline, it does not corrode steel parts, and it also has surface cleaning properties to tolerate small amounts of organic
impurities. However, due to the cyanide content, the bath is extremely toxic, which causes high rinse water and waste costs. At high
current densities, the current efficiency is low, and the risk of hydrogen embrittlement for heat-treated and carbonitrided parts is
high. The other downsides of cyanide baths are poor microthrowing power and difficulties in the plating of cast iron. This is due to
the graphite on the surface of cast iron and its lower hydrogen overvoltage in cyanide solution compared to zinc, which will lead to
hydrogen evolution on graphite spots instead of zinc precipitation (9).
www.iran-mavad.com
Zinc exists in cyanide baths as a tetracyano complex Zn(CN)2 4 , which dissociates at the same speed as the Zn
2þ
ions are
precipitated at the cathode (10). The association with the cyanide complex, which as anionic ion causes a large concentration
polarization, enhances the macrothrowing power of the bath, but makes the current efficiency low. The microthrowing power
(leverage) of a cyanide bath is poor. Cyanide complex makes the crystal structure of the precipitated metal fine, which gives a good
basis for brighteners to make bright coatings, although the deposit brightness from a cyanide bath alone is not as good as that from
other bath types. In alkaline cyanide baths, the zinc ions are also associated with hydroxyl ions as Zn(OH)2 4 , and there is a balance
between cyanide and hydroxyl complexes, which depends on the amount of sodium hydroxide added to the bath.
Due to the adverse effects of toxic cyanide, its amount in the bath has varied from high-cyanide to middle- and low-cyanide
baths. The total sodium cyanide content of the high-cyanide bath is 75–115 g l1, while in middle-cyanide baths it is 35–55 g l1
and in low cyanide 6–15 g l1 (10–12). The drag-out and waste disposal costs are lower in low-cyanide baths, but the requirements
for pretreatment and operating parameter variation are higher. The sodium hydroxide content should be high enough,
80–100 g l1, to have good conductivity and anode dissolution and to produce good brightness.
The brighteners added to the cyanide bath are usually organic, since the concentration of metal brighteners is too critical.
Usually, primary and secondary brighteners exist. The most common primary brightener is polyvinyl alcohol (PVA), while the
typical secondary brighteners are smaller molecules that contain unsaturated bonds and polar groups, for example, aromatic
aldehydes and pyridines.
The cyanide baths also contain sodium carbonate, which is formed when cyanide molecules are oxidized by the oxygen in air.
Some carbonate is required to form a dense coating layer, but an excessive amount, over 70–80 g l1, must be removed from the
bath by lowering the bath temperature below the solubility limit of sodium carbonate.
Important operating parameters are relations of NaCN and Zn concentrations and cathodic current density. The higher the
amount of Zn, the lower is the factor NaCN/Zn, which enhances the current efficiency. However, the macrothrowing power will be
poorer, and operation of the plating process is more demanding. The selection of a cathodic current density depends on concen-
trations, brighteners used, temperature, and agitation. Typical values are 2–5 Adm2. Higher current densities will lower the
cathodic current efficiency. The cyanide baths are operated at room temperatures.
The purity of an anode material is important for cyanide baths, since it affects the brightness of the coating layer. Usually, the
anode material is over 99.99% pure zinc, either in ball or bar form. When the bath is unused, the zinc will dissolve to the bath; to
avoid excessive zinc buildup, the anodes should be removed.
Alkaline noncyanide zinc baths may be prepared by dissolving zinc oxide ZnO to sodium hydroxide and adding brighteners. The
bath will contain zinc 8–10 g l1 and NaOH 90–120 g l1, and the zinc will be in the solution as Zn(OH)2 4 -ions (11,12). Most
properties of the bath are determined by the brightener systems, which are usually patented and may include, for example, PVA and
imines. The good properties of the baths are low metal content and usually inexpensive effluent treatment, good brightness, good
macrothrowing power, and moderate microthrowing power. The danger of hydrogen embrittlement for high heat-treated steels is
lower compared to cyanide baths. The bath requires better operation control than cyanide baths; in particular, the metal content
must be controlled during idle periods, since too high metal concentrations will deteriorate the brightness. However, too low
a metal content will lower the current efficiency. The bath is operated at the room temperature.
Weakly acidic chloride solutions have gained more market share as the properties of the solutions have been developed. The
bath contains zinc 15–30 g l1 and either ammonium or potassium chloride, so that the chloride content of the bath is
110–150 g l1 (10,11). Organic brighteners must be included in the bath. The optimal pH range is 4.5–5.5. Zinc will be dissolved
into the solution as Zn2þ-ions. The precipitate will be columnar and rough unless brighteners are used. Very bright deposits can be
achieved by the use of brighteners, and weakly acidic solutions are generally considered to give the best brightness among the zinc
plating processes. The current efficiency is at the same time very high, around 95–98% at a large current density area, which is
beneficial for the plating of carbonitrided steel and cast iron. The microthrowing power is very good, but as the current efficiency is
high, the macrothrowing power is poor. This, along with the bath chemistry, requires that the parts be cleaned carefully before
plating, and much attention must be paid to the plating fixtures. The adverse effect of the acidic chloride solution is the potential
corrosion of the substrate materials and staining of the coating unless it is carefully rinsed off. Therefore, spot welded or other parts
with lamellar structures are not suitable for the solution. The effluent treatment is usually easy, although ammonium may disturb
the operation of an effluent plant by forming complexes with heavy metals. The bath is operated at room temperature, and current
densities up to 6 Adm2 can be used without significant drop in the current efficiency.
4.11.3.1.1 Comparison of Different Zinc Plating Baths

There is no significant difference in corrosion resistance between the coatings deposited from different bath types.
Cathodic current efficiency is quite high, over 90%, with every type of bath when the current density is small. When the current
density is raised, the efficiency with weakly acidic baths will stay at a high level, but the cyanide and alkaline baths will have
efficiency drop under 90% when the current density is raised from 2 to 4 Adm2. Raising temperature and zinc concentration will
raise the current efficiency of alkaline and cyanide baths. Temperature has little effect on acidic baths, but on contrary zinc
concentration it has a significant effect.
The macrothrowing ability of the cyanide bath is usually higher than that of acidic baths, since the cyanide complex will raise the
concentration polarization. Zincate baths have the best macrothrowing power. Slightly acidic baths have poor macrothrowing
power unless brighteners that increase polarization are added. Elevated zinc concentration will diminish the macrothrowing power
due to depolarization.
www.iran-mavad.com
Cyanide baths have poor microthrowing power, but slightly acidic and modern zincate baths have good microthrowing power
and the crystal structure of the layer is fine. Thus the brightest surfaces can be deposited from acidic baths.
The danger for hydrogen embrittlement is higher in cyanide baths compared to slightly acidic baths, because the hydrogen atoms
are not easily combined to evolve hydrogen gas on surfaces but can be diffused into the metal lattice. However, the corrosion of the
components is higher in acidic baths.
4.11.3.1.2 Post-Treatment of Zinc Coating

Corrosion-resistant coatings of zinc and zinc alloys are usually formed by combining two or three coating layers: At first there is the
zinc or zinc alloy layer, followed by a conversion coating and an optional seal or topcoat. Zinc coatings are usually made passive
after plating with a conversion coating to enhance the corrosion protection of a newly formed coating against the formation of
‘white rust’, which is a loose layer of zinc hydroxide and carbonate. It does not ruin the corrosion protection properties of the
coating, but brightness and good appearance will be lost. Passivation enhances brightness, gives a color for the surface, enhances
the performance of zinc as a corrosion protective layer, reduces the sensitivity for fingerprints, and improves the adhesion of the
following oil, wax, paint, or lacquer layers.
Previously, passivation was practically a synonym for chromate coating because the layer was done using hexavalent chromium
solutions. During the first decade of 2000, legislation both in the United States and Europe forced manufacturers to replace hex-
avalent chromium in a number of applications due to its adverse effect on health and the environment. Restriction of Hazardous
Substance (RoHS) (13), Waste Electric and Electronics Equipment (14), and End-of Life Vehicle (ELV) (15) Directives were
implemented in 27 European Union member states by the end of 2007. Many U.S. states and Canada, China, Japan, and Taiwan are
also enacting similar kind of regulations.
Many alternative coatings have been developed that have been based on cobalt, zirconium, titanium or rare earth metal salts,
iron, manganese or chromium phosphates, organic polymers, permanganates silanes or carboxyl (16,17). In the post-treatment of
zinc and zinc alloys, products with trivalent chromium have been most successful in replacing hexavalent chromium. Trivalent
chromium compounds are readily available; moreover, they are much less toxic, noncarcinogenic, and safe to handle. Their
performance is in most properties equal or better. The trivalent option can be found in almost all colors and thickness ranges; the
corrosion protection properties as measured by neutral salt spray test are equal or better; and their endurance for heat treatment is
much better. The color is, however, achieved by using dyes, which tend to weaken the corrosion resistance of the coating. Almost the
only respect in which the trivalent products are not as good as hexavalent is the self-repair property of mechanical scratches, which
requires more care during manufacturing and handling.
The mechanism for forming a protective layer is almost similar to that with trivalent and hexavalent chromium compounds
(18). When a part is immersed in a solution containing hexavalent chromium, zinc is oxidized by reacting with Cr6þ, which
turns to Cr3þ. Zinc oxide dissolves to the acidic solution and reacts with trivalent chromium to form zinc chromium oxide
compounds, which will precipitate on the surface, forming a gelatinous film. The film will consist of hydrated chromic-
chromate, chromium hydroxide, zinc, and other metal oxides. Some soluble hexavalent chromium is trapped in the film from
the solution and is leachable, which is the cause for self-healing scratches. The reaction with trivalent chromium is similar
except for the oxidizing step by hexavalent chromium. The trivalent chromium reacts directly with dissolved zinc. The bone
structure of the coating layer is in both cases similar, consisting of stable trivalent chromium oxides, but when prepared from
trivalent bath, the hexavalent chromium is missing. Similar layer thicknesses, however, can be achieved, although the kinetics
of the reaction is different.
The barrel-plated work of trivalent passivation is vulnerable for mechanical damages due to the lack of self-healing properties. It
can be overcome by the use of sealers or topcoats. Post-passivation treatments are used to enhance lubricity, torque–tension
properties for fasteners, and the outlook of colored coatings.
The hexavalent passivation layers lose their properties when heated to over 55 C over 5 min. The trivalent layers are much more
durable against heat treatment; they can be heated to 200 C for extended periods and still maintain the better part of their
properties. The heat endurance property is beneficial when parts must be heat treated for hydrogen embrittlement relief.
Thin layers of trivalent passivation, 60–80 nm, typically produce iridescent blue color. Thicker layers, 300–350 nm, produce pale
green or yellow colors, much depending on the alloying of zinc and the alloying element. The yellow color is not inherent for
trivalent layers as it is for hexavalent, and the color is achieved by dyes or other metal oxides. Also, black color can be achieved with
the help of cobalt and iron oxides.
The layer thickness of the post-treatment affects the corrosion protection abilities comparable to hexavalent chromium. A mild
effect on layer color also exists.
Post-passivation treatments may be necessary to modify the color, enhance corrosion resistance, or provide better lubrication or
torque–tension properties for the surface. These include silicate, phosphate, or transition metal sealers and organic lacquer wax or
oil topcoats, lubricants, and color dyes.
4.11.3.2 Chromium Plating

The chromium plating deposits can be split into two main categories: decorative and functional (hard, engineering) chromium
plating. The decorative type is reflective and gives corrosion protection for the underneath layer, which is typically nickel plating. The
functional is used for engineering applications due to its good friction, wear, and heat-resistance properties (19). Both coatings can
www.iran-mavad.com
be prepared from a similar hexavalent solution of chromic acid using sulfuric acid as a catalyst (20). Current density and operating
temperature are different, however, and to get an optimal plating result differing chemistries should also be employed. For
decorative plating underneath nickel, activation and good coverage are required from the bath, when plating speed and the
properties of the coating are important for the hard chromium.
Trivalent solutions are also commercially available for decorative applications, for which they provide an equal or applicable
coating. For functional applications, the trivalent baths have not yet been developed.
4.11.3.2.1 Bath Chemistry

4.11.3.2.1.1 Hexavalent baths
Chromium plating is traditionally made from a solution of chromic acid (CrO3 which forms H2CrO4 after dissolving in water) and
sulfuric acid (H2SO4) using insoluble anodes. The chromium is reduced to metal from the hexavalent state due to the catalytic effect
of sulfate ions. There are also other catalysts in addition to sulfate, for example, fluorides and fluorosilicates. Sometimes several
catalysts, for example, sulfate and fluoride, are used together, and those baths are called mixed-catalyst baths. The process with
mixed-catalyst baths is faster, has better coverage, current efficiency, and larger operating range, is less sensitive to impurities and
current interruptions, and is harder and more corrosion and wear resistant in hard chromium applications, but it is more expensive
to use and fluorides are more corrosive, which may be detrimental for uncovered areas of the parts to be plated (21). Another type of
a chromium bath is self-regulating high-speed chemistry, which contains strontium and calcium salts to get the bath components
beyond the saturated state so that part of the salts are in solid form at the bottom of the tank (22). Thus, there is far less need to
analyze and make additions to the bath compared to the traditional ones. They also use mixed catalysts. There are also proprietary
baths that corroding fluoride catalysts have been eliminated.
The mechanism of the precipitation from hexavalent ion is complex since six electrons for each deposited chromium ion must be
involved. When there are sulfates in the hexavalent solution, the intermediate anionic chrome species Cr3O 8 will react first with
hydrogen ions to bivalent chromium hydroxide, which then forms a polarized complex bisulfate ion (23). The chromium end of the
complex has a positive charge. In the bisulfate complex, chromium ion is able to go through the Helmholtz double layer, adsorb
specifically on the surface, and reduce to metallic chromium, releasing sulfate ion back to the solution to continue working as
a catalyst. Sulfate has a double action since it prevents aquo complexes from forming by shielding chromium atoms.
Other reactions that occur are hydrogen evolution and formation of trivalent chromium. The latter will migrate on an anode and
oxidize back to a hexavalent form with the catalytic help of PbO2 film at the lead anode. To keep the trivalent species concentration
low enough, the anode area must be large compared to the cathode area; typically, proportion 2:1 or more is considered adequate.
Trivalent ion will suppress the deposition rate in hexavalent baths and is thus considered an impurity. The cathode reactions
consume hydrogen ions, which will lead to high pH rise at the surface film of a cathode. This is why trivalent baths have not long
been successful, since trivalent hydroxides are formed in locally high pH environment. They will form polymeric compounds
through oleation and eventually sediment of insoluble chromium hydroxide (24). When the polymers or sediment cover cathode
surfaces, no metallic chromium can be precipitated and only thin layers of chromium have been capable of precipitating from
trivalent baths. Recently, plating current modulation techniques that can overcome this problem have been developed (25). The
trivalent ion also forms a relatively stable complex aquo ion (Cr(H2O)6)3þ at pH values higher than 4.5, hindering the precipitation,
which also must overcome before a metallic coating can be precipitated. Many organic acids, for example, formic, oxalic, and
malonic acid as well as urea and glycine form thermodynamically more stable complexes than the aquo complex and can be used in
trivalent plating bath chemistry.
A typical chromium plating bath consists of 250 g l1 chromic acid and 2.5 g l1 (1:100) sulfuric acid. Higher chromic acid
concentrations, up to 500 g l1, have been used to give better macrothrowing power and resistance to impurities, but the greater the
concentrations the greater the environmental impact will be in the form of drag-out and mist in exhaust ventilation. If a mixed-
catalyst bath is used, the sulfate concentration is lower. The current density may vary in large area, but for decorative chromium it is
lower, typically from 7.5 to 17.5 Adm2. The bright plate area depends on bath concentration, temperature, and applied current
density. When hard chromium is done, the current density is between 30 and 60 Adm2. Elevated temperature of the bath is used,
around 40 C for decorative and 50 C for hard chromium.
The deposit outlook can vary depending on the bath temperature, current density, and bath chemistry. If high current densities
are used, the deposit will be ‘burnt’; that is, it looks dark gray and rough. A lot of chromium hydroxide and oxide are included in
the coating layer. Dull and gray deposits will be formed in too low temperatures. The coating includes a lot of chromium hydrides
and is relatively soft. No crack pattern exists. When the temperature and current density are in the right envelope, the deposit will
be bright. The grain structure is fine, but some nodules may be formed when layer thickness exceeds 20 mm. Therefore, the layer
thickness in bright chromium should be low. Dim chrome plating will be formed when the current density and temperature are
both at the upper part of the operating envelope. The surface is throughout covered with nodules, and it is usually ground before
being put into operation. The wear resistance is high, and the crack pattern that exists has no adverse effects but helps the lubricant
stay on the surface. If the temperature is too high in relation to the current density, a milky and relatively soft deposit will be
formed.
The crystal structure of chromium in thermal equilibrium is body-centered cubic structure (bcc). In electroplating, there will be
formed besides bcc lattice some amount of unstable hexagonal crystals, which are caused by the formation of chromium hydrides.
The hydrides will form in the presence of a large excess of hydrogen atoms. After plating, the hydrogen is gradually depleted from the
coating, and shrinkage in specific volume will lead to cracking of the coating. If the base metal has a tendency toward hydrogen
www.iran-mavad.com
embrittlement, care must be taken to release the internal stresses before plating and to postplate the embrittlement relief heat
treatment.
The current efficiency of the hexavalent chromium reduction to metallic chromium is low, about 10–35%. About 10% of the
current goes to the formation of trivalent chromium, and the rest, which is the greatest portion, goes to hydrogen evolution. There
is a current threshold under which no chromium will be deposited. It depends on catalyst ions and cathode potential. Once the
plating process has been started, it may continue with smaller potential. This is why higher current is used at the start of the
plating.
4.11.3.2.1.2 Trivalent baths

Due to the toxic and environmentally unfriendly impacts of hexavalent chromium, the trivalent chromium ion began to replace it
a couple of decades ago. The outlook varies from metallic white deposit to deeper-looking pewter or stainless steel, depending on
the manufacturing process. The advantages of trivalent process over hexavalent are improved macrothrowing power, that is, more
uniform thickness and better coverage. Auxiliary anodes or robbers needed for the hexavalent process are not required. Also the
microthrowing power (leveling) is better, and waste treatment costs are much lower.
The make-up salts contain a source of trivalent chromium, conductivity salts, complexing, and wetting agents that help the
chromium to plate and reduce the surface tension (26). The latter also prevents mist formation. The trivalent chromium
solutions are based on sulfate or chloride chemistry. At the anode, hexavalent chromium will be formed unless it is elimi-
nated. The hexavalent chromium in bath hinders trivalent chromium from precipitating. Double-cell arrangements in which
isolated lead anodes are used are now almost completely replaced by single-cell arrangements where insoluble graphite or
metallic catalytic composite anodes are used. The bath maintenance that is done based on pH, specific gravity, and ampere-
hours measurements are easier than maintenance of hexavalent baths. The chromium content in trivalent bath is 4–25 g l1,
and its pH is 2–4 (21). Room temperatures or elevated temperatures up to 50 C can be used. Typical current density varies
between 5 and 16 Adm2.
Trivalent baths are much more sensitive to metallic contamination than hexavalent baths. Metallic impurities make the deposit
dark and affect the micro- and macrothrowing power. Resin treatment can be used to remove the impurities from the bath, and the
bath life has proven to be unlimited.
4.11.3.2.2 Decorative Chromium Plating

Decorative chromium plating has a highly reflective and characteristic bluish haze, which makes it desirable for automotive parts,
bicycles, furniture and appliances, and plumbing fixtures. Decorative chromium plating is in most cases deposited on electroplated
multilayer nickel or copper–nickel coating as a thin layer to prevent coloring or staining of the underneath layer and to give good
wear and scratch resistance to the surface. The base material can be almost any metallic surface suitable for plating, typically steel,
aluminum, copper alloys, and zinc die castings. Also, plastics are plated with the help of catalysts that set the surface active for
electroless nickel plating prior to electrolytic plating.
The chromium layer thickness is in the range of 0.13–1.3 mm. Thicker layers would not be bright enough and could contain visible
cracks. The thin layer replicates the surface appearance, which must be well finished. That is also why nickel coating is used
underneath, since finishing the substrate material to the desired look could be difficult and expensive, although chromium could be
deposited directly on steel. The thin layer is brittle and stressed internally. The crack pattern affects the corrosion resistance of the layer
system. The more dispersed the crack pattern, the more uniform the corrosion and the slower the corrosion penetrates into the
surface. Exposed nickel is corroding sacrificially, and the surface area ratio of nickel and chromium is more equal. Microporous
chromium is also used for this reason. Hexavalent chromium needs a specific treatment to be microporous. It can be done, for
example, by depositing the chromium over a special thin nickel layer that contains inert, nonconducting particles. The special nickel
layer being applied on top of either bright or satin nickel for microporous chromium contains a minimum of 10 000 pores per cm2
(27). In spite of the need for special treatment, the microporous deposit is used more frequently due to specification requirements.
Trivalent chromium deposits are microdiscontinuous as plated. Deposits under 0.6 mm tend to be microporous, and thicker deposits
are microcracked (21). The chromium layer is most resistant to wear when it is bright, whereas matt or burnt surfaces are less tolerant.
In decorative plating, care must be taken to keep the nickel layer active prior to chromium plating. If the nickel layer becomes
passive, for example, due to the drying of the surface, the result may be a whitish chromium deposit or no deposit at all. In this case,
the nickel layer must be activated in hydrochloric acid or some other solution that dissolves oxides. The nickel layer must not have
internal stresses that could lead to flaking of the chromium plating.
The additives of nickel bath, which include, for example, sulfur compounds that are used as brighteners, will make the nickel
layer electrochemically more active. It will deteriorate the corrosion properties of the layer, which is the reason to use multilayer
nickel deposit. The first layers against steel substrate are purer, but they are only semibright. The bright layer is deposited between
semibright and chromium layers. In this system, the corrosion into the intermediate layer will stop, which gives cathodic protection
for the underneath layer.
Hexavalent chromium baths have been used to make the bluish white decorative coatings. The same appearance can be achieved
with trivalent chromium baths, and depending on trivalent bath, also a deeper blue or pewter-looking appearance can be obtained.
Trivalent chromium processes do not give any protection to unplated areas, as hexavalent process may do by forming a conversion
layer. A postchromium inhibiting treatment must be used for gaining equivalent protection. Trivalent chromium processes are
much less sensitive for current interruptions, which enables their use in the barrel plating of small parts.
www.iran-mavad.com
4.11.3.2.3 Hard Chromium Plating

Functional or hard chromium plating is extremely hard and corrosion resistant. The hardness is typically between 750 and 1050 HV.
The coating has low coefficient of friction and excellent wear properties. The layer thicknesses may vary from a few micrometers up
to 500 mm, but thicknesses over 250 mm are not recommended due to the brittleness of the layer and the adverse effect on the fatigue
properties of the base metal. The layers are typically ground after plating to get a smooth surface that will fit into measure tolerances.
Hard chromium is used on rolls and roll journals, molding dies, tools, shafts, cylinders, and the like. It can also be used to salvage
worn parts.
Fixtures and good plating tools are necessary for hard chromium due to the poor macrothrowing power of the plating bath and
requirement for thick deposits. Special computer programs have been developed for fixture and rack design. With the help of good
tooling that uses current shields and thieves, shorter plating and grinding time, better quality and productivity can be achieved. In
many cases, the part must be partially plated, for example, the cylinder part of a sliding member. See also Figure 3. Therefore stop-
offs are required. Several kinds of tapes, lacquers, and waxes are available for this purpose.
Due to the poor current efficiency and hydrogen evolution, heavy misting of the plating bath takes place during plating.
Therefore, proper tank ventilation must be applied, and the mist must be eliminated from the exhaust air by a mist eliminator and
scrubber. Fume suppressants may also be added in the plating solution. Water used for rinsing the plated parts can be partly
returned to the plating tank to replace the losses due evaporation, but the rest must be carefully treated to fulfill regulatory and
environmental requirements.
4.11.3.2.4 Trends in Chromium Plating

The driving force behind changes in chromium plating is a legal necessity to replace hexavalent chromium. Use of hexavalent
chromium will be more strongly regulated because of the occupational and environmental risks. In decorative plating, trivalent
chromium has already replaced hexavalent baths, and the trend will continue. In functional plating, a replacement is still to be
developed. Pulse plating is a promising technique. Before any industrially acceptable replacement can be available, lowering
chromium concentration in hexavalent baths will help to fulfill occupational and environmental requirements. Other methods, that
is, thermal spray, will continue to replace hard chromium plating when applicable.
Bulk plating in barrels or trays with trivalent baths will be more popular in the future, since the quality problem like burning or
gray deposits that follow from poor cathode contact are avoided by using trivalent baths.
4.11.3.3 Nickel Plating

Nickel plating is used for decorative, engineering, and electroforming purposes. The typical base metals for nickel plating are
steel, copper alloys, and zinc alloys. Steel and copper alloy substrates are usually plated directly, but copper underplating is
required for plating on zinc. Stainless steel, aluminum, magnesium, and plastics can also be nickel plated, but they require special
treatment before plating. Nickel coatings for engineering applications are usually smooth and matt, while the decorative coatings
are lustrous, smooth, and bright. Bright nickel coatings are plated with thin chromium coating to prevent nickel tarnishing (see
discussion of decorative chromium plating above). Engineering nickel may also be plated with chromium for better wear
resistance. Corrosion resistance is important in both applications as well as durability or wear resistance. Engineering applica-
tions may be used to enhance erosion resistance, hardness, lubricity, magnetic, reflectivity, emissivity, or absorptivity properties.
Nickel coating is used in electronics as an underlayer for precious metal deposits to prevent interdiffusion of substrates and
coating.
The corrosion-resistance properties of the decorative coating have been developed by using multilayer coatings. The bright nickel
deposit contains the sulfur remains that come from the brighteners in the bath. Sulfur deteriorates the corrosion resistance, and thus
a single layer coating can serve only in mild conditions. To have a more corrosion-resistant coating, a double- or triple-layer coating
should be employed.
Double-layer nickel coatings have a highly leveled semibright layer under fully bright nickel. The semibright layer is typically
two-thirds of the total coating thickness, and it is sulfur free. The impurities in the layer have a huge effect on the electrochemical
potential. The semibright nickel is nobler than the bright nickel. Thus the corrosion will not easily penetrate into the semibright
layer but will propagate horizontally in bright nickel layers. In triple-layer nickel a thin, very reactive high-sulfur layer is deposited
between semibright and bright nickel. Each layer has a different electrochemical potential. When the corrosion reaches the inter-
mediate layer, it will spread laterally under bright nickel. The corrosion of the semibright layer and base metal will be delayed.
Double-layer nickel is easier to apply than triple-layer coatings and is more widely used.
Corrosion properties can also be enhanced by making the chromium layer microporous (see above). If the
corrosion environment is very severe, a nickel top layer can also be done microporous, which further reduces the penetration
rate of corrosion. It can be done over double- or triple-layer nickel; in the triple-layer case, it may be called quad nickel
plating.
4.11.3.3.1 Processes
The nickel salts that are used for plating are nickel sulfate, chloride, sulfamate, fluoroborate, and carbonate. The solutions for
decorative nickel plating are different from the solutions used for engineering or electroforming applications, since they differ in
concentration and contain organic addition agents, brighteners. Brighteners control the nucleation and growth of the deposit. The
www.iran-mavad.com
other groups of constituents are boric acid as a buffering chemical, conducting salts and wetting agents. The use of boric acid is
essential to prevent pH from rising on the cathode, since the deposition of basic nickel salts and hydroxides would lead to brittle
and ‘burnt’ deposit. Wetting agents prevent the hydrogen bubbles from attaching on the cathode surface, which would lead to
a porous coating.
The main bath types are Watts type, high or all chloride, and sulfamate nickel. Other bath types include fluoroborate and black
nickel baths. Nickel strike is often used as a prime layer to obtain sufficient adhesion to the substrate. The most often used strike
baths are Woods type all chloride bath and the sulfamate strike bath.
The Watts-type nickel is a mixture of nickel sulfate and chloride salts. It contains nickel sulfate NiSO4$7H2O about 300 g l1,
nickel chloride NiCl2$6H2O about 60 g l1, and boric acid H3BO3 about 40 g l1. The total nickel content of the bath is about
80 g l1. The pH of the bath is kept between 3 and 4.5. Operating temperature is 40–70 C and applied current density 1–10 Adm2
(28–30). It has been in use for almost a century. Iron is a main impurity in this bath and can lead to a rough deposit. Most of the
bright nickel baths are Watts-type baths.
High or all-chloride baths are modified Watts type. They have better conductivity, macrothrowing power, anode dissolution
ability, and tolerance for impurities, but the solution is more corrosive and gives a more brittle deposit. The typical parameters
include 200–300 g l1 NiCl2$6H2O, boric acid 40 g l1, pH 3–4, and temperature 50–70 C. If sulfate is used, its concentration is
around 100 g l1.
The nickel sulfamate is another very widely spread bath type besides the Watts-type bath. The bath is typically used to grow thick
nickel deposits. A typical bath contains nickel sulfamate Ni(NH2SO3)2 350–450 g l1 and boric acid 30–40 g l1. It can also contain
a small amount of nickel chloride to enhance anode dissolution. However, the chloride content must be carefully controlled, since it
causes internal stresses in the coating. One reason to use nickel sulfamate is its low internal stresses. Another reason is the high
solubility of the nickel salts to sulfamic acid, so high nickel concentration can be made up and thus high current density, up to
60 Adm2 can be used to speed up the layer thickness growth. The current efficiency is above 97%. The pH of the bath is kept
between 3 and 4, and the operating temperature is 30–50 C.
The nickel strike bath is used to provide a good adhesion of the coating on stainless steel and other metals that passivize. The
bath contains a high concentration of hydrochloric acid, which dissolves metal oxides. The Woods-type nickel bath contains about
240 g l1 nickel chloride and about 150 ml l1 hydrochloric acid (30%). A high-chloride sulfamic acid-based bath can also be used.
Bright nickel baths contain primary and secondary brighteners. Primary brighteners provide a bright deposit over a bright surface
and keep the deposit ductile, which is important for applying secondary brighteners. Primary brighteners provide compressive
stresses in the coating. Secondary brighteners enhance leveling and provide mirror brightness. Because they make the deposit brittle
and internally stressed, they must be used with the primary brighteners. Sulfur codeposits from brighteners, which is the main factor
for bright nickel being less corrosion resistant than semibright nickel.
Anodes in nickel plating must be soluble into the bath. Pure nickel may be used with chloride-containing baths, but when
chlorides are not used as, for example, in sulfamate baths a special nickel anode metal must be used, usually anode nickel con-
taining sulfur, to improve the anode solubility. However, the sulfur is as nickel sulfide NiS and does not dissolve into the solution.
Therefore anode bags must be used.
4.11.3.3.2 Properties
The current density does not have a large effect on the current efficiency in nickel baths, so the macrothrowing power is inherently
poorer than in cyanide baths; see Section 4.11.3.1 on zinc plating. However, the good leveling properties of bright and semibright
baths can be achieved by additives.
Internal stress often determines the selection of bath formula and parameters. Usually, tensile stress is harmful for the plated
component. The internal stress depends on several factors: notably, bath type, concentrations, pH, temperature, current density,
and impurities. Generally, the internal stresses from different bath types are from lowest to highest in the following order:
sulfamate < fluoroborate < Watts type < high chloride. The typical value for a sulfamate bath is 40 MPa and for high chloride,
220 MPa (29). However, the stress can vary over a wide range; for example, the Watts-type coating can have values from 125 to
185 MPa. Even compressively stressed nickel can be plated when sulfur-containing additives or brighteners are added and sulfur is
codeposited with nickel.
The hardness of the coating also depends on bath type, applied current density, temperature, and especially chloride
content. The hardness of a Watts-type coating is around 150 HV, while that from a nonchloride sulfamate bath is around
300 HV.
The properties of nickel coatings are specified in standard EN ISO 1456: metallic and other inorganic coatings. Electrodeposited
coatings of nickel, nickel plus chromium, copper plus nickel, and of copper plus nickel plus chromium. Respectively, for specifying
electroplated coatings of nickel for engineering purposes there is the standard EN ISO 4526.
4.11.3.3.3 Trends in Nickel Plating

Again, the driving force for changes is legislation, which is expected to become stricter. Nickel compounds and boric acid will be
under evaluation for their hazardous properties. It may have an effect on the consumption of nickel as replacement coatings are
sought. The boric acid replacements are also being searched. Carboxylic acids are able to complex nickel ions and have been
a promising candidate for a replacement; for example, malic acid has been a good substitute for boric acid in Watts-type baths,
giving a good appearance and comparable hardness for Ni coating (31).
www.iran-mavad.com
4.11.3.4 Copper Plating

Electroplated copper is used both for decorative and engineering purposes. The base metals include iron and steel, zinc and zinc die
castings, aluminum, magnesium, copper and nickel alloys, and even plastics after activation and electroless nickel strike. The
applications include printed circuit manufacturing, electrical connectors, decorative or functional plating in automotive, appliance,
plumbing fixtures, knobs, and various goods. Copper is also used as an undercoat for several other metallic coatings. Copper is used
on steel because it is more easily polished by buffing than steel (32).
There are three commonly used bath types for copper electroplating: acid sulfate, pyrophosphate copper, and cyanide copper.
They are all used in several areas. Acid sulfate solutions are the most common in printed circuit industry, but in other areas its
corrosive nature may be preventive while pyrophosphate is used when good macrothrowing power and less corrosive solution are
required (33). Cyanide copper solutions have been in use for a long time, but it has become less popular following the development
of nickel and other methods to provide the necessary brightness and other properties.
The acid sulfate bath consists of copper sulfate, which is dissolved as CuSO4$5H2O into a sulfuric acid solution. The amount of
copper sulfate is 150–250 g l1, and concentrated sulfuric acid 30–75 g l1 in a conventional solution. Chlorides as NaCl may be
added 30–150 mg l1 to promote anode dissolution and tensides for good wetting. There are a number of patented additives for
improving brightness, hardness, fine grain structure, surface smoothing, and the like (34). Cathodic current density is in the range
1–20 Adm2, but most plating is done with 2–3 Adm2. Operation is at room temperature, but temperatures up to 45 C are also
common. Higher concentrations and elevated temperature enable the use of current densities at the high end of the range. The bath
is easy to maintain, and the current efficiency is close to 100%, which makes it a favorable choice for thick deposits. By using
additives, acid sulfate baths may have good microthrowing power for leveling the rough surfaces. Also, smaller copper content and
more concentrated sulfuric acid enhance throwing power but reduce the current efficiency. Acid sulfate baths may produce a deposit
with poor adhesion on steel, zinc, and aluminum substrates by an exchange reaction unless a copper strike is applied. It may be
done from cyanide bath.
The chloride content must be kept within limits. Too much chloride will lead to matt deposits and very high amounts, to
precipitation of copper chloride on anodes, which will be polarized. Copper sulfate is a very corrosive solution for many metals, and
the anode baskets and hooks should be titanium. The anodes are high-purity copper with 0.02–0.08% phosphorus to ensure
dissolution. Bagged anodes and filtration of the solution are recommended, especially if bright plating is produced. Air agitation is
required with high production rates. Pulsed current techniques have been applied to acid sulfate copper plating to improve the
mechanical and physical properties of the coating, for example, finer grain structure, increased hardness, reduced porosity, and
improved leveling. Pulsed current techniques have also been applied for through-hole plating in printed board manufacturing.
Pyrophosphate copper baths consist of copper(II) pyrophosphate Cu2P2O7$3H2O and potassium (or sodium) pyrophosphate
K4P2O7, ammonia and citrates or oxalates. Copper is as a pyrophosphate complex. The ratio of copper to pyrophosphate content is
critical. The amount of copper metal is 22–38 g l1, pyrophosphate ion (P2O4 1
7 ) is 150–250 g l , and a typical weight ratio of
pyrophosphate to copper is 7–8 (34,35). Excess pyrophosphate is required to keep copper in solution and to enhance conductivity.
Ammonia is used to improve deposit uniformity and brightness. Nitrates may also be added to reduce polarization, and citrates or
oxalates act as a buffer. Orthophosphates will be formed into the solution by hydrolysis of pyrophosphate. It enhances anode
corrosion and acts as a buffer, but the solution must be discarded if the concentration exceeds 100 g l1. pH of the solution is
typically 8.2–8.8. At pH values lower than 7, the pyrophosphate complex will be destroyed, and copper pyrophosphate may be
precipitated. At pH values above 11, copper hydroxide may be precipitated.
The greatest advantage of the pyrophosphate bath is that the solution is almost neutral, so it is suitable for easily corroded
substrates. The cathodic current density is between 0.5 and 8 Adm2, and the current efficiency is almost 100%. Copper strike may
be necessary unless a dilute bath is used. Control of additives is essential for the proper operation of the bath. Many organic and
metallic additives are available to improve the properties of the coating, but they will decompose during operation, which may
have adverse effects, for example, making the deposit brittle. The bath is also more sensitive for organic impurities than acid sulfate
baths.
Cyanide baths contain copper cyanide CuCN as a copper source. The selection of the amount of copper is dependent on the
desired production rate and layer thickness; a typical amount is 75 g l1 CuCN (32,34). There is an excess of cyanide in the form of
potassium or sodium cyanide which forms water-soluble complex ions with copper cyanide. A typical amount is 130 g l1 KCN. The
excess also helps in anode dissolution and makes the coating quality better. As for the alkaline cyanide, potassium or sodium
hydroxide is added to the bath for increasing the conductivity and alkalinity of the solution and for reducing the decomposition of
cyanide (36). A typical amount is 30 g l1 KOH. A small amount, about 15 g l1 of alkali carbonate, is added for buffering purposes.
However, carbonate will form due to the cyanide breakdown when it is oxidized by the effect of oxygen in air. Carbonate will build
up into the solution and must be removed when its amount exceeds about 90 g l1.
The selection of the salt depends on price, desired production rate, and maintenance practice of the bath. Potassium salts are
more conductive, allow higher current densities, and produce more even coating distribution, but they are also more expensive to
buy and to maintain since the excess carbonates forming into the bath cannot be frozen out, but either renewal of the bath or more
difficult chemical precipitation by calcium salts must be performed from time to time.
A strike solution must be used to prevent a loosely attached film formation by exchange reaction. Anodes must be pure
copper, and no phosphorus should be alloyed. Pyrophosphate and ammonia will dissolve anodes. The anode to cathode ratio
should be 2:1.
www.iran-mavad.com
4.11.3.5 Silver Plating

The decorative applications for plating flatware, ornaments, and so on, have long been the largest application since the development
of the silver plating method. There are also many engineering applications in electronics, the semiconductor industry, and machine
parts (i.e., as a bearing surface).
Typically, cheaper metal-like iron or steel, copper, or brass is plated with silver in decorative applications. Also, nonconductors
such as glass or ceramics can be plated after the surface has been treated to be conductive.
Almost all industrial silver plating is done from alkaline cyanide solutions. A large amount of research work has been done to
develop noncyanide replacements. Some systems based on iodide, trimetaphosphate, thiosulfate, and succinimide solutions have
been found to work, but all of them have some drawback, i.e., poor adhesion or coating quality that has prevented them from
becoming a widely accepted replacement (37,38).
The silver bath contains silver cyanide salt, alkali cyanide, alkali carbonate, alkali hydroxide, and, optionally, brighteners.
Sodium salts could be used, but they are less conductive than potassium salts, which would lead to smaller current densities. Some
yellowing of the coating can also occur with the use of sodium salts, so the prevalent chemicals are based on potassium salts. The
silver is usually added as potassium silver cyanide KAg(CN)2, and the typical amount of silver metal in solution is 10–40 g l1
(37,39). Silver metal forms complexes with cyanide. The dissociation constant of the complexes is very low, which will lower the
electrode potential substantially. Only a few free silver ions are in the solution, and the deposition happens directly from cyanide
complexes.
Besides the conductivity enhancement, alkali cyanide is added to the bath to keep the cyanide level high enough for anode
dissolution and to replace losses caused by dissociation reactions with the oxygen and carbon dioxide in the air. Potassium or
sodium cyanide is used, depending on the selection of salts. Alkali carbonates raise the conductivity and enhance the throwing
power of the bath. The amount should be at least 15 g l1, but the maximum amount due the crystallization into the coating is
about 50 g l1 with sodium and about 100 g l1 with potassium carbonate. If too much carbonate is present, the coating will be
rough. The excess of sodium carbonate can be frozen out of the solution, but the higher solubility of potassium carbonate makes
that treatment impossible. Barium cyanide could be used for precipitation, but it is very impractical; therefore potassium baths are
usually replaced when the amount of carbonate is too high. In this way, other impurities will also be gone. Alkali hydroxides are
added to get harder and thicker deposits and to enable high current densities. The pH of silver baths is usually over 11.
Brighteners are used in decorative plating to avoid the need for mechanical polishing. Various organic and inorganic sulfur and
metal compounds have been used for brighteners, but nowadays organic sulfur compounds are the most important ones. Some
sulfur from brighteners will codeposit with silver.
The macrothrowing power of the silver bath is very good and is also leveling with small current densities. However, the coverage
is relatively poor.
The cyanide bath is usually operated at room temperature, but it can be heated up to 35 C to enhance current efficiency. The
current density used is typically 1–5 Adm2. Parts should be moved during the plating operation. Anodes, see Figure 11, should be
high purity (>99.9%) silver anodes and the anode to cathode ratio should be at least 1, but higher ratios are recommended. Anodes
should be removed from the bath in longer nonoperational periods to avoid excess silver buildup to the bath.
The plating process should start with silver strike, where a less-silver-containing bath is used to prevent an exchange reaction of
silver and the base metal. A loosely adherent deposition would be formed if the exchange reaction is allowed to happen. The time
required to silver strike is short, typically under 1 min, and the coating thickness is under 0.25 mm. No rinsing is required before the
actual plating tank.
Figure 11 Jewelry in silver plating tank. Silver anodes are bagged.
www.iran-mavad.com
Post-treatment of silver plating is usually done to prevent tarnishing. Bright lacquer or physical vapor deposition (PVD) coatings
may be used if electrical conductivity is not required. Chromate conversion coatings have also been used, but the need to replace
hexavalent chromium makes that unfavorable and has never been an option for silverware. A thin rhodium electroplate may be
used, but the high price of rhodium is preventing the greater use of the method, and the white outlook of silver suffers from the
treatment.
The electric conductivity of silver metal is the best among all metals, but tarnishing and brighteners will lower it to some 55–60%
of that for pure silver. Silver also has good lubricating properties in low-lubricity fluids and antigalling and antiseizing properties at
elevated temperatures, which has been the reason for applications in jet engines. The reflecting properties of the coating is also good:
It can reflect 90–95% of the visible and UV-range light. The antibacterial properties of silver are well known, and some of the most
recent applications rely on the capability to enhance high-frequency properties of communication devices. The standard EN ISO
4521 gives requirements for coating properties and guidelines for the measurement methods.
The use of electroplated silver in electronics will likely grow after the ban to use lead and due to the rising prices of gold.
4.11.3.6 Gold Plating

Gold is one of the most noble metals and has many different applications in various industries. The two most common applications
are decorative gold or gold alloy finishing, see Figure 12, and industrial electronic coatings. Electronic contacts and semiconductor
devices that utilize the ability of low-temperature eutectic of gold with silicon are the main applications in industrial coatings. The
high price of gold limits its applications for wider use. Layer thickness in decorative applications is typically 0.05–0.1 mm, while in
industrial applications the layer thickness is within the 0.5–10 mm range.
Gold electroplating is a much more diverse subject than silver plating. There are at least four different types of electroplating
baths: acidic cyanide solutions, neutral cyanide, alkaline cyanide, and noncyanide baths (40,41). There are, however, hundreds of
different formulations for these bath types. Cost is the main selection criterion besides the properties of the coating.
Most often, the base metal for gold plating is copper or copper alloy. When gold is plated on copper, a barrier layer must be
plated between copper and gold to prevent copper diffusion in gold and eventual migration on gold surface. The electrical
conductivity would otherwise drop drastically, and the corrosion products of copper would cover the surface. A sulfamate nickel
plating is used most often, but for jewelry it should not be used due to the potential allergy hazard of nickel.
The potassium gold cyanide K(Au(CN)2) is stable until pH 3, which enables the use of acidic cyanide baths. The result is a hard
and bright coating that is used in decorative and industrial applications. A typical bath contains 6–30 g l1 gold as potassium gold
cyanide, one or two buffer salt 50–150 g l1, which are typically potassium salts of citric and/or phosphoric acids. These anions also
help to keep the metals in solution. No free cyanide is in the bath. Cobalt, nickel, or iron salts are used 0.2–3 g l1 to improve
coating hardness and wear resistance by transforming the metal structure from equiaxial to lamellar. One or two complexing agents
are added to make the bath less sensitive for metallic impurities. The pH of the bath is typically between 3.5 and 5. Hard gold
density is around 17.5 g cm3, while pure gold deposit density is 19.3 g cm3 (42).
Bath temperature is between 30 and 60 C. The coating will form brighter in low temperature, which is why it is used for
decorative applications, but when a higher deposition rate is wanted, a higher temperature is also used. The other parameters such as
the amount of buffer and metal salts affect the current efficiency, while low pH will increase the hardness of the coating but lower
the current efficiency.
Gold anodes or platinum-plated titanium anodes can be used, but the platinum anodes are more common. Platinum can
dissolve with high current density, which can be prevented by plating the anode with 10 mm gold plate. It will dissolve slowly, since
the anodic current efficiency is only about 1%.
Figure 12 Gold-plated jewelry. Acidic cyanide plating solution with platinized titanium anodes.
www.iran-mavad.com
The plating process starts with a gold strike treatment to prevent the exchange reaction. The same strike bath can be used for the
first rinse after the gold plating. The current density depends a lot on the processing method; that is, 0.2–0.4 Adm2 is used for barrel
plating and 0.5–1 Adm2 for rack plating.
Neutral cyanide baths are used for decorative applications and industrial high-purity soft gold deposits. The primary use is in the
semiconductor industry to do bonding. The bath type is more sensitive for impurities than acid cyanide, and sometimes extra
complexing agents are added to improve its tolerance. The operating temperature of the bath is quite high, 70–80 C. Somewhat
lower current densities should be used compared to the acidic cyanide bath. Also, pulsed current can be used.
Alkaline cyanide baths for gold and gold alloy plating are the oldest types of gold baths, but their importance in industrial
coating has been diminished. The coating is matt or semibright. Gold is added as potassium gold cyanide, and some free potassium
cyanide and carbonate exists. Phosphates and hydroxides can be added to improve conductance. When gold content is 8 g l1 or
more, the cathodic current efficiency is almost 100%, but often gold content is kept lower due to the drag-out losses. The bath is
quite sensitive for organic impurities, and a frequent active carbon treatment may be necessary. Bath temperatures of 50–70 C are
used. The higher the bath temperature, the better is the coating quality and the plating speed. However, free cyanide breakdown will
also get higher. Insoluble anodes should be used to prevent gold buildup in solution.
Noncyanide baths include sulfite and chloride-based baths. There are other chemistries, but they probably have no practical
significance. Noncyanide baths are used for decorative and heavy industrial coatings. Sulfite baths that are based in sodium gold
sulfite Na3Au(SO3)2 have good micro- and macrothrowing power, and they can be used to make coatings from very pure gold up to
50% gold alloys. Bright plating can be achieved by using metallic brighteners; the coating will be amorphic and ductile but porous.
However, brighteners should not be used when pure gold would be required. To avoid too much porosity, a low current density,
under 0.5 Adm2, should be used. The current efficiency is close to 100%. The bath includes gold typically about 10 g l1 as sulfite
complex; its pH is between 6.5 and 10 and temperature 40–60 C. Chloride-based baths are mainly used as a strike bath for
a stainless steel base when no nickel strike is allowed. Chloride baths contain mainly gold as trivalent chloride AuCl3 and
hydrochloric acid.
The standard EN ISO 27874 specifies and gives test methods for electrodeposited gold and gold-alloy coatings for electrical,
electronic, and engineering purposes.
The trends in industrial applications are to develop the speed and uniformity of gold plating to reduce any extra amount of the
coating. The high price of gold drives industry to use selective plating and to replace gold by silver-filled epoxy bonding in semi-
conductors and palladium or palladium alloys with a top layer of gold in connectors.
4.11.3.7 Tin Plating

Tin electroplating is widely used in several industries. Actually, tin is the only metal for which electroplating is its largest use. The
biggest application is food processing equipment and containers; the second largest use is soldering applications; and the third largest
is use as an alloying component in alloys such as bronze. Electroplated tin is also used in electronics, pump parts, and engine parts. In
electronics it is widely used to protect base metals such as copper or nickel from oxidation, thus preserving their solderability.
In food processing, tin is excellent due to its nontoxicity, corrosion resistance, and ductility properties. When protecting steel, the
coating must be pore-free, since it is a more noble metal than steel. Layer thicknesses up to 30 mm are used in food processing
applications, but in the steel sheet coating layer thickness is only about 1–3 mm. If layer thickness is more than 12 mm, it is
considered practically nonporous. To get thinner nonporous layers, a reflowing treatment is performed by raising the temperature
slightly over the melting point of tin after plating. Reflowing also helps to remove whiskers from a layer for preventing the danger of
shortcuts in closely packed electronic devices. In engineering applications, thicker layers are used, for example, 50–250 mm on pump
parts or piston rings (43).
4.11.3.7.1 Processes
Electrolytic tin can be processed from alkaline, neutral, or acidic electrolytes. Neutral electrolytes are used almost exclusively in
electronics. Alkaline stannate electrolyte represents the bath type from the alkaline group as the use of other alkaline bath types is
now marginal. From the acid group, the acid stannate, fluoroborate, and sulfonate baths are used.
Tin is in tetravalent sodium or potassium stannate form in the alkaline stannate bath. Potassium stannate K2Sn(OH)6 is used
when high-speed plating is wanted. Its solubility is much higher than sodium stannate Na2Sn(OH)6, and the conductivity of
potassium salts is also higher. The tetravalent form is used because the bivalent form is unstable and would decompose to metallic
tin and tetravalent form at elevated temperatures. The coating would be spongy and rough due to the disproportion to metallic tin.
In a typical sodium stannate bath, its concentration is 65–110 g l1, and thus tin content 30–50 g l1 (43,44). The current density in
sodium bath is typically 1–2.5 Adm2 and temperature 70–80 C. Potassium stannate baths can be divided into low-, medium-,
and high-tin baths. In low-tin baths, the potassium stannate concentration is 100–115 g l1, leading to tin content 40–50 g l1.
Potassium hydroxide concentration is 12–15 g l1, current density 3–10 Adm2, and operating temperature 65–90 C. In the
medium-tin bath, the potassium stannate concentration is, respectively, 200–220 g l1, tin content 80–88 g l1, potassium
hydroxide concentration is 15–30 g l1, current density below 16 Adm2 and operating temperature 70–90 C. In high-tin bath the
potassium stannate concentration is 390–450 g l1, tin content 155–180 g l1, potassium hydroxide concentration is 15–30 g l1,
current density below 40 Adm2, and operating temperature 80–90 C. The lower concentration baths would typically be used for
barrel plating. Elevated temperatures are required in all bath types.
www.iran-mavad.com
The throwing power of the alkaline process is superior. Sodium (or potassium) carbonate will be formed into the bath by the
following reaction:
Na2 SnðOHÞ6 þ H2 CO3 /SnðOHÞ4 þ Na2 CO3 þ H2 O
The carbonate will enhance throwing power, but at higher concentrations it will lower cathodic current efficiency, which is
normally 80–90%. The excess sodium carbonate can be frozen out from the solution, but it is better to keep extra amounts of
sodium (potassium) hydroxide in solution since besides enhancing conductivity, it will favor the equilibrium of the following
reaction to the direction of stannate:
Na2 SnðOHÞ6 ¼ 2 NaOH þ SnðOHÞ4
Alkalinity also helps to dissolve anodes, but in great excess it will lower cathodic current efficiency because it will hinder the
stannate complex dissociation at the cathode surface and partly because of the coating dissolution. Therefore the amount of
alkalinity must be carefully kept in limits.
Another critical control issue is anodes. Tin anodes will dissolve according to the following reaction:
Sn þ 6 OH /SnðOHÞ6 2 þ 4 e
However, most of the problems with stannate tin baths are caused by problems with anode control, and the outcome is rough and
porous deposit. In addition to the above reaction, tin may be dissolved in bivalent stannite Sn(OH)2
4 form when the potential is low,
causing the problems discussed above. Stannites can cause problems in quite low concentrations, so if they have been formed into
bath, they must be removed by oxidation with, for example, hydrogen peroxide. A plater can ensure that the potential is high enough
and that anodes will dissolve in tetravalent form when a yellow-green film has been formed on the anode surface. If such a film does
not exist, a temporary high current must be driven through anodes. Care should be taken that the anodes do not passivate too much.
The coating is formed through the following reaction:
K2 SnðOHÞ6 þ 4 e /Sn þ 2 KOH þ 4 OH
Before the precipitation, the stannate complex must be dissociated, which polarizes the cathode quite extensively. Therefore the
coating structure will be fine-grained, and no brightener is required for the alkaline stannate bath. As can be noticed, the precip-
itation reaction releases four hydroxyl groups per each precipitated tin atom. They are consumed in anode dissolution, but if the
anode surface turns black it is too passive; if insoluble anodes are used, the amount of alkalinity will increase, which lowers the
cathodic current efficiency.
The acid processes have several advantages over the alkaline stannate bath. The tin ion is in bivalent form in the electrolyte, so the
same amount of electric current will deposit twice that much coating. Moreover, the cathodic current efficiency is near 100%, which
further increases the speed of operations. Elevated bath temperatures are not required, the anode dissolution is not an issue, and
lower voltages can be used. The drawbacks are lower throwing power and higher demands for pretreatment. The electrolyte is more
corrosive, and organic additives must be used to prevent whisker growth.
Stannous sulfate baths are generally used since they are easy to maintain, and with the control of brighteners deposits from matt
to semibright or bright can be plated. A typical bath contains 7–50 g l1 stannous sulfate SnSO4 and 50–150 g l1 sulfuric acid.
Cathodic current density is 1–10 Adm2 depending on agitation and the plating setup, but in rack plating 4 Adm2 would be
a maximum current. The anode:cathode ratio should be 1:1. A number of proprietary organic additives can be classified into the
following groups: colloids like gelatin, cyclic compounds that contain hydroxyl group like 2-naphthol, nonionic tensides, sulfonic
acids and secondary or tertiary amines. A semibright deposit can be obtained by using gelatin and 2-naphthol. Bright tin coatings
can be done by using aliphatic aldehydes and aromatic aldehydes or tensides as primary brighteners (44,45). The bright coating
properties include good corrosion resistance compared to matt coatings, less porosity, and better solderability.
Tin fluoroborate baths are capable of coating all kinds of coatings. Due to the high solubility and electric conductivity of flu-
oroborate, a high current density, up to 100 Adm2, can be applied. Anode and cathode efficiencies are near 100%. A typical
fluoroborate bath contains 75–115 g l1 tin fluoroborate Sn(BF4)2 and 50–150 g l1 fluoroboric acid. Thus the tin content of the
bath is 30–45 g l1. Organic additives such as gelatin and 2-naphthol must be used as grain refiners. The bath may be operated at
room temperature.
Another acid type of bath is based on sulfonates when the use of boron has been more regulated. The solutions contain no
fluorides and are less corrosive than fluoroborate solutions. High plating speed can be achieved with these solutions, but the make-
up cost is much higher than with sulfate or fluoroborate electrolytes.
The requirements and test methods for tin coatings are specified in standard ISO 2093.
It is estimated that the amount of tin plating will further grow along with the growth of the electronics industry. The less harmful
stannate, acid sulfate, and sulfonate methods will replace the fluoroboric methods and of course the tin–lead and tin–bismuth
methods. Tin–copper and tin–silver alloy plating also has potential properties for electronics.
4.11.3.8 Other Metals

A number of other useful plating methods exist which cannot be treated in the space of this text. There are also coatings that would
have vast potential applications, but either the coating method is not feasible or the price for materials is too high. An example is
www.iran-mavad.com
electroplated aluminum, which unfortunately cannot be plated from aqueous solution. Platinum group metals are an example of
the latter type of materials. Platinum, palladium, iridium, rhodium, and ruthenium belong to the group, and all of these metals can
be electroplated (7). Only palladium is used in larger amounts due to its lower price compared to gold. Platinum price has risen
because of its use in automobile catalyzers. Palladium can replace hard gold and soft gold in some applications. Its materials
properties are comparable, and it can be alloy plated with nickel and silver. In electrical contacts that are exposed to high wear, the
metal is even more durable than hard gold. Palladium can be electroplated from ammonia-based or organic amine-based solutions.
Also, sulfonic acid-based solutions are used in palladium–silver alloy plating. In some applications, palladium is plated with gold
flash coating. Rhodium is a very white color metal that resists wear and tarnishing. Therefore, it has been used in jewelry. Rhodium
has been the most expensive platinum group metal.
There are also metals like lead and cadmium which have been used extensively in the past, but because of their toxic nature and
environmental impact their use is either banned or strongly restricted. Cobalt has also been used a lot, but it is suspected to be
carcinogenic. Therefore, electroplating with these metals or their alloys will not be discussed further in this text.
4.11.4 Alloy Plating

4.11.4.1 Principles
To plate two or more metals simultaneously from a solution containing metal ions, the potentials at which the metals are deposited
must be at the same region. The individual electrode potential and deposition from a single metal solution does not, however,
ensure that the metal is deposited as an alloy. At the cathode it is actually the rule that several reduction reactions occur at the same
time. Usually a part of the applied current goes for hydrogen evolution. Similarly, if the two metal depletion potentials are near, the
metals can be codeposited as an alloy. In that case, the alloy composition is dependent on the kinetics of the metal deposition
reactions and the applied voltage.
In many cases, the depletion potentials are too far away from each other, which leads to complete depletion of the more noble
metal before the less noble will precipitate, or at least the quality of the coating will be low when the more noble metal will be far
beyond the limiting current density and be ‘burnt’. The situation can be changed by using complexing agents.
The electrode potential is dependent on the activity of metal ion. The dependence is given by the Nernst equation
E ¼ E0 RTz1 F 1 ln aMe
where
E is electrode potential and E0 is the standard electrode potential,
R is the general gas constant,
T is the temperature,
z is the charge of the metal ion,
F is the Faraday constant, and
aMe is the activity of the metal ion.
When a complexing agent is introduced into the system, a new equilibrium will be formed, for example, according to the
following simple equation:
2þ
Me2þ þ Xcomp ¼ MeXcomp
where Xcomp is the complexing agent. The reaction has an equilibrium constant Kf, which is dependent on the activities of the ions by
1 1
Kf ¼ MeXcomp Me2þ Xcomp
where activities are denoted as brackets. By solving the metal activity from the latter equation and substituting it to the Nernst
equation, we get
1
E ¼ E0 RTz1 F 1 lnKf þ RTz1 F 1 ln MeXcomp $ Xcomp
Depending mostly on the value of Kf, we can see that the potential of a metal ion can be substantially shifted to the negative
direction and thus a more noble metal can deplete near the potential of the less noble. If, for example the Kf value is 1019, its effect
on the potential in 20 C is 0.56 V when z ¼ 2. See Figure 13 for a schematic representation of alloy plating with complexing
agents. Mea is a less noble metal than metal Meb. Without a complexing agent, only the less noble metal Mea would precipitate.
When a complexing agent is used, the metals are depleted at potential E1 with currents ia and ib, which determines the alloy
composition. Tin, copper, zinc, iron, and nickel alloys are typical in alloy plating.
4.11.4.2 Brass Plating

Brass plating is the most widespread alloy plating method due to its color and decorative outlook. Applications vary from fasteners,
lighters, and fittings to other hardware and even functional applications such as corrosion-resistant coatings in marine environ-
ments. Usually, brass plating is post-treated with lacquer to prevent any tarnishing or discoloration.
www.iran-mavad.com
E
Me b2+ + 2e– = Me b
without complexing
agents
Me b2+ + 2e– = Me b
with complexing
agents
Mea2+ + 2e– = Me a
E1
log i a log i b log i
Figure 13 Effect of complexing agents in alloy plating.
A brass-plating process is done mainly from cyanide solutions. The solution contains a mixture of copper and zinc cyanides,
although zinc can exist in other forms as well. The amount of the bath constituents can vary in great amounts. The copper to zinc
ratio is the most important parameter to control the deposit composition (46), which is typically held at about 4:1. The free cyanide
and pH affect the color of the deposit and coating uniformity. Ammonia can be used to enhance the throwing power and color of
the coating. Organic or metallic brighteners are used to increase brightness and improve grain refinement. Tartrates or citrates are
used to prevent anode polarization.
An example of a solution used to make yellow color deposit is a bath containing about 30 g l1 CuCN, 10 g l1 Zn(CN)2, 50 g l1
NaCN, 8 g l1 Na2CO3, 10 g l1 NaHCO3 and 3–5 ml l1 ammonia (47). pH is maintained around 10, temperature at 25–35 C,
and current density at 0.3–1 Adm2. If higher current density, that is, the production rate is desired, a bath at an elevated
temperature above 60 C with higher metal concentrations may be used. Plating can be done by rack or barrel operation. Typically,
barrel solutions are more dilute than the solutions used in rack operations.
Anode material can be as close to the desired coating composition as possible, for example, 75% Cu and 25% Zn. Also, separate
anodes with separate current control can be used. Anode material should be clean from impurities, and the surface area is such that
the current density will be under 0.5 Adm2.
Noncyanide formulations will need to be developed to be more stable in order to gain further acceptance by operators. A
number of such formulations have been developed, for example, sulfate-based formulas containing glycerol. Another trend is
a substitution of brass plating by electrolytic dyes and resins.
4.11.4.3 Zinc–Nickel Plating

The corrosion properties of electroplated zinc will be enhanced when nickel is codeposited with it. The zinc–nickel layer will last
5–7 times longer than an equivalent layer of pure zinc. It will give cathodic protection for steel, when the nickel content is below
15% (10). Beyond that, the sacrificial properties will be lost since the coating will be passive and more noble than steel. The other
properties that are better than that for pure zinc are better formability and weldability. The coating can be phosphatized before
lacquering.
The best performance is obtained from coatings that contain 10–15% nickel (48). The nickel content depends mainly on bath
type, but temperature, current density, and pH also have an effect. Nickel content, however, does not have a significant impact.
www.iran-mavad.com
Zinc–nickel coating can be deposited from various solutions, that is, sulfate, chloride sulfamate, pyrophosphate, and cyanide
solutions. Today acid and alkaline noncyanide solutions are the most frequently used. Higher nickel content coatings can be
achieved by acid baths, usually 10–15%. Alkaline baths yield 6–9% coating composition. The plating rate is higher from acid baths,
but the thickness distribution is poor. The current efficiency of commercial acid baths is near 100%, which is good for parts at risk
for hydrogen embrittlement.
The typical acidic bath contains zinc chloride ZnCl2 60–120 g l1, nickel chloride NiCl$6H2O 80–160 g l1, ammonium or
potassium chloride around 200 g l1, and primary and secondary brighteners. The pH of the bath is 5.5–6, temperature slightly
elevated and current density 0.5–5 Adm2. Ammonium-free acid electrolytes have also been developed, which have better corrosion
properties (49).
As can be seen from bath composition, the plating mechanism is not based on complex agents, but the two metals are near
enough in electrochemical series and kinetic factors affect the mechanism as well. At first, the steel surface favors nickel deposition,
but after the first atomic layers the situation will be changed and zinc will start to precipitate (50).
Zinc and nickel anodes are used, or if only zinc anodes are used the nickel must be added as nickel chloride. Pulse and pulse-
reverse plating are likely to gain more acceptance since they enhance corrosion-resistance properties by refining the grain structure of
the coating.
4.11.4.4 Other Alloy Coatings

Zinc–cobalt, zinc–iron, and tin–zinc alloys can also be plated. Zinc–cobalt alloys are plated from acid chloride or sulfate baths, and
the amount of cobalt in coating is low, typically 0.3%, and may be up to 1% (7,10). After that, the corrosion-resistance properties
will not be further enhanced. The amount of cobalt metal in baths is about one-tenth of the amount of zinc. Cobalt should be
precipitated first since it is nobler, but the zinc hydroxide layer on the surface inhibits the precipitation, which leads zinc to
precipitate first.
Zinc–iron deposits are mainly done from alkaline solutions. The coating is used for its black conversion coating color and its
resistance to corrosive fuel blends (10). Tin–zinc coatings contain about 80% of tin and the rest of zinc. Its corrosion-resistance
properties are fairly good, although it does not give cathodic protection for steel as it is just slightly less noble than steel. The
resistance against the sulfur dioxide atmosphere is excellent. Its friction properties are also good, electrical resistance is low, and it
does not cause bimetallic corrosion. Therefore, it has been used in fasteners, for example, on steel fasteners that attach aluminum
panels. Coating can also be soldered. However, the coating is very soft and thus easily damaged mechanically. Tin–zinc coatings can
be plated from acid, alkaline, and neutral baths.
Nickel–iron coating contains 10–30% iron. Besides a little savings in material costs, the coating has some other attractive
properties: it has excellent leveling properties, it is ductile, and it can be readily chromium plated. However, the bath needs
good pH control and filtration and is not suitable for substrates other than iron and steel (51). The bath composition is
otherwise similar to the regular Watts-type nickel, but iron is added as iron sulfate FeSO4$7H2O about 20 g l1, which
corresponds to about 4 g l1 iron in solution. Moreover, a stabilizator to keep iron in solution and prevent it from oxidizing to
trivalent is required.
4.11.5 Aluminum Finishing
Aluminum finishing is usually performed by anodizing or applying a conversion coating. Aluminum can also be electroplated with
other metallic coatings by using intermediate zincate or stannate layers to help keep the surface active for subsequent coatings, but
the process is not easy to control and the applications are quite limited. However, the applications of anodized aluminum are very
wide, covering building parts and facades, aircraft, and other equipment parts, decorative items, and the like. The wide application
range is based on the hard, abrasion, and corrosion-resistant properties of the coating. Conversion coating is used as an undercoat
for paints to get a good base for painting and better corrosion-resistance properties.
4.11.5.1 Anodizing
In anodizing, the oxide layer that will form from metallic aluminum is made much thicker than the layer that is formed
spontaneously by the oxygen in air. A typical layer thickness for outdoors is about 20–25 mm. Thinner layers are adequate for
indoor use. Hard anodized layers for tooling and the like may be around 50 mm thick. The oxide layer that forms in an electrolyte
is an insulator, but it will also dissolve in the electrolyte, so the layer will be microporous. That is perquisite for thicker layers.
Eventually the resistance of the layer will be high enough to prevent any further thickness growth. If the oxide does not dissolve
into the electrolyte, the layer remains very thin, but also if the dissolution rate is too high no layer will be formed. Below the
porous layer there is a 10–30 nm thick nonporous layer (52). The porosity makes it possible to absorb dyes or metal salts into the
layer, which is important for decorative applications. The layer is usually growing somewhat outward as well, since the specific
volume of aluminum oxide is higher than metallic aluminum. As a rule of thumb, one-third of the layer thickness will grow
outward unless the anodizing time is prolonged or the temperature is too high to dissolve the layer more than the specific volume
difference.
www.iran-mavad.com
Anodizing happens according to the following reaction at the anode

2Al þ 3H2 O/Al2 O3 þ 6Hþ þ 6e
where the aluminum anodizing is occurring through the breakdown of water molecules enforced by the external current. At the
cathode hydrogen ions are reduced, leading to hydrogen evolution.
Sulfuric, chromic, phosphoric, or oxalic acid electrolytes may be used in anodizing. Sulfuric acid is the most often used process
for anodizing (53). Typically, the electrolyte consists of 15 w-% sulfuric acid. The process can be driven by voltage control, which is
about 15–20 V, resulting in a current density of about 1–1.5 Adm2. The higher the voltage, the less porous the oxide layer is. The
solution temperature control is also essential, since the temperature should remain under 22 C. Unless cooled, the temperature will
get higher due to the ohmic resistance of the solution, and above 22 C the dissolution rate of the formed aluminum oxide layer will
be higher than the rate of formation and finally no anodized layer will be formed. Dissolved aluminum in the bath makes the
conductivity worse, and the bath must be replaced if the aluminum concentration gets higher than 15 g l1. Acid concentration is
not critical, but higher concentrations will dissolve the layer faster. Chlorides will cause pitting in the oxide layer, so the amount of
chlorides must be controlled. Generally, the lower the acid concentration and the temperature, the more dense and thick layers can
be done. Soft and powdery layers indicate uncontrolled parameters.
The aluminum alloy has an effect on its anodizing. Pure aluminum is easy to anodize, and it gives a better layer than alloys.
Iron in aluminum alloys is an impurity that deteriorates the brightness. As low as 0.3% iron content in aluminum has a negative
effect. Silicon will make the layer gray from 0.8% up to 5%; the more silicon the alloy contains, the grayer the layer color will be.
Many casting alloys are affected. Titanium will also reduce the brightness, but it also gives better grain structure for the coating.
Magnesium has no negative effects under 5% concentrations, and a slight loss of brightness can be detected thereafter. Copper
has adverse effects from 2% concentrations up. It will color the layer from gray to brown gray. Also, hardness and corrosion
resistance will suffer. Zinc has a negligible effect. The heat treatment and grain structure of an alloy may also affect the brightness
and color.
The fixtures in anodizing must be done from aluminum or titanium. Care must be taken to prevent the parts to move during
operation since it will halt the process as the electrical contact will be lost.
After anodizing, the coating may be colored by dipping the parts into a dye solution. A number of organic and inorganic dyes are
available. Organic dyes usually are less resistant to UV light than inorganic dyes. The dye is absorbed in the porous layer, which is
thereafter sealed. Another way to dye the layer is to use metal salts in the bath, and by applying an alternating current, the layer will
be dyed. The selection of the color is far more limited; usually dark yellow, brown or black colors are available. The color will remain
intact in UV light, which makes it desirable for architectural applications. A third way to dye the layer is to use the alloying elements
of the aluminum. The anodizing is done in an organic acid, that is, maleic acid, using higher current density than in ordinary sulfuric
acid anodizing. The resulting dyeing differs from the absorption dyeing since the layer is through colored instead of color in pores.
After anodizing, the layer is usually sealed to provide corrosion and dirt resistance. It is accomplished by treating the parts in
deionized water over 95 C temperature. The aluminum oxide will react to aluminum oxihydrate, which will seal the pores. Dyed
surfaces are generally sealed in a nickel salt solution, which prevents the bleeding of the dye. Hard anodized layers are usually not
sealed because sealing may deteriorate the hardness and abrasion resistance of the layer.
When a thicker, harder, and more corrosion-resistance anodizing layer is desired, hard anodizing may be applied. It is done by
lowering the temperature of the bath and using higher current density and voltage. Bath temperature may be around 5 C and the
voltage 20–70 V. The electrolyte can be the same as in sulfuric acid anodizing, or an oxalic acid addition can be done. Also bare
oxalic acid baths can be used, which have the benefit that the temperature does not have to be lowered as much as with sulfuric acid.
Alloying elements may also make the process more difficult. The modifications to the process make it much more expensive, and its
use is limited to the special needs of aerospace and other demanding applications.
Chromic acid electrolyte has been used mostly for aerospace applications. It is done in 30–50 g l1 chromic acid solution at an
elevated temperature of 35–40 C. It produces thinner film, about 3–5 mm, which has excellent corrosion-resistance properties and
is less brittle than sulfuric acid anodized coating. The layer is a good base for painting or adhesive bonding. The properties are
beneficial for structures that are exposed to fatigue. The environmental and occupational hazards have reduced the use and
importance of chromic acid anodize.
There are also many variations of these electrolytes and parameters. Phosphoric acid electrolyte has been used in aerospace
because of the excellent bonding properties of the coating. Its use has been increasing as a bonding method in aerospace
applications.
The trends in anodizing aluminum may be expected to be quite stable. The environmental effect of hexavalent chromium may
lead to replacement of chromic acid anodizing with other methods.
4.11.5.2 Conversion Coatings of Aluminum

Traditionally, the conversion coating of aluminum has been made from hexavalent chromium solutions. Besides hexavalent
chromium compounds, the solutions contain agents that activate the surface and accelerate the conversion reactions. The main
purpose of the coating is to enhance-corrosion resistance properties. In case the surface is painted, the conversion coating is a good
base for paints. The application can be done by dipping, spraying, or brushing, so the method is good for protecting damaged or
scratched anodized coatings. The layer does not have good wear resistance. It doesn’t insulate the surface electrically, but it does not
www.iran-mavad.com
work as a good current conductor either. However, it can be used when electrical contact is needed between different parts of
a structure, for example, in grounding the structure (54,55).
The color of a chromate conversion layer indicates its thickness. The most common is a yellow layer, whose mass per unit area is
0.2–3 g m2 that corresponds to a coating thickness of 0.5–2 mm (52). Bright layers are thinner and green layers are thicker. The layer
structure is amorphous, and it consists of mixtures of chromic oxides and chromates. Elevated temperatures up from 70 C will spoil
the properties of the layer.
Since 2006, hexavalent chromium has been forced to be replaced in many industry areas, due to the RoHS and ELV directive. An
aluminum conversion coating that has been done with hexavalent chromium was at first replaced with a combination of organic
polymers and fluoride compounds. They have excellent paint-adhesion characteristics but practically no corrosion-resistance
properties. Therefore, other products have been under development. Research has been ongoing, and one promising group that has
been found is cerium-based systems (56). Other systems like lithium-based or silicon-based sol–gel systems are also under
development. One of the challenges at the moment is to have the rate of the conversion reaction high enough. Trivalent chromium
compounds have also been found to be very good as a drop in replacement (57). However, the use of trivalent systems require more
careful preparation, and additives are required to meet strict corrosion-resistance requirements and to get desired colors.
4.11.6 Electroless Plating Solutions
The principle of electroless plating is discussed above. The most important property of the electroless plating is the uniformity of
layer thickness in corners, cavities, and other areas that are difficult for electroplating. The solutions used for electroless plating
contain a metal salt to be plated, reducing agents, metal complex builders, buffering agents, and stabilizers. Brighteners may also be
used if a bright coating is required. The deposition rate depends on metal concentration, reducing agent concentration, pH, and
temperature. Sometimes the amount of free complexing agent in solution has an effect, but the type of ligand certainly affects the
plating speed (58). Typical plating speeds are 3–5 mm h1, but with electroless nickel even 20 mm h1 can be obtained.
Electroless plating solutions have limited life since the reduction reaction products accumulate in the solution, and finally the
salt content will be too high for proper plating. Sometimes accumulated impurities also shorten the solution life. Usually, the life is
defined in terms of metal turnover (MTO), which means the number of initial metal content that can be precipitated from the
solution after replenishments before the plating rate will be too low or the quality of the coating will be deteriorated. A typical MTO
number is at least 5–7, but it can be as high as 10–20.
The stability of the solution is essential. The start of the reduction reaction is activation limited; hence higher metal concentration
and temperature make the solution more unstable. Also, particle impurities will easily offer a starting place for reactions. Therefore,
the concentration and temperature are kept as low as possible to obtain a reasonable growth rate. The solution must be filtered.
There are two types of stabilizers: catalytic inhibitors and oxidizers. The inhibitors are, for example, sulfur compounds, heterocyclic
compounds with nitrogen or sulfur atom in it, or some metals. Oxidizers are believed to passivate the surface of metal particles, thus
hindering the growth. Lead has been used as an efficient stabilizer. Previously, the lead content of the bath was typically
0.035–0.085 w-% depending on bath type, but after the RoHS directive took effect, the lead content was reduced below 0.0003 w-%.
In the following section, two of the most common electroless plating methods, nickel and copper, will be discussed further.
However, other metals such as cobalt, palladium, gold, silver, and even nickel- and cobalt-based alloys can be plated with the
electroless method.
4.11.6.1 Electroless Nickel

4.11.6.1.1 Bath Types
Nickel metal works as a catalyst for electroless nickel plating. If noncatalytic surfaces are to be plated, they must first be treated with
a catalyst. Polymers usually also require etching in an acidic solution before catalyzing for providing anchor spots for catalyst metal
and wetting ability for aqueous catalyst solutions, see Figures 14 and 15.
After etching, the surface is sensitized, which means absorbing ions that can be oxidized by the actual catalyst ions. For example,
acidic Sn(II) or Sn(IV) solutions or more seldom silver nitrate or gold(III) chloride is used to sensitize the surface that is absorbing the
metal ions. Thereafter a catalyst treatment with, for example, palladium chloride solution is required. Palladium would settle as
a metallic form with sensitized surface by an exchange reaction with the ion used in sensitizing. If the metal that is to be plated also
reacts by an exchange reaction (i.e., copper), no separate catalyst treatment is required. Another popular way is to use the colloidal
technique to catalyze and activate surfaces (Figure 16). Typical polymers that are metal plated with an electroless method are ABS,
ABS/PC, Nylon, and POM (59–61).
There are a number of formulations for the electroless nickel bath. Nickel metal content is achieved by adding nickel sulfate
NiSO4$6H2O or chloride NiCl2$6H2O into the bath, which is similar for almost all formulations. The amount of the salts typically
varies between 20 and 45 g l1, and a typical nickel metal content in a bath is 6 g l1. Nickel metal is complexed with a combination
of lactate, citrate, glycolic, or malic acids. The amount of complexing agents may be 30–50 g l1 (62).
There are several possible reducing agents. The most common one is sodium hypophosphite NaH2PO3$H2O, but also sodium
borohydride NaBH4 or dimethylamine borane (CH3)2NHBH3 could be used. The reducing agent affects the properties of the
coating, since part of it will deposit into the coating.
www.iran-mavad.com
Figure 14 A microscopic picture of an etched ABS plastic surface. Reprinted from McCaskie, J. E. Plating on Plastics. Met. Finish. May 2006, 31–39, with
permission from Elsevier.
Figure 15 A schematic figure of an etched ABS plastic surface. Reprinted from McCaskie, J. E. Plating on Plastics. Met. Finish. May 2006, 31–39, with
permission from Elsevier.
The summary reaction from hypophosphite bath can be written as

2 H2 PO2 þ 2 H2 O þ Ni2þ /Ni þ H2 þ 4 Hþ þ 2 HPO3 2
Hydrogen evolution is occurring as a side reaction. From hypophosphite baths, phosphorus is codepositing along with nickel. In
heat treatment it will form the NiP compound. The amount of phosphorus depends on the bath formulation and pH. Usually, the
bath types are classified as high-phosphorus (10–14% P), middle-phosphorus (4–9% P), and low-phosphorus (below 3% P). The
amount of hypophosphite may vary widely, 10–50 g l1, depending on formulation (63). The pH has an effect on phosphorous
content, acidic, that is, low-pH, baths yield high-phosphorous coatings, and at the alkaline side the phosphorous content is low,
even below 1%. The acidic baths are more common, and they are typically operated at a pH range of 4–6, while alkaline baths are
operated at 9–11. The temperature is at the range of 60–90 C. Acidic baths are usually operated at slightly higher temperatures of
85–95 C than alkaline baths.
www.iran-mavad.com
Figure 16 Plating of a plumbing fixture part; treatments cumulating from left to right are nontreated plastic, etched and activated, electroless
nickel þ electrolytic copper, bright nickel, and bright chromium electroplating.
Sodium borohydride and dimethylamine borane baths yield boron into the coating. Usually, there are two types of baths, which
produce 0.2–2 w-% or 3–8 w-% boron into the coating. For borohydride baths, the required pH is 12–14 to avoid nickel boride
precipitation (63). It limits the use of the bath to the materials withstanding high alkalinity. Thallium is used as a stabilizer in some
formulations, but it may also codeposit in substantial amounts, which is a drawback from an environmental standpoint. Bath
temperature is 90–95 C. The coating contains about 3–8% boron. The coating is harder than nickel–phosphorous coatings, but
corrosion resistance is not as good. It does not, however, deteriorate after heat treatment. The dimethylamine or diethylamine
borane reducing agents can be done in both acid and alkaline solutions with a varying pH in the range 5–11. Boron content varies
between 0.1 and 5 w-%, and the electric conductivity of the coating is much better than that of nickel–phosphorous coatings. When
hydrazine is used, the coating contains Ni–N deposits. The amount of nitrogen, however, is some tenth of a percent. The
commercial use of hydrazine baths is negligible, and there are potential work safety issues with hydrazine since it is carcinogenic.
Metal complex builders or chelating agents are used to keep the nickel metal in solution and to help maintain the stable
operation conditions. They control the precipitation speed and help to keep pH in the specified range. The common complexing
agents used in electroless nickel baths are citrates, acetates, ammonia, and pyrophosphate.
Stabilizers are used to prevent the bath from ‘growing wild’ that is, from experiencing a sudden depletion of the bath when the
metal grows on fine particles within the solution. The stabilizers may be group IV compounds like sulfur and selenium or
unsaturated organic acids like maleic acid. Stabilizers may also be heavy metal cations like Sn2þ. Previously, lead and cadmium were
also used, but their amount has been lowered from some hundreds of ppm to below 1–3 ppm to make the baths compatible with
the RoHS directive. Cadmium has also been used as a brightener. The fourth group of stabilizers consists of oxygen-containing
compounds like MoO2 4 which passivate the surface of metal particles in baths.
The pH has a great effect on the deposition rate and the phosphorous content of the coating; therefore it is usually kept as stable
as possible. The reaction scheme of the deposition is complex as discussed above, but there will form 4 mol of hydrogen ions for
1 mol of deposited nickel, so buffering agents are essential to keep the deposition rate and coating composition within specification.
The most common compounds used are acetic, propionic, and succinic acids. Boric acid and amines are formed in reactions of
borane baths, and they work as buffering salts as well, which leads to a long operating life of that bath type.
The deposition rate depends on several factors: bath temperature, complexing agents and stabilizers, pH, the concentration of
the reducing agent, and nickel concentration. A usual production rate is 15–20 mm h1. The more the bath is used, the more reaction
products will accumulate into the solution, which will slow down the growth rate. The amount of reaction products is about 5 g
sodium orthophosphate per 1 g of coating. Bath life is usually 6 to 8 metal turnover (MTO). After that, besides the slower deposition
rate, tensile stresses will form into the coating, which will cause deterioration of the corrosion-resistant properties.
4.11.6.1.2 Properties
The most important property of the electroless nickel is its extremely good macrothrowing power and coverage. When the solution
is well agitated, all surfaces that are in contact with a solution will have a uniform deposit thickness. Depletion of the solution in
holes or other poorly agitated areas, gas pockets, hydrogen bubbles, or temperature differences may worsen throwing power or
coverage. An electroless nickel bath usually has poor leveling properties; therefore all scratches and dents will be seen on the surface
after plating.
The adhesion of an electroless coating is at the same level as electrodeposited coatings. On iron, steel, and copper alloys, the
proper adhesion is about 350–450 MPa (62). The structure of the coating is as deposited in a metastable amorphous state (58).
Phosphorus or boron is in an oversaturated state in nickel matrix. When heat treated to 300 C or more, the metal structure begins to
crystallize, and Ni3P or Ni3B compounds will precipitate in grain boundaries. The low-phosphorous coatings (2–3% P) are the
www.iran-mavad.com
hardest in the as-plated state, and their hardness is about 700–750 HV, while mid-phosphorous coating (7–9% P) hardness is in the
range of 450–550 HV and the high-phosphorous coatings (11–13% P) have a hardness 450–500 HV. The nickel–boron coatings are
slightly harder as plated, about 700 MPa, and the hardness does not depend on boron content. With heat treatment at 400 C, the
maximum hardness of about 1000 MPa can be reached with high-phosphorous coatings. The hardness depends on phosphorous
content, temperature, and heat treatment time. With higher temperatures or longer treatment times, the hardness will drop again.
Along with the hardness comes the resistance to abrasive wear. The wear depends on the lubrication of the surfaces, and
electroless nickel has a low friction coefficient. High-phosphorous coatings are usually recommended. Making thicker coatings will
lower the wear rate. Electroless nickel is considered to be more wear resistant than electroplated nickel, and it also lowers the
amount of fretting corrosion.
The corrosion-resistance properties of the electroless nickel in neutral or alkaline environments are excellent. Below pH 4, the
corrosion rate of the coating will increase rapidly. Heat treatment does have an adverse effect on the corrosion properties; this is due
to the difference in volumes of the nickel–phosphorus or nickel–boron compounds that lead to internal stresses and microcracking.
However, when heat treated over 500 C, the corrosion-resistance properties are improved, probably due to the improved bonding.
The coating is more noble than steel, and therefore it has to be virtually without pores or cracks. To have good corrosion resistance,
the coating thickness should be about 30 mm and should have a high phosphorous content. Electroless nickel is widely used for
corrosion protection of aluminum and steel.
The electric and thermal conductivity of the coating in an as-plated condition is higher than after heat treatment. This is due to
the more crystalline structure that develops after heat treatment. Mid- and high-phosphorous coatings are nonmagnetic as plated,
and the coating thickness can be measured with a common magnetic method. After heat treatment, the coating will turn slightly
ferromagnetic.
The requirements for electroless nickel are determined in standard EN ISO 4527: Metallic coatings – Autocatalytic (electroless)
nickel–phosphorus alloy coatings – Specification and test methods.
A common problem in operations of electroless nickel is a coating growth on plating equipment. The surface of plastic tanks,
filter pumps, and the like will be rougher, and there are places for deposit growth. This problem can be avoided with stainless steel
tanks and equipment that use anodic protection. If the plated components have some areas that are specified to be left bare, heat-
resistant lacquer or tape may be used.
4.11.6.2 Electroless Copper

Electroless copper is used mainly in the printed circuit board industry to metallize the holes in two or multilayer boards and
metallize other plastics for decorative and electromagnetic shielding applications.
Sodium hypophosphite or dimethylamine borane can be used as a reduction agent for the electroless copper solutions, but the
most important one is formaldehyde, which is used for making conductive through-hole surfaces for plastics in preparation for
electroplating. The overall reaction for electroless copper deposition from that solution is
Cu2þ þ 2 HCHO þ 4OH /Cu þ 2 HCOO þ 2 H2 O þ H2
The detailed reaction scheme and its kinetics can be found from reference (64). The pH in the process is 11–14, and in alkaline
solution copper must be kept in the solution with complexing agents, such as tartrates or EDTA (57,62).
Electroless copper is composed of copper sulfate CuSO4$5H2O 5–15 g l1, sodium potassium tartrate (or EDTA) 25–50 g l1,
sodium hydroxide 5–20 g l1, formaldehyde (40%) 10–40 ml l1, and additives up to 2 g l1. It can be operated in room
temperature, slightly elevated or at 70 C, depending on the composition. The deposition rate is 0.5–5 mm h1. Baths that are
depositing copper with a lower rate are used for undercoating applications, while faster depositing baths may produce with the help
of more ductile additives and a strong coating for engineering applications.
Use of electroless nickel and copper will grow when new applications are invented. They will also replace some methods that for
occupational or environmental reasons have to be substituted.
See also: Protective Transition Metal Nitride Coatings; Large-Scale Industrial Coating Applications and Systems
References
1. Schlesinger, M.; Paunovic, M. Modern Electroplating, 5th ed.; John Wiley & Sons Inc.: New Jersey, 2010.
2. Anonymous. Kemiallinen ja sähko¨kemiallinen pintakäsittely, osa 2; Suomen Galvanotekninen yhdistys ry: Helsinki, 1999.
3. Valero, G., Ed. Metal Finishing 73rd Guidebook and Directory Issue; Elsevier: New York, 2005.
4. Browning, M. E., Ed. ASM Handbook. ASM International: USA, 2007; pp 165–329.
5. Durney, L. J., Ed. Electroplating Engineering Handbook, 4th ed.; Van Nostrand Reinhold: New York, 1984.
6. Oliveira, E. M.; Finazzi, G. A.; Carlos, I. A. Influence of Glycerol, Mannitol and Sorbitol on Electrodeposition of Nickel from a Watts Bath and on the Nickel Film Morphology. Surf.
Coat. Technol. 2006, 200, 5978–5985.
7. Rao, C. R. K.; Trivedi, D. C. Chemical and Electrochemical Depositions of Platinum Group Metals and Their Applications. Coord. Chem. Rev. 2005, 249, 613–631.
8. Standard EN ISO 2081:2008. Metallic and Other Inorganic Coatings. Electroplated Coatings of Zinc with Supplementary Treatments on Iron or Steel.
9. Tencer, M. Electrical Conductivity of Chromate Conversion Coating on Electrodeposited Zinc. Appl. Surf. Sci. 2006, 252, 8229–8234.
www.iran-mavad.com
10. Winand, R. Electrodeposition of Zinc and Zinc Alloys. In Modern Electroplating; Schlesinger, M., Paunovic, M., Eds., 5th ed.; John Wiley & Sons Inc.: New Jersey, 2010;
pp 285–307.
11. Schneider, S. Zinc Plating. Plat. Surf. Finish. August 2003, 46–47.
12. Anonymous. Sinkki, Kemiallinen ja sähkökemiallinen pintakäsittely, osa 2; Suomen Galvanotekninen yhdistys ry: Helsinki, 1999; pp 240–283.
13. Restriction of the Use of Certain Hazardous Substances in Electrical and Electronic Equipment Directive (RoHS) 2002/95/EC.
14. Waste Electrical and Electronic Equipment Directive (WEEE) 2002/96/EC.
15. End of Life Vehicles Directive (ELV) 2000/53/EC.
16. Bibber, J. An Overview of Nonhexavalent Chromium Conversion Coatings – Part II: Zinc. Met. Finish. February 2002, 100 (2), 98–102.
17. Tomachuk, C. R.; Elsner, C. I.; Di Sarli, A. R.; Ferraz, O. B. Morphology and Corrosion Resistance of Cr(III)-Based Conversion Treatments for Electrogalvanized Steel. J.
Coat. Technol. Res. 2010, 7 (4), 493–502.
18. Zaki, N. Trivalent Conversion Coating for Zinc and Zinc Alloys. In Metal Finishing 73rd Guidebook and Directory Issue; Valero, G., Ed.; Elsevier: New York, 2005;
pp 380–389.
19. Lindsay, J. H. Decorative & Hard Chromium Plating. Plat. Surf. Finish. August 2003, 22–24.
20. Newby, K. R. Functional Chromium Plating. Met. Finish. 2007, 105 (10), 182–191.
21. Snyder, D. L. Decorative Chromium Plating. Met. Finish. 2007, 105 (10), 173–181.
22. Anonymous. Kromi, Kemiallinen ja sähkökemiallinen pintakäsittely, osa 2; Suomen Galvanotekninen yhdistys ry: Helsinki, 1999; pp 55–91.
23. Mandich, N. V.; Snyder, D. L. Electrodeposition of Chromium. In Modern Electroplating; Schlesinger, M., Paunovic, M., Eds., 5th ed.; John Wiley & Sons Inc.: New Jersey,
2010; pp 205–248
24. Giovanardi, R.; Orlando, G. Chromium Electrodeposition from Cr(III) Aqueous Solutions. Surf. Coat. Technol. 2011, 205, 3497–3955.
25. Hall, T. D.; Taylor, E. J.; Inman, M. Functional Trivalent Chromium Electroplating. Plat. Surf. Finish. 2010, 11, 42–48.
26. Schario, M. Decorative Trivalent Chromium Plating. Met. Finish. 2008, 106 (6), 66–68.
27. Gardner, A. Decorative Trivalent Chromium Plating. Met. Finish. 2006, 104 (11), 41–45.
28. Di Bari, G. A. Nickel Plating. Plat. Surf. Finish. August 2003, 31–34.
29. Anonymous. Nikkeli, Kemiallinen ja sähkökemiallinen pintakäsittely, osa 2; Suomen Galvanotekninen yhdistys ry: Helsinki, 1999; pp 166–191.
30. Di Bari, G. A. Electrodeposition of Nickel. In Modern Electroplating; Schlesinger, M., Paunovic, M., Eds., 5th ed.; John Wiley & Sons Inc.: New Jersey, 2010; pp 79–114.
31. Saito, K.; Kishimoto, K.; Nobira, Y.; Kobyakawa, K.; Sato, Y. Nickel Electroplating Bath Using Malic Acid as a Substitute Agent for Boric Acid. Met. Finish. 2007, 105 (12),
34–60.
32. Anonymous. Kupari, Kemiallinen ja sähkökemiallinen pintakäsittely, osa 2; Suomen Galvanotekninen yhdistys ry: Helsinki, 1999; pp 121–137.
33. Hamilton, A. C. Acid Sulfate & Pyrophosphate Copper Plating. Plat. Surf. Finish. August 2003, 12–13.
34. Dini, J. W.; Snyder; Donald, L. Electrodeposition of Copper. In Modern Electroplating; Schlesinger, M., Paunovic, M., Eds., 5th ed.; John Wiley & Sons Inc.: New Jersey,
2010; pp 33–78.
35. Barauskas, R. Copper Plating. In Metal Finishing 73rd Guidebook and Directory Issue; Valero, G., Ed.; Elsevier: New York, 2005; pp 175–187.
36. Horner, J. Cyanide Copper Plating. Plat. Surf. Finish. August 2003, 19–21.
37. Schlesinger, M. Electroless and Electrodeposition of Silver. In Modern Electroplating; Schlesinger, M., Paunovic, M., Eds., 5th ed.; John Wiley & Sons Inc.: New Jersey,
2010; pp 131–138.
38. Anonymous. Hopea, Kemiallinen ja sähkökemiallinen pintakäsittely, osa 2; Suomen Galvanotekninen yhdistys ry: Helsinki, 1999; pp 26–41.
39. Blair, A. Silver Plating. Plat. Surf. Finish. August 2003, 42–47.
40. Anonymous. Kulta, Kemiallinen ja sähkökemiallinen pintakäsittely, osa 2; Suomen Galvanotekninen yhdistys ry: Helsinki, 1999; pp 94–116.
41. Kohl, P. A. Electrodeposition of Gold. In Modern Electroplating; Schlesinger, M., Paunovic, M., Eds., 5th ed.; John Wiley & Sons Inc.: New Jersey, 2010; pp 131–138.
42. Blair, A. Gold Plating. Plat. Surf. Finish. August 2003, 29–30.
43. Carano, M. Tin Plating. Plat. Surf. Finish. August 2003, 43–45.
44. Anonymous. Tina, Kemiallinen ja sähkökemiallinen pintakäsittely, osa 2; Suomen Galvanotekninen yhdistys ry: Helsinki, 1999; pp 291–307.
45. Hirsch, S. Tin, Lead and Tin-Lead Plating. In Metal Finishing 73rd Guidebook and Directory Issue; Valero, G., Ed.; Elsevier: New York, 2005; pp 240–254.
46. Kowalski, A. J. Brass Plating. Plat. Surf. Finish. August 2003, 15–16.
47. Anonymous. Tina, Kemiallinen ja sähkökemiallinen pintakäsittely, osa 2; Suomen Galvanotekninen yhdistys ry: Helsinki, 1999; pp 138–150.
48. Anonymous. Sinkki-nikkeli- ja sinkkikobolttiseokset, Kemiallinen ja sähkökemiallinen pintakäsittely, osa 2; Suomen Galvanotekninen yhdistys ry: Helsinki, 1999.
pp 284–290.
49. Sonntag, B.; Dingwerth, B. Investigation of High-Performance Acid Zinc-Nickel Electrolyte. Met. Finish. 2011, 109 (3), 24–28.
50. Changa, L. M.; Chena, D.; Liub, J. H.; Zhangb, R. J. Effects of Different Plating Modes on Microstructure and Corrosion Resistance of Zn–Ni Alloy Coatings. J. Alloys
Compd. 2009, 479, 489–493.
51. Anonymous. Nikkeli-rautapinnoitteet, Kemiallinen ja sähkökemiallinen pintakäsittely, osa 2; Suomen Galvanotekninen yhdistys ry: Helsinki, 1999; pp 211–216.
52. Moritz, G. Alumiinin pintakäsittely, Kemiallinen ja sähkökemiallinen pintakäsittely, osa 1; Suomen Galvanotekninen yhdistys ry: Helsinki, 1996; pp 157–185.
53. Grubbs, C. A. Anodizing of Aluminum. In Metal Finishing 73rd Guidebook and Directory Issue; Valero, G., Ed.; Elsevier: New York, 2005; pp 350–366.
54. Eppensteiner, F. W. Chromate Conversion Coatings. In Metal Finishing 73rd Guidebook and Directory Issue; Valero, G., Ed.; Elsevier: New York, 2005; pp 367–379.
55. Bibber, J. An Overview of Nonhexavalent Chromium Conversion Coatings – Part I: Aluminum and Its Alloys. Met. Finish. 2001, 99 (12), 15–22.
56. O’Keefe, M.; Geng, S.; Joshi, S. Cerium-Based Conversion Coatings as Alternatives to Hex Chrome. Met. Finish. 2007, 105 (5), 25–28.
57. Bhatt, H.; Manavbasi, A.; Rosenquist, D. Trivalent Chromium for Enhanced Corrosion Protection on Aluminum Surfaces. Met. Finish. 2009, 107 (7–8), 31–37.
58. Vakelis, A. Electroless Plating. In Coatings Materials and Surface Coatings; Tracton, A. A., Ed.; Taylor & Francis: Boca Raton, 2007; pp 28-1–28-12.
59. McCaskie, J. E. Plating on Plastics. Met. Finish. May 2006, 31–39.
60. Schlesinger, M. Electroless Deposition of Nickel. In Modern Electroplating; Schlesinger, M., Paunovic, M., Eds., 5th ed.; John Wiley & Sons Inc.: New Jersey, 2010;
pp 447–458.
61. Paunovic, M.; Schlesinger, M. Fundamentals of Electrochemical Deposition, 2nd ed.; John Wiley & Sons: New Jersey, 2006.
62. Baudrand, D. Electroless Plating. Plat. Surf. Finish. August 2003, 27–28.
63. Anonymous. Autokatalyyttinen nikkeloïnti, Kemiallinen ja sähkökemiallinen pintakäsittely, osa 2; Suomen Galvanotekninen yhdistys ry: Helsinki, 1999; pp 192–210.
64. Paunovic, M. Electroless Deposition of Copper. In Modern Electroplating; Schlesinger, M., Paunovic, M., Eds., 5th ed.; John Wiley & Sons Inc.: New Jersey, 2010; pp 433–446.
www.iran-mavad.com
4.12 Plasma Sources in Thin Film Deposition
Y Setsuhara, Osaka University, Osaka, Japan

4.12.2 Pressure and Power Source Frequency of Plasma Sources 307
4.12.3 Plasma Sources for CVD Processes 308
4.12.3.1 DC Glow and DC Arc 308
4.12.3.2 Atmospheric Pressure Nonequilibrium Plasmas 310
4.12.3.3 Microplasmas 310
4.12.3.4 Capacitively Coupled RF Plasmas 311
4.12.3.5 Inductively Coupled RF Plasmas 312
4.12.3.6 Microwave Plasmas (ECR and SWP) 314
4.12.4 Plasma Sources for PVD Processes 315
4.12.4.1 Cathodic Arc 315
4.12.4.2 Magnetron Sputtering 316
4.12.5 Plasma Source for Large-Area Process 320
4.12.6 Summary 322
References 322
4.12.1 Introduction
Plasma processes have been intensively developed for more than three decades and have made outstanding progress as key
manufacturing technologies for development of a variety of industrial applications (1) ranging from surface modification of
materials (2), hard coatings (3), protective coatings for mechanical and tribological applications (4,5) to the state-of-the-art
advanced technologies including flat-panel displays (FPDs) (6), thin film photovoltaic cells (7), biomaterials (8), and microelec-
tronics (9,10).
Research and development of the plasma-process technologies are still in progress to challenge novel applications, including
nanostructure materials formation and nanoprecision control of materials surface (4) toward new materials synthesis in future. In
other words, it may not be exaggerated to say that the plasma-process technologies have been developed and are to be developed as
key technologies for providing technological solutions to overcome needs and/or problems for a variety of innovations on
a background of economical and social trends on a global scale. Especially, large-area processes are required for enhancement of
production efficiency and/or cost reduction (5,6).
Furthermore, low-damage and low-temperature processing of soft materials (polymers, biomaterials) (7) is required for
development of next-generation device technologies in organic–inorganic hybrid materials, which are expected to offer a wide range
of applications, including flexible electronic devices (11) and advanced ultra large-scale integration (ULSI) for low-k interlayer and
extreme ultraviolet lithography (12). As a means of low-temperature processes and nanofabrication methods, plasma processes are
considered to be one of the most promising technologies for device fabrications with organic–inorganic hybrid materials.
Among the wide range of plasma-process applications, thin film deposition can offer innovations with a variety of functionalities
via film formation to attain desired properties. Major roles of plasma sources in thin film deposition involve enhancement and
control of reactivity during film growth via production of ions and radicals in the plasmas (formation of high-temperature phases at
low substrate temperature) and supply of deposition species (especially metallic atoms with high melting point) via sputtering with
ions and/or evaporation with intense high-temperature discharge such as vacuum cathodic arc. Thin film deposition processes with
plasmas can be categorized into chemical vapor deposition (CVD) and physical vapor deposition (PVD).
In the following sections of this chapter, plasma sources for plasma-enhanced CVD and PVD processes are reviewed in terms of
discharge characteristics and specific issues in thin film deposition applications.
4.12.2 Pressure and Power Source Frequency of Plasma Sources
Most of the plasma sources for processing of materials are based on gas discharge, mainly because the plasmas can be sustained and
easily controlled by electric power sources. Historically, a variety of power sources have been employed for plasma production; from
direct current (DC) and alternating current (AC) voltage sources to high frequency (HF) and microwave power sources (1).
Plasma sources used for thin film deposition are summarized in Figure 1, showing typical regions of discharge conditions in
terms of gas pressure and frequency of the power source. The designated terms for the plasma sources are named in accordance with
the discharge regimes. With DC voltage power sources, a variety of plasmas can be sustained via DC glow, DC arc, DC magnetron,
and vacuum cathodic arc discharges. With higher frequencies, nonequilibrium (nonthermal) and thermal plasmas can be sustained

www.iran-mavad.com
308 Plasma Sources in Thin Film Deposition
10
10
9 ECR SWP
Frequency of power source (Hz)

10
8
10
7 CCP
10 RF magnetron
6 ICP
10
5
10 Pulsed magnetron
(unipolar/bipolar)
4
10 Dual magnetron
3
10 Atmospheric pressure plasmas
2 Thermal plasmas
10 Nonequilibrium (DBD, APGD, etc.)
1 Microplasmas
10
0
10 Cathodic arc
DC arc
DC DC magnetron DC glow
–4 –3 –2 –1 0 1 2 3 4 5
10 10 10 10 10 10 10 10 10 10
Pressure (Pa)
Figure 1 Typical regions of discharge conditions in terms of gas pressure and frequency of power source for a variety of plasmas in thin film deposition.
at gas pressures as high as atmospheric pressure. Pulsed magnetron plasmas are sustained with unipolar and bipolar DC pulses with
repetition frequency of several kilohertz to several hundred kilohertz. Power sources with frequencies as high as the radio frequency
(RF) can be used for sustaining RF magnetron plasma, capacitively coupled plasma (CCP), and inductively coupled plasma (ICP).
The RF power sources are also employed to generate thermal plasmas through inductive coupling. Microwave power sources are
employed to generate electron cyclotron resonance (ECR) plasma and surface wave plasma (SWP). The microwave power sources
can also be used for production of thermal plasmas and microplasmas. As can be seen from Figure 1, even at the same power-source
frequency, typical pressure regions for sustaining plasmas depend on the discharge regimes.
In the following sections, discharge characteristics and specific issues in thin film deposition are reviewed in accordance with
major applications to the CVD and PVD processes.
4.12.3 Plasma Sources for CVD Processes
In the plasma-enhanced CVD process, the thin film deposition process depends significantly on (1) sticking probability of
molecules dissociated mainly via electron-impact dissociation in the plasma as precursor, (2) migration of the precursors affected by
the substrate temperature, and (3) chemical reactions of excited species (radicals and ions) on the growing film (13). For formation
of good-quality films with high packing density, production of radical precursors with moderate sticking probability and
enhancement of the precursor migration are essential. In other words, production of precursors with sticking probability as high as
1 tends to result in deposition of films with lower packing density and rough surface due to insufficient migration of the precursors
on the growing films; moreover, cluster formation in the gas phase may also result in degradation of the film quality (defect
formation in the film). Furthermore, termination of dangling bonds via chemical reactions with excited species is also important.
In case of the hydrogenated amorphous silicon (a-Si:H) film deposition via plasma CVD with hydrogen diluted silane (SiH4)
gas, it is well known that the sticking probability of SiH3, which is produced via electron-impact dissociation of SiH4, is about 0.1;
however, radical species with excessive dissociation (SiH2, SiH, and Si) are about 1. Therefore, high-density production of SiH3
radicals with suppression of excessive dissociation of the molecule is very important for high-quality a-Si:H film formation (13).
Here it is important to be noted that the electron-impact dissociation process significantly depends on the electron temperature or,
more exactly, the electron energy distribution function of the plasma.
4.12.3.1 DC Glow and DC Arc

When a DC voltage applied between electrodes exceeds the breakdown voltage for the gas, a discharge can be ignited and plasma can
be sustained. This breakdown voltage depends on the pd product, where p is the gas pressure and d is the distance between the
electrodes. The dependence of this breakdown voltage on the pd product is well known as Paschen’s Law (10,14,15), which implies
that with increasing gas pressure a narrower gap distance is favored to attain a practical breakdown voltage for ignition of the gas
discharge.
Variations of the plasmas sustained at a lower gas pressure with the DC power source depend on the voltage–current relation of
the DC discharge. At a sufficiently low current region, faint discharge (called a Townsend discharge) is initiated. With increasing
current, discharge is sustained at almost constant voltage (normal glow). With successive increase of current to a certain value, there
www.iran-mavad.com
Plasma Sources in Thin Film Deposition 309
is a region where the voltage increases with current (abnormal glow). And at higher current region, the discharge tends to be highly
conductive and voltage drops to attain arc discharge. In low-pressure regions, plasmas sustained as normal glow and abnormal glow
are used for materials processing. At an atmospheric pressure, spark ignition proceeds directly to arcing (16).
In weakly ionized glow discharges sustained at low pressure, the discharge is sustained by ionization of gas molecules via
collision with electrons, which are accelerated by static electric field with the applied DC voltage. Typical electron density (or plasma
density) of the glow discharge at low pressure is in the range of 108–1013 cm3 (16) and the electron–molecule collision is not
sufficient enough to attain thermal equilibrium. Thus the electron temperature Te is higher than the gas temperature Tg and the ion
temperature Ti by one or two orders of magnitude, as illustrated in Figure 2 for a discharge with a mercury and rare gas mixture
(16,17). This feature of the glow discharge is referred to as ‘nonequilibrium plasma’ or ‘cold plasma.’
With increasing pressure at a constant current, enhancement of energy exchange via electron collisions with gas molecules and
ions leads to thermal equilibrium of electrons, gas molecules, and ions to attain a condition of Te w Ti w Tg as illustrated in
Figure 2. This condition of the discharge is designated as ‘thermal plasma,’ and the attained electron density (or plasma density) is
in the range of 1016–1019 cm3 (16). The DC arc sustained at pressures as high as atmospheric pressure is referred to as ‘plasma
torch’ due to its nature of directed plasma flow through a nozzle.
For the transition region from the nonequilibrium plasma at low pressure to the thermal plasma at high pressure, the medium-
pressure plasma in the 0.1–10 Torr (101–103 Pa) range has been defined by Yoshida et al. as ‘mesoplasma’ due to the advantages of
both nonequilibrium plasma at low pressure and thermal plasma (18). Materials processes with mesoplasma (mesoplasma pro-
cessing) are considered a suitable tool, particularly for large-scale thin film electronics, for which high-rate processing of materials is
an essential requirement for economical feasibility (18). Some excellent examples of mesoplasma processing are described in the
ICP section.
A typical design of the plasma torch consists of a cylindrically shaped cathode, an outer grounded and water-cooled shield as the
anode, and, for applications in thin film deposition, the plasma torches contain a port for injecting precursor compounds that are
used in thin film deposition (16). By feeding gas flow between the cathode and the shield, and by applying DC power of up to
200 kW, an arc between the electrodes may be ignited and sustained (16). Thermal plasma processing for film deposition is referred
to as injection plasma processing (IPP) for coating, which has been characterized into three types as thermal plasma CVD (injection
of gaseous or liquid reactants into the plasma), thermal plasma flash evaporation method (injection of powders as raw materials
into the plasma), and plasma spraying (injected powder size must be large enough not to be vaporized, and small enough to be
completely melted) (19). In the literature, Yoshida pointed out that an RF plasma torch and a hybrid plasma torch are strongly
recommended, because the most important key to success for IPP in various coating applications depends significantly on the
development of plasma torches that make it possible to generate relatively large-size plasma, and to sustain a stable plasma, in
particular, when raw materials are being axially injected (19).
Thermal plasma processing was originally developed as an effective coating technology primarily due to its highly collisional
characteristics (i.e., its extremely high-temperature reaction environment), and thus high processing rates are attained in general (18).
In the IPP, the precursors, in the form of solid pellets, powder, or volatile molecules, are injected just at the downstream of the
arc, where they are vaporized and/or dissociated into reactive species, which are sprayed onto a substrate, thereby coating it with
a film at rates up to 10 mm min1 (16). Films successfully deposited so far via IPP with plasma torches include SiC (20), SiN (21),
TiO2 (22), Y–Ba–Cu–O (29,23), Al2O3 (24), and diamond (25–29).
1
10
Te of Nonequilibrium plasma
Nonequilibrium plasma
and microplasma
Te
Temperature (eV)
0 Thermal plasma
10 Mesoplasma
–1
10
Tg
Tg of Nonequilibrium plasma
and microplasma
–2
10 –2 –1 0 1 2 3 4 5
10 10 10 10 10 10 10 10
Pressure (Pa)
Figure 2 Schematic of gas-pressure dependence of electron temperature Te and gas temperature Tg in plasma discharges sustained with DC power
source at a constant current. Reproduced from Kubota, N.; Ayabe, M.; Fukuda, T.; Akashi, M. Mater. Res. Soc. 1997, 458, 449–452.
www.iran-mavad.com
4.12.3.2 Atmospheric Pressure Nonequilibrium Plasmas

As described previously, the electric discharge sustained at an atmospheric pressure tends to be in thermal equilibrium state, in
which the electron temperature tends to decrease and the gas temperature tends to be elevated due to enhancement of energy
exchange via electron collisions with gas molecules and ions. When the substantial discharge duration sd and the residence time of
the source gas sres are considerably shorter than the electron energy relaxation time srelax (sd << srelax and/or sres << srelax), however,
it may be possible to suppress energy exchange of the electrons with the gas molecules and ions. Consequently, it becomes possible
to maintain the nonequilibrium state (Te >> Tg), as shown in Figure 2, even in the discharges sustained at pressures as high as the
atmospheric pressure (30).
Discharge conditions to maintain the nonequilibrium state (Te >> Tg) as mentioned are known to be attained with dielectric
barrier discharge (DBD) (31,32). The source configuration for this discharge regime consists of two parallel electrodes, on which at
least one electrode is covered with a dielectric insulator, as schematically shown in Figure 3. In the DBD regime, due to the charge
accumulation on the dielectric insulator, the substantial duration of the discharge sd is effectively suppressed so as to be consid-
erably shorter than the electron energy relaxation time srelax, as sd << srelax. Furthermore, in this discharge regime, the plasma is
generated through a number of filamentary microarcs or streamers, lasting for 10–100 ns, with sizes around 10–100 mm in
diameter, and the streamers randomly distributed in space and time (33–35). Historically, DBD has been used for ozone production
since its initial development in 1857 (36); it has also been examined for the plasma-assisted CVD of polymers (33). However, since
the plasma is not uniform, its use in deposition is limited to cases where the surface need not be smooth (16). Recently, Nozaki and
coworkers proposed that this feature of rough surface formation in CVD with DBD is partly attributed to particle formation in the
gas phase due to ionization in the streamers in a wide range of gas gap (37).
In the late 1980s, Okazaki and her coworkers (38) discovered that a stable and uniform discharge analogous to a low-pressure
glow discharge was sustained even at an atmospheric pressure under selected conditions: (1) helium was used as dilute gas, (2) an
insulating plate was set on the lower electrode plate as in the DBD regime, (3) the brush-style electrode was better for the upper
electrode, and (4) the continuous and stable discharge was created by using a power source at a frequency of 3 kHz or RF
(13.56 MHz) and the discharge was not stable when using an AC power source at 50 Hz.
The discharge regime to attain the stable and radially uniform plasma has been designated as atmospheric pressure glow (APG)
discharge (38,39). The discovery by Okazaki and her coworkers is considered to be significant in that the stable and radially uniform
plasma can be sustained with a very similar discharge configuration to that of DBD (one electrode covered with the dielectric
insulator) when selected conditions described above are employed for the kind of dilute gas (He) and the frequency of power
(3 kHz or RF). APG is operated near breakdown field; thus current density per voltage half cycle is uniquely determined by gap
separation and dielectric properties (37).
The distinct difference between DBD and APG in terms of the discharge structure has brought about remarkable inno-
vations on materials processing with atmospheric pressure plasmas. Recent studies on CVD of carbon nanotubes (CNT)
by Nozaki and coworkers have shown that multiwall CNTs were successfully deposited when the APG regime was employed
for CVD process; however, the CVD with the DBD regime resulted in formation of carbonaceous materials, but CNTs could
not be grown, even though the deposition experiments with APG and DBD regimes were carried out using the same
apparatus (37).
Furthermore, it has been reported that the APG with low-frequency (10–100 kHz) power source requires a dielectric barrier on at
least one side of the electrodes for stable and radially uniform operation; however, when the operating frequency is increased to RF
at 13.56 MHz, the plasma stability is greatly improved and consequently the dielectric barrier is no longer required for stable
operation (40). This feature of the stable production of atmospheric pressure nonequilibrium plasmas may be common with the
plasma jet (16) and the cold plasma torch (41), in which RF power at 13.56 MHz was employed to generate nonequilibrium
plasmas at an atmospheric pressure. The plasma jet has been applied to plasma-assisted CVD of SiO2 films (42) and the cold plasma
torch has been applied to CVD of SiO2 and TiO2 films (43,44).
4.12.3.3 Microplasmas
As described in the previous subsection, when the substantial duration of the discharge period and the residence time of the source
gas are considerably shorter than the electron energy relaxation time (sd << srelax and/or sres << srelax), it may be possible to
Electrode
Dielectric insulator
V
Dielectric insulator
Source gas Electrode
Figure 3 Schematic illustration of source configuration for dielectric barrier discharge (DBD) and atmospheric pressure glow (APG) discharge.
www.iran-mavad.com
suppress energy exchange of the electrons with the gas molecules and ions, and thus it will be possible to maintain the
nonequilibrium state (Te >> Tg), as shown in Figure 2, even in the atmospheric-pressure discharges (30). The physical situations in
plasma production with reduced feature size down to the micrometer–millimeter range may meet the above-mentioned conditions
because the residence time sres of source gas in the microplasma is as short as 1 ms to 1 ms in a small size at a moderate gas flow rate,
and the discharge is often operated in a pulsed mode, of which the pulse duration is shorter than the electron energy relaxation time
srelax, as sres << srelax (30).
Another feature of the microplasma is that the optimum gas pressure p for attaining breakdown voltage as small as possible in
accordance with Paschen’s Law (10,14,15) becomes a few tens of kPa to MPa due to scaling down of d from millimeter to
micrometer sizes. This implies that microplasma is characterized as ‘collision-dominated high-density plasma’ (30).
The microplasmas are reported to generate in a wide frequency range of the power sources from DC to GHz (30). In the plasma
production in DC operation, the most popularly used design is the hollow cathode type, which has been developed by
Schoenbach’s group recently for miniaturized structures and integrated assemblies (45,46). Engemann’s group has demonstrated
a microplasma production with DBD scheme (47) with a quite simple structure, in which the capacitively coupled ring electrodes
are driven by a low-frequency power source (1–10 kHz), keeping the upstream side grounded. With increasing pressure for
sustaining discharge, on the contrary, power sources with higher frequency are favorable considering increases in the electron
collision frequency (30). There have been reports with RF at 13.56 MHz (48,49) and even at ultra-high frequency (UHF) at 450 MHz
(50). Microplasmas will be applied to localized and high-rate deposition of films.
4.12.3.4 Capacitively Coupled RF Plasmas

CCP can be sustained by applying RF voltage at frequencies in the region of 10–100 MHz to parallel electrode plates (1,10), as
schematically shown in Figure 4. Electrons are accelerated via static electric field capacitively coupled to the RF voltage. In
a capacitively coupled discharge, typical electron density is in the range of 108–1010 cm3 in gas pressures in the range of
101–103 Pa. In these typical discharge conditions, electron temperature (1–10 eV) is much higher than gas temperature, and thus the
CCPs are in nonequilibrium state. In most of the discharge conditions, the electron plasma frequency (0.1–1 GHz), which is in
proportion to square root of electron density, is much higher than the frequency of the RF power source (10–100 MHz), and the
static electric field cannot penetrate deeply into plasma region due to skin effect and is applied only in the skin sheath region.
A characteristic feature of CCP is that the ions are accelerated toward the powered electrode due to negative DC potential
(self-bias) that develops between the plasma and the powered electrode as a consequence of (1) the use of a coupling capacitor
between the output of the impedance matching network and the powered electrode to make the powered electrode at floating
potential and (2) the use of appropriately shaped areas of the (smaller) powered electrode and the (larger) grounded electrode
(1). The ion energy near the powered electrode can reach energies of a few hundred electronvolt due to the self-bias. The self-bias
voltage increases with increasing RF power; however, the plasma density and consequently the ion current density increase at the
same time. Thus for independent control of the ion flux and the ion energy, another RF power source can be used for bias energy
control.
Successful applications of CCP include etching processes in microelectronic device fabrications (9,10,12) and silicon film
deposition for photovoltaic cells (7,13) and thin film transistors of FPDs (6). This regime of plasma production is also used in high
pressure up to atmospheric plasmas as mentioned in the preceding sections.
Impedance
matcher
Blocking
capacitor
Powered electrode
RF power
source
Grounded electrode
Figure 4 Schematic illustration of source configuration for CCP.
www.iran-mavad.com
Figure 5 Schematic illustrations of source configurations for ICPs: (a) conventional downstream type, (b) planar antenna configuration, and (c) internal
antenna configuration.
4.12.3.5 Inductively Coupled RF Plasmas

ICP can be sustained by applying RF current at frequencies of the order of 1–100 MHz to inductive antenna conductor (1,10,51), as
schematically shown in Figure 5. Electrons are accelerated by an inductive electric field Eind, which is induced by a sinusoidal
temporal variation of the RF magnetic field Brf following the vector differential equation representing Faraday’s law of induction;
i.e., rot Eind ¼ vBrf/vt. The vector differential equation implies that (1) the direction of the induced electric field is the rotation
direction of the temporal variation of the magnetic field, and (2) the strength of the induced electric field is in proportion to the
temporal variation of the magnetic field strength. In other words, (1) the electric field induced by the RF antenna conductor is in
parallel along the antenna conductor and (2) the strength of the induced electric field is in proportion to the RF current and the
frequency of the RF current. Typical electron density of ICP is in the range of 1010–1012 cm3 (approximately one or two orders of
magnitude higher than that in CCP) in gas pressures in the range of 101–103 Pa. As in the case with CCP, in these typical discharge
conditions, electron temperature (1–10 eV) is much higher than gas temperature, and thus the ICPs generated in low pressure
regions are in nonequilibrium state. Whereas, the ICPs sustained at higher pressures tend to be thermalized to approach thermal
equilibrium with increasing pressure, as shown in Figure 2. Also, as in the case with CCP, in most of the discharge conditions, the
electron plasma frequency is of the order of 1–10 GHz, which is much higher than the frequency of the RF power source
(1–100 MHz). This implies that the induced electric field cannot penetrate deeply into plasma region due to the skin effect and is
applied only in the skin sheath region.
Variations of ICP reactors with different antenna configurations are schematically shown in Figure 5. The downstream type of
reactor shown in Figure 5(a) employs a helical coil wound around a dielectric tube. In the planar antenna configuration shown in
Figure 5(b), a helical coil wound on a plane is employed for large-area processing. In these antenna configurations, the RF antenna
is located in air and the induction electric field is applied to the skin sheath region through a dielectric window. Here one of the
major problems associated with the dielectric window in these ICP configurations (external supply of discharge power) is degra-
dation of the source efficiency because the field strength for sustaining the discharge tends to be reduced by the thick dielectric
window (52).
From this viewpoint, an internal antenna configuration, as shown in Figure 5(c), to produce ICP has attracted great interest as
a promising candidate as an efficient high-density source, as well as an alternative to avoid problems associated with dielectric
windows in conventional sources. The internal antenna-driven ICPs have been also studied as one of the methods for enhancement
of the planner magnetron discharge to increase sputtered metal ionization (53,54). In the internal antenna configurations, however,
it is essential to minimize the electrostatic coupling of RF antenna voltages to the plasma, which results in anomalous rise of plasma
potential (52–54).
Figure 6 shows an electrical circuit representation of an ICP antenna and schematic illustrations of RF voltage distribution along
the antenna. When a sinusoidal RF current Irf is applied to the antenna with an inductance L, the RF voltage arising at the power end
of the antenna V1 is given by V1 ¼ uLIrf, where u is an angular frequency of the RF current. Assuming a simple case where the RF
current leakage to the plasma is neglected, the RF voltage may be linearly distributed along the antenna electrode as schematically
illustrated in Figure 6. Here it is pointed out that the RF voltage is always present at the powered end of the antenna, and thus the
electrostatic coupling as in the CCP is always present in generation of ICPs.
For ICP generation with lowered plasma potential, which is significant for low-damage high-quality processing of materials,
minimization of the electrostatic coupling of RF antenna voltages to the plasma is very important. Furthermore, considering the
antenna designs for the ICP production to satisfy the requirement for enlargement of plasma sources toward a meter scale, increase of
an antenna inductance and hence increase of an RF-voltage amplitude arising at the antenna terminals cannot be avoided with
increasing source size, when the source employs a large loop-shaped antenna. These constraints associated with large-area and/or large-
volume ICP sources may be solved by employing the antenna configuration that allows low-voltage operation of the ICPs (52).
Feasibility of high-density plasma production with suppressed electrostatic coupling by the low-voltage operation of the internal
antenna has been approached by Setsuhara (52) via (1) the employment of antenna configurations with reduced inductance, which
is roughly proportional to the square of turn numbers and the area size of the loop, (2) minimization of the RF-voltage amplitude
by the antenna termination with a blocking capacitor, and (3) dielectric isolation of the antenna conductor from plasma. As a novel
www.iran-mavad.com
Figure 6 Electrical circuit representation of ICP antenna and schematic illustrations of RF voltage distribution along the antenna.
technology to achieve low-voltage operation of ICPs suitable for producing large-area source with low-damage process capability,
an internal antenna configuration with multiple low-inductance antenna (LIA) units, as schematically shown in Figure 7, has been
proposed (52). The ICP production using the LIA units has resulted in stable plasma production at RF input powers of 3 kW to attain
densities approaching as high as 1 1012 cm–3 at argon pressures around 1 Pa with simultaneous achievement of the suppression of
the electrostatic coupling (52).
Furthermore, the reduction of the antenna inductance for sustaining ICPs has exhibited an additional favorable feature to
perform low-damage plasma processes via low-voltage operations of ICPs. Figure 8 shows the ion energy distribution function
(IEDF) attained in an ICP sustained with LIA at Ar at a pressure of 13 Pa and with 13.56 MHz RF power of 1 kW. A typical IEDF for
conventional RF plasma source (CCP mode with larger potential fluctuation) is also shown for comparison. The peak ion energy in
the IEDF for the ICP sustained with the LIA is shown to be significantly lowered as being equivalent to or less than 5 eV as compared
with conventional RF plasma sources. Moreover, the IEDF for the ICP sustained with the LIA showed a very narrow energy width of
about 2 eV, which is significantly narrower than that for the conventional RF plasma sources.
Figure 7 Schematic illustration of low-inductance antenna for ICP generation.
www.iran-mavad.com
Figure 8 Ion energy distribution function (IEDF) attained in an ICP sustained with LIA at Ar at a pressure of 13 Pa and with 13.56 MHz RF power of 1 kW.
Typical IEDF for conventional RF plasma source (CCP mode with large potential fluctuation) is also shown for comparison.
Plasmas sustained with the ICP regime have been widely employed for a variety of materials processing due to capability of high-
density plasma production (1,10). The feature of low-damage processing with the LIA has enabled high-quality and large-area
deposition of microcrystalline silicon films; the bottom-gate thin film transistors formed via plasma-enhanced CVD has shown
enhanced mobility as well as significant advantages in thin-film transistor (TFT) reliability with considerably suppressed threshold-
voltage shifts less than 0.5 V (55).
The ICP regime of plasma production has also been used at high pressure up to atmospheric pressure, as mentioned in the
preceding sections.
4.12.3.6 Microwave Plasmas (ECR and SWP)

Plasma generation with microwaves (microwave plasmas), which can be operated at a considerably wide range of gas pressures
from high vacuum at pressures as low as 103 Pa to even atmospheric pressure, is widely employed in many applications (1). In the
microwave plasmas, the microwave power is transferred via a waveguide (or via a coaxial cable) and is launched to the plasma
generation region through a dielectric window, which is configured at one face of the waveguide (or through a dielectric tube
surrounding the inner electrode of the coaxial cable). Among a variety of the microwave plasmas, the ECR plasma and the SWP,
which are schematically shown in Figure 9, are most commonly used in materials processing for etching and CVD.
The ECR plasma is a typical example of microwave plasma generation in the presence of external static magnetic field. As
schematically shown in Figure 9(a), the ECR plasma can be generated via electron cyclotron wave, which is a right-hand
circular wave excited in magnetized plasmas, when (1) the electron cyclotron frequency with the static magnetic field strength
Microwave
Dielectric window
Electromagnets Dielectric window
Microwave
Surface wave
Electron
ECR region cyclotron
wave
(a) ECR plasma (b) Surface wave plasma
Figure 9 Schematic illustrations of source configurations for microwave plasmas: (a) ECR plasma and (b) surface wave plasma.
www.iran-mavad.com
satisfies the resonance condition (ECR region in Figure 9(a)) with the frequency of the microwave, and (2) the microwave
launching to the plasma satisfies the non-cut-off condition in the dispersion relation of the wave. For example, the magnetic
field strength of 875 G corresponds to the electron cyclotron frequency of 2.45 GHz. A characteristic feature of the ECR
plasma is that discharges with high plasma density can be efficiently sustained even at high vacuum at pressures as low
as 103 Pa.
In the absence of the external magnetic field, on the contrary, the microwave power launched through the dielectric window can
excite a surface wave to generate SWP (56,57), as schematically shown in Figure 9(b). The surface wave can propagate along the
boundary between the dielectric window and the plasma. Recently, Sugai and coworkers developed a large-area source for silicon
film deposition (58). When the SWP is generated with a long waveguide to sustain large-area plasmas, the microwave power density
is attenuated along the waveguide, in general, due to microwave power absorption to the plasmas. Therefore, in large-area plasma
generation with SWP, the waveguide is configured with a certain number of slots, through which the microwave power is launched
toward the plasma chamber through the dielectric window and the power distribution along the microwave propagation is
controlled (58).
4.12.4 Plasma Sources for PVD Processes
In the plasma-enhanced PVD processes, it has been generally recognized that the bombardment of a growing film with energetic
ions results in beneficial modifications in attainment of desired properties and characteristics of the deposited films (59). One of the
effects of ion bombardment on the microstructure of the deposited films is well depicted by a modified structure diagram (60) as
a function of the particle’s energy, which was originally proposed by Thornton (61).
Furthermore, the ion bombardment has significant effects on the formation of crystalline films and/or the metastable phase at
lower substrate temperature than is expected in the thermal equilibrium state. As well as this significant advantage, phase formation,
preferential crystallographic orientation, and residual stress in the films can also be controlled primarily by the bombardment
energy and the ratio of the ion flux fi to the flux of metal atoms fm transported to a growing film on the substrate (59). Key
parameters for the ion bombardment processes are (1) flux ratio of ions to deposition metal atoms fi/fm and (2) ion energy Ei
bombarded onto the growing film via substrate bias. Here it is recognized that the product of the flux ratio and the ion energy
(Ei fi/fm) corresponds to the average energy supplied to the growing film per a deposition metal atom.
In the considerably low limit of the flux ratio (fi/fm << 1), the deposition process is not considerably affected by the ion
bombardment; however, with increasing flux ratio, improvement of the film quality (adhesion, film packing density, high-
temperature phase formation, crystallization, and preferential orientation control) can be expected. One of the well-known
examples is the deposition process of cubic boron nitride (c-BN) films, in which bombardment with ions with a certain critical
kinetic energy is required for c-BN nucleation on non-c-BN substrates (62). Especially, the deposition process with the flux ratio of
fi/fm > 1 is referred to as ionized PVD (IPVD) (63). It is also noted that excessive amount of the flux ratio may result in resputtering
of the deposition atoms and no film can be formed. Thus, in the PVD processes, proper control of ion density (plasma density) is
one of the key issues for the plasma source.
4.12.4.1 Cathodic Arc

The cathodic arc, which is also referred to as vacuum arc, is a discharge between two metallic electrodes (cathode and anode) in
vacuum, in which arc spots with 1–10 mm size are concentrated on the cathode surface (64–67). The cathodic arc plasma source
consists of a metallic cathode as a source material, an anode, an arc ignitor, and a means of arc confinement (66). The current density
in the spot is of the order of 106–108 A cm2, in which about 10% of the discharge current is the ion current from the cathode (64).
Since the spots move rapidly and randomly across the surface of the cathode, arc spot confinement is a key issue in arc source design
and configuration with magnetic field for stable and uniform deposition (66).
The plasma produced at cathode spots expands rapidly into the vacuum and the final ion velocities have been found in the range
1–2 104 m s1, which are almost independent of the cathode material and ion charge state (67). This fact has been attributed to
electron–ion coupling (67,68). The directed ion velocities correspond to approximately 20 eV for light elements and 200 eV for
heavy elements (67). These high ion energies are favorable for the different film properties obtained by cathodic arc deposition.
Randhawa et al. (65) have summarized the characteristic features of the cathodic arc plasma and the cathodic arc plasma
deposition (CAPD) process as follows: (1) the core of the CAPD process is the arc spot that generates material plasma, (2) a high
percentage (30–100%) of the material evaporated from the cathode surface is ionized, (3) multiply charged ions are produced in
the plasma, and (4) the kinetic energies of the ions are typically in the range 10–100 eV.
Despite the desirable features of the CAPD process with a high degree of ionization and moderately high ion energy, the plasma
source suffers from problems associated with simultaneous production of small liquid droplets commonly referred to as macro-
particles, with sizes ranging from 0.2 to 100 mm (64). In the case of carbon, the particles take the form of solid debris (64). In order
to avoid problems with the macroparticles in thin film deposition, considerable efforts have been made in recent years using
a variety of macroparticle filters (64,67). Thin film deposition methods with macroparticle filters are referred to as filtered cathodic
arc deposition (68), which has been applied to a variety of advanced coatings. Successful applications of the cathodic arc include
deposition of metallic films, nitrides, carbides, oxides, and diamond-like carbon (DLC) (64–69).
www.iran-mavad.com
4.12.4.2 Magnetron Sputtering

Magnetron sputtering discharge (1,63,70,71) can be sustained by applying negative voltage Vd to a sputtering target, which is
configured with a component of external magnetic field line B0 overlaid in parallel with target surface, as schematically shown in
Figure 10. Since the static electric field E0 with a direction perpendicular to the target is applied in the vicinity of the target surface
when the negative voltage Vd is applied to the target, the magnetic configuration provides a region in which the static electric field E0
is perpendicular to the external magnetic field line B0. In this region, electrons accelerated by the static electric field perform a drift
motion in the direction of E0 B0, which is perpendicular to both E0 and B0 and is in parallel with the target surface. This region in
which the static electric field E0 is perpendicular to the external magnetic field line B0 is called the E B (E cross B) region.
Consequently, the electrons are effectively confined in the E B region. Electrons confined in the E B region ionize the working
gas molecule via electron-impact ionization. The negative voltage to the target can be applied by employing a negative DC power
source for DC magnetron (70,71) and an RF power source for RF magnetron (1). In addition to those conventional magnetron
sputtering, discharge regimes with DC pulse source at repetition rate of the order of 10–100 kHz (72) have been developed recently.
The ions are accelerated toward the target surface by the static electric field and collide with the target surface. In the ion collision
with the target surface, secondary electrons are emitted from the target surface and are accelerated back to the discharge region by the
electric field E0 to gain kinetic energy of eVd, which is sufficiently high enough to ionize the gas molecules in the discharge region, in
the same manner of sustaining DC glow discharge (known as g effect). Due to the effective confinement of the accelerated electrons
in the E B region, high-density plasmas with electron density of the orders of 1012–1013 cm3 can be easily sustained in the
magnetron discharge in argon gas pressure of 101–102 Pa by applying DC target voltage of 300 to 500 V.
In the ion collision onto the target surface, atoms in the vicinity of the target surface are recoiled by the ion collision to gain
a kinetic energy to form collision cascade with atom displacements. When the target atoms at the very surface have a kinetic energy
higher than the bonding energy of the target material, the atoms can escape from the target. This phenomenon of atom ejection
from solid matter is called sputtering (73,74). Here it is important to note that the number of atoms sputtered out of the solid per an
incident ion is defined as sputtering yield, which increases with increasing ion energy up to several tens of keV (73,74). This means
that the atom flux from the target is determined as the product of the ion current to the target and the sputtering yield. Thus, in
conventional magnetron sputtering, increases in the target voltage result in both the increase in the ion current due to enhancement
of plasma density and the increase in the sputtering yield. As a consequence, considering that the flux ratio of ions to deposition
metal atoms fi/fm to the substrate is significant for the sputter deposition process, it is noted that the ion flux and the atom flux
cannot be controlled independently in the conventional magnetron sputtering process.
Magnetron sputtering has become one of the conventional technologies for thin film deposition, which has been widely
employed in a variety of applications (1) for more than three decades; however, further technological developments have been and
are still in progress to overcome the following issues: (1) enhancement of the metal ionization and increase of the ion current to the
substrate for enhancement of reactivity during deposition via substrate biasing (63,70,71); (2) enhancement of deposition rate in
reactive sputter processes, in which the supply of reactive gas for compound film formation (especially oxides and nitrides) suffers
degradation of the deposition rate due to compound formation on the target surface (known as target poisoning) (72,75–81); and
(3) utilization of high-power pulsed DC bias to the target for enhancement of metal ionization and/or reactivity of the deposition
process (63,72,82–85). The problem in (2) the reactive sputter process with DC and RF magnetron has been characterized as the
hysteresis of the reactive-gas partial pressure vs. flow rate due to compound layer formation on the target called target poisoning, in
which the deposition rate drops considerably (72,75–77). The reactive DC and RF sputtering process has been improved via
ExB region
S N Target
S
Vd
E0
B0 B0
Substrate
Vb
Figure 10 Schematic illustration of magnetron sputtering.
www.iran-mavad.com
stabilization of the deposition process by precise control of reactive gas flow and pumping speed, plasma emission monitoring, and
arc detection (72,75–77). After this early stage of the technological development with the DC and RF sputtering process, deposition
methods with unipolar and bipolar pulsed magnetron sputtering techniques with single and dual targets have been developed for
further improvement of the reactive sputter deposition with enhanced deposition rate (72,78–81). Furthermore, (3) utilization of
high-power pulsed DC bias to the target, which is referred to as high-power pulsed magnetron sputtering (HPPMS), is a relatively
new technique first proposed by Kouznetsov et al. (82,83) for generation of highly ionized plasmas and as a measure to minimize
the effects of arcing in the reactive sputter deposition process (63,72,82–85). In this novel pulsed magnetron discharge technique of
HPPMS, while the average power is kept at as low as several W cm–2, the discharge is concentrated in short pulses with peak power
densities up to several kW cm2, which is considerably higher than that for a conventional DC magnetron, less than a few tens
W cm2 (83,85).
Enhancement of the ionization in the magnetron sputtering (issue (1) above) can be attained either by improvement of plasma
confinement or by an installation of additional discharge source to assist the magnetron discharge (70,71). The unbalanced
magnetron and the magnetron sputtering using a magnetic multipolar plasma confinement are typical examples of the systems
based on the improvement of plasma confinement (70,71). Furthermore, the sputtering systems with enhanced ionization have
been developed with additional discharge, including ECR plasmas and ICP (63,70,71).
Figure 11 shows a schematic illustration and a photograph for the typical configuration in the ICP-assisted magnetron
(53,54,63). An example of the axial distribution of the plasma density with and without ICP assistance is shown in Figure 12
(53,54). In this example, a magnetron source with an Al disc target of 64 mm in diameter has been assisted by installation of ICP
sustained with an antenna coil located at axial region of z ¼ 40 to 0 mm and discharge experiments were preformed in argon gas at
a pressure of 5.2 Pa, where the sputtering target was applied with DC voltage of 200 V and the ICP was sustained with 13.56 MHz
RF power of 400 W. In the conventional magnetron regime (magnetron only in Figure 12), the electron density in the region apart
from the target surface was considerably low, on the order of 109 cm3, even though the plasma density in the E B region in the
vicinity of the target surface was expected to be considerably higher than this. On the contrary, the electron density with ICP (ICP
only in Figure 12) in the region between the target and the substrate showed by more than two orders of magnitude higher density
than that with the magnetron only. Furthermore, the electron density of the magnetron discharge assisted with the ICP
(magnetron þ ICP in Figure 12) showed enhancement of the electron density in region near to the target. As a consequence, target
discharge current was observed to increase from 0.4 A for the case of magnetron only to 2.7 A for the case of magnetron þ ICP (53).
Furthermore, the enhancement of the plasma density in the substrate region via additional installation of ICP is significant for
utilizing the ions (both Ar and Al ions in this example) in ion-assisted deposition process or IPVD (53) with substrate biasing since
the enhancement of the plasma density directly corresponds to the enhancement of the transport ratio of the ion flux to the
deposited atoms on a growing film.
In addition, the ion-to-neutral emission intensity ratios are compared for Ar and Al species. Figure 13 shows the ion-
to-neutral intensity ratios of optical emission lines of Al and Ar measured from (1) the upper (z ¼ 46 mm) and (2) the lower
(z ¼ þ50 mm) windows as a function of RF power (53). Here, the ICP-assisted sputtering discharge was generated in an argon
pressure of 5.2 Pa with a target voltage of 300 V. The emission lines used for this investigation are the Arþ ion line at 480.6 nm,
the Ar neutral line at 420.0 nm, the Alþ ion line at 390.1 nm, and the Al neutral line at 394.4 nm. The results indicate that the
energy of electrons in the discharges sustained with ICP is primarily consumed for the ionization of the Al atoms, with the first
ionization potential (w6.0 eV) substantially smaller than that for Ar atoms (w15.8 eV), and thus the ionization of the sputtered
Al atoms are effectively enhanced in the high-density plasma mainly sustained with ICP in the region between the target and the
substrate.
Figure 11 Schematic illustration of ICP-assisted magnetron.
www.iran-mavad.com
Figure 12 Axial distribution of the electron density in presence and absence of the ICP sustained at an argon pressure of 5.2 Pa. The antenna coil
was located in the axial region between z ¼ 40 mm and z ¼ 0 mm. The sputter discharge experiments were preformed at an argon pressure of
5.2 Pa with a target voltage of 200 V. The ICP was sustained with 13.56 MHz RF power of 400 W.
(a) Upper window (z = –46 mm)

Ar pressure = 5.2 Pa, Target voltage = –300 V
Intensity ratio (arb. units)
100
+ *
Ar / Ar
10–1
+ *
Al / Al
10–2
10–3
0 100 200 300 400
RF power (W )
(b) Lower window (z = +50 mm)

Ar pressure = 5.2 Pa, Target voltage = –300 V
Intensity ratio (arb. units)
100 + *
Ar / Ar
10–1
10–2
+ *
Al / Al
10–3
0 100 200 300 400
RF power (W )
Figure 13 Variation in the intensity ratios of ion to neutral lines of Al and Ar measured from (a) the upper (z ¼ 46 mm) window and (b) the
lower (z ¼ þ50 mm) window, as a function of RF power. Here the sputtering discharge was generated at an argon pressure of 5.2 Pa with a target voltage
of 300 V. The emission lines used for this investigation are the Arþ ion line at 480.6 nm, the Ar neutral line at 420.0 nm, the Alþ ion line at 390.1 nm,
and the Al neutral line at 394.4 nm. Reproduced from Setsuhara, Y.; Kamai, M.; Miyake, S.; Musil, J. Jpn. J. Appl. Phys. 1997, 36, 4568.
www.iran-mavad.com
Another example of ICP-assisted magnetron sputtering is introduced here for the configuration schematically shown in
Figure 14 (86), in which the plasma density distribution obtained from a plasma fluid simulation is also depicted. In this example,
the magnetron sputter discharge is assisted with a new type of LIA located beside the magnetron target. In this new type of LIA, an RF
antenna conductor with a length much shorter than the RF wavelength is embedded in a hall region dug in the chamber wall and the
dielectric window plate is installed at the same height of the chamber-wall inner surface as an insulator for shielding the antenna
conductor from the plasma. Since the conventional LIA shown in Figure 7 has the shape that is salient toward the plasma region
from the chamber-wall inner surface, this new type of the LIA shown in Figure 14 is called as inner-type LIA (86). A special feature of
the ICP-assisted magnetron with the inner-type LIA is that the ICP antenna is located at the same height or beneath the target surface
of the magnetron. In this new configuration, the plasma in the region between the target and the substrate is maintained by
diffusion of the plasma sustained via RF power absorption in the vicinity of the inner-type LIAs.
In the conventional designs of the plasma-assisted magnetrons (70), additional plasma sources are located in the region between
the target and the substrate; thus the target–substrate distance is elongated due to the installation of the additional plasma source. As
a consequence, the conventional configuration of the plasma-assisted magnetron suffers loss of deposition rate because a consid-
erable portion of the sputtered metal flux from the target could not contribute to the film deposition due to an increase in collisions
of the metal atoms with ambient gas molecules in the elongated target–substrate distance.
This new type of the ICP-assisted magnetron has been applied to film deposition of transparent amorphous oxide semi-
conductor, a-InGaZnO4 (a-IGZO), which is expected as a new material for the next-generation thin film transistor, which is essential
for fabrication of advanced devices including high-speed FPDs and flexible devices, due to considerably higher mobility
(>10 cm2 (Vs)1) than that with amorphous silicon widely used in the conventional devices (87–90). Typical target discharge
characteristics (target current as a function of target voltage) are shown in Figure 15 for the magnetron discharge with an IGZO disc
Figure 14 Schematic illustration of new type of ICP-assisted magnetron sputter system, in which the magnetron sputter discharge is assisted with ICP
sustained with inner-type LIA.
www.iran-mavad.com
Figure 15 Target current as a function of target voltage for DC magnetron sputter discharge assisted with ICP sustained with inner-type LIA.
A magnetron source with an InGaZnO4 rectangular target was assisted by installation of ICP sustained with two inner-type LIAs located beside the sputter
target and discharge experiments were preformed in argon–oxygen mixture gas with oxygen partial pressure of 20% at a pressure of 2.0 Pa, where the
sputtering target was applied with DC voltage and the ICP was sustained with 13.56 MHz RF power in the range of 0–500 W.
target biased with DC voltage with ICP assistance. In this example, a magnetron source with an InGaZnO4 rectangular target has
been assisted by installation of ICP sustained with two inner-type LIAs located beside the sputter target, and discharge experiments
were preformed in argon–oxygen mixture gas with oxygen partial pressure of 20% at a pressure of 2.0 Pa, where the sputtering target
was applied with DC voltage and the ICP was sustained with 13.56 MHz RF power in the range of 0–500 W. In absence of the ICP
assistance (RF power ¼ 0 W), target discharge could not be sustained with DC negative bias up to 500 V due to high electrical
resistivity of the target material. On the contrary, when the ICPs were additionally sustained, a substantial amount of sputter
discharge was observed with a target current sufficiently high for the IGZO film deposition. Furthermore, the target discharge current
has exhibited voltage–current characteristics analogous to the characteristics of the ion saturation current, implying that the plasma
density in the vicinity of the target is primarily determined by the discharge with the ICP; i.e., the target current does not increase
significantly with increasing negative bias to the target when the RF power is kept constant. This feature of the voltage–current
characteristic can offer flexible and independent control capability in the flux ratio of the reactive species to the deposited atoms. In
other words, the flux of the deposited atoms can be controlled independently via target voltage (the sputtered atom flux increases
with increasing voltage due to increase in sputtering yield), while the flux of the reactive species (in this case, oxygen radicals and
ions) can be kept almost constant in the ICP sustained with the constant RF power. Especially, this feature of the independent and
fine control capability of the flux ratio is very significant for deposition of IGZO films, since the optimum deposition condition for
high-quality films can be obtained only in a very narrow window of flux ratio of the oxygen radicals and ions to the deposited atoms
(In, Ga, and Zn atoms) (88). Here it should be noted that independent control of the plasma density and the target current cannot
be attained in the conventional magnetron sputtering regimes (DC magnetron, RF magnetron, and pulsed magnetron) because the
variation of the target discharge power results in variations of the target current and the plasma density; both the target current and
the plasma density are raised with increasing target discharge power in the conventional regimes. With the new method described
above, a-IGZO films with mobility as high as 14 cm2 (Vs)1 have been successfully formed without substrate heating.
4.12.5 Plasma Source for Large-Area Process
Fabrications of FPDs are trended toward enlargement of substrate size (recently exceeding 3 m in the tenth generation: G10) and
high throughput, primarily due to enhancement of production efficiency and cost reduction. These trends require development of
meters-scale/ultralarge-area plasma reactors to realize uniform plasma process over ultralarge-area substrates, including a variety of
film deposition (microcrystalline silicon, a-IGZO, transparent conductive oxides, dielectric materials for passivation), etching, and
substrate cleaning.
Major specifications for plasma sources to meet these trends for large-area processing include (1) controllability of plasma
profile to ensure uniformity over large areas, (2) high-density plasma production for high throughput and enhanced processing
efficiency, (3) low plasma potential for high-quality processing with suppressed plasma damage, and (4) low-pressure pro-
cessing for enhanced utilization efficiency of materials–gas molecules, controllability of plasma and radical production, and
suppression of dust-particle formation (91). Among them, the specifications (2)–(4) are mainly dependent on the scheme of
www.iran-mavad.com
discharge (method of power coupling for plasma production) and/or the plasma excitation frequency, which are to be selected
according to the processing requirements. Therefore many efforts so far have been focused on the issues to scale up the
conventional sources, which have been well developed in silicon device technologies and a variety of coating technologies. In
other words, most of the large-area sources have been developed in some ways of scaling up the conventional sources toward
meters in size by enlargement of power coupling elements (capacitive electrodes, inductive antennas, microwave transmission
lines, and so on).
Plasma sources (1,10) developed so far for large-area processing include CCP, ICP, and SWP. However, considering design
issues for plasma production with high-frequency power sources to satisfy the requirements for enlargement of source size
exceeding a meter, power deposition profiles and hence the plasma distributions become inherently nonuniform especially due
to standing wave effects, which cannot be avoided with increasing source size when the source employs power coupling
elements with a scale length equivalent to or as long as the one-quarter wavelength of the high-frequency power transmission
(91–94).
In order to provide straightforward and promising solutions to overcome the problems described, a series of our works
(52,91,95–98) have been carried out on the basis of the plasma generation and profile-control technologies employing
multiple LIA modules, whose antenna size is substantially shorter than the propagation wavelength, to sustain meters-scale
ICP, whose concepts are schematically shown in Figure 7. For active control of power deposition profiles and hence the plasma
uniformity, Setsuhara and coworkers (91,95) developed plasma control technologies, in which each LIA module is driven by
independent RF power source and operation conditions (RF power, phase, frequency, pulse duration, and timing) are
controlled with an integrated controller. With this configuration, the RF power supplied to each of the LIA modules can be
controlled actively. Here it should be noted that the plasma generation scheme employing the LIA modules is considered to be
a unique solution that enables active control of the power deposition profiles even at the elevated excitation frequency without
any problems associated with the standing wave effects; i.e., the present scheme is the only state-of-the-art technology that
allows source operation and profile control at excitation frequency of 27 MHz or higher in the plasma reactors with a scale as
large as meters (95).
Examples of large-area process reactors are shown in Figure 16 for the reactors with (1) the conventional LIAs as shown in
Figure 7 and (2) the inner-type LIAs as shown in Figure 14. An excellent feature in the reactor with the inner-type LIAs is that the
distance between the top flange and the substrate can be shortened and the overall flat structure can be realized at the top flange.
Additionally, a plasma-density profile simulated for a reactor of 4200 mm 4200 mm is shown in Figure 17. The plasma-density
Figure 16 Schematic illustrations of large-area plasma reactor configurations with (a) conventional LIAs as shown in Figure 7 and (b) inner-type LIAs as
shown in Figure 14.
www.iran-mavad.com
Figure 17 Plasma density profile numerically simulated for a reactor of 4200 mm 4200 mm.
profile shown in Figure 17 is obtained from the plasma fluid simulation for the Ar plasma sustained at a pressure of 0.65 Pa with
multiple inner-type LIAs. The density profile implies that the eleventh-generation ultralarge glass substrate (3000 mm 3200 mm)
can be successfully processed with a plasma-density uniformity of 4.9%.
4.12.6 Summary
Plasma sources in thin film deposition, namely the plasma-enhanced CVD and PVD processes, have been reviewed in terms of
discharge characteristics and specific issues in thin film deposition applications. Typical regions of discharge conditions are
summarized in terms of gas pressure and frequency of the power source.
In the plasma-enhanced CVD processes, thin film deposition process depends significantly on (1) sticking probability of
molecules dissociated mainly via electron-impact dissociation in the plasma as precursor, (2) migration of the precursors affected by
the substrate temperature, and (3) chemical reactions of excited species (radicals and ions) on the growing film. Therefore, for
formation of good-quality films with high packing density, production of radical precursors with moderate sticking probability and
enhancement of the precursor migration are essential. For the plasma-enhanced CVD processes, the discharge characteristics and
properties have been reviewed for DC glow, DC arc, atmospheric pressure nonequilibrium plasmas (DBD and APG), microplasmas,
capacitively coupled RF plasmas, inductively coupled RF plasmas, and microwave plasmas (ECR and SWP).
In the plasma-enhanced PVD processes, importance of the ion bombardment during film growth has been described for
attainment of desired properties and characteristics of the deposited films. Thus the proper control of ion density (plasma density) is
one of the key issues for the plasma source. For the plasma-enhanced PVD processes, the discharge characteristics and properties
have been reviewed for cathodic arc and magnetron sputtering.
Specifications and requirements for the large-area plasma source have been briefly reviewed and examples are shown for the
large-area plasma production of ICPs sustained with LIAs.
References
1. Conrads, H.; Schmidt, M. Plasma Sources Sci. Technol. 2000, 9, 441–454.

2. CzerwiecU, T.; Renevier, N.; Michel, H. Surf. Coat. Technol. 2000, 131, 267–277.
3. Mayrhofer, P. H.; Mitterer, C.; Hultman, L.; Clemens, H. Prog. Mater. Sci. 2006, 51, 1032–1114.
4. Robertson, J. Surf. Coat. Technol. 1992, 50, 185–203.
5. Bhushan, B. Diam. Relat. Mater. 1999, 8, 1985–2015.
6. Kuo, Y. J. Electrochem. Soc. 1995, 142, 2486–2507.
7. Goetzbergera, A.; Heblinga, C.; Schock, H.-W. Mater. Sci. Eng. R 2003, 40, 1–46.
8. Chua, P. K.; Chena, J. Y.; Wanga, L. P.; Huang, N. Mater. Sci. Eng. R 2002, 36, 143–206.
9. Mogab, C. J.; Adams, A. C.; Flamm, D. L. J. Appl. Phys. 1978, 49, 3796.
10. Lieberman, M. A.; Lichtenberg, A. J. Principles of Plasma Discharges and Materials Processing; Wiley: New York, 1994.
11. Choi, M.-C.; Kim, Y.; Ha, C.-S. Prog. Polym. Sci. 2008, 33, 581.
12. International Technology Roadmap for Semiconductors, 2007 edition, Executive Summary.
13. Matsuda, A. Jpn. J. Appl. Phys. 2004, 43, 7909–7920.
14. Paschen, F. Wied. Ann. 1889, 37, 69–96.
www.iran-mavad.com
15. Torres, J.-M.; Dhariwal, R. S. Microsyst. Technol. 1999, 6, 6–10.

16. Schütze, A.; Jeong, J. Y.; Babayan, S. E.; Park, J.; Selwyn, G. S.; Hicks, R. F. IEEE Trans. Plasma Sci. 1998, 26, 1685–1694.
17. Elenbaas, W. The High Pressure Mercury Vapor Discharge; North-Holland: Amsterdam, 1951.
18. Kambara, M.; Yagi, H.; Sawayanagi, M.; Yoshida, T. J. Appl. Phys. 2006, 99, 074901-1–074901-7.
19. Yoshida, T. Pure Appl. Chem. 1994, 66, 1222–1230.
20. Yoshida, T. Mater. Trans. JIM 1990, 31, 1–11.
21. Chang, Y.; Young, R. M.; Pfender, E. Plasma Chem. Plasma Process. 1989, 9, 277–289.
22. Kubota, N.; Ayabe, M.; Fukuda, T.; Akashi, M. Mater. Res. Soc. 1997, 458, 449–452.
23. Cunha, A. G.; Orlando, M. T. D.; Emmerich, F. G.; Larica, C.; Baggio-Saitovitch, E. M. Physica C 1997, 282–287, 489–490.
24. Fincke, J. R.; Swank, W. D.; Haggard, D. C. Plasma Chem. Plasma Process. 1993, 13, 579–600.
25. Lu, Z. P.; Stachowicz, L.; Kong, P.; Heberlein, J.; Pfender, E. Plasma Chem. Plasma Process. 1991, 11, 387–394.
26. Lu, Z. P.; Heberlein, J.; Pfender, E. Plasma Chem. Plasma Process. 1992, 12, 35–53.
27. Olson, J. M.; Dawes, M. J. J. Mater. Res. 1996, 11, 1765–1775.
28. Baldwin, S. K., Jr.; Owano, T. G.; Kruger, C. H. Plasma Chem. Plasma Process. 1994, 14, 383–406.
29. Baldwin, S. K., Jr.; Owano, T. G.; Zhao, M.; Kruger, C. H. Diam. Relat. Mater. 1997, 6, 202–206.
30. Tachibana, K. IEEJ Trans. 2006, 1, 145–155.
31. Kogelschatz, U.; Eliasson, B.; Egli, W. Pure Appl. Chem. 1999, 71, 1819–1828.
32. Liu, D.; Yu, S.; Ma, T.; Song, Z.; Yang, X. Jpn. J. Appl. Phys. 2000, 39, 3359–3360.
33. Eliasson, B.; Kogelschatz, U. IEEE Trans. Plasma Sci. 1991, 19, 1063–1077; Kogelschatz, U. IEEE Trans. Plasma Sci. 2002, 30, 1400–1408.
34. Salge, J. Surf. Coat. Technol. 1996, 80, 1–7.
35. Nozaki, T.; Unno, Y.; Okazaki, K. Plasma Sources Sci. Technol. 2002, 11, 431–438.
36. Kogelschatz, U. Plasma Chem. Plasma Process. 2003, 23, 1.
37. Nozaki, T.; Kimura, Y.; Okazaki, K. J. Phys. D Appl. Phys. 2002, 35, 2779–2784.
38. Okazaki, S.; Kogoma, M.; Uehara, M.; Kimura, Y. J. Phys. D Appl. Phys. 1993, 26, 889–892.
39. Massines, F.; Rabehi, A.; Decomps, P.; Gadri, R. B.; Segur, P.; Mayoux, C. J. Appl. Phys. 1998, 83, 2950–2957.
40. Nozaki, T.; Okazaki, K. Plasma Process. Polym. 2008, 5, 300–321.
41. Koinuma, H.; Ohkubo, H.; Hashimoto, T.; Inomata, K.; Shiraishi, T.; Miyanaga, A.; Hayashi, S. Appl. Phys. Lett. 1992, 60, 816–817.
42. Babayan, S. E.; Jeong, J. Y.; Tu, V. J.; Park, J.; Selwyn, G. S.; Hicks, R. F. Plasma Sources Sci. Technol. 1998, 7, 286–288.
43. Inomata, K.; Ha, H.; Chaudhary, K. A.; Koinuma, H. Appl. Phys. Lett. 1994, 64, 46–48.
44. Ha, H.; Yoshimoto, M.; Koinuma, H.; Moon, B.; Ishiwara, H. Appl. Phys. Lett. 1996, 68, 2965–2967.
45. Schoenbach, K. H.; El-Habachi, A.; Shi, M.; Ciocca, M. Plasma Sources Sci. Technol. 1997, 6, 468.
46. Stark, R. H.; Schoenbach, K. H. J. Appl. Phys. 1999, 85, 2075.
47. Kedzierski, J.; Engemann, J.; Teschke, M.; Korzec, D. Solid State Phenom. 2005, 107, 119.
48. Yoshiki, H.; Ikeda, K.; Wakaki, A.; Togashi, S.; Taniguchi, K.; Horiike, Y. Jpn. J. Appl. Phys. 2003, 42, 4000.
49. Sladek, R. E. J.; Stoffels, E. J. Phys. D Appl. Phys. 2005, 38, 1716.
50. Ito, T.; Terashima, K. Appl. Phys. Lett. 2002, 80, 2648.
51. Hopwood, J. Plasma Sources Sci. Technol. 1992, 1, 109.
52. Setsuhara, Y.; Shoji, T.; Ebe, A.; Baba, S.; Yamamoto, N.; Takahashi, K.; Ono, K.; Miyake, S. Surf. Coat. Technol. 2003, 174–175, 33.
53. Setsuhara, Y.; Kamai, M.; Miyake, S.; Musil, J. Jpn. J. Appl. Phys. 1997, 36, 4568.
54. Yamashita, M.; Setsuhara, Y.; Miyake, S.; Kumagai, M.; Shoji, T.; Musil, J. Jpn. J. Appl. Phys. 1999, 38, 4291.
55. Takahashi, E.; Nishigami, Y.; Tomyo, A.; Fujiwara, M.; Kaki, H.; Kubota, K.; Hayashi, T.; Ogata, K.; Ebe, A.; Setsuhara, Y. Jpn. J. Appl. Phys. 2007, 46, 1280.
56. Moisan, M.; Baudry, C.; Leprince, P. IEEE Trans. Plasma Sci. 1975, PS-3, 1004.
57. Nagatsu, M.; Xu, G.; Yamage, M.; Kanoh, M.; Sugai, H. Jpn. J. Appl. Phys. 1996, 35. L341.
58. Ishijima, T.; Nojiri, Y.; Toyoda, H.; Sugai, H. Jpn. J. Appl. Phys. 2010, 49, 086002.
59. Smidt, F. A. Int. Mater. Rev. 1990, 35, 61.
60. Messier, R.; Giri, A. P.; Roy, R. A. J. Vac. Sci. Technol. A 1984, 2, 500.
61. Thornton, J. A. Annu. Rev. Mater. Sci. 1977, 7, 239.
62. Hofsäss, H.; Eyhusen, S.; Ronning, C. Diam. Relat. Mater. 2004, 13, 1103–1110.
63. Helmersson, U.; Lattemann, M.; Bohlmark, J.; Ehiasarian, A. P.; Gudmundsson, J. T. Thin Solid Films 2006, 513, 1–24.
64. Martin, P. J.; Bendavid, A. Thin Solid Films 2001, 394, 1–15.
65. Randhawa, H.; Johnson, P. C. Surf. Coat. Technol. 1987, 31, 303–318.
66. Randhawa, H. Thin Solid Films 1988, 167, 175–185.
67. Sanders, D. M.; Anders, A. Surf. Coat. Technol. 2000, 133–134, 78–90.
68. Anders, A. Vacuum 2002, 67, 673–686.
69. Hantzsche, E. IEEE Trans. Plasma Sci. 1995, 23, 893–898.
70. Musil, J. Vacuum 1998, 50, 363–372.
71. Musil, J. Surf. Coat. Technol. 1998, 100–101, 280–286.
72. Sproul, W. D.; Christie, D. J.; Carter, D. C. Thin Solid Films 2005, 491, 1–17.
73. Sigmund, P. Phys. Rev. E 1969, 184, 383–416.
74. Laegreid, N.; Wehner, G. K. J. Appl. Phys. 1961, 32, 365–369.
75. Kadlec, S.; Musil, J.; Vyskocil, J. Vacuum 1987, 37, 729–738.
76. Sproul, W. D. Surf. Coat. Technol. 1987, 33, 73–81.
77. Brauer, G.; Szczyrbowski, J.; Teschner, G. J. Non-Cryst. Solids 1997, 218, 19–24.
78. Schiller, S.; Goedicke, K.; Reschke, J.; Kirchhoff, V.; Schneider, S.; Milde, F. Surf. Coat. Technol. 1993, 61, 331–337.
79. Safi, I. Surf. Coat. Technol. 2000, 127, 203–219.
80. Musil, J.; Baroch, P.; Vlcek, J.; Nam, K. H.; Han, J. G. Thin Solid Films 2005, 475, 208–218.
81. Musil, J.; Baroch, P. IEEE Trans. Plasma Sci. 2005, 33, 338–339.
82. Kouznetsov, M.; Macak, K.; Schneider, J. M.; Helmersson, U.; Petrov, I. Surf. Coat. Technol. 1999, 122, 290.
83. Macak, K.; Kouznetsov, V.; Schneider, J.; Helmersson, U. J. Vac. Sci. Technol. A 2000, 18, 1533–1537.
84. Ehiasarian, A. P.; Hovsepian, P. Eh.; Hultman, L.; Helmersson, U. Thin Solid Films 2004, 457, 270–277.
85. Vlcek, J.; Kudlacek, P.; Burcalova, K.; Musil, J. J. Vac. Sci. Technol. A 2007, 25, 42–47.
86. Setsuhara, Y. J. Plasma Fusion Res. 2011, 87, 24–33 (in Japanese).
87. Nomura, K.; Ohta, H.; Takagi, A.; Kamiya, T.; Hirano, M.; Hosono, H. Nature 2004, 432, 488.
www.iran-mavad.com
88. Nomura, K.; Kamiya, T.; Ohta, H.; Hirano, M.; Hosono, H. Appl. Phys. Lett. 2008, 93, 192107.
89. Hosono, H.; Nomura, K.; Ogo, Y.; Uruga, T.; Kamiya, T. J. Non-Cryst. Solids 2008, 354, 2796.
90. Kamiya, T.; Nomura, K.; Hosono, H. Sci. Technol. Adv. Mater. 2010, 11, 044305.
91. Setsuhara, Y.; Takenaka, K.; Ebe, A.; Nishisaka, K. Plasma Process. Polym. 2007, 4, S628–S632.
92. Perrin, J.; Schmitt, J.; Hollenstein, C.; Howling, A.; Sansonnens, L. Plasma Phys. Controlled Fusion 2000, 42, B353.
93. Wu, Y.; Lieberman, M. A. Plasma Sources Sci. Technol. 2000, 9, 210.
94. Khater, M. H.; Overzet, L. J. Plasma Sources Sci. Technol. 2000, 9, 545.
95. Setsuhara, Y.; Takenaka, K.; Ebe, A. Surf. Coat. Technol. 2008, 202, 5225–5229.
96. Tsuda, O.; Ishihara, M.; Koga, Y.; Fujiwara, S.; Setsuhara, Y.; Sato, N. J. Phys. Chem. B 2005, 109, 4917.
97. Deguchi, H.; Yoneda, H.; Kato, K.; Kubota, K.; Hayashi, T.; Ogata, K.; Ebe, A.; Takenaka, K.; Setsuhara, Y. Jpn. J. Appl. Phys. 2006, 45, 8042.
98. Setsuhara, Y.; Takenaka, K.; Ebe, A.; Nishisaka, K. Solid State Phenom. 2007, 127, 239.
www.iran-mavad.com
4.13 Hard Nanocomposite Coatings
J Musil, P Zeman, and P Baroch, University of West Bohemia, Plzen, Czech Republic
4.13.1 Present State of the Knowledge in the Field of Nanocomposite Coatings 325
4.13.1.1 Origin of Enhanced Hardness in Nanostructured and Nanocomposite Coatings 325
4.13.1.2 Classification of Hard Nanocomposite Coatings 326
4.13.1.3 Enhanced Hardness Induced by Nanostructure 327
4.13.1.4 Transition Regions between Crystalline and Amorphous Materials 327
4.13.1.5 Energy and Its Effect on the Growth of Coatings 327
4.13.1.6 Formation of High-Temperature Phases at Low Temperatures 330
4.13.1.7 Limits of Hard Nanocomposite Coatings 331
4.13.2 Advanced Hard Nanocomposite Coatings 333
4.13.2.1 Nanocomposite Coatings with High Thermal Stability 333
4.13.2.1.1 Si–Zr–O Coatings Thermally Stable up to 1400 C 333
4.13.2.2 Coating Protecting Substrate against Oxidation above 1000 C 333
4.13.2.2.1 Hard Amorphous Nonoxide Nanocomposite Coatings 334
4.13.2.2.2 Hard Amorphous Oxide-Based Composite Coatings 337
4.13.2.3 Nanocomposite Coatings Composed of Nanograins Dispersed in Amorphous Matrix 339
4.13.2.3.1 Hard nc-MeC/a-C Composite Coatings with Low Friction m and Low Wear k 339
4.13.2.4 Hard Coatings Resistant to Cracking in Bending 341
4.13.2.4.1 Hard Al–Cu–O Coatings with Enhanced Resistance to Cracking 341
4.13.2.4.2 Hard Zr–Al–O Coatings with Enhanced Resistance to Cracking 343
4.13.2.4.3 Hard Al–O–N Coatings with Enhanced Resistance to Cracking 344
4.13.3 Trends of the Present and the Next Development of Hard Nanocomposite Coatings 345
4.13.3.1 Hard Coatings with Enhanced Toughness 345
4.13.3.2 Formation of Hard Nanocomposite Coatings on Unheated Substrates 346
4.13.3.3 Very-High-Rate Deposition of Hard Oxide Coatings 348
4.13.3.4 Summary 349
Acknowledgments 349
References 350
4.13.1 Present State of the Knowledge in the Field of Nanocomposite Coatings
Nanocomposite coatings represent a new generation of materials. They are composed of at least two separated phases with
a nanocrystalline (nc-) and/or amorphous (a-) structure. The nanocomposite materials, due to very small (10 nm) grains and
a significant role for boundary regions surrounding individual grains, exhibit enhanced or even completely new properties, and
also behave in a strongly different manner compared to the conventional materials composed of larger (100 nm) grains. This
difference in the behavior of the nanocomposite materials is caused by the increase of the ratio of the surface S of the grain to its
volume V (S/V > 0.1), resulting in (1) increasing dominance of the grain boundary regions with decreasing grain size d,
(2) reduced action of processes inside the grain volume (e.g., stopping the generation of dislocations), and (3) the promotion of
new processes such as grain boundary sliding and/or grain boundary enhancement due to an interatomic interaction between
atoms of neighboring grains. The dramatic change in the behavior of the nanocomposite materials compared to that of the
conventional materials is a result of (1) the strong change of the micro- and/or nanostructure of material, particularly the size d of
grains, the shape of crystallites, and the separation distance w between grains; and (2) the enhanced chemical bonding between
atoms of neighboring grains.
Enhanced or new and unique physical and functional properties of nanocomposite coatings have comprised a main driving
force stimulating their development. In recent years, great progress in the development of nanocomposite thin coatings and coatings
has been made ((1–70) and references therein). At present, there are a huge number of papers devoted to nanocomposite coatings,
and thus we can refer to only some of them; we apologize to all authors whose papers are not cited in this chapter due to its limited
extent. Many papers clearly demonstrated that nanocomposite coatings can exhibit different enhanced properties (mechanical,
electrical, optical, magnetic, photocatalytic, and others) compared to bulk materials and single-phase thin coatings and coatings.
4.13.1.1 Origin of Enhanced Hardness in Nanostructured and Nanocomposite Coatings

The hardness H of the nanocomposite coating can be more than two times greater than that of its harder component.
Main mechanisms, which are responsible for the hardness enhancement, are (1) compressive macrostress s generated in

www.iran-mavad.com
326 Hard Nanocomposite Coatings
Figure 1 Hardness H of the coating as a function of the size d of grains. Adapted after Musil, J.; Miyake, S. Nanocomposite Coatings with Enhanced
Hardness. In Novel Materials Processing (MAPEES’04); Miyake, S., Ed.; Elsevier Ltd.: Amsterdam, 2005; pp 345–356.
the coating during its growth, (2) the dislocation-induced plastic deformation, (3) the nanostructure of materials, (4) the
solid solution hardening, and (5) cohesive forces between atoms of neighboring grains. The compressive macrostress s can
be removed if deposition parameters used in the formation of the coating are correctly selected (54,64). It enables one to
investigate the enhancement of hardness H induced by the nanostructure. The dislocation-induced plastic deformation
dominates in the materials composed of large grains of d 10 nm, in which dislocations can be generated and can
propagate. On the contrary, the nanostructure is dominant in the materials composed of small grains of d 10 nm. In
cases when elements forming the material are soluble, one-phase nanostructured solid solution materials are formed.
However, in the case when the amount of the soluble element in the compound surpasses a solubility limit, two- or
multiphase nanocomposite materials are formed. Therefore, the hardness enhancement of coating strongly depends on the
grain size d; see Figure 1. This figure shows that there is a critical value of the grain size dc z 10 nm at which a maximum
value of the hardness Hmax of coating is achieved. This hardness is called the enhanced hardness. The region of grains
with d around dc corresponds to a continuous transition from the activity of the intragranular processes at d > dc,
dominated by the dislocations and described by the Hall–Petch law (H w d1/2) (71,72), to the activity of the inter-
granular processes at d < dc, dominated by interactions between atoms of neighboring grains and/or by the small-scale
sliding in grain boundaries.
4.13.1.2 Classification of Hard Nanocomposite Coatings

According to the hardness H, the nanocomposite coatings can be divided into three groups: (1) the hard coatings with H < 40 GPa,
(2) the superhard coatings with H ranging from w40 to w80 GPa, and (3) the ultrahard coatings with H > 80 GPa. It was found
that there are two groups of two-phase hard nanocomposite coatings, those composed of
1. Two hard phases called hard phase/hard phase composites (e.g., nc-Me1N/a-Me2N); or
2. One hard and one soft phase called hard phase/soft phase composites (e.g., nc-Me1N/Me3).
Here Me1 ¼ Ti, Zr, W, Ta, Cr, Mo, and Al are the elements forming hard nitrides, and Me2 ¼ Si, B, and so on and Me3 ¼ Cu, Ni,
Ag, Au, and Y are soft metals.
The possibility to form hard materials with enhanced hardness in both hard–hard and hard–soft composite systems clearly
indicates that the enhanced hardness of the nanocomposite coating is a result of the change of the geometry of the crystallites, and
particularly a dramatic reduction of the grain size d.
Very often, Me2 atoms can be dissolved into Me1 atoms, and the Me1(Me2) solid solution is formed. The solid solution is formed
as far as the amount of Me2 is smaller than that corresponding to its solubility limit of Me2 in Me1. Therefore, according to the
solubility limit Me2s (the maximum dissolution of Me2 atoms in Me1 atoms), two different materials can be formed in the system
Me1–Me2–N when both the Me1 and Me2 atoms form nitride: (1) the one-phase Me1(Me2)N solid solution nitride if Me2 < Me2s;
and (2) the two-phase nc-Me1(Me2)N–Me2N composite if Me2 > Me2s. Similarly, two different materials can be formed in the
system Me1–Me3–N when the Me1 atoms form nitride and the Me3 atoms do not create Me3N nitride: (1) the one-phase Me1(Me3)
N solution nitride if Me3 < Me3s; and (2) the two-phase nc-Me1(Me3)N–Me3 composite coatings if Me2 > Me2s (73). It means that
according to the amount of Me2 or Me3 atoms in the material, either the one-phase solid solution nitride or the two-phase
composite coating is formed. The process resulting in the enhanced hardness of the solid solution nitride is called the solid solution
hardening. The same situation takes place in the case when, instead of pure metallic Me2 atoms, Si, C, or other metalloid and
nonmetallic atoms are used (27,74–77).
www.iran-mavad.com
Hard Nanocomposite Coatings 327
Figure 2 Schematic illustration of four nanostructures of the nanocomposite coating with enhanced hardness: (a) Nanosized bilayers, (b) columnar
nanostructure, (c) nanograins surrounded by a tissue phase, and (d) mixture of nanograins with different crystallographic orientation. Reproduced from
Musil, J. Hard Nanocomposite Coatings: Thermal Stability, Oxidation Resistance and Toughness, Surf. Coat. Technol. 2012, 207, 50–65.
4.13.1.3 Enhanced Hardness Induced by Nanostructure

It was found that there are at least four types of nanostructures that result in the enhanced hardness of nanocomposite coatings: (1)
bilayers with nanosize period l, (2) the columnar nanostructure, (3) nanograins surrounded by very thin (w1 to 2 ML) tissue phase,
and (4) the mixture of nanograins with different crystallographic orientations and/or different phases; see Figure 2. Here
l ¼ h1 þ h2, where h1 and h2 are the thickness of the first and second layers in the bilayer, respectively, and ML denotes the
monolayer.
The nanocomposites composed of grains with d 10 nm can be divided into four groups:
1. Multilayers composed of nanosized bilayers.
2. Nanocomposites composed of nanocolumns or nanowalls.
3. Nanocomposites composed of nanograins surrounded by a tissue phase.
4. Nanocomposites composed of mixture of nanograins with different crystallographic orientation.
Individual nanostructures of nanocomposites are formed under different conditions using a sequential deposition of individual
layers in the nanosized bilayers, or a deposition in transition regions where the coating structure changes from crystalline through to
nanocrystalline and then amorphous. In the transition regions, nanostructured and nanocomposite coatings can be formed. In the
reactive sputtering of nanocomposite coatings, three transition regions – (1) the transition from the crystalline material to the X-ray
amorphous material; (2) the transition between two crystallographic orientations of grains of the same material; and (3) the
transition between two crystalline phases of different materials were found (54,58,64).
4.13.1.4 Transition Regions between Crystalline and Amorphous Materials

Nanostructured or nanocomposite A1xBxN nitride coatings can be created in three transition regions. These regions are sche-
matically displayed in Figure 3. The nanostructured coating is one-phase material and is created in the case when the element B
added in the AN nitride dissolves in the nitride. The nanocomposite coating is two-phase material and is created in the case when
(1) the amount of B element nB added to the AN nitride overpasses a solubility limit nBs; or (2) the element B is immiscible in the
AN nitride, and a second phase is formed. The transition from the crystalline to the amorphous phase, displayed in Figure 3(a), is formed
when nB > nBs, the BN nitride is amorphous, and the amount of the element B gradually increases and the BN nitride (amorphous
phase) dominates over the AN nitride (crystalline phase). A typical example is the Ti1xSixN nitride. The transition between the same
materials composed of grains of different crystallographic orientation is formed when the amount of nitrogen N incorporated in the
A1xBxNy is gradually increased from the A1xBx alloy to the A1xBxNy nitride; see Figure 3(b). Here, y ¼ N/A1xBx is the stoichi-
ometry of the nitride. Typical examples of these materials are TM(Me)Nx nitrides with TM ¼ Zr, Ti, and so on and a low amount
(<10 at.%) of added metal Me ¼ Ni, Fe (78–81). In this case, the enhancement of the coating hardness H takes place not only inside
the transition region but also at the edges of this region. The coatings with a strong preferred crystallographic orientation exhibit
a columnar microstructure with columns perpendicular to the coating–substrate interface; this microstructure of coating also results
in the enhancement of the coating hardness as shown in Figure 2. The transition between crystalline phases of two different
materials, displayed in Figure 3(c), is formed in the case when the amount of B element added to the A1xBxN nitride is gradually
increased and one crystalline nitride (AN) is converted to the second crystalline nitride (BN). Typical examples of these materials
are, for example, Ti1xAlxN coatings (82–88). More details are given in (54,58,64,90).
4.13.1.5 Energy and Its Effect on the Growth of Coatings

The addition of a selected element into a base material, as shown in this chapter, is one of many possible methods used to control
the structure of nanocomposite coatings. It is based on the stopping of grain growth in one phase by a segregation in the second
www.iran-mavad.com
Figure 3 Schematic illustration of three transition regions of A1xBxN compounds: (a) Transition from crystalline to amorphous phase; (b) transition
between grains of the same material, with two different preferred crystallographic orientations; and (c) transition between crystalline phases of two
different materials (e.g., Ti1xAlxN results in TiN nitride for x ¼ 0 and AlN nitride for x ¼ 1). Reproduced from Musil, J. Hard Nanocomposite Coatings:
Thermal Stability, Oxidation Resistance and Toughness, Surf. Coat. Technol. 2012, 207, 50–65.
Figure 4 Effect of amount of added element into one-phase coating according to Barna–Adamik’s structural zone model (91,92) and comparison with
the real microstructure of Zr–Cu–N composite coating with low (1.2 at.%) and high (20 at.%) amounts of Cu (28,93); here, Ts and Tm are the substrate
temperatures during coating deposition and the melting temperature of coating, respectively. Reproduced from Musil, J.; Vlcek, J. Magnetron
Sputtering of Hard Nanocomposite Coatings and Their Properties. Surf. Coat. Technol. 2001, 142, 557–566.
phase to grain boundaries in the formation of nanocrystalline composite materials, or is based on suppression of the formation of
the nanocrystalline phase at the expense of the dominant formation of the amorphous phase in the formation of the X-ray
amorphous composite materials. The effect of the amount of added element in the base material on its structure is illustrated in
Figure 4. According to the Barna–Adamik structural zone model (SZM), such composite materials can be formed only in the case
when a low-energy E, not sufficient to induce a nanocrystallization of the amorphous phase and/or a growth of the nanocrystals, is
delivered to the growing coating. Therefore, the effect of the energy E delivered to the growing coating on its structure needs to be
assessed.
The energy delivered to the growing coating has a crucial effect on its structure, elemental and phase composition, and physical
properties (81,88,94,95). The energy can be delivered by (1) the substrate heating Esh; (2) the conversion of the kinetic energy Ep of
bombarding ions (Ebi) and fast neutrals (Efn) incident on the surface of growing coating; (3) the heat evolved during formation of
the compound Ech (the energy released in exothermic chemical reactions); (4) the heating from the sputtered magnetron target Emt,
which is almost always not perfectly cooled; and (5) the radiation from the plasma Erad. The total energy ET delivered to the growing
coating can be expressed by the following formula (95):
ET ¼ Esh ðTs ; td Þ þ Ep ðUs ; is ; aD ; pT ; td Þ þ Ech ðtd ; Tch Þ þ Emt ðWd ; td ; dst Þ þ Erad ðtd Þ [1]
www.iran-mavad.com
where td is the time of the coating deposition; Tch is the temperature of the chemical reaction; pT ¼ pAr þ pRG is the total pressure of
the sputtering gas mixture; pAr and pRG are the partial pressures of argon and reactive gas, respectively; Wd ¼ UdId/S is the magnetron
target power density; Id and Ud are the magnetron current and voltage, respectively; S is the whole area of the magnetron target; and
ds–t is the substrate-to-target distance. The energy delivered to the growing coating by incident particles Ep is composed, as was
already given in this chapter, of two terms:
Ep ¼ Ebi þ Efn [2]
In the simplest case of a collision-less, fully ionized plasma when Efn ¼ 0, the energy Ep can be expressed in the following form:

Ep J cm3 ¼ Ebi ¼ Ei ðvi =vca Þ ¼ Us Up is =aD zðUs is Þ=aD [3]
where Ei is the energy of the ion; ni and nca are the flux of ions and condensing atoms, respectively; Us is the substrate bias; Up is the
plasma potential; is is the substrate ion current density; and aD is the deposition rate of the coating.
Equation [3] shows two important facts:
1. The energy Ebi delivered to the growing coating by bombarding ions can be easily calculated from measured deposition
parameters (Us, is), and the deposition rate aD ¼ h/td calculated from the measured coating thickness h and deposition
time td.
2. The energy Ebi strongly depends not only on Us and is but also on aD. This fact is of extraordinary importance, particularly in
(1) reactive sputtering of compounds such as nitrides, oxides, and other compounds; and (2) high-rate sputtering of coatings.
The decrease of aD in the reactive sputtering of the compounds results in the increase of the energy Ebi delivered to the growing
coating due to a slower growth of the coating that contributes to the nanocrystallization of amorphous coatings and to the
improvement of crystallinity of the coatings. On the contrary, the high-rate sputtering results in the decrease of the energy Ebi
delivered to the growing coating and slows down its crystallization.
In the case of higher sputtering pressures pT, the energy Ebi is lower than that defined by eqn [3] in consequence of collisions
between particles, and has to be determined from eqn [4]:

Ebi J cm3 zðUs is ÞexpðL=li Þ=aD [4]
where L is the thickness of the voltage sheath, and li is the ion mean free path for collisions leading to losses of the ion energy in the
sheath. The ion mean free path can be calculated from the Dalton law as li (cm) z 0.4/pT (Pa) (96). The thickness of the high-
voltage (Us >> Ufl) sheath can be calculated from the Child–Langmuir equation for the direct current (dc) sheath; here, Ufl is the
floating potential. The sheath thickness L can be calculated from the formulas given in (97). More details are given in
(54,81,88,94,95).
Generally, sputtering discharges are not fully ionized (in particular, the ionization of sputtered atoms is low) and contain
fast neutrals with energies ranges from several eV to several tens of eV. Therefore, at low (0.1 Pa) sputtering pressures, the
coating grows under a combined action of the energies delivered to the growing coating by both the bombarding ions and the
fast neutral atoms; according to the Dalton law, the mean free path lmfp z 40 mm at p ¼ 0.1 Pa. Hence, when the sputtering is
carried out at low pressures p 0.1 Pa and small substrate-to-target distances ds–t z 50 mm, the energy of sputtered atoms does
not decrease because there are no collisions with the background gas atoms and sufficiently energetic atoms arrive at the surface
of growing coating. This energy is identical to or even higher than the surface energy Esurf of atoms and thus also strongly
influences the coating growth (98); Esurf varies in the range from 1 to 8 eV for the majority of materials (99). At present, only
the first two terms in eqn [1] (i.e., Es (Ts) and Ep (Us, is, aD, pT, td)), are considered as main sources of the energy that influence
the coating growth. The third term is expected to be utilized in the development of new technological processes of formation of
new advanced materials. The last two terms in eqn [3] play important roles in the formation of coatings on heat-sensitive
substrates such as polymer foils, polycarbonate, and substrates for flexible electronics. The target power density must be limited
to a value at which the heat transferred from the target to the substrate does not cause the destruction of the heat-sensitive
substrate.
The effect of the ion energy Ei and the flux of ions ni on the coating microstructure is illustrated in Figure 5. This figure displays
the cross-section microstructure and the surface morphology of the TiN coatings reactively sputtered in a mixture of argon and
nitrogen at three values of (1) the substrate bias Us, and (2) the substrate ion current density is. Both the microstructure and the
surface morphology of the coating dramatically change with increasing Us and is. Already, small increases of Us and is are sufficient
for the conversion of a columnar microstructure to a dense featureless microstructure. The increase of Us results in gradual
smoothening of the coating surface. When the substrate bias Us is high (100 V in our case), the ion bombardment of growing
coating is strong and the growing coatings are partially resputtered. It results in slowing the growth of sputtered coating and its
smaller thickness h when the coating is deposited at the same deposition time. All these changes of the microstructure of the coating
also result in changes to its mechanical properties. Therefore, process parameters used in the coating deposition must be carefully
selected during the formation of the coating with required properties.
To conclude this section, it is worthwhile to note that hard low-stress sputtered coatings can be formed if low-energy ions
(negative biases (jUsj 50 V)) and high ion fluxes (is 1 mA cm2) are used (54,101). The substrate ion current density is must be
increased if the sputtering gas pressure increases and the energy Ei decreases due to ion collisions with atoms. The magnitude of ion
flux is limited by the substrate temperature Ts permitted in a given application.
www.iran-mavad.com
Figure 5 Effect of substrate bias Us and substrate ion current density is on the microstructure of 3 mm thick TiN coatings reactively sputtered at
pressure p ¼ 5 Pa and target-to-substrate distance ds–t ¼ 200 mm. Reproduced from Kadlec, S.; Musil, J.; Munz, W.-D.; Hakanson, G.; Sundgren, J.-E.
Reactive Deposition of TiN Films Using Unbalanced Magnetron. Surf. Coat. Technol. 1989, 39–40, 487–497.
4.13.1.6 Formation of High-Temperature Phases at Low Temperatures

The structure of sputtered coatings can be strongly influenced not only by the temperature T under which the coating is formed but
also by the (1) heating rate (i.e., the rate of temperature increase from RT to T) and (2) cooling rate (i.e., the rate of temperature
decrease from T to RT). According to the phase diagrams, different phases can be formed at different temperature T, and, moreover,
the phase formed at the highest T is most stable. As example, the binary phase diagram of the Ti–Fe alloy is given in Figure 6. This
diagram describes the formation of a Ti–Fe alloy by the equilibrium formation process (i.e., by a slow heating from RT to a given
temperature T followed by a slow cooling of the alloy from T to RT). Figure 6 shows that at first, a low-T (aTi) phase is formed when
T increases. The high-T (bTi) phase can be formed at high temperatures T 882 C only. However, the high-T (bTi) phase is
converted during a slow cooling backward to the low-T (aTi) phase. This conversion takes place also in the case of a Ti–Fe alloy with
22 at.% Fe.
A completely different situation takes place in the case when nonequilibrium formation processes that allow rapid heating and
rapid cooling are used. In this case, the Ti coating with a c-bTi high-T phase can be formed on an unheated substrate. The high-T
(bTi) phase can be created if (1) a small (<22 at.%) amount of Fe is added to Ti; (2) the temperature T is rapidly increased to values
corresponding to the region where there is a high-T (bTi) phase, shown by the red region in Figure 6; and (3) the created high-T
(bTi) phase is very rapidly cooled down to RT. These requirements are fulfilled in the sputter deposition of the coating performed at
the atomic level. In sputtering, the coating is formed by the condensation of atoms with the energy several eV. This energy dissipates
in a place of the atom incident and easily increases the local temperature in this place to a value necessary to form the high-T
c-bTi(Fe) phase. Moreover, a very fast dissipation of this energy in close vicinity of every incident atom results in (1) extremely fast
cooling (1013 K s1 (103–105)) and (2) freezing of the high-T c-bTi(Fe) phase in the sputtered Ti–Fe alloy coating. This extremely
fast cooling of the coating material results in an extremely short time t that is available and thus is insufficient to form metastable
phases during cooling from high T to RT. Therefore, the Ti(Fe) alloy coating with the high-T c-bTi(Fe) phase can be formed on an
unheated substrate. The process of freezing the high-T phase is schematically illustrated in Figure 7. A pure high-T phase creates at
cooling times tcool t0 only; t0 is the longest time allowing the formation of the high-T phase on an unheated substrate. Materials
created at slower cooling rates (t > t0) are mixtures of the high-T phase and the low-T phase. The amount of the low-T phase in this
mixture increases with increasing cooling time tcool.
www.iran-mavad.com
Figure 6 Binary phase diagram of a Ti–Fe alloy (102). Adapted after In Smithells Metal Reference Book; Brandes, E. A., Ed., 6th ed., Butterworths: London;
pp 8–24.
Figure 7 Schematic illustration of freezing of the high-T phase in coating as a function of cooling time tcool.
The possibility of the formation of alloy coatings with the high-T phase on unheated substrates is demonstrated on sputtering of
the Ti–Fe alloy coating with the high-T c-bTi(Fe) phase; see Figure 8. Pure Ti coatings with a high-T c-bTi phase are not formed on
the unheated substrate by sputtering. The pure Ti coatings with a low-T h-aTi phase are sputtered only; see Figure 8(a). However,
a small addition of Fe allows one to produce the Ti(Fe) coating with a high-T c-bTi(Fe) phase. It is an important finding. More
details are given in (106).
4.13.1.7 Limits of Hard Nanocomposite Coatings

Until recently, the main attention in the investigation of hard nanocomposite coatings concentrated on the hardness of materials
(coatings), the methods of hardness enhancement, and the achievement of hardness approaching or even exceeding that of dia-
mond. Hard nanocomposite coatings based on nitrides, particularly the composites of the type nc-MeN/a-Si3N4 with low
(10 at.%) Si content, were successfully developed. These nanocomposites exhibit the enhanced hardness, achieving up to
50–70 GPa, but none of them exhibited hardness H approaching that of diamond. The ultrahardness of nc-TiN/a- and nc-TiSi2
nanocomposites with H ¼ 80 to 105 GPa reported by Veprek et al. (17,24) was most probably incorrectly measured (89,194).
www.iran-mavad.com
Figure 8 XRD pattern from the coating sputtered from a (a) pure Ti (99.5) target and (b) TiFe (90 wt % Ti, 10 wt % Fe) target on an unheated glass
substrate. Reproduced from Musil, J.; Bell, A. J.; Vlcek, J.; Hurkmans, T. Formation of High Temperature Phases in Sputter Deposited Ti-Based Films
Below 100 C. J. Vac. Sci. Technol. 1996, A14 (4), 2247–2250.
Despite a huge effort in many labs around the world, no samples with such high hardness have been prepared so far. It means that
diamond still remains the hardest material.
Simultaneously, it is well known that hard materials are brittle and their brittleness increases with increasing hardness. The high
brittleness of hard coatings strongly limits their practical utilization. This concerns mainly hard ceramic materials based on nitrides,
oxides, carbides, borides, other compounds and their combinations, which are widely used in many applications as either protective
or functional coatings. This is the main reason why now many labs around the world try to develop new advanced hard ceramic
coatings with enhanced toughness that exhibit enhanced resistance to cracking and, simultaneously, sufficiently high (20 GPa)
hardness (47,50,60,61,64,90). In spite of the fact that hardness is one of the most important mechanical properties of the material, it
is insufficient to select the material for a given application according to its hardness alone. The material must exhibit not only a high
hardness H but also a sufficient toughness because the coating toughness can be for many applications more important than its
hardness. This is why it is vitally important to master the formation of hard coatings with high toughness. Hard coatings with high
toughness represent a new class of advanced nanocomposite coatings; see Figure 9.
Figure 9 Classification of nanocomposites according to their hardness and toughness. Adapted after Musil, J.; Vlcek, J. Magnetron Sputtering of Alloy
Based Films and Its Specificity. Czech. J. Phys. 1998, B48, 1209–1224.
www.iran-mavad.com
New advanced hard nanocomposite coatings not only should be hard and tough, but also should exhibit high thermal stability
and oxidation resistance, low friction, high wear resistance, high resistance to cracking, high corrosion resistance, and other
functions important for many applications. Therefore, Section 4.13.2 is devoted to advanced hard nanocomposite coatings. Special
attention is devoted to the thermal stability and oxidation resistance of hard nanocomposite coatings at temperatures above
1000 C, and to their tribological properties and resistance to cracking in bending.
4.13.2 Advanced Hard Nanocomposite Coatings

4.13.2.1 Nanocomposite Coatings with High Thermal Stability
Unique properties of nanostructured coatings are due to their nanostructure. However, the nanostructure is a metastable phase.
It means that in the case when the temperature T, under which the coating is operated, achieves or exceeds the temperature
Tnc stab corresponding to the thermal stability of the nanostructure, the material of the coating starts to convert to the
conventional crystalline material due to grain growth. The grain growth results in destruction of the coating nanostructure and
in the formation of new crystalline phases (25,39,52,56,59,62,65,70,101,107). This is the reason why nanocomposite coatings
lose their unique properties at temperatures T > Tnc stab. These changes result in changes of the coating properties. Therefore,
the thermal stability of the nanocomposite coating is usually determined by the stability of its properties with increasing
temperature T. In summary, we can say that the thermal stability of nanocomposite coatings is of key importance for many
applications.
The thermal stability of the hard nanocomposite coatings produced so far is, however, relatively low, being less than 1000 C. It
is a reason why these nanocomposite coatings cannot be used in many new advanced applications such as high-speed cutting (108),
thermal protection of some mechanical parts (turbine blades, etc.), rockets and space vehicles for orbital flights (109), and others.
Therefore, it is vitally important to develop new hard nanocomposite materials that are thermally stable at temperatures T above
1000 C. In principle, two kinds of materials meet this requirement: (1) materials with a thermally stable crystalline structure, and
(2) amorphous materials thermally stable above 1000 C.
4.13.2.1.1 Si–Zr–O Coatings Thermally Stable up to 1400 C

Thermal annealing is a process at which the temperature T of the part coated by a coating is at first increased from RT to a maximal
(operation) temperature Tmax, and later, after finishing the operation (e.g., cutting), T is decreased back to RT. No change of the
coating properties is required during this process. It is possible only in the case when Tmax is smaller than the temperature Tnc stab
corresponding to the thermal stability of the nanostructure (i.e., Tmax < Tnc stab and/or no change in the crystalline structure takes
place). This fact is illustrated in the evolution of X-ray diffraction structure of the 7 mm thick, amorphous a-(Si–Zr–O) coating
containing a small (5 at.%) amount of Zr with increasing maximum annealing temperature Ta max; see Figure 10.
The as-deposited Si31Zr5O64 coating, prepared by magnetron sputtering at the substrate temperature Ts ¼ 500 C and the
substrate held on a floating potential Ufl (i.e., at the substrate bias Us ¼ Ufl), is X-ray amorphous. After deposition, the Si31Zr5O64
coating was thermally annealed in the flowing air to different selected maximum temperatures Ta max with 10 C min1 heating rate
and 30 C min1 cooling rate; after reaching Ta max, the coating was immediately cooled to RT. The Si31Zr5O64 coating
was deposited on the Si(100) substrate for annealing at Ta max 1300 C and on the sapphire substrate for annealing at
Ta max > 1300 C. More details are given in (68).
From Figure 10, the following conclusions can be drawn: (1) the nanocrystallization of a-Si31Zr5O64 coating starts at T > 900 C;
(2) at first, the ZrO2 phase crystallizes and small t-ZrO2 grains occur at Ta max z 1000 C; (3) small roughness of thermally annealed
Si31Zr5O64 coating and the superposition of low-intensity reflections from t-ZrO2 grains on the broad X-ray peak of the amorphous
a-SiO2 phase indicate that the t-ZrO2 grains are embedded in the a-SiO2 matrix; (4) the increase of intensity and the decrease of width
of X-ray reflections from t-ZrO2 grains with increasing Ta max indicate the growth of ZrO2 grains; (5) the t-ZrO2 phase is thermally stable
(there is no conversion to m-ZrO2) in a wide range of Ta max from RT up to w1500 C; (6) the growth of intensities of the X-ray
reflections from t-ZrO2 grains stops at w1500 C due to (a) the strong interaction between the SiO2 phase of coating and the Al2O3
substrate and (b) the formation of new phases; and (7) a mixture of t-ZrO2 þ m-ZrO2 þ Al6Si2O13 þ SiO2 phases is formed at
Ta max > 1500 C, and it results in a strong increase in the surface roughness Ra of coating (here, Al6Si2O13 ¼ 3Al2O3$2SiO2 is the
mullite phase).
Besides, it was found that the mechanical properties (i.e., the microhardness H and the effective Young’s modulus E* ¼ E/
(1 n2) of the Si31Zr5O64 coating) remain almost constant during thermal annealing from RT to 1400 C to RT with the annealing
temperature Ta > 1000 C for 4 h in the flowing air because of a perfect stability of the t-ZrO2 phase in the whole temperature range
(RT to 1400 C) after its creation at Ta z 1000 C (68). This result is of key importance because it demonstrates that the mechanical
properties of the hard protective coating do not change during thermal annealing as far its structure remains unchanged. This conclusion is of
general validity.
4.13.2.2 Coating Protecting Substrate against Oxidation above 1000 C

The protection of the substrate against oxidation requires completely avoiding a direct connection of the external atmosphere with
the substrate surface. It cannot be achieved with crystalline coatings, as shown in Figure 11(a). Only amorphous coatings and those
www.iran-mavad.com
Figure 10 Evolution of X-ray diffraction structure and surface roughness Ra of the 7 mm thick Si31Zr5O64 coating with Ta max increasing

from 900 to 1700 C. Reproduced from Musil, J.; Satava,
V.; Zeman, P.; Cerstvý, R. Protective Zr-Containing SiO2 Coatings Resistant to Thermal Cycling in
Air up to 1400 C. Surf. Coat. Technol. 2009, 203, 1502–1507.
Figure 11 Forms of substrate protection against oxidation from the external atmosphere using (a) crystalline coating, (b) amorphous coating,
and (c) composite coating composed of well-separated grains embedded in an amorphous matrix; d is the grain size, w is the distance between grains,
and ML denotes the monolayer. Reproduced from Musil, J., Hard Nanocomposite Coatings: Thermal Stability, Oxidation Resistance and Toughness.
Surf. Coat. Technol. 2012, 207, 50–65.
in which well-separated grains are embedded in an amorphous matrix ensure a perfect protection of the substrate against oxidation;
see Figures 11(b) and 11(c). Therefore, amorphous coatings thermally stable above 1000 C needs to be developed. Both hard
amorphous non-oxide-based and oxide-based nanocomposite coatings thermally stable above 1000 C are briefly described in this
chapter.
4.13.2.2.1 Hard Amorphous Nonoxide Nanocomposite Coatings

Recently, two new classes of hard amorphous nanocomposite coatings with thermal stability above 1000 C were developed.
www.iran-mavad.com
1. a-(Si3N4/MeNx) composite coatings with high (50 vol.%) content of the a-Si3N4 phase; here, Me ¼ Ta, Zr, Ti, Al, Mo, W,
and so on and x ¼ N/Me is the stoichiometry of nitride (59,65,110–117).
2. a-(Si–B–C–N) coatings with strong covalent bonds (52,118–120).
These amorphous coatings were reactively sputtered using an unbalanced magnetron. They exhibit (1) high hardness ranging
from w20 to w40 GPa; and (2) excellent oxidation resistance even at temperatures considerably exceeding 1000 C. Physical and
mechanical properties, and the resistance to oxidation of these amorphous coatings, strongly depend on their thermal stability. As
was given in this chapter, the thermal stability of the amorphous coating depends on three factors: (1) a thermal stability of
individual components (phases) from which the amorphous coating is composed, (2) a composition of annealing atmosphere, and
(3) an interaction of the coating with the substrate during annealing. The effect of these three factors on the coating stability is
further illustrated in the example of the Me-Si–N coating during its thermal annealing.
4.13.2.2.1.1 Thermal stability of a-(Si3N4/MeNx) composite coatings

1. Effect of stoichiometry of metal nitride MeNx
The hard amorphous a-(Si3N4/MeNx) composite coating is composed of two phases: (1) the MeNx metal nitride phase
and (2) the Si3N4 phase. Generally, the stoichiometry x of the metal nitride can be continuously varied from sub-
stoichiometric (x < 1) through stoichiometric (x ¼ 1) to overstoichiometric (x > 1) with an increasing amount of N incor-
porated into the coating. The Si3N4 phase is more stable than the MeNx phase, which is in good agreement with heats of their
formation: DHSi3 N4 ¼ 745:1 kJ mol1 , DHZrN ¼ 365.5 kJ mol1, DHTiN ¼ 336.6 kJ mol1, DHTaN ¼ 252.4 kJ mol1, and
so on (101). The crystallization temperature depends, however, on the stoichiometry x of the MeNx phase. It is illustrated in
Figure 12, where differential scanning calorimetry (DSC) curves of as-deposited a-(Si–Zr–N) coatings containing the sub-
stoichiometric (x < 1) and overstoichiometric (x > 1) ZrNx nitride are displayed. The crystallization temperature Tcr was
determined from the exothermic peak (increase of heat flow) on the DSC curve and is defined as the temperature corre-
sponding to the maximum of the DSC peak. The amorphous a-(Si3N4/ZrNx¼0.8) coating with substoichiometric (x < 1)
zirconium nitride crystallizes at lower (Tcr z 1175 C) temperature T compared to the amorphous a-(Si3N4/ZrNx¼1.2) coating
with overstoichiometric (x > 1) zirconium nitride (Tcr z 1560 C). It means that the a-(Si–Zr–N) composite coating with
overstoichiometric ZrNx>1 nitride is thermally more stable than the a-(Si–Zr–N) composite coating with substoichiometric
ZrNx<1 nitride. The main finding of this investigation is the fact that the as-deposited amorphous a-(Si3N4/ZrNx>1) with
overstoichiometric ZrNx>1 and high (>50 vol.%) content of the Si3N4 phase is thermally stable up to w1500 C. More details
are given in (59,65).
2. Effect of composition of annealing atmosphere
The crystallization of the a-(Si3N4/MeNx) coating sputtered on Si(100) substrate can be also influenced by elements incorpo-
rated in it during annealing from the annealing atmosphere. It is illustrated in Figure 13 in which XRD patterns of the same
a-(Zr15Si25N60) coating annealed in flowing argon and vacuum (102 Pa) are displayed. The coatings were annealed at 1100 C for
30 min with heating and cooling rates of 10 and 30 K min1, respectively. Almost no crystallization of coating takes place in the case
of annealing in argon. On the contrary, the coating annealed in vacuum crystallizes. It is due to an interaction of oxygen from
residual atmosphere with the coating. It means that already a small amount of O2, which incorporates into the coating from the
residual atmosphere (vacuum with p0 z 102 Pa), is already sufficient to reduce the crystallization temperature Tcr of the coating
material. More details are given in (59).
Figure 12 Determination of crystallization temperature Tcr of (a) an a-(Si/ZrNx¼0.8) coating with substoichiometric (x < 1) zirconium nitride; and (b) an
a-(Si/ZrNx¼1.2) coating with overstoichiometric (x > 1) zirconium nitride during thermal annealing (Ta) in flowing Ar using DSC. Reproduced from Musil,
J.; Zeman, P. Hard Amorphous a-Si3N4/MeNx Nanocomposite Coatings with High Thermal Stability and High Oxidation Resistance. Solid State Phenom.
2007, 127, 31–36.
www.iran-mavad.com
Figure 13 XRD patterns of a-(Zr15Si) coating thermally annealed at temperature Ta ¼ 1100 C for 30 min in (a) flowing argon and (b) vacuum (102 Pa).
Adapted after Musil, J.; Zeman, P. Hard Amorphous a-Si3N4/MeNx Nanocomposite Coatings with High Thermal Stability and High Oxidation Resistance.
Solid State Phenom. 2007, 127, 31–36.
3. Effect of interaction of coating with substrate

The thermal stability of the a-(Si3N4/MeNx) coating can be also influenced by an interaction of the coating with the substrate.
That is illustrated in Figure 14 in which XRD patterns of the a-(Ti11Si26N63) coating on Al2O3 (sapphire) substrate (1) in the
as-deposited state at Ts ¼ 500 C, and (b) after thermal annealing in flowing air to Ta ¼ 1700 C are compared. The XRD pattern of
the as-deposited coating contains only two Al2O3 (006) reflections from the sapphire substrate. On the contrary, the XRD pattern of
the annealed coating exhibits, besides reflections from the substrate, a reflection from the Al6Si2O13 (¼3Al2O3 þ 2SiO2) mullite
phase. The formation of the mullite phase is clear evidence of a strong interaction of the Al2O3 substrate with SiO2 from the oxidized
a-(Ti11Si26N63) coating leading finally to the coating’s destruction. To suppress the interaction between the coating and the
substrate, a barrier interlayer must be deposited.
4.13.2.2.1.2 Oxidation of a-(Si3N4/MeNx) composite coatings

The oxidation of the a-(Si3N4/MeNx) composite coating is a two-step process. At first, the MeNx metal nitride phase is decomposed
during heating or thermal annealing at a temperature Tdec satisfying the eqn [5]; Tdec is the temperature at which the MeNx nitride is
decomposed. For simplicity, we assume that the MeNx¼1 nitride is the stoichiometric compound.
MeN/Me1 N1 þ Me2 þ Ng [5]
Here, Me ¼ Me þ Me , N ¼ N þ Ng, and the index g denotes the gas phase. The nitrogen gas Ng escapes from the coating, and
1 2 1
free Me2 metal atoms are formed. Then, the free Me2 atoms are oxidized by oxygen contained in the oxidation atmosphere, and
Me2O2 oxide is formed according to the following reaction:
Me2 þ O2 /Me2 O2 [6]
This is why the a-(Si3N4/MeNx) coating is covered by a thin oxide layer when it is thermally annealed in air. It is, however,
worthwhile to note that (1) some metals form solid oxides (e.g., ZrO2, TiO2, and Ta2O5), and other metals form volatile oxides (e.g.,
WOx and MoOx); and (2) the microstructure and surface morphology of the solid and volatile oxides strongly differs. It means that
the microstructure and surface morphology of an a-(Si3N4/MeNx) coating operated in air at high temperatures strongly depend on
the selection of Me.
Figure 14 XRD pattern of a-(Ti11Si) coating sputtered on Al2O3 (sapphire) substrate (a) in an as-deposited state at Ts ¼ 500 C, and (b) after thermal

annealing in flowing air to Ta ¼ 1700 C. Reproduced from Musil, J.; Zeman, P.; Dohnal, P.; Cerstvý, R. Ti-Si-N Nanocomposite Films with a High Content
of Si. Plasma Process. Polym. 2007, 4, S574–S578.
www.iran-mavad.com
Figure 15 Scanning electron microscopy (SEM) cross-section images and atomic force microscopy (AFM) images of the surface morphology of
amorphous (a) Si–Ta–N and (b) Si–W–N coatings deposited on Si(100) substrates after thermal annealing in flowing air to Ta ¼ 1300 C. Adapted after
Musil, J.; Baroch, P.; Zeman, P. Hard Nanocomposite Coatings: Present Status and Trends. In Plasma Surface Engineering Research and Its Practical
Applications; Wei, R., Ed.; Research Signpost Publisher: Kerala, India, 2008; pp 1–34.
4.13.2.2.1.3 Microstructure and surface morphology of an a-(Si3N4/MeNx) composite coatings after thermal annealing in flowing air
The as-deposited a-(Si–Me–N) composite coating with the metal Me forming a solid MeOx oxide is covered with a thin continuous
oxide layer during thermal annealing in air. This oxide layer acts as a barrier for penetration of the oxygen from the external
atmosphere into the bulk of coating. The bulk of the coating remains amorphous because of a high thermal stability of the a-Si3N4
phase. On the contrary, the Si–Me–N coatings with the metal Me forming a volatile MeOx oxide exhibit a porous microstructure that
expands through the whole coating from its surface up to the coating–substrate interface if the annealing temperature Ta is suffi-
ciently high and the annealing time sufficiently long. The MeOx created at the surface of Si–Me–N composite coating also strongly
influences its surface morphology. The coating containing the solid oxide exhibits a smooth surface. On the contrary, the coating in
which the free metal Me is converted into a volatile oxide escaping from the coating exhibits a rough surface. Typical representatives
of these two qualitatively different composite coatings are a-(Si–Ta–N) and a-(Si–W–N) coatings with high content of the Si3N4
phase. The cross-sections and surface morphology of a-(Si–Ta–N) and a-(Si–W–N) coatings after their annealing in flowing air are
displayed in Figure 15(a) and 15(b), respectively.
During thermal annealing, the temperature Ta was increased from RT to 1300 C and then immediately decreased to RT, with the
heating and cooling rates between 10 and 30 K min1 respectively; no isothermal heating at Ta ¼ 1300 C was carried out. More
details are given in (64,65,110,113).
4.13.2.2.1.4 The present state of the art in oxidation resistance of hard nonoxide coatings based on nitrides
The present status in the oxidation resistance of hard coatings is summarized in Figure 16. Here, the mass gain Dm of the coating as
a function of the annealing temperature Ta is displayed. The temperature Ta corresponding to the sharp increase of the coating mass
Dm is defined as a maximum temperature Tmax that still avoids the oxidation of the coating. The oxidation resistance is higher as
Tmax becomes higher. For all these coatings, the oxidation resistance of crystalline coatings is lower than 1000 C (curves in blue).
This fact is not surprising because the coatings composed of grains always allow fast interdiffusion of oxygen from an external
atmosphere and elements of the coating via grain boundaries. This phenomenon dramatically decreases the oxidation resistance of
the bulk of coating and thus its barrier action. Some improvement of the oxidation resistance could be expected to be achieved if an
intergranular glassy phase is used. A strong improvement of oxidation resistance is, however, achieved if protective coatings are
X-ray amorphous (curves in red). At present, the X-ray amorphous coatings thermally stable above 1000 C exhibit the highest high-T
oxidation resistance.
4.13.2.2.2 Hard Amorphous Oxide-Based Composite Coatings

The pure amorphous oxide coatings are relatively soft materials with hardness H of about 10 GPa only, and they easily crack.
Therefore, these coatings are hardened by hard grains embedded into a soft amorphous matrix. These nc-Me1Ox/a-Me2Oy nano-
composite coatings exhibit up to w2 times higher hardness compared with amorphous oxide coatings and moreover have an
www.iran-mavad.com
Figure 16 Oxidation resistance of selected hard binary, ternary, and quaternary nitrides and hard amorphous a-(Si/MeNx) and a-(Si,B,C) composite
nitrides represented as Dm ¼ f(Ta); here, Dm is the increase in mass of coating due to its oxidation. Adapted after Musil, J.; Zeman, P. Hard Amorphous a-
Si3N4/MeNx Nanocomposite Coatings with High Thermal Stability and High Oxidation Resistance. Solid State Phenom. 2007, 127, 31–36.
enhanced resistance to cracking; Me1 and Me2 can be either different or the same metals (e.g., nc-(g-Al2O3)/a-Al2O3). Both
properties are of key importance when the nc-oxide/a-oxide composite coatings are used as protective coatings. Also, it is worth-
while to note that mechanical properties of the nc-Me1Ox/a-Me2Oy nanocomposite coating do not change during thermal annealing
as long as the structure of grains and the ratio of both phases (nc-Me1Ox and a-Me2Oy) in the coating do not change. It was
demonstrated for the nc-ZrO2/a-SiO2 nanocomposite coating; see Section 4.13.2.2 and (68).
4.13.2.2.2.1 Oxidation of amorphous oxide-based composite coatings

The oxide-based composite coatings composed of stoichiometric oxides do not oxidize in the presence of oxygen because the
material is already fully oxidized. Therefore, there is no increase of the mass Dm of such coatings during thermal annealing as
long as their structure remains amorphous. On the contrary, the under-stoichiometric oxide coating oxidizes, its mass increases
during thermal annealing in the presence of oxygen, and it results in a change of its properties. The properties of the stoichio-
metric amorphous oxides also change in the case when the amorphous material starts to crystallize. Therefore, it is of key
importance to show in what temperature range the properties of amorphous oxide-based composite coatings can be thermally
stable.
Similarly, as in the case of amorphous nitride-based coatings, a reaction of the coating elements with the substrate elements must
also be avoided in the case of amorphous oxide-based composite coatings. Once the annealing temperature Ta achieves the value at
which elements from the substrate diffuse in the coating and mutually react, the phase composition of the coating is changed
because new phases are created. It immediately results in a change of the mechanical properties of the coating, cracking occurs due to
different densities of new phases, and the protective function of the coating is lost. It means that the protection of the substrate
against oxidation is determined by the following processes: (1) an interaction of the coating with the substrate, (2) a formation of
new phases with different densities, and (3) a decrease of the crystallization temperature of the coating. The last two processes result
in connection of the external oxidation atmosphere with the surface of the substrate and its oxidation via cracks and/or grain
boundary regions.
The effect of the elemental composition of the substrate is illustrated in Figure 17. Here, the oxidation protection of the Si and
a-Al2O3 substrates by the Al31Ti7O62 coating is compared. The sputtered Al31Ti7O62 coating is the nc-(g-Al2O3)/a-(Al–Ti–O)
composite. During thermal annealing, the nc-(g-Al2O3)/a-(Al–Ti–O) composite transforms into a mixture of rutile TiO2 þ a-Al2O3
(121). This structural transformation results in cracking of the Al31Ti7O62 coating and the penetration of the external oxidation
atmosphere to the surface of the substrate. It results in oxidation of the Si surface and the increase of mass Dm, as shown in
Figure 17. No mass increase (Dm ¼ 0) occurs in the case when the Al31Ti7O62 coating is deposited on a sintered polycrystalline
corundum Al2O3 substrate because the substrate is already fully oxidized.
In summary, it is possible to conclude that for a reliable operation of the oxide coatings at high temperatures, it is necessary
(1) to avoid cracking of the coating and/or (2) to separate the coating from the substrate by an interlayer that prevents their mutual
interaction.
www.iran-mavad.com
Figure 17 Oxidation of the 745 nm thick Al31Ti7O64 composite coating sputtered on (1) Si(100) and (2) a-Al2O3 substrate as a function of maximum
annealing temperature Ta. Adapted after Musil, J.; Satava, V.; Cerstvy, R.; Zeman, P.; Tolg, T. Formation of Crystalline Al-Ti-O Thin Films and Their
Properties. Surf. Coat. Technol. 202 (208), 6064–6069.
4.13.2.3 Nanocomposite Coatings Composed of Nanograins Dispersed in Amorphous Matrix

DNG/AM nanocomposite coatings are composed of a small amount of nanograins dispersed in an amorphous matrix, as shown in
Figure 11(c). They can be formed in two ways: (1) by the dispersion of nanograins (one phase) in the amorphous matrix (second
phase) during the coating deposition (e.g., by the cosputtering of two elements in a mixture of argon and reactive gas), or (2) by the
postdeposition nanocrystallization of the amorphous material. A correct selection of (1) the elements from which the coating is
formed; and (2) the deposition conditions, particularly the substrate temperature Ts and the energy E delivered to the growing
coating by bombarding ions and/or condensing atoms, is needed to form suitable nanograins. Namely, different compounds (e.g.,
TiNx, ZrO2, TiC, SiO2, Al2O3, and Si3N4) differ in values of the crystallization temperature Tcr and the energy E needed for their
crystallization. For instance, the crystalline Si3N4, SiO2, and Al2O3 coatings can be formed at high values of substrate temperature
Ts 800 C due to high-crystallization temperatures Tcr of these compounds. On the contrary, TiNx, TiC, and ZrNx nitrides easily
crystallize at low values of Ts 500 C. It is well known that the addition of selected elements into the amorphous coating can either
decrease or increase its crystallization temperature. The elements that accelerate the crystallization are called the promoters of
crystallization (e.g., Zr added to HfO2 (122)), and the elements that slow the crystallization (e.g., Al added to HfO2 (122)) are called
the inhibitors of crystallization. The formation of nanograins plays a dominant role in production of the NG/AM nanocomposite
coatings. The element added to the amorphous matrix strongly influences not only properties of the NG/AM nanocomposite
coating but also the conditions of its growth and/or the stabilization of phases from which it is composed (e.g., Y stabilizes all forms
of zirconia (123–125) and Ge stabilizes the tetragonal t-ZrO2 phase in oxidized Zr–Ge–N coatings (126)).
The NG/AM nanocomposite coatings are interesting from both the scientific and application points of view. The most
remarkable feature of NG/AM nanocomposite coatings is the fact that a very low (5 at.%) amount of the element added in the
amorphous material results in an enhanced and/or unique property of the base material. The physics behind these enhanced and/or
new properties of alloyed coatings need to be understood. Properties of the NG/AM nanocomposites are determined not only by the
size d of nanograins and the separation distance w between nanograins but also by their structure, elemental composition of both
the amorphous matrix and nanograins, and mutual interaction between the nanograins. The selection of elements and the control
of the amount of individual phases are very efficient ways to optimize the physical and functional properties of the NG/AM
nanocomposite coating. As an example, the formation of the NG/AM nanocomposite with optimized tribological properties (i.e.,
with low friction and low wear) is given in this section.
4.13.2.3.1 Hard nc-MeC/a-C Composite Coatings with Low Friction m and Low Wear k
It is well known that MeC/a-C nanocomposite coatings exhibit excellent tribological properties (i.e., low coefficient of friction m and
coefficient of wear k; here, Me ¼ Ti, Zr, W, Mo, etc.). The values of m and k can be controlled by the amount of Me atoms incor-
porated in MeC grains embedded in the amorphous a-C matrix. The structure of the MeC/a-C coating gradually changes from
crystalline through to nanocrystalline and then to amorphous with increasing content of C atoms in the coating (i.e., with an
increasing C/Ti ratio); see Figure 18.
The MeC/a-C coating with structure corresponding to a transition region between the nanocrystalline and amorphous structure
usually exhibits good tribological properties. However, no optimization of m and k can be performed because it is very difficult to
distinguish the effects of small changes in the nanocrystalline and/or amorphous structure and correlate these changes with m and k.
On the other hand, already small changes in the structure of the MeC/a-C coating result in clear changes of its mechanical properties
(i.e., the hardness H, effective Young’s modulus E*, and elastic recovery We). The changes in H, E*, and We can be easily correlated
with m and k, and this way the tribological properties of coating can be optimized; see Figures 19 and 20.
www.iran-mavad.com
Figure 18 Evolution of the structure of a TiC/a-C coating with an increasing C–Ti ratio. The dotted lines show positions of X-ray reflections from

a powder TiC standard. Adapted after Musil, J.; Novák, P.; Cerstvý, R.; Soukup, Z. Tribological and Mechanical Properties of nc-TiC/a-C Nanocomposite
Thin Films. J. Vac. Sci. Technol. 2010, 28 (2), 244–249.
Figure 19 Coefficient of friction (CoF) m of the 3 mm thick TiC/a-C coating with low (2) and high (>2) values of the C/Ti ratio as a function of

(a) hardness H, and (b) effective Young’s modulus E*. Reproduced from Musil, J.; Novák, P.; Cerstvý, R.; Soukup, Z. Tribological and Mechanical
Properties of nc-TiC/a-C Nanocomposite Thin Films. J. Vac. Sci. Technol. 2010, 28 (2), 244–249.
Figure 20 Wear k of the 3 mm thick Ti–C coating with low (2) and high (>2) values of the C–Ti ratio as a function of (a) hardness H and (b)

effective Young’s modulus E*. Reproduced from Musil, J.; Novák, P.; Cerstvý, R.; Soukup, Z. Tribological and Mechanical Properties of nc-TiC/a-C
Nanocomposite Thin Films. J. Vac. Sci. Technol. 2010, 28 (2), 244–249.
www.iran-mavad.com
The following important issues can be drawn from Figure 16: (1) the TiC/a-C composite coating can exhibit two values of m for
the same hardness H that show that the hardness H is not a sufficient condition to reach low friction m; (2) the friction m decreases
with decreasing value of E*; and (3) the TiC/a-C composite coating with (a) an X-ray amorphous structure and (b) H ¼ 15.5 GPa,
E* ¼ 138 GPa (H/E* ¼ 0.12) exhibits the lowest value of the friction m. Similar conclusions are valid also for the wear k of the
TiC/a-C composite coating; see Figure 20.
Besides, it was found that the TiC/a-C composite coatings with low E* satisfying the ratio H/E* > 0.1 are highly elastic coatings
with the elastic recovery We 70%. The enhanced elasticity of the TiC/a-C composite coating is also very important in the formation
of tribological coatings with reduced friction m and wear k (69,127–129). The need for high value of the elastic recovery We of the
coating for the reduction of its friction m and wear k was demonstrated also for a-(Ti–C–N) (128) Mo–C (130), and Mo–O–N (130).
However, here it is worthwhile to note that the formation of the coating exhibiting simultaneously sufficiently high hardness
H z 16–20 GPa, low effective Young’s modulus E* satisfying the ratio H/E* > 0.1, and high elastic recovery We 70% is difficult
and, in some combination of elements, an unrealizable task. In summary, it can be concluded that besides a friction phase, which is
always the most important part of every tribological coating, its mechanical properties (H, E*, and We) can strongly contribute to
the reduction of friction m and wear k if they are optimized as shown in this chapter. The nc-TiC/a-C coatings with low friction m and
low wear k are typical example of the NG/AM nanocomposites. The finding that coatings with the ratio H/E* > 0.1 are highly elastic
(the elastic recovery We 70%) is also of key importance for the development of hard coatings with enhanced resistance to
cracking.
4.13.2.4 Hard Coatings Resistant to Cracking in Bending

Hard coatings with enhanced resistance to cracking represent a new class of hard coatings. These coatings are composites composed
of grains dispersed or embedded in a matrix of the base material. They can be composed of nitrides, carbides, oxides, a combination
of nitrides and oxides and other compounds selected according to applications for which they are developed. Recently, it has been
found that hard coatings resistant to cracking exhibit a low effective Young’s modulus E* satisfying the ratio H/E* 0.1 and a high
elastic recovery We 60% (131–133). This fact was demonstrated for the oxide–oxide (Al–Cu–O (131,133)) and nitride–oxide
composite (Al–O–N (132)) coatings. The production of hard coatings with H/E* > 1 and We 60% is a quite difficult but possible
task. The easiest way allowing one to control effectively the mechanical properties (H, E*, We) of the hard coating is by adding
a selective element and controlling its amount in the base material, for instance, Cu or Zr or N in the Al–O system, as demonstrated
in this section.
4.13.2.4.1 Hard Al–Cu–O Coatings with Enhanced Resistance to Cracking

The evolution of the hardness H, the effective Young’s modulus E*, the ratio H/E*, and the elastic recovery We of the Al–Cu–O
coating with increasing amount of Cu is illustrated in Figure 21. From this figure it is seen that while H of the Al–Cu–O coating
strongly increases with a small addition of Cu (up to w5 at.%) from H z 7.5 GPa (pure Al2O3) to H z 17 GPa at w5 at.% Cu, the
increase of E* is much smaller. In consequence of this fact, the Al–Cu–O coatings with 3 at.% Cu exhibit a high ratio H/E* > 0.1
that results in a strong increase of We up to w75% at 10 at.% Cu. It is, however, very important because a low value of E* satisfying
the ratio H/E*0.1 and high values of We strongly increase the resistance of the Al–Cu–O coating to cracking. It was demonstrated
by an indentation test when the diamond indenter was impressed into the coating under a high load L ¼ 1 N. The resistance of the
coating to cracking was determined from a rise of cracks formed during the impression of diamond in the coating as a function of H,
H/E*, We and of the amount of Cu in the coating (60). Results of this test are displayed in Figure 21.
Figure 21 (a) Hardness H and effective Young’s modulus E*; and (b) elastic recovery We and H/E* ratio of the Al–Cu–O coating as a function of Cu
content. Adapted after Blazek, J.; Musil, J.; Stupka, P.; Cerstvy, R.; Houska, J. Properties of Nanocrystalline Al-Cu-O Films Reactively Sputtered by DC
Pulsed Dual Magnetron. Appl. Surf. Sci. 2011, 258, 1762–1767.
www.iran-mavad.com
Figure 22 SEM images of diamond indenter impression images into (a) nc-(g-Al2O3)/a-Al2O3 and (b)–(e) nc-(Al82x,Cu3x)O12/a-(Al–Cu–O) composite
coatings created under load L ¼ 1 N. Reproduced from Blazek, J.; Musil, J.; Stupka, P.; Cerstvy, R.; Houska, J. Properties of Nanocrystalline Al-Cu-O Films
Reactively Sputtered by DC Pulsed Dual Magnetron. Appl. Surf. Sci. 2011, 258, 1762–1767.
Figure 22 shows that only the pure Al2O3 coating and the Al–Cu–O coating with 2 at.% Cu exhibit clear radial cracks. No cracks
are formed in the coatings with a higher (3 at.%) content of Cu. The rise and disappearance of cracks during the indentation
correlate well with We and H/E* ratio of the Al–Cu–O coating. No cracks are formed at We 60% and H/E* 0.1. It indicates that
high values of We 60% and H/E* 0.1 are the main parameters characterizing the coating with enhanced resistance to cracking.
Here, it is necessary to note that the formation of cracks in the coating during an indentation test can also be influenced by other
factors, particularly by (a) the macrostress s generated in the coating during its growth and (b) the mechanical properties of the
substrate. The effect of s on the resistance of the Al–Cu–O coating to cracking is illustrated in Figure 23. Two coatings with lower
(1.5 GPa) and higher (2.2 GPa) compressive stress are compared. Both coatings exhibit H/E* > 0.1, which has almost the same
(1) hardness H, (2) elastic recovery We and (3) content of Cu but different compressive macrostress s because they were sputtered at
different pressures p; see Table 1. As expected, the coating sputtered at a lower value of pressure p exhibits a higher macrostress s. The
Al–Cu–O coating with lower compressive stress exhibits clear radial cracks. On the contrary, the Al–Cu–O coating with higher
compressive stress exhibits no cracks. This fact indicates that a higher compressive stress helps to close cracks in the coating after
unloading the diamond indenter. This is a very important finding that has already been demonstrated for Zr–Si–N, W–Si–N, and
Mo–Si–N coatings (60). Moreover, it was found that both coatings exhibit almost the same value of the ratio H3/E*2 (resistance to
plastic deformation) but very different values of E* and H/E*. The coating with a higher value of H/E* and lower E* exhibits no
cracks. It indicates that the H/E* ratio plays a greater role in the determination of the resistance of coating to cracking compared to
the H3/E*2 ratio, which characterizes the resistance of the coating to plastic deformation.
It is very probable that the lowest absolute value of E* of the coating material with a given hardness H is a key parameter
that decides on the resistance of the coating to cracking. To confirm this hypothesis, however, a special investigation needs to be
carried out.
Recently, as expected, it has been found that the mechanical properties of the substrate (Hs and Es*) can also strongly influence
the assessment of the resistance of the coating by the indentation test (60). For instance, it was demonstrated that a geometrical form
of cracks depends on the substrate hardness Hs. The cracks are radial for hard substrates (Hs 0.5 Hf); here, Hs and Hf are the
hardness of the substrate and the coating, respectively. On the contrary, the cracks are circular for soft substrates (Hs < 0.5 Hf), e.g.,
15 330 steel with Hs ¼ 2.9 GPa. It means that the resistance to cracking of hard coatings deposited on soft substrates is difficult to
determine by the indentation test. It is one of the reasons why new methods suitable for the determination of the resistance of the
coating to cracking need to be develop. In Section 4.13.2.4.2, a method based on the generation of cracks in the coating deposited at
a metallic strip during its bending along a cylinder is given.
In summary, we can conclude that the resistance of thin protective coatings against cracking depends on (1) the coating structure
(e.g., crystalline or amorphous), (2) the mechanical properties of both the coating (Hf, E*, f We, Hf/Ef*, and s) and the substrate
(Hs and Es*), and (3) the macrostress s (tensile and compressive) in the coating. More details are given in (60,134).
Figure 23 SEM images of diamond indenter impression images into nc-(Al8–2x,Cu3x)O12/a-(Al–Cu–O) composite coatings with almost the same
(z6.4 at.%) Cu content but with (a) high s ¼ 2.2 GPa and (b) low s ¼ 1.5 GPa compressive macrostress formed under load L ¼ 1 N. Mechanical
properties of these coatings are given in Table 1. Adapted after Blazek, J.; Musil, J.; Stupka, P.; Cerstvy, R.; Houska, J. Properties of Nanocrystalline Al-Cu-
O Films Reactively Sputtered by DC Pulsed Dual Magnetron. Appl. Surf. Sci. 2011, 258, 1762–1767.
www.iran-mavad.com
Table 1 Hardness H, effective Young’s modulus E*, elastic recovery We, and H/E* and H3/E*2 ratio of nc-(Al8–2x,Cu3x)O12/a-(Al–Cu–O) composite
coatings deposited at two values of total pressure pT ¼ pAr þ pO2 of sputtering gas
pT (Pa) Cu (at.%) h (nm) H (GPa) E* (GPa) We (%) H/ E* () H3/ E* 2 (GPa) s (GPa)
0.5 6.2 2020 16.7 135 71.0 0.124 0.256 2.2

1.5 6.5 1850 19.4 170 70.0 0.114 0.252 1.5
Adapted after Blazek, J.; Musil, J.; Stupka, P.; Cerstvy, R.; Houska, J. Properties of Nanocrystalline Al-Cu-O Films Reactively Sputtered by DC Pulsed Dual Magnetron. Appl. Surf.
Sci. 2011: 258, 1762–1767.
4.13.2.4.2 Hard Zr–Al–O Coatings with Enhanced Resistance to Cracking

Hard Zr–Al–O coatings prepared by sputtering are two-phase composites composed of a crystalline ZrO2 phase and an amorphous
Al2O3 phase (133). Compared to the Al–Cu–O coating discussed in this chapter, the mechanical properties of the Zr–Al–O coating
were controlled by the ratio Zr/Al (i.e., by the ratio of the crystalline ZrO2 and amorphous Al2O3 phases).
The Zr–Al–O coatings with Zr/Al > 1 and Zr/Al < 1 were prepared by reactive sputtering using alternating current (ac) pulse dual
magnetrons. The magnetrons were equipped with a target composed of an Al plate (B ¼ 50 mm) fixed to the magnetron cathode by
a Zr fixing ring with inner diameter Bin. The content of Al in the Zr–Al–O coating was controlled by Bin.; more details are given in
(133). By this way, it was possible to control the ratio of the crystalline ZrO2 phase and the X-ray amorphous Al2O3 phase in the
Zr–Al–O coating and thereby its structure. It was found that only the Zr–Al–O coatings with Zr/Al > 1 meet requirements necessary
for enhanced resistance of coating to cracking defined in this chapter. It is illustrated in Figure 24, where the evolution of H, E*,
H/E*, and We of the Zr–Al–O coating with Zr/Al > 1 with increasing partial pressure of oxygen pO2 used in its sputtering is shown.
From this figure, it is seen that (1) pO2 very effectively controls the hardness H, effective Young’s modulus E*, and elastic recovery
We of the Zr–Al–O coating; and (2) Zr–Al–O coatings with relatively high hardness H z 18 GPa, H/E* > 0.1, and We 70% can be
formed at pO2 0.15 Pa. The Zr–Al–O coatings with H/E* > 0.1 and We 70% are highly elastic and exhibit an enhanced resistance
to cracking.
The resistance of coating to cracking was tested by a bending test, which imitates the coating deformation in a more realistic way.
The principle of the bending test is shown in Figure 25. The coating is deposited on a metallic strip, and the coated strip is bended
along a fixed cylinder of diameter Bfc using a moving cylinder of diameter Bmc up to an angle ac at which cracks in the coating
occurred.
Figure 24 Evolution of H, E*, We, and H/E* of Zr–Al–O coatings rich in Zr (Zr/Al > 1) with increasing partial pressure of oxygen pO2 used in their
sputtering at Ts ¼ 500 C, Ida ¼ 2 A, Us ¼ Ufl, ds–t ¼ 80 mm, and pT ¼ pAr þ pO2 ¼ 1 Pa. MM, TM, and OM denote the metallic, transition, and oxide
modes in reactive sputtering, respectively. More details are given in. Reproduced from Musil, J.; Sklenka, J.; Cerstvy, R. Transparent Zr-Al-O
Nanocomposite Coatings with Enhanced Resistance to Cracking. Surf. Coat. Technol. 2012, 206, 2105–2109.
www.iran-mavad.com
Figure 25 Schematic illustration of apparatus used to form cracks in coating deposited on metallic strip during bending of coated strip (135). Adapted
after Musil, J.; Sklenka, J.; Cerstvy, R. Transparent Zr-Al-O Nanocomposite Coatings with Enhanced Resistance to Cracking. Surf. Coat. Technol. 2012, 206,
2105–2109 and Musil, J. Hard Nanocomposite Coatings: Thermal Stability and Toughness. Surf. Coat. Technol., submitted for publication.
Figure 26 Surface morphology of (a) the Mo strip prior to coating deposition; (b) an as-deposited 3.3 mm thick Zr–Al–O coating with low hardness
H ¼ 7.1 GPa, low ratio H/E* ¼ 0.06, and low We ¼ 44%; and (c) a coating morphology after bending the coating–substrate couple along a steel cylinder of
diameter B ¼ 25 mm to angle a ¼ 30 . Photos were taken in back-scattering electron mode. Adapted after Musil, J.; Sklenka, J.; Cerstvy, R. Transparent
Zr-Al-O Nanocomposite Coatings with Enhanced Resistance to Cracking. Surf. Coat. Technol. 2012, 206, 2105–2109.
The resistance to cracking of the Zr–Al–O coating with Zr/Al > 1 was investigated by bending of the Mo strip
(80 15 0.1 mm3) coated with several micrometer thick Zr–Al–O coatings exhibiting different values of H, E*, H/E*, and We
along a cylinder (B ¼ 25 mm). The formation of cracks is illustrated in Figure 26. Here, the surface morphology of (1) the Mo strip
prior to the coating deposition; (2) an as-deposited 3.3 mm thick Zr–Al–O coating with low hardness H ¼ 7.1 GPa, low ratio
H/E* ¼ 0.06, and low We ¼ 44%; and (3) a coating morphology after bending the coating–substrate couple to angle a ¼ 30 are
compared. These coatings easily crack in bending. Cracks created in the coating during bending are clearly seen in Figure 26 and are
perpendicular to the direction of bending. It means that a bending test can be used to assess the resistance of the coating against
cracking.
On the contrary, highly elastic, hard Zr–Al–O coatings with H ¼ 18–19 GPa, high ratio H/E* 0.1, and high We 70%,
prepared at pO2 0.1 Pa as shown in Figure 24, are more resistance to cracking compared to the coatings with low ratio H/E* < 0.1
and low We 70%. It is due to the fact that the material with H/E* > 0.1, which exhibits a lower value of E* at a given hardness H, is
much more elastic compared to the coating with H/E* 0.1, distributes the applied load over a wider area, and in this way strongly
increases the resistance of the coating against cracking. The coatings with low E*, high ratio H/E* > 0.1, and high We 70% exhibit
no cracks even after bending at a ¼ 180 . This finding is of great importance for many applications.
The results obtained from the bending test agree well with those obtained from the indentation test, that is, several micrometers
thick, highly elastic, hard coatings with H z 20 GPa, low effective Young’s modulus E* satisfying the ratio H/E* 0.1, and high
value of elastic recovery We > 60% exhibit enhanced resistance to cracking.
4.13.2.4.3 Hard Al–O–N Coatings with Enhanced Resistance to Cracking

Hard coatings with enhanced resistance to cracking can be not only the oxide–oxide composite coatings produced by (1) the
addition of selected element into a base oxide (e.g., Al–Cu–O) and (2) a combination of the crystalline oxide with the amorphous
www.iran-mavad.com
Figure 27 H, E*, We, and H/E* of the Al–O–N coatings sputtered at Ts ¼ 500 C as a function of oxygen duty cycle sO2 =TO2 . Reproduced from Musil, J.;
Meissner, M.; Jilek, R.; Tolg, T.; Cerstvy, R. Two-Phase Single Layer Al-O-N Nanocomposite Films with Enhanced Resistance to Cracking. Surf. Coat.
Technol. 2012, 206, 4230–4234.
oxide (e.g., Zr–Al–O) but also a nitride–oxide composite coating. A typical example of the latter type of hard coating with enhanced
resistance to cracking can be a two-phase single-layer Al–O–N nanocomposite coating. These nanocomposite coatings can be
prepared by reactive sputtering using a dual magnetron in a mixture of N2 þ O2 with a pulsed inlet of oxygen; more details are given
in (132).
The mechanical properties of the Al–O–N coating can be effectively controlled by the ratio of the amount of the nitride and
oxide phase in the coating, which strongly changes its structure. The structure of Al–O–N coating is controlled by the ratio sO2 =TO2 ,
as shown in Figure 27; here, sO2 and TO2 are the length of the oxygen pulse and the period of oxygen pulses, respectively. This figure
shows correlations between the mechanical properties (H,E*) and H/E* ratio of the Al–O–N coatings sputtered at Ts ¼ 500 C and
their structure. This figure shows that (1) the nc-AlN/a-(Al–O–N) and nc-(g-Al2O3)/a-(Al–O–N) nanocomposite coatings have
a higher hardness H compared with that of the c-AlN (sputtered at sO2 =TO2 ¼ 0) and the amorphous a-(Al–O–N) coatings
(sputtered at sO2 =TO2 ranging from w0.18 to w0.3); (2) two kinds of nanocomposite coatings with H/E* > 0.1 are formed: (a) the
nc-AlN/a-(Al–O–N) nanocomposite coatings in the interval sO2 =TO2 ranging from w0.1 to w0.18 where the hardness H is almost
constant and E* decreases, and (b) the nc-(g-Al2O3)/a-(Al–O–N) nanocomposite coatings at sO2 =TO2 0.3; and (3) while both the
nc-nitride/a-oxide and the nc-oxide/a-oxide nanocomposite coatings exhibit relatively high hardness H ¼ 15–20 GPa, low Young’s
modulus E* resulting in high ratio H/E* > 0.1, and high elastic recovery We > 60%, the X-ray amorphous a-(Al–O–N) coatings
exhibit much lower hardness H z 10 GPa, low ratio H/E* < 0.1, and low We ¼ 50%.
This experiment shows that the Al–O–N coatings with different values of H, H/E*, and We can be prepared if the amounts of N and
O are correctly controlled. Similarly, as in the case of Al–Cu–O and Zr–Al–O composite coatings, it was also found that the highly
elastic hard Al–O–N coatings with H/E* > 0.1 and a high elastic recovery We 60% exhibit an enhanced resistance to cracking.
The resistance of coating to cracking was tested by bending of the Mo strip coated with Al–O–N coating; for more details, see (132).
In summary, it can be concluded that the incorporation of nanograins into the matrix of a base material of coating can strongly
increase not only the hardness H of the composite coating but also its elasticity and resistance to cracking. Further detailed study is,
however, necessary to understand fully correlations between the elemental and phase composition of the composite coating and its
resistance to cracking.
4.13.3 Trends of the Present and the Next Development of Hard Nanocomposite Coatings
Before formulating the main directions of a next development of the hard coatings, we briefly describe a present activity in this field.
The main activity is concentrated mainly on a solution of the following problems: (1) Formation of hard coatings with enhanced
toughness, (2) formation of crystalline coatings on unheated substrates, and (3) formation of hard oxide coatings with a very high
deposition rate aD 1000 nm min1. Ideas about how to solve these problems are briefly described in this section.
4.13.3.1 Hard Coatings with Enhanced Toughness

Hard nanocomposite coatings with enhanced toughness are coatings that are simultaneously hard and tough. Such coatings should
be very elastic and exhibit a low plastic deformation, resilient properties when the plastic deformation is zero, and an enhanced
resistance to cracking.
The way how to produce hard, tough, and resilient coatings is indicated by Hooke’s law s ¼ E$3; here, s is the stress (load) and
3 is the strain (deformation). If we need to form the material that exhibits a higher elastic deformation (higher value of 3) at a given
value s, its Young’s modulus E must be reduced. It means that materials with the lowest value of the Young’s modulus E at a given
hardness H (s ¼ const) need to be developed. It is a simple solution but a very difficult task.
www.iran-mavad.com
Figure 28 Schematic illustration of stress s versus strain 3 curves of superhard (brittle), hard (resilient), hard, and hard (tough) coatings. Resilient
coatings exhibit no plastic deformation (line 0A). Reproduced from Musil, J. Hard Nanocomposite Coatings: Thermal Stability, Oxidation Resistance and
Toughness. Surf. Coat. Technol. 2012, 207, 50–65.
The stress s versus strain e dependences for brittle, tough, and resilient hard coatings are schematically displayed in Figure 28.
Superhard materials are very brittle, exhibit almost no plastic deformation, and have very low strain 3 ¼ 31. Hard and tough
materials exhibit both the elastic and plastic deformation. The material, which withstands a higher strain 3 ranging from 31 to 3max
without cracking, exhibits a higher toughness. The hardness of tough materials is higher in the case when 3max is achieved at higher
values of smax (H z 3s). On the contrary, fully resilient hard coatings exhibit, compared to hard and tough materials, a lower
hardness H, no plastic deformation (line 0A), and high elastic recovery We. The hardness H of hard, tough, and well-resilient
coatings, ranging from about 15 to 25 GPa, is, however, sufficient for many applications. The main advantage of these coatings is
their enhanced resistance to cracking. These are reasons why in the very near future, hard, tough, and fully resilient hard coatings will
be developed. These coatings represent a new generation of advanced hard nanocomposite coatings.
Therefore, a new task in the development of advanced hard nanocomposite coatings with enhanced toughness and enhanced
resistance to cracking is to produce coatings with (1) a low value of the Young’s modulus E* satisfying the H/E* 0.1 ratio and (2)
a high value of the elastic recovery We. Experiments indicate that the values of E* and We can be controlled by the elemental and
phase composition of the coating. As an example, the effect of the addition of a selected element in a base material on its H/E* ratio
is illustrated in Figure 29.
Figure 29 displays H ¼ f(E*) dependences of five Ti–N, Ti–Al–N, Zr–N, Zr–Cu–N, and Al–Cu–N nitride coatings prepared by
magnetron sputtering (28,36). Also, in this figure a straight line H/E* ¼ 0.1, which divides the H–E* plane in two regions with
H/E* > 0.1 and H/E* < 0.1, is displayed. From this figure it is seen that experimental points corresponding to individual nitrides are
quite well distributed along mutually separated straight lines. This figure clearly shows that (1) the coating material with the same
hardness H and different elemental composition can exhibit different values of the effective Young’s modulus E*; (2) the value of E*
of the Me1–Me2–N coating depends not only on the element Me2 added to the Me1N binary nitride but also on the element Me1,
which forms the binary nitride; (3) not all nitrides exhibit H/E* > 0.1; and (4) the coating material with the ratio H/E* > 0.1 can be
achieved only in the case when both elements Me1 and Me2 are correctly selected. The last fact represents a huge potential for new
industrial applications, particularly for the improvement of properties of the binary nitrides and the development of new advanced
protective coatings, such as the improvement of cutting properties and lifetime of cutting tools.
The preparation of the coatings with H/E* > 0.1 is a complex and difficult task because the hardness H and the effective Young’s
modulus E* are two mutually coupled quantities. The magnitudes of H and E* depend on the deposition parameters used in the
preparation of coating and are controlled not only by its elemental composition, as shown in this chapter, but also by its structure,
phase composition, and microstructure (i.e., by the energy delivered to the growing coating, particularly by bombarding ions and
condensing atoms). At present, there are no general rules that allow one to predict how to prepare the coatings with H/E* > 0.1.
4.13.3.2 Formation of Hard Nanocomposite Coatings on Unheated Substrates

At present, there is an urgent need to master the formation of hard nanocomposite thin coatings on heat-sensitive substrates such as
polymer foils, textiles, plastics, polycarbonate, and so on. It is a very difficult task because many coatings and particularly those
based on oxides exhibit an amorphous structure when they are deposited (sputtered) without a conventional substrate heating.
Therefore, another kind of energy delivered to the growing coating on an unheated substrate to stimulate its nanocrystallization or
full crystallization has to be used.
The energy E delivered to the growing coating is composed of two main components (i.e., E ¼ Es þ Ep), where Es is the energy
delivered by substrate heating and Ep is the energy delivered by bombarding particles (ions Ebi and fast neutrals Efn)
www.iran-mavad.com
Figure 29 Control of the effective Young’s modulus E* of the binary nitrides by addition of selected elements. Adapted after Musil, J.; Vlcek, J.
Magnetron Sputtering of Hard Nanocomposite Coatings and Their Properties. Surf. Coat. Technol. 2001, 142, 557–566 and Musil, J.; Vlcek, J.; Regent, F.;
Kunc, F.; Zeman, H. Hard Nanocomposite Coatings Prepared by Magnetron Sputtering. Key Eng. Mater. 2002, 230–232, 613–632.
Figure 30 Schematic illustration of the two-dimensional (2D) structural zone model of J. A. Thornton extended in a region of low-pressure sputtering.
Here, Ts and Tm are the substrate temperature and the melting temperature of coating material, respectively. Adapted after Musil, J. Low-Pressure
Magnetron Sputtering. Vacuum 1998, 50 (3–4), 363–372.
(i.e., Ep ¼ Ebi þ Efn). Therefore, in the case when crystalline coatings are required to be deposited on unheated substrates, it is
necessary to replace Es with Ep. This is the principle of the formation of nanocrystalline coatings on unheated substrate; see
Figure 30. It was demonstrated experimentally that crystalline, high-quality, dense sputtered coatings with a smooth surface and
featureless microstructure are produced in zone T of the SZM developed by J. A. Thornton (136).
From Figure 30, it is clearly seen that just the zone T expands to low values of the ratio Ts/Tm / 0 with decreasing argon pressure
pAr. This means that the crystalline coatings from the zone T of SZM can be formed without the substrate heating thanks to the
energy Ep because the sputtering at low pressures pAr 0.1 Pa is a collision-less deposition process (l << ds–t, where l is the mean
free path of particles and ds–t is the substrate-to-target distance) without loss of energy of bombarding and/or condensing particles.
The whole kinetic energy of these particles is transferred to the growing coating and stimulates its crystallization. This means that the
formation of nanocrystalline sputtered coatings on unheated substrates is based on the replacement of the equilibrium substrate
and coating heating (Ts/Tm / 0) with the nonequilibrium atomic scale heating (Ep) in which the energy necessary for the coating
nanocrystallization is delivered by condensing and bombarding particles (137). The maximum energy Ep max delivered to the
coating is, however, limited to a value that still prevents a thermal destruction of the heat-sensitive substrate.
www.iran-mavad.com
4.13.3.3 Very-High-Rate Deposition of Hard Oxide Coatings

Hard oxide coatings represent a huge application potential. Their utilization in many applications is, however, limited by a very low
deposition rate. It is well known that the reactive magnetron deposition of oxides in the oxide mode of sputtering is very slow (only
about several nm min1). This fact prevents one from producing high-quality extremely thick (tens to hundreds mm) protective
coatings that could be, for instance, operated as effective thermal barriers. These are the main reasons why recently strong attention
has focused on the development of new ways of the high-rate deposition of oxide coatings (138–193).
At present, the following four methods make it possible to deposit oxide coatings with a very high deposition rate, and they are
the subject of intensive development: (1) gas flow sputtering (159–166), (2) high-power pulsed magnetron sputtering (HPPMS)
(167–173), (3) ionized evaporation (174–183), and (4) magnetron sputtering from a molten target (184–193). The principles of
these methods are schematically shown in Figure 31.
The gas flow sputtering is based on a hollow cathode or magnetron high-density discharge combined with a high flow of the
inert gas along cathodes that results in a transport of the sputtered material to the substrate and a high-rate deposition of coatings
(w120 nm min1 kW1 for alumina coatings (160)) at relatively high pressures of about 40 Pa. The HPPMS achieves a high-rate
deposition of coatings due to extremely high target power densities (1000 W cm2). The main advantage of this process is a high
degree of ionization of sputtered atoms (80%). It means that the HPPM sputtering is an excellent alternative for the cathodic arc
evaporation (CAE), mainly due to avoiding a generation of macroparticles that (1) are easily formed in CAE, and (2) strongly
deteriorate a quality of coatings produced by CAE. A maximum deposition rate aD of oxide coatings deposited by reactive HPPM
sputtering achieves about 200 nm min1 (Al2O3 (150) and Ta–O–N (158)). The ionized evaporation is the process with the highest
deposition rate of oxide coatings. The material is evaporated by an electron beam (EB) and ionized by an electric discharge; thus,
this process is called ionized evaporation or ionized EB evaporation. The plasma activation of the vapor of evaporated material
considerably improves properties of the deposited coating due to densification of its microstructure. The deposition rate of the
Al2O3 coating with a dense microstructure and a smooth surface of about 6000 nm min1 is reported in (180). The magnetron with
molten target combines the evaporation of deposited material and the ionization of vapor of evaporated material in one system.
This system is operated without EB, and conversion of a solid target to a molten one is achieved by a combined action of
(1) increased target power loading and (2) limited cooling of the magnetron target. The magnetron with a molten target is suitable
not only for very-high-rate deposition of metals and alloys with low melting points (e.g., In (156 C), Sn (232 C), and Zn (419 C),
which easily evaporates) but also for high-rate deposition of oxides and other compounds. The magnetron with a molten target is
Figure 31 Schematic illustration of (a) gas flow sputtering, (b) high-power pulsed magnetron sputtering, (c) ionized evaporation, and (d) magnetron
sputtering with a molten target.
www.iran-mavad.com
Figure 32 Deposition rate aD of transparent Si–Zr–O oxide coatings, reactively deposited by an ac-pulsed dual magnetron operated in bipolar mode at
Ts ¼ 500 C, Us ¼ Ufl, ds–t ¼ 80 mm, and pT ¼ pAr þ pO2 ¼ 1 Pa in the transition mode of sputtering (close to the metallic mode of sputtering), as
a function of target power density Wta averaged over the pulse period. Adapted after Musil, J.; Satava, V.; Baroch, P. High-Rate Reactive Deposition of
Transparent SiO2 Films Containing Low Amount of Zr from Molten Magnetron Target. Thin Solid Films 2010, 519, 775–777.
now under development, and in first experiments extremely high deposition rates for Si–Zr–O coatings of about w550 nm min1 at
the target distance ds–t ¼ 80 mm (800 nm min1 at ds–t ¼ 60 mm) were already achieved (188); see Figure 32.
Figure 32 shows that the deposition rate aD of the Si–Zr–O coating increases as expected with increasing target power density Wta.
The increase of aD is relatively slow at Wta 55 W cm2 but strongly increases at Wta 60 W cm2. This change in the increase of aD
is connected with a heating of the sputtered target. The cold target is gradually converted to a hot target and molten target as Wta
increases. In our experiment, the target is cold up to Wta z 55 W cm2, hot for Wta ranging from w55 to w75 W cm2 and molten
for Wta 75 W cm2. The deposition rate aD from the cold target is controlled by sputtering, from the hot target by the combined
action of sputtering and sublimation, and from the molten target mainly by evaporation. This is the reason why the deposition rates
aD of the coating deposited from the hot and molten targets are higher than that of the coating sputtered from the cold target. The
highest deposition rate aD of the coating is achieved in the evaporation from the molten target because during evaporation, atoms
are evolved from the target more effectively than by sputtering and sublimation. Here, it is worthwhile to note that the deposition of
coating from the molten magnetron target is not simply evaporation but also ionized evaporation because evaporated atoms pass
through a magnetron discharge where they are effectively ionized. The magnetrons with hot and molten targets are new sputtering
systems not mastered yet. Both systems are now a subject of intensive development in many labs (184–193).
The development of magnetron sputtering systems with a molten target is of key importance, particularly for a very-high-rate
deposition of extremely thick (100 mm) coatings with a dense microstructure, smooth surface, and enhanced resistance to
cracking.
4.13.3.4 Summary
In summary, it can be concluded that the next research activity in the field of hard nanocomposite coatings is expected to
concentrate mainly on solutions to the following problems: (1) the development of hard coatings with enhanced toughness and
increased resistance to cracking; (2) the investigation of DNG/AM composite coatings composed of a small amount of nanograins
dispersed in the amorphous matrix, with the aim to develop new coatings with unique physical and functional properties; (3) the
investigation of the electronic charge transfer between nanograins with different chemical composition and different Fermi energies
in nanocomposite coatings, with the aim to understand its effect on the functional properties of coating; (4) the nanocrystallization
of amorphous materials at temperatures of about or less than 100 C for flexible electronics; (5) the formation of high-temperature
phases at temperatures T 500 C using super-fast heating and cooling at the atomic level; (6) the development of nanocomposite
coatings that are thermally stable above 1500 C and protect the substrate against oxidation at temperatures up to w2000 C; (7)
the formation of multilayers composed of nanobilayers; (8) the high-rate reactive deposition of hard coatings based on oxides with
a deposition rate aD exceeding 10 000 nm min1; and (9) the development of new physical vapor deposition systems for the
production of new advanced coatings under new physical conditions, such as the system for high-power pulsed magnetron
sputtering, a magnetron with hot and molten targets, a magnetron with regulated plasma potential, and sputtering systems with
enhanced sputtering gas ionization for operation at low pressures and very-high-rate reactive deposition of oxide coatings.
Acknowledgments
This work was supported in part by the Grant Agency of the Czech Republic (GACR) under Project No. P108/12/0393.
www.iran-mavad.com
References
1. Gleiter, H. Nanocrystalline Materials. Prog. Mater. Sci. 1989, 33, 223–315.

2. Birringer, R. Nanocrystalline Materials. Mater. Sci. Eng. 1989, A117, 33–43.
3. Barnett, S. A. Deposition and Mechanical Properties of Superlattice Thin Films. In Physics of Thin Films; Fracombe, M. H., Vossen, J. A., Eds.; Academic Press: New York,
1993; pp 1–73.
4. Siegel, R. W. What Do We Really Know about the Atomic-Scale Structures of Nanophase Materials? J. Phys. Chem. Solids 1994, 55 (10), 1097–1106.
5. Siegel, R. W.; Fougere, G. E. Grain Size Dependent Mechanical Properties in Nanophase Materials. In Grant, H. J., Amstrong, R. W., Otooni, M. A., Ishizaki, K., Eds., Vol. 362;
Materials Research Society: Arrendale, PA, 1995; pp 219–229.
6. Veprek, S.; Reiprich, S. A Concept for the Design of Novel Superhard Coatings. Thin Solid Films 1995, 265, 64–71.
7. Gleiter, H. Nanostructured Materials: State of the Art and Perspectives. Nanostruct. Mater. 1996, 6, 3–14.
8. Lu, K. Nanocrystalline Metals Crystallized from Amorphous Solids: nanocrystallization, Structure, and Properties. Mater. Sci. Eng. 1996, R16, 161–221.
9. Liu, B. X.; Jin, O. Formation and Theoretical Modeling of Non-Equilibrium Alloy Phases by Ion Mixing. Phys. Stat. Sol. (a) 1997, 161, 3–33.
10. Vaz, F.; Rebouta, L.; da Silva, M. F.; Soares, J. C. Thermal Oxidation of Ternary and Quaternary Nitrides of Titanium, Aluminium and Silicon. In Protective Coatings and Thin
Films; Paleau, Y., Barna, P. B., Eds.; Kluwer Academic Publisher, 1997; pp 501–510.
11. Yip, S. The Strongest Size. Nature 1998, 391, 532.
12. Veprek, S.; Nesladek, P.; Niederhofer, A.; Glatz, F.; Jilek, M.; Sima, M. Recent Progress in the Superhard Nanocrystalline Composites: Toward Their Industrialization and
Understanding of the Origin of Superhardness. Surf. Coat. Technol. 1998, 108–109, 138–147.
13. Voevodin, A. A.; Zabinski, J. S. Superhard, Functionally Gradient, Nanolayered and Nanocomposite Diamond-Like Carbon Coatings for Wear Protection. Diam. Rel. Mater.
1998, 7, 463–467.
14. Musil, J.; Vlcek, J. Magnetron Sputtering of Films with Controlled Texture and Grain Size. Mater. Chem. Phys. 1998, 54, 116–122.
15. Kim, H. S. A Composite Model for Mechanical Properties of Nanocrystalline Materials. Scripta Mater. 1998, 39 (8), 1057–1061.
16. Veprek, S. The Search for Novel, Superhard Materials. J. Vac. Sci. Technol. 1999, A17, 2401–2420.
17. Niederhofer, A.; Moto, K.; Nesladek, P.; Veprek, S. Diamond is not the Hardest Material Anymore: Ultrahard Nanocomposite nc-TiN/a- & nc-TiSi2 Prepared by Plasma CVD. In
Hrabovsky, M., Konrad, M., Kopecky, V., Eds., Vol. 3; Institute of Plasma Physics AS CR: Prague, Czech Republic, 1999; pp 1521–1525.
18. Musil, J.; Zeman, P.; Hruby, H.; Mayrhofer, P. H. ZrN/Cu Nanocomposite Film – A Novel Superhard Material. Surf. Coat. Technol. 1999, 120, 179–183.
19. Musil, J. Hard and Superhard Nanocomposite Coatings. Surf. Coat. Technol. 2000, 125, 322–330.
20. Musil, J.; Hruby, H. Superhard Nanocomposite TiAl-N Films Prepared by Magnetron Sputtering. Thin Solid Films 2000, 365 (1), 104–109.
21. Gleiter, H. Nanostructured Materials: Basic Concepts and Microstructure. Acta Mater. 2000, 48, 1–29.
22. Voevodin, A. A.; Zabinski, J. S. Superhard Wear-Resistant Coatings with “Chameleon” Surface Adaptation. Thin Solid Films 2000, 370, 223–231.
23. Gleiter, H. Tuning the Electronic Structure of Solids by Means of Nanometer-Sized Microstructures. Scripta Mater. 2001, 44, 1161–1168.
24. Veprek, S.; Niederhofer, A.; Moto, K.; Bolom, T.; Mannling, H.-D.; Nesladek, P.; Dollinger, G.; Bergmaier, A. Composition, Nanostructure and Origin of the Ultrahardness in nc-
TiN/a-Si3N4/a- and nc-TiSi2 Nanocomposites with HV ¼ 80 to 105 GPa. Surf. Coat. Technol. 2000, 133–134, 152–159.
25. Hultman, L. Thermal Stability of Nitride Thin Films. Vacuum 2000, 57 (1), 1–30.
26. Conrad, H.; Narayan, J. On the Grain Size Softening in Nanocrystalline Materials. Scripta Mater. 2000, 42, 1023–1030.
27. Hauert, R.; Patscheider, J. From Alloying to Nanocomposites – Improved Performance of Hard Coatings. Adv. Mater. 2000, 2 (5), 247–259.
28. Musil, J.; Vlcek, J. Magnetron Sputtering of Hard Nanocomposite Coatings and Their Properties. Surf. Coat. Technol. 2001, 142, 557–566.
29. Veprek, S.; Argon, A. S. Mechanical Properties of Superhard Nanocomposites. Surf. Coat. Technol. 2001, 146–147, 175–182.
30. Patscheider, J.; Zehnder, T.; Diserens, M. Structure-Performance Relations in Nanocomposite Coatings. Surf. Coat. Technol. 2001, 146–147, 201–208.
31. Badzian, A. Diamond Challenged by Hard Materials: A Reflection on Development in the Last Decade. Mater. Chem. Phys. 2001, 72, 110–113.
32. Sung, J. The Design of Exotic Superhard Materials. Mater. Chem. Phys. 2001, 72, 141–146.
33. Brazhkin, V. V.; Lyapin, A. G.; Hemley, R. J. Harder than Diamond: Dreams and Reality. Philos. Mag. 2002, A82 (2), 231–253.
34. Veprek, S.; Argon, A. S. Towards the Understanding of Mechanical Properties of Super- and Ultrahard Nanocomposites. J. Vac. Sci. Technol. 2002, B20 (2), 650–664.
35. Musil, J.; Kunc, F.; Zeman, H.; Polakova, H. Relationships between Hardness, Young’s Modulus and Elastic Recovery in Hard Nanocomposite Coatings. Surf. Coat. Technol.
2002, 154 (2–3), 304–313.
36. Musil, J.; Vlcek, J.; Regent, F.; Kunc, F.; Zeman, H. Hard Nanocomposite Coatings Prepared by Magnetron Sputtering. Key Eng. Mater. 2002, 230–232, 613–632.
37. Gleiter, H.; Fichtner, M. Is Enhanced Solubility in Nanocomposites an Electronic Effect? Scripta Mater. 2002, 46, 497–500.
38. Zhang, S.; Sun, D.; Fu, Y. Superhard Nanocomposite Coatings. J. Mater. Sci. Technol. 2002, 18, 485–491.
39. Andrievski, R. A. Thermal Stability of Nanomaterials. Russ. Chem. Rev. 2002, 71 (10), 853–866.
40. Patscheider, J. Nanocomposite Hard Coatings for Wear Protection. M.R.S. Bulletin 2003, 28 (3), 180–183.
41. Zhang, S.; Sun, D.; Fu, Y.; Du, H. Recent Advances of Superhard Nanocomposite Coatings: A Review. Surf. Coat. Technol. 2003, 167, 113–119.
42. Veprek, S.; Mukherjee, S.; Karvanková, P.; Mannling, H.-D.; He, J. L.; Moto, K.; Procházka, J.; Argon, A. S. Limits to the Strength of Super- and Ultrahard Nanocomposite
Coatings. J. Vac. Sci. Technol. 2003, A21 (3), 532–544.
43. Demyashev, G. M.; Taube, A. L.; Siores, E. E. Superhard Nanocomposite Coatings. In Nalwa, H. S., Ed., Vol. 1; American Scientific Publishers, 2003; pp 1–61.
44. Demyashev, G. M.; Taube, A. L.; Siores, E. E. Superhard Nanocomposites. In Nalwa, H. S., Ed., Vol. 10; American Scientific Publishers, 2003; pp 1–46.
45. Musil, J. Hard Nanocomposite Films Prepared by Magnetron Sputtering. In Voevodin, A. A., Levashov, E. A., Shtansky, D. V., Moore, J. J., Eds., Vol. 155; Kluwer Academic
Publishers B.V.: Dordrecht, The Netherlands, 2004; pp 43–56.
46. Leyland, A.; Matthews, A. Design Criteria for Wear-Resistant Nanostructured and Glassy-Metal Coatings. Surf. Coat. Technol. 2004, 177–178, 317–324.
47. Kuntz, J. D.; Zhang, G. D.; Murherjee, A. K. Nanocrystalline-Matrix Ceramic Composites for Improved Fracture Toughness. M.R.S. Bulletin Jan 2004, 22–27.
48. Musil, J.; Miyake, S. Nanocomposite Coatings with Enhanced Hardness. In Novel Materials Processing (MAPEES’04); Miyake, S., Ed.; Elsevier Ltd.: Amsterdam, 2005;
pp 345–356.
49. Andrievski, R. A. Nanomaterials Based on High-Melting Carbides, Nitrides and Borides. Russ. Chem. Rev. 2005, 74 (12), 1061–1072.
50. Zhang, S.; Sun, D.; Fu, Y.; Du, H. Toughening of Hard Nanostructured Thin Films: A Critical Review. Surf. Coat. Technol. 2005, 198, 2–8.
51. Musil, J. Nanocomposite Coatings with Enhanced Hardness. Acta Metall. Sin. (Engl. Lett.) 2005, 18 (3), 433–442.
52. Vlcek, J.; Potocký, S; Cízek, J.; Houska, J.; Kormunda, M.; Zeman, P.; Perina, V.; Zemek, J.; Setsuhara, Y.; Konuma, S. Reactive Magnetron Sputtering of Hard Si-B-C-N Films
with a High-Temperature Oxidation Resistance. J. Vac. Sci. Technol. 2005, 23 (6), 1513–1522.
53. DeHosson, J. T. M., Cavaleiro, A., Eds. Nanostructured Coatings; Springer ScienceþBusiness Media, LCC: New York, U.S.A, 2006.
54. Musil, J. Physical and Mechanical Properties of Hard Nanocomposite Films Prepared by Reactive Magnetron Sputtering. In Nanostructured Coatings; DeHosson, J. T. M.,
Cavaleiro, A., Eds.; Springer ScienceþBusiness Media, LCC: New York, U.S.A, 2006; pp 407–463.
55. Lu, C.; Mai, Y. W.; Shen, Y. G. Recent Advances on Understanding the Origin of Superhardness in Nanocomposite Coatings: A Critical Review. J. Mater. Sci. 2006, 41,
937–950.
56. Mayrhofer, P. H.; Mitterer, C.; Hultman, L. Microstructural Design of Hard Coatings. Prog. Mater. Sci. 2006, 51, 1032–1114.
www.iran-mavad.com
57. Zhang, S., Ali, N., Eds. Nanocomposite Films and Coatings; Imperial College Press: London, UK, 2007.
58. Musil, J. Properties of Hard Nanocomposite Thin Films. In Nanocomposite Films and Coatings; Zhang, S., Ali, N., Eds.; Imperial College Press: London, UK, 2007;
pp 281–328.
59. Musil, J.; Zeman, P. Hard Amorphous a-Si3N4/MeNx Nanocomposite Coatings with High Thermal Stability and High Oxidation Resistance. Solid State Phenom. 2007, 127,
31–36.
60. Musil, J.; Jirout, M. Toughness of Hard Nanostructured Ceramic Thin Films. Surf. Coat. Technol. 2007, 201, 5148–5152.
61. Zhang, S.; Wang, H. L.; Ong, S. E.; Sun, D.; Bui, X. L. Hard yet Tough Nanocomposite Coatings. Present Status and Future Trends. Plasma Process. Polymer. 2007, 4,
219–229.
62. Raveh, A.; Zukerman, I.; Shneck, R.; Avni, R.; Fried, I. Thermal Stability of Nanostructured Superhard Coatings: A Review. Surf. Coat. Technol. 2007, 201, 6136–6142.
63. Wei, R., Ed. Plasma Surface Engineering Research and Its Practical Applications; Research Signpost Publisher: Kerala, India, 2008.
64. Musil, J.; Baroch, P.; Zeman, P. Hard Nanocomposite Coatings: Present Status and Trends. In Plasma Surface Engineering Research and Its Practical Applications; Wei, R.,
Ed.; Research Signpost Publisher: Kerala, India, 2008; pp 1–34.
65. Musil, J.; Vlcek, J.; Zeman, P. Advanced Amorphous Non-Oxide Coatings with Oxidation Resistance above 1000 C. Adv. Appl. Ceram. 2008, 107 (3), 148–154.
66. Pogrebnjak, A. D.; Shpak, A. P.; Azarenkov, N. A.; Beresnev, V. M. Structures and Properties of Hard and Superhard Nanocomposite Coatings. Phys. Usp. 2009, 52,
29–54.
67. Matthews, A.; Leyland, A. Materials Related Aspects of Nanostructured Tribological Coatings. SVC Bulletin Spring 2009, 40–44.
68. Musil, J.; Satava, V.; Zeman, P.; Cerstvý,
R. Protective Zr-Containing SiO2 Coatings Resistant to Thermal Cycling in Air up to 1400 C. Surf. Coat. Technol. 2009, 203,
1502–1507.
69. Musil, J.; Novák, P.; Cerstvý, R.; Soukup, Z. Tribological and Mechanical Properties of nc-TiC/a-C Nanocomposite Thin Films. J. Vac. Sci. Technol. 2010, 28 (2), 244–249.
70. Musil, J. Nanocomposite Coatings with Thermal Stability and Protection of Substrate against Oxidation above 1000 C. Part 1 Galvanotechnik 2010, 101 (8), 1856–1867.
Part 2 Galvanotechnik, 2010, 101 (9), 2116–2121.
71. Hall, E. O. The Deformation and Ageing of Mild Steel: III Discussion and Results. Proc. Phys. Soc. 1951, B64, 747–753.
72. Petch, N. J. The Cleavage Strength of Polycrystals. J. Iron Steel Inst. 1953, 174, 25–28.
73. Andronis, M.; Jimenez, O.; Leyland, A.; Matthews, A. The Morphology and Structure of PVD ZrN-Cu Thin Films. J. Appl. Phys. D 2009, 42, 085308.
74. Sandu, C. S.; Medjani, F.; Sanjines, R.; Karimi, A.; Levy, F. Structure, Morphology and Electrical Properties of Sputtered Zr-Si-N Thin Films: From Solid Solution to
Nanocomposite. Surf. Coat. Technol. 2006, 201, 4219–4229.
75. Sandu, C. S.; Sanjines, R.; Berkahoul, M.; Medjani, F.; Levy, F. Formation of Composite Ternary Nitride Thin Films by Magnetron Co-Deposition. Surf. Coat. Technol. 2006,
201, 4083–4089.
76. Lu, Y. H.; Shen, Y. G. Nanostructure Transition: From Solid Solution Ti(N, C) to Nanocomposite nc-Ti(N, C)/a-(C, CNx). Appl. Phys. Lett. 2007, 90, 221913.
77. Sandu, C. S.; Medjani, F.; Sanjines, R. Optical and Electrical Properties of Zr-Si-N Thin Films: From Solid Solution to Nanocomposite. Rev. Adv. Mater. Sci. 2007, 15,
173–178.
78. Musil, J.; Regent, F. Formation of Nanocrystalline NiCr-N Films by Reactive DC Magnetron Sputtering. J. Vac. Sci. Technol. 1998, A16, 3301–3304.
79. Musil, J.; Karvankova, P.; Kasl, J. Hard and Superhard Zr-Ni-N Nanocomposite Films. Surf. Coat. Technol. 2001, 139, 101–109.
80. Suna, J.; Musil, J.; Ondok, V. Enhanced Hardness in Sputtered Zr-Ni-N Films. Surf. Coat. Technol. 2006, 200, 6293–6297.
81. Musil, J.; Polakova, H.; Suna, J.; Vlcek, J. Effect of Ion Bombardment on Properties of Hard Reactively Sputtered Ti(Fe)Nx Films. Surf. Coat. Technol. 2004, 177–178,
289–298.
82. Munz, W. D. Titanium Aluminium Nitride Films: A New Alternative to TiN Coatings. J. Vac. Sci. Technol. 1986, A4 (6), 2717–2725.
83. Wahlstrem, L.; Hultman, L.; Sundgren, J.-E.; Abidi, F.; Petrov, I.; Greene, J. E. Crystal Growth and Microstructure of Polycrystalline Ti1xAlxN Alloy Films Deposited by
Ultra-High Vacuum Dual Target Magnetron Sputtering. Thin Solid Films 1993, 235, 62–70.
84. Suzuki, T.; Huang, D.; Ikuhare, Y. Microstructure and Grain Boundaries of (Ti, Al)N Films. Surf. Coat. Technol. 1998, 107, 41–47.
85. Zhou, M.; Makino, Y.; Nose, N.; Nogi, K. Phase Transformation and Properties of Ti-Al-N Films by rf-Plasma Assisted Magnetron Sputtering Method. Thin Solid Films
1999, 339, 203–208.
86. Musil, J.; Hruby, H. Superhard Nanocomposite TiAl-N Films Prepared by Magnetron Sputtering. Thin Solid Films 2000, 365, 203–208.
87. Hasegawa, H.; Kimura, A.; Suzuki, T. Ti1–x AlxN, Ti1–xZrxN and Ti1–xCrN Films Synthesized by the AIP Method. Surf. Coat. Technol. 2000, 132, 76–79.
88. Polakova, H.; Musil, J.; Vlcek, J.; Allart, J.; Mitterer, C. Structure-Hardness Relations in Sputtered Ti-Al-V-N Films. Thin Solid Films 2003, 444, 189–198.
89. Musil, J.; Zeman, H.; Kunc, F.; Vlcek, J. Measurement of Hardness of Superhard Films by Microindentation. Mater. Sci. Eng. 2003, A340, 281–285.
90. Musil, J.; Vlcek, J. Magnetron Sputtering of Alloy Based Films and Its Specificity. Czech. J. Phys. 1998, B48, 1209–1224.
91. Barna, P. B.; Adamik, M. Growth Mechanisms of Polycrystalline Thin Films. In Science and Technology of Thin Films; Matacotta, F. C., Ottaviani, G., Eds.; World Scientific
Publishing Co: Singapore, 1995; pp 1–28.
92. Barna, P. B.; Adamik, M. Formation and Characterization of the Structure of Surface Coatings. In Protective Coatings and Thin Films; Pauleau, Y., Barna, P. B., Eds.;
Kluwer Academic Publishers: the Netherlands, 1997; pp 279–297.
93. Musil, J.; Vlcek, J.; Zeman, P.; Setsuhara, Y.; Miyake, S.; Konuma, S.; Yabuta, K.; Saito, H.; Kumagai, M. Morphology and Microstructure of Hard and Superhard Zr-Cu-N
Nanocomposite Coatings. Jpn. J. Appl. Phys. 2002, 41 (Part 1, No. 11A), 6529–6533.
94. Musil, J.; Suna, J. The Role of Energy in Formation of Sputtered Nanocomposite Films. Mater. Sci. Forum 2005, 502, 291–296.
95. Musil, J.; Sicha, J.; Herman, D.; Cerstvy, R. Role of Energy in Low-Temperature High-Rate Formation of Hydrophilic TiO2 Thin Films Using Pulsed Magnetron Sputtering.
J. Vac. Sci. Technol. A 2007, 25 (4), 666–674.
96. Vlcek, J.; Rusnak, K.; Hajek, V.; Martinu, L. Reactive Magnetron Sputtering of CNx Films: Ion Bombardment Effects and Process Characterization Using Optical Emission
Spectroscopy. J. Appl. Phys. 1999, 86, 3646–3655.
97. Lieberman, M. A.; Lichtenberg, A. J. Principles of Plasma Discharges and Materials Processing; Wiley Interscience: New York, 1994.
98. Suna, J.; Musil, J.; Dohnal, P. Control of Macrostress s in Reactively Sputtered Mo-Al-N Films by Total Gas Pressure. Vacuum 2006, 80, 588–592.
99. Iriarte, G. F.; Engelmark, F.; Katardiev, I. V. Reactive Sputter Deposition of Highly Oriented AlN Films at Room Temperature. J. Mater. Res. 2002, 17 (6), 1469–1475.
100. Kadlec, S.; Musil, J.; Munz, W.-D.; Hakanson, G.; Sundgren, J.-E. Reactive Deposition of TiN Films Using Unbalanced Magnetron. Surf. Coat. Technol. 1989, 39–40,
487–497.
101. Mitterer, C.; Mayrhofer, P. H.; Musil, J. Thermal Stability of PVD Hard Coatings. Vacuum 2003, 71, 279–284.
102. In Smithells Metal Reference Book; Brandes, E. A., Ed., 6th ed., Butterworths: London; pp 8–24.
103. Marks, N. A. Evidence for Subpicosecond Thermal Spikes in the Formation of Tetrahedral Amorphous Carbon. Phys. Rev. A 1988, 38, 3098.
104. Houska, J.; Bilek, M. M. M.; Warschkow, O.; McKenzie, D. R.; Vlcek, J. Ab Initio Simulations of Nitrogen Evolution in Quenched CNx and SiBCN Amorphous Materials. Phys.
Rev. B 2005, 72, 0542041–0542046.
105. Houska, J.; Klemberg-Sapieha, J. E.; Martinu, L. Atomistic Simulations of the Characteristics of TiSiN Nanocomposites of Various Compositions. Surf. Coat. Technol. 2009,
203, 3348–3355.

106. Musil, J.; Bell, A. J.; Vlcek, J.; Hurkmans, T. Formation of High Temperature Phases in Sputter Deposited Ti-Based Films Below 100 C. J. Vac. Sci. Technol. 1996, A14 (4),
2247–2250.
www.iran-mavad.com
107. Hultman, L.; Mitterer, C. Thermal Stability of Advanced Nanostructured Wear-Resistant Coatings. In Nanostructured Coatings; DeHosson, J. T. M., Cavaleiro, A., Eds.;
Springer ScienceþBusiness Media, LCC: New York, U.S.A, 2006; pp 464–510.
108. Erkens, G. New Approaches to Plasma Enhanced Sputtering of Advanced Hard Coatings. Surf. Coat. Technol. 2007, 201, 4806–4812.
109. Tului, M.; Marino, G.; Valente, T. Plasma Spray Deposition of Ultra Temperature Ceramics. Surf. Coat. Technol. 2006, 201, 2103–2108.
110. Zeman, H.; Musil, J.; Zeman, P. Physical and Mechanical Properties of Sputtered Ta-Si-N films with a High (40 at.%) Content of Si. J. Vac. Sci. Technol. 2004, A22 (3),
646–649.
111. Musil, J.; Daniel, R.; Zeman, P.; Takai, O. Structure and Properties of Magnetron Sputtered Zr-Si-N Films with a High (25 at.%) Si Content. Thin Solid Films 2005, 478,
238–247.
112. Musil, J.; Dohnal, P.; Zeman, P. Physical Properties and High-Temperature Oxidation Resistance of Sputtered Si3N4/MoNx Nanocomposite Coatings. J. Vac. Sci. Technol.
2005, B23 (4), 1568–1575.
113. Musil, J.; Daniel, R.; Soldán, J.; Zeman, P. Properties of Reactively Sputtered W-Si-N Films. Surf. Coat. Technol. 2006, 200, 3886–3895.
114. Zeman, P.; Musil, J. Difference in High-Temperature Oxidation Resistance of Amorphous Zr-Si-N and W-Si-N Films with a High Si Content. Appl. Sur. Sci. 2006, 252,
8319–8325.
115. Zeman, P.; Musil, J.; Daniel, R. High-Temperature Oxidation Resistance of Ta-Si-N Films with a High Si Content. Surf. Coat. Technol. 2006, 200, 4091–4096.
116. Daniel, R.; Musil, J.; Zeman, P.; Mitterer, C. Thermal Stability of Magnetron Sputtered Zr-Si-N Films. Surf. Coat. Technol. 2006, 201, 3368–3376.
117. Musil, J.; Zeman, P.; Dohnal, P.; Cerstvý, R. Ti-Si-N Nanocomposite Films with a High Content of Si. Plasma Process. Polym. 2007, 4, S574–S578.
118. Vlcek, J.; Potocký, S; Houska, J.; Zeman, P.; Perina, V.; Setsuhara, Y. New Quaternary Si-B-C-N Films Prepared by Reactive Magnetron Sputtering. Trans. Mater. Res. Soc.
Jpn 2006, 21, 447–451.
119. Houska, J.; Vlcek, J.; Potocký, S; Perina, V. Influence of Substrate Bias Voltage on Structure and Properties of Hard Si-B-C-N Films Prepared by Reactive Magnetron
Sputtering. Diam. Rel. Mat. 2007, 16, 29–36.
120. Houska, J.; Vlcek, J.; Hreben, S.; Bílek, M. M. M.; McKenzie, D. R. Effect of B and Si/C Ratio on High-Temperature Stability of Novel Si-B-C-N Materials. Europhys. Lett.
2006, 76, 512–518.
121. Musil, J.; Satava, V.; Cerstvy, R.; Zeman, P.; Tolg, T. Formation of Crystalline Al-Ti-O Thin Films and Their Properties. Surf. Coat. Technol. 202 (208), 6064–6069.
122. Hong, Y. E.; Kim, Y. S.; Do, K.; Lee, D.; Ko, D. H.; Ku, J. H.; Kim, H. Thermal Stability of Al- and Zr- Doped HfO2 Thin Films Grown by Direct Current Magnetron Sputtering.
J. Vac. Sci. Technol. A 2005, 23 (5), 1413–1418.
123. Badwal, S. P. Zirconia-Based Solid Electrolytes: Microstructure, Stability and Ionic Conductivity. Solid State Ionics 1992, 52, 23–32.
124. Chen, M.; Hallstedt, B.; Gauckler, L. J. Thermodynamic Modelling of the ZrO2-YO1.5 System. Solid State Ionics 2004, 170, 255–274.
125. Briois, P.; Lapostolle, F.; Demange, V.; Djurado, E.; Billard, A. Structural Investigation of YSZ Coatings Prepared by DC Magnetron Sputtering. Surf. Coat. Technol. 2007,
201, 6012–6018.
126. Pilloud, D.; Pierson, J. F.; Cavaleiro, A.; Marco de Lucas, M. C. Effect of Germanium Addition on the Properties of Reactively Sputtered ZrN films. Thin Solid Films 2005,
492, 180–185.
127. Musil, J.; Hromádka, M.; Novák, P. Effect of Nitrogen on Tribological Properties of Amorphous Carbon Films Alloyed with Titanium. Surf. Coat. Technol. 2011, 205 (Suppl.
2), S84–S88.
128. Leyland, A.; Mathews, A. On the Significance of the H/E Ratio in Wear Control: A Nanocomposite Coating Approach to Optimized Tribological Behavior. Wear 2000, 246,
1–11.
129. Leyland, A.; Matthews, A. Optimization of Nanostructured Tribological Coatings. In Nanostructured Coatings; DeHosson, J. T. M., Cavaleiro, A., Eds.; Springer
ScienceþBusiness Media LLC: New York, USA, 2006; pp 511–538.
130. Musil, J.; et al. Mo-C and Mo-O-N Nanocomposite Coatings, unpublished results.
131. Blazek, J.; Musil, J.; Stupka, P.; Cerstvy, R.; Houska, J. Properties of Nanocrystalline Al-Cu-O Films Reactively Sputtered by DC Pulsed Dual Magnetron. Appl. Surf. Sci.
2011, 258, 1762–1767.
132. Musil, J.; Meissner, M.; Jilek, R.; Tolg, T.; Cerstvy, R. Two-Phase Single Layer Al-O-N Nanocomposite Films with Enhanced Resistance to Cracking. Surf. Coat. Technol.,
submitted for publication.
133. Musil, J.; Sklenka, J.; Cerstvy, R. Transparent Zr-Al-O Nanocomposite Coatings with Enhanced Resistance to Cracking. Surf. Coat. Technol. 2012, 206, 2105–2109.
134. Jirout, M.; Musil, J. Effect of Addition of Cu into ZrOx Films on Their Properties. Surf. Coat. Technol. 2006, 200, 6792–6800.
135. Musil, J. Hard Nanocomposite Coatings: Thermal Stability and Toughness. Surf. Coat. Technol., submitted for publication.
136. Thornton, J. A. High Rate Thick Films Growth. Ann. Rev. Mater. Sci. 1977, 7, 239–260.
137. Musil, J. Low-Pressure Magnetron Sputtering. Vacuum 1998, 50 (3–4), 363–372.
138. Scherer, M.; Wirtz, P. Reactive High-Rate DC Sputtering of Oxides. Thin Solid Films 1984, 119, 203–209.
139. Schiller, S.; Heisig, U.; Korndorfer, Chr.; Beister, G.; Reschke, J.; Steinfelder, K.; Srumplel, J. Reactive DC High-Rate Sputtering as Production Technology. Surf. Coat.
Technol. 1987, 33, 405–423.
140. Lehan, J. P.; Sargent, R. B.; Klinger, R. E. High-Rate Aluminum Oxide Deposition by MetaMode™ Reactive Sputtering. J. Vac. Sci. Technol. A 1992, 10 (6), 3401–3406.
141. Schneider, J. M.; Sproul, W. D.; Chia, R. W. J.; Wong, M. S.; Matthews, A. Very-High Rate Reactive Sputtering of Alumina Hard Coatings. Surf. Coat. Technol. 1993, 96,
262–266.
142. Hichwa, B. P. Low Temperature, High Rate Sputter Deposition of Metal Oxides and Transparent Conductors. In Proceedings of the 2nd International Symposium on
ISSP’93, Tokyo, 1993; pp 61–69.
143. Sproul, W. D. High-Rate Reactive DC Magnetron Sputtering of Oxide and Nitride Superlattice Coatings. Vacuum 1998, 51 (4), 641–646.
144. Kharrazi, M.; Azens, A.; Kullman, L.; Granqvist, C. G. High-Rate Dual-Target DC Magnetron Sputter Deposition of Electrochromic MoO3 Films. Thin Solid Films 1997, 295,
117–121.
145. Olsson, M. K.; Macak, K.; Helmersson, U.; Hjorvarsson, B. High Rate Reactive DC Magnetron Sputter Deposition of Al2O3 Films. J. Vac. Sci. Technol. A 1998, 16 (2),
639–643.
146. Nadel, S. J.; Greene, P. Strategies for High Rate Reactive Sputtering. Thin Solid Films 2001, 392, 174–183.
147. Boling, N.; Wood, B.; Morand, P. A High Rate Reactive Sputtering Process for Batch, In-Line or Roll Coaters; Deposition Sciences Inc.: Santa Rosa, CA, 2002, 1–4.
148. Snyders, R.; Gouttebaron, R.; Dauchot, J. P.; Hecq, M. Increasing the Deposition Rate of Oxide Films by Increasing the Plasma Reactivity. Surf. Coat. Technol. 2005, 200,
448–452.
149. Musil, J.; Herman, D.; Sicha, J.; Baroch, P. High-Rate Reactive Magnetron Sputtering of Al2O3 films. In Proceedings of the 17th International Symposium on Plasma
Chemistry; Mostaghimi, J.; Coyle, T. W.; Pershin, V. A.; Salimi Jazi, H. R., Ed.; Toronto, Canada, 2005; pp 130–131.
150. Sproul, W. D.; Chistyakov, R.; Abraham, B. Important Developments in High Power Pulsed Magnetron Sputtering. Bull., Soc. Vac. Coaters 2006, 35–37.
151. Sicha, J.; Herman, D.; Musil, J.; Stryhal, Z.; Pavlik, J. High-Rate Low-Temperature DC Pulsed Magnetron Sputtering of Photocatalytic TiO2 Films: The Effect of Repetition
Frequency. Nanoscale Res. Lett. 2007, 2 (3), 123–129.
152. Posadowski, W. M.; Wiatrowski, A.; Dora, J.; Radzimski, Z. J. Magnetron Sputtering Process Control by Medium-Frequency Power Supply Parameter. Thin Solid Films
2008, 516, 4478–4482.
153. Hoshi, Y.; Ishikara, D.; Sakai, T.; Kamiya, O.; Lei, H. High Rate Reactive Deposition of TiO2 Films Using Two Sputtering Sources. Vacuum 2010, 84, 1377–1380.
www.iran-mavad.com
154. Baer, D. R.; Engelhard, M. H.; Lea, A. S.; Nachimuthu, P.; Droubay, T. C.; Kim, J.; Lee, B.; Matthews, C.; Opila, R. L.; Saraf, L. V.; Stickle, W. F.; Wallace, R. M.;
Wright, B. S. Comparison of the Sputter Rates of Oxide Films Relative to the Sputter Rate of SiO2. J. Vac. Sci. Technol. A 2010, 28 (5), 1060–1072.
155. Anders, A. Deposition Rates of High Power Impulse Magnetron Sputtering: Physics and Economics. J. Vac. Sci. Technol. 2010, A28 (4), 783–790.
156. Musil, J.; Satava, V.; Baroch, P. Reactive Magnetron Sputtering of Thin Films and Advanced Nanocomposite Coatings. In 117th Technical Meeting of Division of Sputtering
and Plasma Processes of the Vacuum Society of Japan, Tokyo, Japan, Sputtering & Plasma Processes, 2010; Vol. 25 (1); pp 33–44.
157. Musil, J.; Baroch, P. High-Rate Pulsed Reactive Magnetron Sputtering of Oxide Nanocomposite Coatings. Vacuum 2013, 87, 96–102.
158. Rezek, J.; Vlcek, J.; Houska, J.; Cerstvy, R. High-Rate Reactive Deposition of Multifunctional Ta-O-N films Using High Power Impulse Magnetron Sputtering. In 10th
International Conference on Reactive Sputter Deposition (RSD 2011), Linkoping, Sweden, 8–9 December 2011.
159. Ishii, K. High-Rate Low Kinetic Energy Gas-Flow-Sputtering System. J. Vac. Sci. Technol. 1989, 7, 256–258.
160. Jung, T.; Westphal, A. High Rate Deposition of Alumina Films by Reactive Gas Flow Sputtering. Surf. Coat. Technol. 1993, 59, 171–176.
161. Kalber, T.; Jung, T. A Novel Low-Cost Process for the Deposition of Metallic and Compound Thin Films on Plastics. Surf. Coat. Technol. 1998, 98, 1116–1120.
162. Birkholz, M.; Albers, U.; Jung, T. Nanocomposite Layers of Ceramic Oxides and Metals Prepared by Reactive Gas-Flow Sputtering. Surf. Coat. Technol. 2004, 179,
279–285.
163. Ortner, K.; Jung, T.; Klages, C.-P.; Linder, B.; Strauss, B.; Samann, N. Dynamic PACVD Coating of Strip Metal with Silicon Oxide, Titanium Oxide and Amorphous Carbon
Films. Surf. Coat. Technol. 2005, 200, 976–979.
164. Bandorf, R.; Bloche, A.; Ortner, K.; Luthje, H.; Jung, T. High Rate Deposition of Magnetic Material by Gas Flow Sputtering. Plasma Process. Polym. 2007, 4 (3),
S129–S133.
165. Ortner, K.; Koessler, D.; Jung, T.; Jacobsen, H.; Quenzer, H.-J. Novel Method for High Rate Deposition of Lead Zirconate Titanate Piezoelectric Films. Plasma Process.
Polym. 2007, 4 (3), S134–S138.
166. Tang, S.; Schulz, U. Gas Flow Sputtering – An Approach to Coat Complex Geometries and Non Line Sight Areas. Surf. Coat. Technol. 2009, 204, 1087–1091.
167. Sproul, W. D.; Christie, D. J.; Carter, D. C. The Reactive Sputter Deposition of Aluminium Oxide Coatings Using High Power Pulsed Magnetron Sputtering (HPPMS). In
Proceedings of the 47th Annual Technical Conference; Society of Vacuum Coaters: Dallas, TX, 2004; pp 96–100.
168. Vlcek, J.; Pajdarova, A. D.; Musil, J. Pulsed DC Magnetron Discharges and Their Utilization in Plasma Surface Engineering. Contrib. Plasma Phys. 2004, 44, 426–436.
169. Vlcek, J.; Pajdarova, A. D.; Belsky, P.; Lukas, J.; Kudlacek, P.; Musil, J. Characterization of High-Power Pulsed DC Magnetron Discharges for Ionized High-Rate Sputtering
of Copper Films. In Proceedings of the 48th Annual Technical Conference; Society of Vacuum Coaters, 2005; pp 465–469.
170. Vlcek, J.; Kudlacek, P.; Burcalova, K.; Musil, J. High-Power Pulsed Sputtering Using a Magnetron with Enhanced Plasma Confinement. J. Vac. Sci. Technol. A 2007,
25 (1), 42–47.
171. Ehiasarian, A. P. Fundamentals and Applications of HIPIMS. In Plasma Surface Engineering Research and Its Practical Applications; Wei, R., Ed.; Research Signpost: Kerala,
2008; pp 35–86.
172. Bradley, J. W.; Welzel, T. Physics and Phenomena in Pulsed Magnetrons: An Overview. J. Phys. D 2009, 42, 093001.
173. Mishra, A.; Kelly, P. J.; Bradley, J. W. The Evolution of the Plasma Potential in a HiPIMS Discharge and Its Relationship to Deposition Rate. Plasma Sources Sci. Technol.
2010, 19, 045014.
174. Movchan, B. A. EB-PVD Technology in the Gas Turbine Industry: Present and Future. JOM Nov 1996, 40–45.
175. Morgner, H.; Neumann, M.; Straach, S.; Krug, M. The Hollow Cathode: A High-Performance Tool for Plasma-Activated Deposition. Surf. Coat. Technol. 1998, 108–109,
513–519.
176. Scheffel, B.; Metzner, Chr.; Goedicke, K.; Heinss, J.-P.; Zywitzki, O. Rod Cathode Arc-Activated Deposition (RAD) – A New Plasma-Activated Electron Beam PVD process.
Surf. Coat. Technol. 1999, 120–121, 718–722.
177. Movchan, B. A. Functionally Graded EB PVD Coatings. Surf. Coat. Technol. 2002, 149, 252–262.
178. Movchan, B. A.; Yakovchuk, K. Y. Graded Thermal Barrier Coatings Deposited by EB-PVD. Surf. Coat. Technol. 2004, 188–189, 85–92.
179. Movchan, B. A. Inorganic Materials and Coatings Produced by EB PVD. Surf. Eng. 2006, 22 (1), 35–45.
180. Plasma-Activated Electron Beam High-Rate Deposition, Bulletin of the Fraunhofer Institut Elekhtronenstral-und Plasmatechnik, Winterbergstrasse 28, 01277 Dresden,
Germany, 2007.
181. Biber, V.; Kuzmichev, A. Double EB Evaporation System for Self-Ion Enhanced PVD. Elektrotechnica & Elektronika (Bulgaria) 2009, 44 (5–6), 118–122.
182. Reinhold, E.; Richter, J.; Seyfert, U.; Wenzel, B. -D. New Aspects of EB-PVD for Large Area Coating of Metal Strips. In Proc. 43rd Ann. Tech. Conf. of the SVC, Denver,
2000; p 153.
183. Reinhold, E.; Faber, J. Large Area Electron Beam Physical Vapor Deposition (EB-PVD) and Plasma Activated (EB) Evaporation – Status and Prospects. Surf. Coat. Technol.
2011, 206, 1653–1659.
184. Brande, P. V.; Lucas, S.; Weymeersch, A. European Patent No. 0780489, Nov 1996.
185. Brande, P. V.; Weymeersch, A. Steel Coating by Self-Induced Ion Plating, A New High Throughput Metallization Ion Plating Technique. J. Vac. Sci. Technol. A 2000, 18 (4),
1555–1560.
186. Brande, P. V.; Lucas, S.; Weymeersch, A. Process for Formation of a Coating on a Substrate, United States Patent No. 6432281 B2, 13 Aug 2002.
187. Contino, A.; Feldheim, V.; Lybaert, P.; Deweer, B.; Cornil, H. Modelling of Continuous Steel Coating by Self-Induced Ion Plating (SIIP). Surf. Coat. Technol. 2005, 200,
898–903.
188. Musil, J.; Satava, V.; Baroch, P. High-Rate Reactive Deposition of Transparent SiO2 Films Containing Low Amount of Zr from Molten Magnetron Target. Thin Solid Films
2010, 519, 775–777.
189. Brauer, G.; Szyska, B.; Vorgohl, M.; Bandorf, R. Magnetron Sputtering – Milestones of 30 years. Vacuum 2010, 84, 1354–1359.
190. Billard, A.; Mercs, D.; Perry, F.; Frantz, C. Influence of the Target Temperature on Reactive Sputtering Process. Surf. Coat. Technol. 1999, 116–119, 721–726.
191. Kawamata, K.; Deguchi, T. Growth Rate Reproducibility for Magnetism Fluoride Films by Sputtering. J. Vac. Soc. Jpn 2003, 46, 624–627.
192. Mercs, D.; Perry, F.; Billard, A. Hot Target Sputtering: A New Way for High-Rate Deposition of Stoichiometric Ceramic Films. Surf. Coat. Technol. 2006, 201, 2276–2281.
193. Vlcek, J.; Zustin, B.; Rezek, J.; Burcalova, K.; Tesar, J. Pulsed Magnetron Sputtering of Metallic Films Using a Hot Target. In Proceedings of the 52nd Annual Technical
Conference; Society of Vacuum Coaters: Santa Clara, CA, 9–14 May 2009; pp 219–222.
194. Fischer-Cripps, A. C.; Bull, S. J.; Schwarzer, N. Critical Review of Claims for Ultra-Hardness in Nanocomposite Coatings. Philos. Mag. 2012, 92, 1601–1630.
www.iran-mavad.com
www.iran-mavad.com
4.14 Protective Transition Metal Nitride Coatings
PH Mayrhofer, Vienna University of Technology, Wien, Austria
R Rachbauer and D Holec, Montanuniversität Leoben, Leoben, Austria
F Rovere, Montanuniversität Leoben, Leoben, Austria and RWTH Aachen University, Aachen, Germany
JM Schneider, RWTH Aachen University, Aachen, Germany

4.14.2 Transition Metal Nitride Coatings 357
4.14.2.1 General 357
4.14.2.2 Bonding and Properties 357
4.14.2.3 Al-Containing Transition Metal Nitrides 360
4.14.2.3.1 Coatings within the Ti–Al–N System 361
4.14.2.3.2 Coatings within the Cr–Al–N System 361
4.14.3 Thermal and Chemical Stability of Nitride Coatings 362
4.14.3.1 Phase Stabilities and Mechanical Properties during Annealing 362
4.14.3.1.1 Ti–Al–N System 362
4.14.3.1.2 Cr–Al–N System 364
4.14.4 Oxidation and Oxidation Protection 365
4.14.4.1 General 365
4.14.4.2 Transport Mechanisms 365
4.14.4.3 Rate Laws 366
4.14.4.4 Scale Adhesion and Separation 368
4.14.4.5 Reactive Element Effects 369
4.14.4.6 Cr2O3–Al2O3 370
4.14.5 Simultaneous Investigation of Nanostructure and Chemistry 371
4.14.5.1 General Considerations 371
4.14.5.2 Atom Probe Tomography and Complementary Techniques 372
4.14.5.2.1 APT of Ti–Al–N Coatings 374
4.14.6 Multinary Transition Metal Aluminium Nitrides 375
4.14.6.1 Modeling of Ti–Al–TM–N Alloys and Comparison to Experiments 376
4.14.6.1.1 Applied Modeling Methodology for Ti–Al–TM–N Alloys 376
4.14.6.1.2 Impact of TM-Alloying on the Phase Stability of Ti–Al–N Coatings 377
4.14.6.1.3 Thermal Stability of Ti–Al–TM–N Coatings 378
4.14.7 Summary and Outlook 382
Acknowledgments 383
References 383
4.14.1 Introduction
As transition metal nitrides are nowadays applied as structural and functional materials in a variety of applications, we attempt to
summarize the current understanding of mechanisms relevant for the thermal and chemical stability of these coatings based on
state-of-the-art experimental and computational data.
Surface and coating technologies employed to obtain materials with properties that are difficult or impossible to achieve for bulk
materials have been a cornerstone of technological development for a long time. According to Pliny the elder (1), the production of
mirrors by application of thin layers of metal to glass sheets, instead of polishing bulk metal surfaces, dates back to the first century
AD. Paint coatings made of tar or other hydrophobic substances were commonly used to provide protection for the wooden hulls of
ancient ships against the chemically aggressive seawater. Yet another example is the application of thin layers of gold as decorative
coatings on top of less precious base materials, a technique which has been practiced for millennia (2,3). Being an effective barrier
against environmental attack, gold coatings assured the persistence of countless cultural treasures to our days. Research and
development of modern application tailored materials implements a tremendous need for a detailed understanding of the intrinsic
material properties and suitable combinations. Unlike diamond, which is still an extremely valued material for many different areas
(not only as jewelry) highest purity and single crystallinity of materials does not always yield optimized properties. Actually, for the
majority of engineering and protective materials this perspective fails, as many important material properties as, e.g., hardness and
ductility (especially if they need to be combined), or electrical, optical, and thermal properties are determined by the incorporation

www.iran-mavad.com
356 Protective Transition Metal Nitride Coatings
of ‘imperfections,’ and tremendously depend on the respective length scale of such ‘defects.’ A good example for the critical
importance of extended defects for the mechanical properties is ‘Damascus steel’ (4,5). Nowadays material design proposals,
especially for nanomaterials, have to include the defect structure. Here, classical engineering concepts are often insufficient for
a description of the resulting properties (6–8).
In the course of technological advance, the possibilities of surface modification in terms of application techniques and
available materials have vastly increased and are today often crucial for innovation. Here, in particular thin films – i.e., material
layers that range from a couple of nanometers to several micrometers in thickness – have been a technological driving force in the
last decades.
Faster, hotter, and lighter requirements constitute challenges for the materials engineer. To realize faster machining speeds in
materials processing industries, higher temperatures in combustion processes, or weight reductions of moving components in
energy conversion applications such as turbine blades, material solutions have to be designed. Hence new materials enable and
contribute to improving the technological and economic efficiency of industrial processes, and thereby also promote a more
efficient and thus more environmentally friendly utilization of resources. However, in most cases these industrial demands heavily
depend on advances in materials development, and increasingly, these challenges can only be met with innovative material
solutions such as composite material concepts including suitable combinations of bulk materials and coatings.
The achievement of particular optical and electronic properties in micro- and optoelectronic industries, the provision of envi-
ronmental protection or wear resistance, e.g., in materials processing, or still simply decorative purposes are only some examples of
the industrial demands that are today met by thin film technologies. State-of-the-art thin film technology benefits from the unique
size-dependent properties at the small end of the scale axis, resulting in a huge range of diverse applications from electronic, optical,
and medical, toward protective devices. Often low coating synthesis temperatures (e.g., for plasma-assisted vapor deposition
techniques the substrate temperature is typically below 0.2 of the melting point in Kelvin) are used (9), leading to limited ad-atom
kinetics during deposition. With the introduction of such nonequilibrium thin film growth techniques like sputtering, the tech-
nological possibilities were further enriched by the ability to synthesize metastable materials that often surpass the physical and
chemical properties of their stable constituents (10). Especially in the field of hard protective coatings, the group of transition metal
nitrides (TMN) attracts industrial interest because of their high melting point, high hardness, and fairly good oxidation and
corrosion resistance. Here, the development of Al-containing TMN led to a further improvement of their already excellent properties
(10–15).
The thermal stability of TMN coatings is strongly affected by the defect structure as recovery (e.g., defect annihilation and atomic
rearrangement resulting in lower stresses), interdiffusion, recrystallization, or phase transformation (13–15) may occur during
annealing. These phenomena are technologically relevant since the resulting structure (on both a micro- and nanoscale) has a large
impact on the film properties. Therefore, the development of advanced hard thin films can be based on phase transformation
concepts from physical metallurgy, such as primary (vapor to solid) and secondary (e.g., precipitation) phase transformations. Here,
a fundamental understanding of the intrinsic properties of single-layer TMN becomes inevitable, being the key for materials
selection and application tailored coating synthesis with respect to more sophisticated architectures, like nanocomposites or
nanolaminates (16).
This chapter thereby concentrates on the development of a state-of-the-art understanding of metastable phases in nitrides,
their decomposition processes to reach equilibrium, and their interaction with oxidative environment. Investigations of the
mechanisms that define the thermal stability of TMNs have to include atomic level information. A combinatorial approach was
chosen, involving theoretical and experimental work, to study the thermally induced isostructural decomposition of Ti1xAlxN
coatings. The spinodal decomposition of Ti1xAlxN hard coatings implements a substantial hardness increase with annealing
temperature or time, respectively, commonly referred to as age hardening. This particular feature is believed to bear a high
potential with respect to the enhancement of tool lifetime, which implements economical but also increasingly environmental
considerations. The goal of improving the thermal stability and oxidation resistance is currently approached by the investigation
of the alloying effect of transition metals (TMs) to Ti1xAlxN. Thereby, density functional theory (DFT) calculations represent
a well-established tool for the prediction or explanation of structure and properties of nanometer-sized features, where the
properties are determined (or critically influenced) by the electronic structure of the solid (17). For application tailored materials
design it is crucial to acquire basic knowledge on the micro- and nanostructural evolution of the metastable phases as a function
of temperature and time.
Cr1xAlxN-based hard coatings offer excellent oxidation and corrosion properties due to the formation of adherent and
protective oxide scales. Therefore, they are considered promising protective coatings for numerous applications in cutting tool,
automotive, and aerospace industries. A systematic investigation of the possibility to further improve the oxidation resistance of
Cr1xAlxN-based thin films by incorporation of the reactive trace element yttrium is a partial discussion of this chapter. Yttrium is
commonly used for improving the oxidation resistance of metallic MCrAlY (M ¼ Ni, Co, Fe) coatings (18) and can also beneficially
affect mechanical properties as it promotes grain refinement (19–21).
In the following sections the emphasis is centered on providing an introduction to the field of TMNs in general. We will use the
model system Ti1xAlxN to discuss the process of spinodal decomposition, which is responsible for the observed hardness increase
during annealing, and the model system Cr1xAlxN to discuss mainly the oxidation processes of such ternary TMNs. For the example
of Y-alloyed Cr1xAlxN coatings we will demonstrate the reactive element effect for such ternary TMNs. Finally, the influence of
alloying elements such as Y, Nb, Ta, Zr, and Hf on the phase formation, structure, mechanical, and thermal properties is discussed
for the model system Ti1xAlxN.
www.iran-mavad.com
Protective Transition Metal Nitride Coatings 357
4.14.2 Transition Metal Nitride Coatings

4.14.2.1 General
Due to attractive mechanical and chemical properties such as high hardness, good abrasive and sliding wear resistance, and
high temperature stability, as well as oxidation and corrosion resistance, TMN thin films have found widely spread techno-
logical use. Typical applications involve wear protective coatings for tools, such as cutting inserts, dies, or molds, or corrosion
and abrasion resistant layers in aerospace and automotive applications. Furthermore, TMNs are used as diffusion barriers, for
instance in microelectronics, or particularly TiN, simply as decorative coatings. While binary TMNs are still widely used, thin
film deposition techniques such as physical vapor deposition, where thermodynamic equilibrium is seldom obtained, allow for
the synthesis of superior metastable coating systems such as the ternary Al-containing cubic TMNs as discussed below. In
general, the increasing industrial demand for advanced coating systems with tailored properties has motivated the continuous
development toward quaternary and multinary TMN-based thin film systems. Besides chemical advances, microstructural
engineering has also yielded significant improvements of their properties, such as higher hardness via layered, two-dimensional
(superlattices) or three-dimensional (3D) nanostructures (nanocomposites, precipitates) formed during deposition or during
annealing (22–33).
4.14.2.2 Bonding and Properties

The attractive mechanical and chemical properties of TMNs mainly arise from the strong covalent component to their chemical
bonding, which can generally be considered as a mixture of metallic, ionic, and covalent contributions. Figure 1 gives an
overview on the classification of hard ceramic materials according to their chemical bonding and the associated change in
properties. TMNs belong to the group of metallic hard materials. The distinct metallic contribution causes electrical conduc-
tivity, as well as a good adhesion and considerable ductility (as compared to ionic hard materials), in addition to the high
hardness and phase stability, caused by covalent and ionic bonding components. The charge transfer from metal to nonmetal
species associated with ionic bonding is however relatively small, as compared to ionic hard materials such as alumina or
zirconia (34–37).
Stoichiometric early TMNs with a metal/nonmetal ratio of 1:1 crystallize in the face centered cubic (fcc) B1 structure (space
group Fm-3m, NaCl prototype). This is the case of Sc-, Ti, V-, Cr-, Y-, Zr-, Nb-, Hf-, and TaN, i.e., 3d-5d TMNs of groups IIIB-VB (VIB).
NbN and TaN are borderline cases as for these compounds – depending on the deposition conditions – either B1 or hexagonal B81
(NiAs prototype) and Bh (WC prototype) is obtained, respectively. MoN and WN, on the other hand, preferentially crystallize in
hexagonal phases (WN is even mechanically unstable in the cubic B1 modification). Finally, the stable form of AlN is hexagonal
close packed (hcp) B4 structure (space group P63mc, wurtzite prototype) (30,38–42). For the Al-containing cubic TMNs discussed
below, particularly the B1 structure is of industrial interest as it is associated with superior film properties, such as higher hardness
and better wear and oxidation resistance (43,44). Consequently, the bonding of this structural configuration is discussed here in
more detail.
The electron population of states at the different energy levels depends on structure and composition of the respective TMN
compound. In the rock-salt (B1) structure, the TM is octahedrally coordinated by the nitrogen ligands. The d-electron states centered
Figure 1 Classification of hard ceramic materials according to their chemical bonding and the corresponding change in properties. Modified after
Mayrhofer, P. H.; Mitterer, C.; Clemens, H. Adv. Eng. Mater. 2005, 7 (12), 1071; Holleck, H. In Surface Engineering: Science and Technology I; Kumar, A.;
Chung, Y. W.; Moore, J. J.; Smugeresky, J. E., Eds.; The Minerals, Metals and Materials Society, 1999; p 207; and Holleck, H. Surf. Coat. Technol. 1990,
43–44, 245.
www.iran-mavad.com
on a metal site can therefore be classified into orbitals with t2g and eg symmetry, the former pointing toward the faces and the latter
pointing toward the corners of the octahedron formed by the nitrogen atoms, see Figure 2. The density of states, as shown in
Figure 3 for TiN, allows discussion of the individual bonding contributions of interacting N and TM electrons.
While the core electrons at energies far below the Fermi level, EF, do not contribute to bonding, the region between w10
and w2 eV can be understood as the bonding states responsible for the covalent character of the compound. The N p-orbitals
interact with the TM d-orbitals, forming a so-called sp3d2 hybridization, which results in an eg symmetry (electron localization along
<100> directions in Figure 4(a)) in fcc crystals. The states from w2 to EF, referred to as the metallic region, however, mainly host
Ti d-electrons. Since the TM d-states localized in this region can be regarded as less strongly bound compared to the energetically
lower bonding states, the corresponding states are understood to mostly account for the metallic bonding character of TMN. The
corresponding electrons of this energy level are predominantly localized along the <110> directions of cubic crystals and form the
t2g symmetry (Figure 4(b)). The bonding sp3d2 hybridized eg states can accommodate six electrons. Consequently, already for 3d-
TMNs two metal-s, one metal-d, and three nitrogen-p electrons account for band filling. Any further valence electron has to populate
the energetically less favorable orbitals with t2g symmetry, or the antibonding states with eg symmetry (45–48).
A local charge enrichment or depletion within the metallic region, e.g., in (non)stoichiometric TMN or due to TM alloying, can
result in a collapse of the structure and the formation of different hybridization schemes (compare Section 4.14.2.3 and 4.14.5.2 as
well as (49–52)). The structural transformation from cubic to hexagonal goes along with a change from a sp3d2- toward the
sp3-hybridization, as apparent in w-AlN, where only Al s- and N p-states remain to interact.
Contrary to the transition of bonding to antibonding states, which often is connected with a well-defined minimum in the
electronic density of states (DOS), the nonbonding or antibonding character cannot be easily distinguished as the corresponding
states essentially overlap in the electronic energy spectrum, see Figure 5. However, for IIIB, IVB, and VB-TMNs, where the TM
exhibits more than one d electron, nonbonding states become populated, which causes an overall decrease in phase stability with
increasing valence of the TM (45,47,53,54). This is consistent with ab initio calculated decreasing energies of formation Ef at zero
Figure 2 Schematic covalent bonds between metal (grey spheres) d and nonmetal (red spheres) p orbitals in the (100) plane of the rock-salt
(B1, NaCl prototype) structure. Modified after Schwarz, K. Crit. Rev. Solid State Mater. Sci. 1987, 13, 211.
Figure 3 Projected density of states (DOS) for fcc-TiN. The electrons interacting in the hybridized region between w8 and w2 eV reflect the
covalent character of the bonding. Higher states up to EF mostly contribute to the metallic bonding of the binary TMN.
www.iran-mavad.com
Figure 4 Ab initio obtained local charge distribution in c-TiN. Interaction of N p- and Ti d-electrons occurs either along (a) the <001> directions
(eg symmetry) while Ti d- and Ti d-electrons interact in (b) the <110> directions (t2g symmetry) in fcc crystals of TMN. The overlap of the different
bonding contributions is responsible for the covalent-metallic properties of TiN.
Figure 5 Ab initio calculated density of states for B1 CrN; bonding and antibonding states are separated by a pseudogap. Population of nonbonding
or antibonding states shifts the Fermi level above the pseudogap.
Kelvin, and the general trend of reported decreasing decomposition temperatures with increasing valence of the TM, see Table 1. The
energy of formation Ef is defined as
Ef ;TMN ¼ Utot;TMN Utot;TM Utot;N ; [1]
where Utot,TMN, Utot,TM, and Utot,N designate the total energies of the TMN, the corresponding pure TM in its equilibrium crystal
structure, and a nitrogen dimer, respectively. The total (internal) energy in turn is the sum of all electronic contributions to the
formation of a compound. Ab initio calculations are discussed in more detail in Ref. (55).
The increase in decomposition temperature and the corresponding small decrease in Ef from YN to ZrN (Table 1) may be
explained by a stronger metal–metal (dds) interaction in ZrN. For pure metals, the cohesive energy varies across the periodic table
of elements and peaks around VB- and VIB-TMs, which is mirrored by the maximum melting temperatures of these TMs. This
www.iran-mavad.com
Table 1 Ab initio calculated energies of formation, Ef, together with decomposition reported temperatures
TMN Ef (eV/atom) (49) Tdec ( C) (40,57)
ScN 2.985 Unknown

TiN 2.704 3290
VN 1.980 2100
CrN 1.413 1050
YN 2.742 2674
ZrN 2.751 3410
NbN 1.990 1370
Reproduced from Rovere, F.; Music, D.; To Baben, M.; Fuss, H.-G.; Mayrhofer, P. H.; Schneider, J. M. J. Phys. D Appl. Phys. 2010, 43 (3), 035302.
implies an increasing metal–metal interaction from IIIB- to VIB-TMNs. On the other hand, as nonbonding states become populated
the metal–metal bond length is increasingly stretched from IIIB to VIB-TMNs and the metal–metal interaction is weakened as
compared to the pure metals. However, the very small difference of only <4.5% (48) in the Zr–Zr bond lengths between the metallic
Zr bulk phase and in its nitride can explain that the increasing metal–metal interaction in fact yields a slightly larger phase stability
of ZrN as compared to YN. Contrariwise, for NbN the population of nonbonding states stretches the metal–metal bond length by
9.2% (48) as compared to the pure metal. Thus, the phase stability decreases significantly (48,56).
It has to be noted that none of the mentioned stoichiometric mononitrides actually exhibits the theoretically assumed ideal B1
structure. In practice, deviations from the ideal stoichiometry always occur, and many nitrides are in fact largely nonstoichiometric
due to considerable vacancy concentrations. Cubic (B1) TiN, for instance, is stable over a broad composition range and exhibits
vacancy concentrations up to very high values, where vacancies can exist either on metal or on nitrogen lattice sites. If certain
threshold values are exceeded, vacancy ordering gives rise to the formation of phases such as Ti2N in the Ti–N system, V32N26 in the
V–N system, or Cr2N in the Cr–N system (13).
The defect structure of TMNs is strongly influenced by the synthesis, which consequently has a strong effect on the diffusivity. In
general, nitrogen diffusion, which occurs via a vacancy mechanism, is significantly faster than the diffusion of the metal, which also
happens via a vacancy mechanism (13,58–60). For that reason TMNs are commonly used as barriers against metal diffusion,
whereas they cannot effectively stop nitrogen diffusion (28,30,61–65).
4.14.2.3 Al-Containing Transition Metal Nitrides

The partial substitution of the TM by Al, originally motivated by the expected formation of a stable alumina layer during thermal load,
was first investigated for TiN by Knotek et al. (10) and Münz (11). In addition to an improved oxidation resistance, the synthesis of
compounds with different electronic configuration and different atomic radii was expected to lead to conditions that can be observed
in solid solution-hardened materials. Indeed, Ti1xAlxN films turned out to exhibit superior hardness, oxidation resistance, and wear
behavior as compared to binary TiN films (10,11,43,66–71). For equilibrium conditions, the solubility of AlN in B1 TiN is very
limited (<2 at% at 1000 C (72)), as TiN crystallizes in the B1, NaCl-prototype structure, whereas AlN crystallizes in the B4, wurtzite-
prototype structure. Nevertheless, physical or chemical vapor deposition processes, such as magnetron sputtering, plasma-assisted
chemical vapor deposition, or various arc evaporation techniques, allow for the synthesis of supersaturated solid-solution Ti1xAlxN
films with AlN mole fractions as high as x ¼ 0.9 (73). The synthesis of such quasibinary metastable films of the form TM1xAlxN is
possible by the kinetic limitations caused by high deposition rates in combination with extremely high cooling rates (w1013 K s1).
In other words, the quenching from the vapor phase does not allow the condensing atoms to reach a stable configuration charac-
terized by a low free energy of formation (74). The supersaturated cubic structure then remains stable due to the large activation
energy required for phase separation. This energetic barrier is mainly made up by the elastic strain energy caused by the formation of
hexagonal AlN precipitates within the cubic TM1xAlxN matrix and the energy necessary to form new surfaces (interfaces, boundaries)
(75). Here, in particular, the strain energy due to the large volume mismatch of about 20% (76) between the B1 and B4 structural
modifications of AlN is an obstacle against chemically driven phase separation (75). The latter however inevitably occurs when at
elevated temperatures diffusion is sufficiently activated and the resistance against plastic deformation is reduced.
Vapor deposition phase fields predicting the stability range for various pseudobinary nitride systems with respect to temperature
and AlN content were first defined by Holleck (74). Since then considerable experimental and theoretical research efforts have been
undertaken to investigate the phase stabilities of TM1xAlxN systems. Here, in particular, Ti1xAlxN (24,77–85) and Cr1xAlxN
(43,86–94) have been studied, but also Zr1xAlxN (44,95–98) or Nb1xAlxN (99,100) receive increasing attention.
In general, the mechanical and chemical properties of TM1xAlxN films improve with the Al content. However, this is only
possible up to a certain critical Al content xcrit, where transformation of the cubic B1 structure into the hexagonal B4 structure occurs
(Figure 6). The onset of B4 crystallization initially causes the simultaneous crystallization of B1 and B4 phases in the experimentally
observed transition region before at higher Al contents only the B4 structure is obtained. Generally, the onset of B4 crystallization is
associated with undesirable coating performances, as it causes a drop in hardness and wear resistance (43,67,80). The occurrence of
a two-phase structure was furthermore shown to cause a decline in oxidation resistance (43,101–103). The critical Al content differs
for the different TMs in TM1xAlxN, which denotes varying driving forces for phase separation in dependence on the TM.
www.iran-mavad.com
Figure 6 Schematic of the structural evolution in TM1xAlxN in dependence on the AlN mole fraction: transition from the fcc B1 structure into the
hcp B4 structure occurs at a critical Al content, which varies for different TMs.
4.14.2.3.1 Coatings within the Ti–Al–N System

Due to the extremely limited kinetics (condensation from the vapor phase can be assigned with cooling rates up to 1013 K s1 (104))
apparent in PVD deposition techniques it is possible to synthesize a supersaturated solid solution of Ti1xAlxN (10,11,105–109),
although c-TiN and w-AlN exhibit essentially no solubility for each other; see the schematic quasibinary phase diagram in
Figure 7(a) (110). The ab initio predicted energies of formation, Ef, for cubic (c or cub) and wurtzite (w or wur) solid solutions of
the resulting metastable Ti1xAlxN films exhibit an energetically preferred cubic structure up to x w 0.6–0.7, see Figure 7(b). In the
region where Ef exhibits almost similar values for cubic and wurtzite structures, a dual phase regime is preferred, followed by
a single-phase wurtzite regime for higher Al contents (79,81,98,111,112). The corresponding metastable phase diagram, adapted
from experimental observations by Spencer (113), is shown for Ti1xAlxN in Figure 7(c).
Ti1xAlxN can be regarded as a model system in materials science of thin films, as it was shown that hardness, oxidation
resistance, and thus thermal stability increases with the Al content, x, within the cubic regime (24,79,80,103,111,114–119). Ternary
Ti1xAlxN coatings form a dense stable Al2O3 oxide scale on top of the film, which effectively reduces the oxidation rate.
The incorporation of Al into other binary TMNs is quite similar with respect to the development of certain element-specific
(meta)solubility limits and superior properties of the ternaries compared to the binary nitrides, as valid also for TM ¼ V, Cr, Zr,
Nb, Hf, Ta, etc. (43,70,95,96,98,99,120–125). With increasing Al content, however, the driving force for decomposition also
increases drastically, which results in extraordinary properties at elevated temperatures (the topic of Section 4.14.3).
4.14.2.3.2 Coatings within the Cr–Al–N System

Ternary Cr1xAlxN- or multinary Cr1xAlxN-based thin films, where Al substitutes for Cr, are of increasing industrial importance due
to good mechanical properties in combination with excellent corrosion and oxidation resistance (43,70,86,120,126–128). There-
fore, they are widely used as abrasion-resistant layers in automotive industry (e.g., valves, tappets, and camshafts) or as protective
coatings for forming and machining tools (94,129–134). Whereas Ti1xAlxN-based films generally exhibit higher hardness values
and benefit from pronounced age hardening during service (22,34,135,136), Cr1xAlxN-based coating systems show a significantly
better oxidation resistance at elevated temperatures (130,137). This is due to the fact that for Cr1xAlxN films, both Cr and Al form
protective oxides, which together lead to the formation of protective mixed Al2O3 þ Cr2O3 scales (138–140). Ti-based nitride thin
Figure 7 (a) Quasibinary phase diagram along the tie line of TiN–AlN (redrawn from Holleck, H. Surf. Coat. Technol. 1988, 36 (1–2), 151). (b) Ab initio
obtained energies of formation, Ef, for the cubic (c or cub) and wurtzite (w or wur) solid solutions of Ti1xAlxN (adapted from Mayrhofer, P. H.; Music, D.;
Schneider, J. M. J. Appl. Phys. 2006, 100 (9), 094906). (c) Metastable PVD-phase diagram (reproduced from Spencer, P. Int. J. Mater. Res. 2001, 92 (10),
1145) based on experiments, confirming the ab initio predicted phase limits.
www.iran-mavad.com
films, on the other hand, commonly form layered or mixed Al2O3 þ TiO2 scales (66), where the presence of TiO2 limits the overall
protection capability (see also Section 4.14.4).
Cr1xAlxN films can be synthesized in the cubic B1 structure up to AlN mole fractions of 75% (x ¼ 0.75), before at higher mole
fractions the B4 structure crystallizes (43,87,89,90,141). Based on ab initio calculations (92,142), thermodynamic modeling (91), or
the calculation of band parameters (143), maximum Al contents of x ¼ 0.48–0.815 are obtained. The lower end of predicted values
originates from calculations based on a large number of randomly chosen (ad hoc) supercells, which were conducted to illustrate
the effect of varying metal sublattice populations on the phase stability of these materials (92). Configurational entropy was hence
suggested as a possible reason for the experimentally observed spread of maximum achievable Al contents for the B1 structure.
Another reason is different stress states corresponding to various deposition conditions, as demonstrated in Ref. (142). However, the
experimentally obtained critical Al contents of x ¼ 0.68 and 0.69 (see Refs. (49,50)) and the value of x ¼ 0.75, as obtained by ab
initio calculations (50), agree well with experimentally and theoretically obtained values reported elsewhere.
4.14.3 Thermal and Chemical Stability of Nitride Coatings
In general, the synthesis of thin films via ‘cold’ deposition processes such as physical vapor deposition causes the formation of
metastable materials far away from their thermodynamic equilibrium. This is particularly true for the supersaturated quasibinary
systems discussed above, but also for binary film systems. At any rate, the energetic particle bombardment during film growth causes
a high defect density in the form of a large number of point defects (interstitial atoms or vacancies) or defect clusters, which give rise
to considerable intrinsic (growth-induced) compressive stresses (15). Consequently, there usually exist large driving forces for
atomistic and structural changes to cause recovery (e.g., relaxation of stress due to defect annihilation and atomic rearrangement) or
recrystallization processes, as well as the phase separation into stable constituents for supersaturated films. These diffusion-driven
processes are activated as soon as enough energy is provided at sufficiently elevated temperatures. Such conditions can occur for
instance during machining. Commonly, relaxation processes will take place upon annealing at temperatures in excess of the
deposition temperature (144,145). Then, point, line, and area defects change into their low energy configurations by defect
redistribution or annihilation, which globally decrease the number of defects. This subsequently causes a significant decline in
hardness, as fewer defects act as obstacles for dislocation movement (13,15,22,146).
For metastable supersaturated films, on the other hand, the decline in hardness due to recovery and recrystallization processes is
counteracted by age-hardening phenomena during the early stages of decomposition. Age-hardening is either caused by the formation
of coherent phases via a spinodal mechanism (147), as it is observed for Ti1xAlxN films (22,34,135,136), or by the precipitation of
small incoherent hcp AlN particles, as it occurs in Cr1xAlxN films (148–150). Small, coherent precipitates hinder dislocation
movement due to the additional energy necessary to cut through the precipitate (cutting), whereas small incoherent particles force
dislocations to bow and curl around them (Orowan mechanism) (6,151). For this reason, cubic (B1) Al-containing metastable
TM1xAlxN films generally exhibit superior thermomechanical properties as compared to their binary counterparts (10,11,43,66–71).
In practice, i.e., during thermal loading in an oxidizing atmosphere, the thermal stability of nitride thin films is governed not
only by recovery, relaxation, or phase separation processes, but also by the oxidation of the material. This particularly important
topic is discussed in more detail in Section 4.14.4. Nevertheless, the properties, as well as the decomposition and oxidation kinetics
of the foremost described Cr1xAlxN-based thin films, are reviewed in Sections 4.14.3.1.2 and 4.14.4.6.
4.14.3.1 Phase Stabilities and Mechanical Properties during Annealing

4.14.3.1.1 Ti–Al–N System
The Ti1xAlxN system exhibits a wide miscibility gap at low temperatures, Figure 7(a), because the preparation of metastable phases
and outstanding performance at high temperatures is possible (75). A simplified picture to understand the supersaturated solid
solution of cubic NaCl-structured Ti1xAlxN (for x w< 0.66) assumes two sublattices: One is fully populated by nitrogen while the
other sublattice is randomly occupied by composition-dependent Al and Ti atoms (78). This assumption enables for the calculation
of Ef and Hmix (being the mixing enthalpy) of the ternary compound, which describe the phase stability of predefined structures.
Also, the driving force for decomposition into the (meta)stable compounds can be expressed as function of Al content, x, and
temperature (see Figure 8) (85,110,152).
Two different regions, called binodal and spinodal, can be identified in the composition–temperature phase diagram according to
Refs. (78,83,85); see Figure 8(a). In the region of the miscibility gaps, where the second derivative of the Gibbs free energy is positive
(v2 G/vc2 > 0), the decomposition of Ti1xAlxN takes place by nucleation and growth (e.g., Ti1xAlxN / c-TiN þ w-AlN), requiring
an activation energy for the formation of w-AlN precipitates. However, within a second region covered by the spinodal line, where
v2 G/vc2 < 0, already small composition fluctuations result in isostructural decomposition toward c-TiN and c-AlN if kinetics allow
(75,147,153). This implies that within the spinodal region, essentially no energy barrier for decomposition exists, apart from thermal
activation enabling for diffusion. The thermally induced spinodal decomposition of such alloys gained great importance in industry,
as it results in an increased hardness. This effect is often referred to as age hardening, due to its dependence on time and temperature.
The positive Hmix over the whole composition range indicates a high driving force for decomposition of c-Ti1xAlxN into the cubic
boundary phases c-TiN and metastable c-AlN (Figure 8(b)). With increasing compressive pressure, this driving force significantly
increases, while Hmix for the decomposition of c-Ti1xAlxN into c-TiN and w-AlN is effectively reduced, compare Figure 8(b) and 8(c).
www.iran-mavad.com
Figure 8 (a) Composition–temperature phase diagram of c-Ti1xAlxN (from Alling, B.; Ruban, A. V.; Karimi, A.; Peil, O. E.; Simak, S. I.; Hultman, L.;
Abrikosov, I. A. Phys. Rev. B 2007, 75, 045123). Mixing enthalpies of Ti1xAlxN as a function of Al content, x, at different pressures (adapted from
Alling, B.; Odén, M.; Hultman, L.; Abrikosov, I. A. Appl. Phys. Lett. 2009, 95 (18), 181906). (b) Hmix for the isostructural decomposition into c-TiN and
c-AlN. (c) Influence of pressure on Hmix with respect to c-TiN and w-AlN.
Materials exhibiting a large single-phase field, such as TiN, suffer from softening in the course of thermal load due to recovery,
recrystallization, and grain growth processes. In contrast, metastable c-Ti1xAlxN shows a substantial hardness increase with
increasing temperature, age hardening, as shown in Figure 9. It was observed that within the single-phase cubic regime, up to
temperatures of w900 C, TiN- and AlN-enriched domains form (22,136), whose locally different lattice parameters cause strain
hardening of Ti1xAlxN due to hindering of dislocation motion (15,153,154). If the chemical composition of stoichiometric AlN is
reached locally, the metastable c-AlN domains transform into the stable w-AlN phase. This process also depends on the local
conditions, as this phase transition is accompanied with a volume increase of w25%. Hence, the matrix needs to be ‘soft’ enough,
and the dislocation activity needs to be high enough, which is generally reached at higher temperatures. Coarsening of the structure,
the loss of coherency strains between c-TiN and w-AlN, which further exhibits inferior mechanical properties compared to c-AlN,
causes steadily decreasing hardness within the dual-phase regime (22).
In the past, computational (78,83,136,152,155) and experimental (22,135) investigations dealt with the effect of, e.g.,
composition, temperature, vacancies, isostatic pressure, internal strain, and surface energies on the extent of the driving force for
spinodal decomposition. In general, the driving force (Hmix) for decomposition into c-TiN and metastable c-AlN increases with
increasing Al content up to the metasolubility limit of the cubic phase field (compare Section 4.14.2.3). Also increasing compressive
stresses – corresponding to isostatic pressures in Refs. (142,152) – in the coating are supposed to promote the isostructural
decomposition, while on the other hand they act as a retarding factor for the precipitation of w-AlN with w25% larger specific
volume than c-AlN (75,136,142,156). Since another contribution to the decomposition driving force is related to the reduction of
Figure 9 Film hardness over annealing temperature for c-TiN and c-Ti0.34Al0.66N (adapted from Mayrhofer, P. H.; Hörling, A.; Karlsson, L.; Sjölén, J.;
Larsson, T.; Mitterer, C.; Hultman, L. Appl. Phys. Lett. 2003, 83 (10), 2049). The hardness increase with temperature is referred to as age hardening.
www.iran-mavad.com
total surface energy and the creation of new interfaces during phase separation, the overall driving force for decomposition is
difficult to determine (34,75) and needs sophisticated experimental proof. A comprehensive overview on the structural evolution of
Ti0.46Al0.54N thin films as a function of temperature and time exploits these issues in detail, see Refs. (157,158).
4.14.3.1.2 Cr–Al–N System

Depending on the structure of the metastable Cr1xAlxN films, either the precipitation of w-AlN (also abbreviated as hexagonal
h-AlN) for the case of a cubic matrix (x < xcrit) or the precipitation of c-CrN for the case of a hexagonal matrix (x > xcrit) initiates the
decomposition of the metastable compound into its stable constituents c-CrN and w-AlN (159), see Figure 10.
During annealing of metastable cubic structured Cr1xAlxN films at higher temperatures (>900 C), the decomposition
continues, as at atmospheric pressure CrN is thermally stable only up to w1000 C (38). Breaking of the thermally unstable Cr–N
bonds is associated with nitrogen loss and causes the formation of hexagonal h-Cr2N as an intermediate step before elemental Cr is
formed. Simultaneously, Cr1xAlxN decomposes into w-AlN, elemental body centered cubic Cr, and N2. For the case of the
industrially preferred cubic Cr1xAlxN films the decomposition sequence may hence be summarized as follows:
c-Cr 1x Alx N/h-Cr 2 N þ w-AlN þ ðgÞN2 /c-Cr þ w-AlN þ ðgÞN2 [2]
The breaking of Cr–N bonds furthermore depends on the nitrogen partial pressure (38). For Cr0.44Al0.56N films, the onset of
nitrogen loss was shown to increase by more than 100 C when annealing was conducted in N2 atmosphere rather than in inert Ar
atmosphere (160). Under atmospheric conditions, the onset of nitrogen loss falls in between these extreme cases but is additionally
superimposed by oxidation processes, see Ref. (140).
The structure evolution with Ta of single-phase cubic Cr1xAlxN coatings is further influenced by the driving force for decom-
position and microstructural features present. This has been studied in detail in Refs. (149,150), showing that single-phase cubic
Cr1xAlxN coatings with Al content close to the metastable solubility experience an earlier phase separation than coatings with lower
Al content. Hence, the coatings Cr0.44Al0.56N and Cr0.32Al0.68N (x ¼ 0.56 and 0.68) also exhibit a different hardness evolution with
Ta, see Figure 11. The cemented carbide (cc) substrate exhibits a hardness of 21 0.6 GPa, which is thermally unaffected up to
Figure 10 Schematic illustration of the individual phase fractions as a function of annealing temperature for (a) B1 c-Cr1xAlxN and (b) B4 h-Cr1xAlxN.
Modified after Mayrhofer, P. H.; Willmann, H.; Reiter, A. E. Surf. Coat. Technol. 2008, 202, 4935.
34
32
30
H (GPa)
28 Cr 0.44 Al 0.56 N
Cr 0.32 Al 0.68 N
26
24
25 100 300 500 700 900
T a (°C)
Figure 11 Hardness of Cr0.44Al0.56N and Cr0.32Al0.68N coatings on cemented carbide after annealing at Ta for 60 min in Ar atmosphere. Error bars
represent the standard deviation of 40 indents. Modified after Willmann, H.; Mayrhofer, P. H.; Hultman, L.; Mitterer, C. J. Mater. Res. 2008, 23 (11), 2880.
www.iran-mavad.com
1000 C. In the as-deposited state, the coatings Cr0.44Al0.56N and Cr0.32Al0.68N have comparable hardness values of H ¼ 30.0 1.1
and 30.1 1.3 GPa, respectively. For Cr0.44Al0.56N, the hardness only slightly decreases from 30.0 1.1 to 29.4 0.9 GPa with
increasing Ta from room temperature to 900 C, respectively. This corresponds well with the structural investigations, indicating
only minor changes during annealing up to 900 C (150).
The higher Al-containing coating, Cr0.32Al0.68N, exhibits a small increase in H from 30.1 1.3 GPa in the as-deposited state to
30.7 0.7 GPa with increasing Ta to 600 C, respectively. For a further increase in Ta to 725 C the hardness shows a more
pronounced increase to the peak value of 31.6 1.4 GPa, see Figure 11. The structural investigations of Al0.68Cr0.32N as a function
of Ta suggest that the temperature for this hardness increase coincides with the onset temperature for the formation and precipi-
tation of h-AlN. Consequently, the additional h-AlN phase formed provides further obstacles for dislocation movement. If
annealing is conducted at temperatures above 725 C, the hardness rapidly decreases to 27.5 1.0 GPa with increasing Ta to 900 C.
This H versus Ta behavior indicates that the formation of h-AlN results in an increased hardness of Cr0.32Al0.68N as long as the
additional phase is small in size, evenly distributed, and of only a small volume fraction. This is valid for Cr0.32Al0.68N if the
annealing temperatures are between 500 and 725 C. With increasing volume fraction and size of the h-AlN phase, having a w21%
lower specific packing-density and a w22% lower bulk modulus than c-Cr1xAlxN (92), the hardness decreases, as obtained for
annealing at Ta 750 C.
The lower Al-containing coating Cr0.44Al0.56N exhibits almost no changes in phase composition and structure for annealing
temperatures Ta below 900 C (150). Hence, also the hardness is nearly unaffected (except for recovery processes, which are
active as soon as the annealing temperature is above the deposition temperature) by annealing treatments below 900 C, see
Figure 11.
4.14.4 Oxidation and Oxidation Protection

4.14.4.1 General
Oxidation in the most general view is a chemical reaction where the oxidized species donates electrons to a second reactant, the
oxidizing agent, which by accepting them reduces itself. The reaction of metals or metallic compounds with oxygen at ambient
temperature, which is commonly referred to as corrosion, has a considerable impact on the world’s economy. In the United
States alone the financial damage connected to corrosion rises to as much as $276 billion per year (161). This is not surprising
if one considers that the standard free energies of formation of all non-noble metals are negative (162). Hence, for a structural
material in an oxygen-containing atmosphere, the question does not stand if oxidation occurs but rather on which time scale it
occurs.
At low temperatures, activation energies are generally large and oxidation reactions are slowed down to long time scales. The
situation is, however, quite different for elevated temperatures where activation energies are readily overcome and oxidation
processes can deteriorate metallic components in much shorter times.
If an oxide is not volatile, it remains at the surface of the metal and forms a continuous oxide scale. If this scale is furthermore
resistant to transport of species to the reaction zone it slows down the ongoing oxidation, where the slowest process becomes rate
determining. Engineering materials, which are designed to withstand high temperature exposure in aggressive environments, rely
consequently on the formation of continuous, adherent, and slowly growing scales. In practice, alloys that form scales based on
Al2O3, Cr2O3, or SiO2, as major constituents, are used to reach acceptable lifetimes of metallic components for high temperature
applications in oxidizing atmosphere (163–165).
4.14.4.2 Transport Mechanisms

High temperature oxidation resistance is based on the possibility of the reactants, namely the metal (alloy) and the gas, to form an
oxide scale that subsequently separates the reactants and thereby retards the transport of the reactants through it. Cr2O3, a-Al2O3,
and SiO2 scales exhibit the smallest relative growth rates as compared to other oxides such as NiO, CoO, or FeO. Consequently,
oxidation protection for engineering applications is in most cases based on alloys that form these scales (164).
In order for the oxidation reaction to proceed, either neutral atoms or ions and electrons must migrate through the
growing scale. Since all metal oxides are of ionic nature, it is reasonable to consider the former case as negligible. Figure 12
presents a simplified model for diffusion-controlled oxidation via ionic transport for the idealized case of a compact, perfectly
adherent scale.
Metal cations (Mnþ) and electrons (e) diffuse outward to the scale–gas interface whereas oxygen anions and electron holes (hþ)
diffuse inward toward the metal(alloy)–scale interface. A precondition for simultaneous migration of ions and electrons is that the
oxides formed during oxidation are nonstoichiometric compounds, which are essentially classified as semiconductors and may in
turn show negative (n-type) or positive (p-type) behavior. Negative semiconductors, where the electrical charge is transferred by
negative carriers, exhibit either an excess of metal or a deficit of nonmetal. For the case of metal excess, interstitial cations cause
electron conduction as thereby excess electrons are excited into the conduction band. For the case of a nonmetal deficit, excitation
of excess electrons in the conduction band also occurs, but the electroneutrality is maintained by vacancies on the anion sublattice.
p-type behavior, on the other hand, where charge transfer occurs through positive carriers, is caused by the formation of vacancies on
the cation lattice together with electron holes. The formation of electron holes is realized by changing the valence state of the metal.
www.iran-mavad.com
Figure 12 Simplified model of diffusion-controlled scale growth for metals. Mnþ designates cations, Vn cation vacancies, e electrons, hþ electron
holes, and O2 oxygen anions. Modified from Birks, N.; Meier, G. H. Introduction to High Temperature Oxidation of Metals. Edwar Arnold Ltd.: London, 1983.
Finally, very few oxides belong to the class of so-called intrinsic semiconductors, where the concentration of electronic defects far
exceeds that of ionic defects. In this case electrons are excited from the valence state to the conduction band, producing a quasifree
electron and a positive electron hole. Although oxides usually belong to one or the other group, it is reasonable to consider that at all
times all of the possible defects are present to some extent, with a certain type being predominant (164). However, for both cases,
n-type and p-type behavior, the vacancy concentration n is depending on the oxygen partial pressure and can be calculated after (166):
1=v 1=v
nzpO2 or nzpO2 ; [3]
for n-conductors and p-conductors, respectively. Here, v depends on the oxide composition and typically exhibits values between 2
and 8. Consequently, the defect structure of the oxide determines which species are mobile during oxidation, with the vacancy
concentration gradient being the driving force for diffusion. For the case of p-type behavior, for instance, the vacancy concentration
according to eqn [3] is highest at the scale–gas interface, which causes cation outward diffusion.
The exact local determination of the dominant point defect in Cr2O3 and a-Al2O3 oxide scales is difficult as the amount of
diffusing species is very small (both are highly stoichiometric oxides). Furthermore, the transport properties of these oxides are not
only microstructure and impurity sensitive but also depend on temperature and oxygen pressure (163,167). Thereof, the consid-
erable scatter in published diffusion data for Cr2O3 and a-Al2O3 may be understood (163,168–175).
Diffusion data for single crystal Cr2O3 reveals that oxygen diffusion is by 3 orders of magnitude smaller than Cr, which strongly
indicates predominant cation disorder and furthermore that Cr2O3 scales grow primarily via outward cation diffusion (172–175).
Kofstad and Lillerud (174–177) could show that the actual cation transport at the Cr/Cr2O3 interface (at low oxygen partial
pressure) occurs via an interstitial mechanism, which is consistent with investigations by Hindam and Whittle (163). At near
atmospheric oxygen partial pressures on the other hand (upper regions of the scales), Cr2O3 is an intrinsic electronic conductor
(175). Nevertheless, for temperatures T sufficiently below the melting temperature Tm (i.e., T < 0.6 Tm), the dominant transport
mechanism in growing oxide scales is so-called short-circuit diffusion, which embraces diffusion of metal or oxygen ions through
extended defects (dislocations, voids, grain boundaries) in the oxide scale rather than through the bulk crystal lattice. For most
practical situations, chromia (Tm ¼ 2424 C (38)), or alumina (Tm ¼ 2054 C (57)) scales grow in such a ‘low’ temperature regime,
when temperatures do not exceed w1600 or 1200 C, respectively (178). Caplan and Sproule (179) have shown that together with
predominant cation transport, inward diffusion of oxygen anions also occurs along scale grain boundaries. This is consistent with
more recent studies by Stroosnijder et al. (180). For chromia, furthermore, the possibility of vapor species transport along cracks or
voids was suggested to contribute to the overall growth kinetics (175,181).
For the above-mentioned reasons, there is a considerable scatter in the data reported for oxygen diffusion in a-Al2O3 single
crystals. However, in polycrystalline specimens, both Al and oxygen are found to be mobile (163,168–170,182). As the oxygen
transport is considerably enhanced by smaller grain size, the oxygen inward diffusion along grain boundaries seems to be the
dominant transport mechanism. Nevertheless, Prescott and Graham (183) pointed out that exclusive inward diffusion of oxygen
might be an oversimplification and that the role of cation diffusion cannot be omitted. Various studies have confirmed the notion
that both oxygen and cations contribute to the overall growth processes of the scales with oxygen inward diffusion along scale grain
boundaries being the dominant transport mechanism (184–187).
Contrary to chromia, which only exists in its stable corundum structure (a-Cr2O3), the situation for alumina is complicated by
the existence of several metastable fast-growing transition alumina next to the stable a-modification. The formation of metastable
alumina polymorphs (mainly the g- and the q-modifications) may be expected at temperatures below 900 C whereas for higher
temperatures slow-growing protective a-Al2O3 forms (188,189). The g- and the q-phase most likely grow via outward transport of
Al3þ ions, with the q-phase having a cation-vacancy network structure (190–192).
4.14.4.3 Rate Laws

When oxidation-prone materials are used in elevated temperature applications, given enough time, inevitably component failure
occurs. Obviously, it is vital to know how fast the oxidation process is going to be. As discussed in the previous section, this is
www.iran-mavad.com
controlled by the transport mechanism of species through the scale formed during oxidation, where the slowest process at all times
is rate determining. The rate of formation of oxide scales on a metal (M) according to typical oxidation reactions,
2M þ O2 / 2MO [4]
M þ O2 / MO2 [5]
2M þ 3O2 / M2 O3 [6]
can in principle be measured by several methods, involving the amount of metal consumed, the amount of oxygen consumed, or
the amount of oxide produced (mass gain). In practice, however, only the measurement of the mass gain and the amount of oxygen
used allow for continuous measurements, of which the former is more common, see Figure 13. By the measurement of the weight
gain Dm over time t, several rate laws can be identified. The linear law,
Dm ¼ kl $t [7]
where Dm, kl, and t designate the mass gain, the linear rate constant, and the time, is obtained when the rate of reaction is inde-
pendent of time. Such behavior occurs when a surface reaction step or the diffusion through the gas phase is the rate-limiting step.
As it is unlikely for very thin scales that ionic transport through the scale is rate limiting, a linear rate law can be expected for the
initial stages of oxidation. At this stage, processes such as adsorption, dissociation, chemisorption, and ionization of oxygen are rate
controlling.
The parabolic law,
ðDmÞ2 ¼ kp $t [8]
where Dm, kp, and t are the mass gain, the parabolic rate constant, and the time, is found to be obeyed when the diffusion through
the scale is the rate determining process and as such the desired behavior for protective oxide scales.
Finally, when metals are oxidized at low temperatures (<400 C) the initially rapid formation of a very thin scale and
a subsequently very low rate of reaction is found to conform to a logarithmic behavior (logarithmic law). The latter is only added for
completeness and is not considered further, as during high-temperature oxidation measurements usually only parabolic or linear
behavior is observed (164,165).
Furthermore, under certain conditions it is also possible that a superposition of different oxidation kinetics is obtained. Cr2O3,
for instance, initially shows parabolic behavior before transformation to linear oxidation kinetics occurs. This is due to the
formation of volatile oxide species most likely according to the reaction
3
Cr 2 O3 s þ O2 g / 2CrO3 g ; [9]
2
when Cr2O3 is heated to above 1000 C (172,181,193). Mass gain data, however, always represent the net change during oxidation.
Consequently, one must be aware of the fact that the interpretation of mass gain data alone without comparative investigational
methods such as x-ray diffraction (XRD) analysis or electron microscopy studies may lead to erroneous conclusions. Another
example for the simultaneous occurrence of different mass increasing and mass decreasing processes during oxidation is the
oxidation of metal nitrides (MeN), where nitrogen is released after, e.g.,
3
2MeN þ O2 /Me2 O3 þ N2 [10]
2
or
2MeN þ 2O2 /2MeO2 þ N2 : [11]
Figure 13 Measurement of oxidation rates and observed rate laws. Modified after Ashby, M. F.; Jones, D. R. H. Engineering Materials 1. Butterworth-
Heinemann: Oxford, 1996.
www.iran-mavad.com
As oxidation is a thermally activated reaction, oxidation rates follow Arrhenius’s law, i.e., the kinetic constants kl and kp increase
exponentially with temperature (165):
kl ¼ Al eQl =RT and kp ¼ Ap eQp =RT [12]
where Al and Ap designate constants and Ql and Qp designate activation energies. Typical ranges of reported apparent parabolic rate
constants obtained for Al2O3, Cr2O3, and TiO2 forming metallic compounds are summarized in Figure 14. Furthermore, oxidation
rates are obviously depending on the oxygen partial pressure in such a way that increasing oxygen partial pressures cause increasing
rates of oxidation (164,165).
Summarizing, for an oxide scale to be protective, a parabolic rate law is necessary, which as discussed above is obtained if
diffusion of species through the oxide scale is the rate-determining step of the oxidation reaction. Protective scales are hence
characterized by low diffusion coefficients that correlate with high melting points. This explains why a-Al2O3 (Tm ¼ 2054 C (57)),
Cr2O3 (Tm ¼ 2424 C (38)) and SiO2 (Tm ¼ 1726 C (194)) are protective scales, whereas Cu2O (Tm ¼ 1229 C (38)) and FeO
(Tm ¼ 1371 C (195)) for instance are less protective. Furthermore, the former films are electrical insulators and efficiently slow
down the migration of electrons through the scale, which as discussed above is a precondition for the oxidation reaction to proceed
(164,165). However, the slowest growing scale can only protect the underlying metal from further oxidation as long as the scale and
the metal maintain contact. Mechanisms that lead to scale-metal separation are discussed in the following section.
4.14.4.4 Scale Adhesion and Separation

Generally, the oxidation of a compound may be divided into three stages: a first transient stage, where the oxides of every active
component are formed; a steady state stage, which corresponds to the formation and the diffusion-controlled growth of a continuous
oxide scale; and finally, the breakaway stage, where the rate of oxidation rapidly increases up to the rate of the unprotected
compound (163).
The breakaway stage is in principle caused by the excess of a critical stress level in the oxide scale. Thereafter, stress relief by
cracking or spallation of the oxide or the plastic deformation of the substrate and/or the oxide occurs, which ultimately causes scale
failure. Stresses arise either in form of growth stresses, which develop during the isothermal formation of the scale, or in form of
thermal stresses, which are caused by the differential thermal expansion/contraction between the metal substrate and the scale during
thermal loading. Growth stresses may occur for different reasons, such as the volume mismatch between oxide and metal from
which it forms, epitaxial stresses, and the oxide formation within the scale. However, when the interfacial adhesion of the scale to
the metal is insufficient to resist the stresses developed, delamination or spallation of the scale occurs (164,189).
For alumina and chromia scales, the formation of new oxide within the oxide scale is believed to be an important mechanism to
explain the compressive stress generation during scale growth (179,189). Compressive stresses may hereafter be generated by the
formation of new oxide in the scale grain boundaries, where outward diffusion of cations and inward diffusion of oxygen anions
occurs (see also Section 4.14.4.2). Consequently, lateral scale growth leads to sliding at the compound/alloy interface, deformation
Figure 14 Arrhenius plot of reported apparent parabolic rate constants kp for alumina, chromia, and titania forming alloys. Modified after Hindam, H.;
Whittle, D. P. Oxid. Met. 1982, 18, 245; Leyens, C. Oxidation and Protection of Titanium Alloys and Titanium Aluminides. In Titanium and Titanium
Alloys; Leyens, C.; Peters, M., Eds; Wiley-VCH: Köln, 2003; p 187; Grabke, H. J. Mater. Sci. Forum 1997, 251–254, 149.
www.iran-mavad.com
Figure 15 Schematic diagram on how the scale growth process of a-Al2O3 or Cr2O3 may affect the growing scale: (a) formation of new oxide within
scale grain boundaries leads to compressive stresses that (b) cause interfacial sliding displacement (lateral scale growth). At stress concentration
such as surface irregularities or interfacial voids (c) localized detachment occurs which eventually causes (d) scale failure. Modified after Stott, F. H.
Mater. Sci. Forum 1997, 251–254, 19.
of the substrate, and the localized detachment (interfacial voids) of the scale from the metal surface at stress concentrations such as
surface irregularities (Figure 15(a)–(c)). Heavily convoluted, partially detached scales then eventually fail as critical loads are
exceeded, e.g., during thermal cycling (Figure 15(d)). The concept of the formation of oxide within the existing oxide scale is
however a matter of some controversy and still lacks experimental proof (164,189,196).
On the other hand, it is generally accepted that nonmetallic impurities and here in particular sulfur, are very detrimental to the
scale adhesion. Sulfur is believed to diffuse to the free surface of emerging voids, where it decreases the metal surface energy and
thereby enhances void nucleation and growth (197–202).
As for the initial formation of voids themselves, which undoubtedly contribute significantly to the overall scale adhesion
properties, vacancy coalescence in the absence of effective sinks is a favored mechanism. Vacancies can be generated either by
outward cation diffusion in p-type scales or simply by the incorporation of metal atoms (cations) from the substrate into the scale, if
the latter does not sufficiently relax to maintain contact with the receding scale (163,164,203,204). Moreover, vacancies can be
generated by the Kirkendall effect, arising from the faster backward diffusion of the more noble metal to the alloy bulk than the
forward diffusion of the more oxidizable metal to the alloy–scale interface. Such a mechanism may be expected in scales with
prevailing inward oxygen transport (205–207). However, the formation of voids and cavities is also consistent with the develop-
ment of lateral compressive stresses due to new oxide formation within the growing oxide scale, as discussed above.
A further source of stress that considerably affects the scale adhesion is the specimen geometry. The thereby induced stresses are
essentially growth induced. Depending on the growth mechanism – cation outward diffusion causing scale growth at the gas–scale
interface or anion inward diffusion causing scale growth at the scale/metal interface – and the type of curvature (convex or concave),
either compressive stresses (cation migration on convex surfaces and anion migration on concave surfaces) or tensile stresses (cation
migration on concave surfaces and anion migration on convex surfaces) develop in the oxide scale. In systems where cations are
mobile, edges and corners do furthermore restrain relaxation of the growing scale. In these geometrical stabilized regions, the scale
has to creep to maintain contact with the metal. As in most cases the scaling rate is too high or the scale plasticity too low to allow for
sufficient creep, the contact between the scale and the metal is eventually lost (164).
4.14.4.5 Reactive Element Effects

Unquestioned, a slow oxidation rate and a sufficient scale adhesion are equally important for the overall oxidation performance of
high temperature materials in oxidizing environments. The incorporation of small amounts of reactive elements (e.g., Y, Zr, Hf, Ce)
or their stable dispersions, is known to beneficially affect the scale adhesion (208,209). Thus, the resistance to scale loss by
spallation, e.g., during thermal cycling, can be significantly reduced.
Although there is still some controversy about the main mechanism of the so-called reactive element effects, it is generally accepted
that the impact on the scale growth mechanism plays a significant role and may even explain all of the observed effects. It is
confirmed experimentally that during oxidation the reactive element segregates to scale grain boundaries and the metal–scale
interface. Thereby, the oxygen potential gradient across the scale is the driving force for the slow out-diffusion of large reactive
element ions. Consequently, short-circuit outward transport of cations along grain-boundaries is inhibited and anion inward
diffusion becomes the predominant transport mechanism, see Figure 16 (180,184,210–212).
www.iran-mavad.com
Figure 16 Schematic diagram on the effect of reactive element incorporation on the scale growth kinetics: (a) cation outward diffusion along scale
grain boundaries is an important (in alumina) or the predominant (in chromia) transport mechanism in undoped scales; (b) reactive element
incorporation blocks short-cut diffusion along grain boundaries and causes prevailing anion inward diffusion.
At this stage it also noteworthy that the inhibiting of cation transport in Cr2O3 has a much greater effect on the parabolic rate
constant than in a-Al2O3, as the latter exhibits an anion inward diffusion controlled scale growth already in the undoped state.
The major benefit of reactive element additions in alumina is hence the improvement of the scale adhesion (163,211,213).
According to the oxide within oxide formation theory (see Section 4.14.4.4), the suppression of cation outward diffusion would
decrease lateral scale growth and consequently also decrease the generation of growth stresses that are known to promote scale
delamination. This is consistent with the fact that scale spallation is rarely observed on alloys containing small amounts of
reactive elements (163,189).
Furthermore, the blocked cation outward diffusion suppresses cation vacancy formation, and their coalescence into voids at the
metal-oxide interface (210,214). In this regard the effect of reactive element additions on sulfur impurities is also believed to be
important. According to more recent investigations that are consistent with experimental observations, the ‘decoration’ of the
metal–scale interface with reactive element ions seems to suppress sulfur segregation to free surfaces (interfacial voids, cavities).
Sulfur segregations are known to reduce the surface energy, which promotes interfacial void growth and ultimately causes a worse
scale adhesion (201,211,215).
Although most of the beneficial effects of reactive element additions can be explained by the changeover in growth mechanism
to predominant inward growth and the suppression of interfacial void growth, it is not entirely excluded that other effects also
contribute to slower growth kinetics and more importantly, to the scale adhesion. For mature scales, for instance, the mechanical
keying of the scale by the formation of oxide ‘pegs’ was suggested to contribute to scale adherence (163,196,216). Furthermore, it
was suggested that the rapid oxidation of reactive element additions promotes the formation of very fine grained protective a-Al2O3
or Cr2O3 scales by providing nucleation sides (217,218). However, this is not necessarily a reactive element related mechanism, as
any heterogeneous addition would have the same effect (211). On the other hand, the maintenance of a small grain size by a solute
drag effect is considered to improve the scale plasticity and consequently the scale adhesion (213,217,219).
In practice, the mere addition of reactive elements alone is of course insufficient, as the amount added is of equal importance as
its distribution (220). Generally, only a sufficient amount of reactive element is necessary to dope the grain boundaries and the
metal–scale interface (211). Excessive addition is in fact detrimental to the oxidation performance as thereby either the formation of
heterogeneous and consequently nonprotective oxide scales is promoted and/or the transformation of fast-growing transition
alumina (g, q) to stable a-Al2O3 is retarded (192,221–223). Whereas small additions, which are unlikely to affect strongly the defect
structure of transition alumina, are believed to accelerate the transformation to the stable a-modification by provision of hetero-
geneous nucleation sites and/or by hampering of the grain growth of unstable alumina modifications, higher contents seem to
retard this transformation. This can be explained by the impact of overdoping on the defect structure of the unstable alumina
polymorphs. More specifically, overdoping reduces the number of anion vacancies, which retards the transformation of transition
alumina to a-Al2O3 as the transformation occurs via an annihilation reaction between cation and anion vacancies accompanied by
structural arrangements (221,222,224).
4.14.4.6 Cr2O3–Al2O3
As the Cr1xAlxN-based coating system forms mixed Cr2O3 þ Al2O3 scales during oxidation, this section briefly discusses the binary
system Cr2O3–Al2O3. It is noted that the stable a-modification of Al2O3 is always considered here.
In principle, the combination of chromia and alumina forming metals offers the advantage of improved self-healing capacity
due to the more pronounced contribution of cation outward migration as compared to pure Al2O3. The latter mainly grows by
anion inward diffusion (see Section 4.14.4.2). Upon the formation of a crack in the scale, the cation outward migration will
overgrow the crack, which is not possible in scales that grow purely by anion inward migration (166). Of course this is at the expense
of slightly increased scale growth rates as compared to pure Al2O3 formation.
www.iran-mavad.com
Figure 17 Calculated Al2O3–Cr2O3 phase diagram by Besmann et al. (Besmann, T. M.; Kulkarni, N. S.; Spear, K. E. J. Am. Ceram. Soc. 2006, 89, 638)
with experimental solidus (;) and liquidus (:) data of Bunting (Bunting, E. N. Bur. Stand. J. Res. 1931, 6, 947) and the miscibility gap (C)
determined by Sitte (Sitte, W. Investigation of the Miscibility Gap of the System Chromia-Alumina below 1300C. In Reactivity of Solids; Barret, P., Dufour,
L.-C., Eds.; Elsevier: Dijon, 1985; p 451.).
Figure 17 shows the calculated Al2O3–Cr2O3 phase diagram by Besmann et al. (225), which also contains experimental solidus
and liquidus data of Bunting (226) and the asymmetric miscibility gap determined by Sitte (227). In the miscibility gap below
1300 C, a chromia lean alumina phase coexists with a chromia rich alumina phase. However, due to low ion diffusivity below
1000 C, the decomposition was reported to be slow and might be difficult to observe (228).
Figure 18 shows a cross-sectional TEM image of the oxide scale formed on a Cr0.45Al0.53Y0.02N coating after isothermal oxidation
at 1000 C for 50 h. As obtained by TEM investigations, a very dense, nanocrystalline mixed Al2O3–Cr2O3 scale that effectively
protects the underlying coating is formed. A more detailed discussion of the oxidation of Cr–Al–Y–N films is given in Ref. (140).
4.14.5 Simultaneous Investigation of Nanostructure and Chemistry

4.14.5.1 General Considerations
The simultaneous investigation of nanostructure and chemistry and its effect on various material properties is an extreme challenge
in materials science. While most studies on structure–property relationships of thin films use single-crystal (sc) substrates, a totally
different situation arises for investigations of structural phenomena on polycrystalline (pc) substrates. Studies employing high-
temperature annealing treatments in particular, as performed to obtain age hardening of Ti1xAlxN, require a sophisticated
investigation concept.
In order to determine the mechanical properties and structural response of Ti1xAlxN coatings on annealing temperature (Ta) and
annealing time (ta), care has to be taken in choosing suitable substrate materials and investigation techniques. On the one hand, single
crystal substrates such as silicon (100, 111) or MgO (100, 111) or equivalent sublayers of, e.g., TiN can predetermine the initial growth
conditions for thin films, resulting in predominant orientations or even epitaxial growth of the developing film. This may strongly
affect the coating properties and stands in general contrast with common deposition techniques for polycrystalline tool materials,
where only local epitaxy (229) is observed on different microstructural features, but in general no epitaxial film growth occurs.
On the other hand, thermally induced substrate interaction, such as substrate material in-diffusion (135,230) or spinel
formation (231,232) at the substrate interface, should be avoided, in order to asses solely the properties of the coating. This
limitation can be overcome by (chemically) removing the substrate after deposition to obtain free-standing thin film material.
Hence, two different polycrystalline substrates, namely low alloy steel foil (thickness 0.2 mm) and polycrystalline Al2O3 platelets
(20 7 0.5 mm), were selected for investigations of the thermal stability of Ti1xAlxN, resembling more the growth conditions on
tools and components than on single crystals. Postdeposition dissolution of the Fe foils in 10 mol% nitric acid results in free-
standing coating material that can be used for the structural investigations of the sole coating material by APT and TEM, or after
grinding to powder for XRD analysis. The thermally stable alumina substrates were used for measuring the electrical resistivity by the
four-point probe technique and mechanical properties by nanoindentation subsequent to the annealing treatments. Further single
www.iran-mavad.com
Figure 18 Overview (a) and detailed (b) TEM cross-section of a Cr0.45Al0.53Y0.02N film after oxidation at 1000 C for 50 h. (c) Selected area diffraction
pattern for detail (b). (d) Energy dispersive x-ray spectroscopy line scan across oxide scale (reproduced from Rovere, F.; Mayrhofer, P. H.; Reinholdt, A.;
Mayer, J.; Schneider, J. M. Surf. Coat. Technol. 2008, 202, 5870). Au (FIB) denotes the Au protective layer added during focused ion beam (FIB)
preparation on top of the sample.
crystalline Si (20 7 0.5 mm) and Al2O3 (10 10 0.5 mm) platelets were chosen for the investigation of mechanical properties
(e.g., biaxial stress measurements, XRD, or TEM) at room temperature.
In order to conduct detailed studies on the combined influence of structure and chemistry on mechanical properties, experi-
mental investigations that enable a 3D understanding are required. Although high-resolution TEM (HR-TEM) is a powerful tool for
analysis of nanostructures and especially orientation relationships, sample thicknesses of <50 nm (to reach electron transparency)
still contain too many overlapping grains in analysis direction to enable accurate distinction between the individual features. In
addition, chemical analysis of 3D structures in TEM by energy dispersive x-ray spectroscopy (EDX) or electron energy loss spec-
troscopy (EELS) is limited for similar reasons. The restricted lateral resolution of secondary ion mass spectroscopy (SIMS) hinders its
applicability for the present studies. The observation of chemical fluctuations at the atomic scale is however a crucial feature in
understanding atomistic processes determining various physical phenomena, such as age hardening, in materials science. Hence, we
want to discuss in more detail atom probe tomography (APT), which is still only rarely used for investigations of hard protective
coatings.
4.14.5.2 Atom Probe Tomography and Complementary Techniques

Based on the principle of field ion microscopy (FIM), invented by Müller in 1951 (233), the direct image formation of specimen
surface ions is combined with a time-of-flight (TOF) mass spectrometer in modern atom probe tomography (234,235). The highest
chemical sensitivity is thus combined with near-atomic resolution in three dimensions (236–238), which seems highly suited for
the investigation of spinodal decomposition, as shown for the Fe–Cr system in Refs. (239–241).
A FIM operates as a lens-less point projection microscope by ionization of image gas atoms (He or Ne) at the surface of a needle-
shaped specimen. Due to a high standing voltage (V), sharp needle-shaped specimens with a tip radius (rtip) of less than 100 nm are
required to achieve a sufficiently high electric field for ionization (F0) of the gas atoms at the apex, according to the formula
(236,237)
F0 ¼ V=k$rtip [13]
where k is a geometrical constant.
www.iran-mavad.com
In contrast to FIM, the working principle of APT, as schematically shown in Figure 19, uses directly evaporated specimen ions,
which are subsequently accelerated toward a position sensitive detector (PSD). In order to break the surface-atom bonds and
1 1 1
achieve ionization, high electric fields of, e.g., FTi
0 ¼ 26 V nm , F0 ¼ 19 V nm , or F0 ¼ 52 V nm
Al W
are required (236,237). The
evaporation process essentially can consist of a two-stage mechanism: First the thermally activated or electric field induced ions
escape over an activation-energy barrier (see Figure 20(a)), and then often post-field ionization processes into higher charge states
can take place (236,237,242,243).
Figure 19 Schematic drawing of the working principle of a state-of-the-art local electrode atom probe (IMAGO LEAP 3000X-HR), as used in our studies
(adapted from Miller, M. K.; Forbes, R. G. Mater. Charact. 2009, 60 (6), 461). On top of a standing voltage (below F0) either voltage pulses VP or
laser pulses with a certain laser energy EL are triggered by a time-of-flight unit to control the field evaporation process at the specimen.
Figure 20 (a) The field evaporation mechanism of surface atoms as adapted from Miller, M. K.; Forbes, R. G. Mater. Charact. 2009, 60 (6), 461. The
energy zero corresponds to a neutral atom (being bonded by L0 ) in remote field-free space. Applying an external electric bias potential lowers the
energy barrier for the atom (Le ) to become integrally charged with increasing distance from its original location. (b) Required voltage for field
evaporation of W as a function of temperature, for a given evaporation rate of 100 at s1 (dc voltage, ball symbols) or 109 at s1 (superimposed voltage
pulses, star symbols). (The data are obtained from Kellogg, G. L. J. Appl. Phys. 1981, 52 (8), 5320). In the case of an electrically nonconductive specimen
(circle) at low temperatures, the necessary evaporation voltage often exceeds the mechanical strength of the material (green arrow). The application of
laser pulses increases the temperature only and thus drastically reduces the required evaporation voltage (violet dashed arrow). (Reproduced from
Miller, M. K.; Forbes, R. G. Mater. Charact. 2009, 60 (6), 461.)
www.iran-mavad.com
Since a sufficiently high electric field would result in continuous but uncontrolled evaporation (see Figure 20(b)), two
approaches were developed with the goal to time-resolve the evaporated species. Initially the standing voltage is kept slightly below
F0. Secondly a pulsing unit, operating in the kHz range, can be used to apply a certain voltage fraction (VP/V ¼ 0.01–0.5) on top of
the standing voltage to control the ionization process (compare Figures 19 and 20(b)). This method is however limited to elec-
trically conductive materials, as the tensile load on the specimen due to the electric field often exceeds the mechanical strength of
brittle materials; e.g., TMN (242,243,244–246). Another possibility to overcome the energetic barrier is thermal activation using
laser pulses in the pico- to femtosecond range (247,248) (see Figure 19). The necessary nominal voltage for evaporation can thus
effectively be reduced due to the laser energy (see arrows in Figure 20(b)) (247–250) and enables for the investigation of low-
conductivity materials, such as oxides and nitrides (251–253).
To prevent surface migration and diffusion of the specimen material, which would corrupt the quality of the measurement, the
whole assembly has to be cooled to cryogenic temperatures (20–80 K) (253,254). By means of either voltage- or laser-pulsed APT it
is possible to measure the TOF of the evaporated specimen ions, which in combination with the atomic coordinates (x, y, see
Figure 19) from the PSD results in almost atom-by-atom chemical sensitivity in three dimensions (255). Computer-aided
reconstruction of the acquired data, e.g., by the software package IVAS (256), further allows for direct and statistical data pro-
cessing, which is further explained in the following section on APT measurements of Ti1xAlxN.
4.14.5.2.1 APT of Ti–Al–N Coatings

Classical preparation techniques for metallic specimens use a two-step chemical etching technique as given in Ref. (234) to obtain
needle-shaped specimens that fulfill all geometric preconditions for APT measurements. Because of the high chemical stability and
geometric limitations due to the limited thickness of thin films on a substrate, this preparation technique is improper in the case of
micrometer thin Ti1xAlxN coatings. As suggested in the literature (257,258), more sophisticated techniques, using focused ion
beam (FIB) microscopy, can be utilized for site-specific specimen preparation of, e.g., grain boundaries and metallic thin films. In
the case of metallic multilayer films for storage applications, the interesting features are planar and thus lie perpendicular to the
measurement direction of APT (259–263). The specimen preparation by deposition of thin films on presharpened Si microtips
(262,263), prior to Ga-beam sharpening in the FIB, appears as a time-saving alternative to the lift-out techniques generally used.
In case of TMN however difficulties arise from direct deposition of, e.g., TiN films on preshaped Si microposts, as visible in the
TEM cross-sections in Figure 21 (264). Although the growth behavior of the films on the Si substrate and Si microposts are
comparable with respect to column width and orientation, the TiN film on the Si micropost exhibits extensive cracks not only within
the Si micropost but also between the TiN columns. The good adherence between the TiN film and the Si micropost causes cracking
below the interface. This is because of the thermal mismatch of the Si substrate and the TiN coating during cooling from deposition
temperature (Tdep ¼ 500 C) to room temperature, which usually results in compressive stresses of the substrate (visible as strain
contrast lines in Figure 21(a)).
Because of the columnar growth of the films during deposition, similar to the TiN films above, common FIB-based techniques
were not suitable for the present study. By virtue of the focus on the 3D nature of spinodal decomposition in Ti1xAlxN thin films,
a FIB-based preparation technique was modified to fulfill our requirements, as explained in detail in Ref. (157). The film growth
direction is usually in line with the APT measurement direction, thus the chemical fluctuations at the column boundaries are
difficult to access. Another point is that the Ti1xAlxN films were annealed at high temperatures, which would result in extensive
substrate material indiffusion, as explained in Refs. (135,264,265). Thus the APT specimen preparation was performed from free-
standing Ti1xAlxN coating material, in the as-deposited state and additionally after annealing to 900 and 1350 C. The crucial issue
is to mount the Ti1xAlxN coating material with the growth direction perpendicular to the measurement direction, which enables
the crossing of column interfaces during the measurement.
Since laser-assisted APT is strongly affected by specimen geometry, laser energy, and material properties (266), the suitability of
steel and Si was studied as support materials for APT of TMN. Moreover, the relatively low thermal, and in case of Ti1xAlxN also
Figure 21 TEM cross-sections of TiN films after deposition on (a) a planar Si (001) substrate and (b) a Si micropost as usually used for sample
carrier for APT. Although the structural features are equal for both substrates, the intercolumnar and interfacial cracks (marked with arrows) to the
Si micropost make APT impossible. Reproduced from Rachbauer, R. Age Hardening in Transition Metal Aluminium Nitride Thin Films Studied at the
Atomic Scale. Ph.D. Thesis, Montanuniversität Leoben, Leoben, AT, July 2011.
www.iran-mavad.com
Figure 22 A 3D-interconnected structure of TiN- and AlN-rich domains has formed in Ti0.46Al0.54N after annealing to 900 C. Isoconcentration surfaces
are drawn at (red) Al and (blue) Ti concentrations of 2, 5, and 7 at% above the average film composition. Empty regions between the TiN- and
AlN-enriched domains correspond to the remaining Ti0.46Al0.54N matrix with lower element concentration fluctuations as the chosen threshold values.
electrical conductivity, causes significant troubles during APT. Thermal ‘tails’ appear in the mass spectrum, which can significantly
reduce the signal-to-noise ratio and corrupt the resolution of the analysis (266). The influence of measurement temperature,
evaporation rate, pulse frequency, and laser energy was explored for this reason in order to achieve optimum conditions for data
acquisition. It turned out that laser energies of 0.7 nJ, pulsing frequencies of 200 kHz, temperatures of 60 K, and the utilization of
predefined Si microposts as support materials reached most suitable results (157,158).
Using APT for the investigation of nanostructures formed during annealing provides essential information for the understanding
of spinodal decomposition and its impact on the mechanical properties of Ti1xAlxN. For the presentation of the APT data, 1D-, 2D-,
and 3D-concentration profiles (236,237,267) were used in combination with statistics (236,237,268,269) on the extent of
decomposition in the various annealing states. This approach delivers information on the location of, e.g., oxygen impurities (157)
or the magnitude of elemental fluctuations on the metallic sublattice of Ti1xAlxN already after deposition (158,270), highlighting
the excellent applicability of APT for the investigations of TMN thin films.
For graphical presentation of the 3D nature of the TiN- and AlN-enriched domains, formed in the course of the annealing
treatments, isosurface plots (271) that connect subsamples (voxels) of the measured sample volume exhibiting equal concentration
of a certain element were used, see e.g., Figure 22. For further information and discussion see Refs. (158,270).
The structural and chemical investigations of annealed c-Ti1xAlxN by XRD, HR-TEM, and APT provide insights in the evolution
of microstructure as a function of annealing temperature and time. The spinodally formed structural features can hence be
correlated with the evolution of mechanical and electrical properties. A UMIS nanoindentation device (272) was used to determine
film hardness and indentation modulus (see Figure 23(a)) for Ti1xAlxN films on pc-Al2O3 and correlated to the microstructural
strain ε (Figure 23(b) – determined by the Williamson-Hall procedure (273) from XRD of powdered Ti1xAlxN films (Figure 23(c))
– after annealing for either 1 or 120 min at Ta.
By means of cross-combining experimental and computational studies, it was shown that the 3D nature of the isostructurally
formed c-TiN and c-AlN domains (compare Figure 22) results in increasing coherency strain between the microstructural features.
In advance of other results on small angle x-ray scattering (SAXS) (136,274), EELS (275), or XRD (22,276), it was possible to identify
the importance of coherent interfaces and the impact on the thermal stability during the first stages of decomposition, before loss of
coherency and the precipitation of w-AlN results in reduced mechanical properties and decreasing strain (see Figure 23). This
approach enabled to draw a comprehensive picture of the impact of temperature and time on the decomposition pathway of
Ti1xAlxN. Further information and experimental details are included in Ref. (158).
4.14.6 Multinary Transition Metal Aluminium Nitrides
As discussed briefly in the introduction, smart alloying concepts enable the creation of enhanced material properties. For this
reason, a lot of research activity was conducted in past decades to overcome certain deficiencies of Ti1xAlxN. In combination with
coating architecture, especially the alloying of Ti1xAlxN with additional elements was regarded as a promising approach to improve
e.g., friction, wear, oxidation resistance, and hardness of Ti1xAlxN.
Nonmetal alloying of C, Si, and B mainly takes place by incorporation of the nonmetals on the N sublattice. The resulting grain
refinement and the formation of various nanostructures achieve increasing hardness and in general improved oxidation resistance.
Other practical aspects like low friction and wear still require improvement (230,277–283).
On the other hand, group III–VI TM-alloying is mainly achieved by substitution of Ti or Al at the metallic sublattice, which
allows for a huge variety in size and different bonding types, due to their different electronic configuration (e.g., valence electron
number, additional d- and f-states). Investigations of Y and Cr have already shown a big potential of TM additions to effectively
enhance the oxidation resistance (284–287) of Ti1xAlxN. In terms of reduced friction especially at elevated temperatures, the
addition of V (288–293) proved to be useful due to the formation of a V2O5-Magnéli oxide phase at high temperatures. Other TM
elements, e.g., W or Mo, yield corresponding results with respect to the development of Magnéli phases, which however suffer from
www.iran-mavad.com
Figure 23 (a) Film hardness of Ti0.46Al0.54N as a function annealing temperature, Ta, and time, ta. (b) Corresponding microstrain determined by
the Williamson-Hall method. (Reproduced from Williamson, G. K.; Hall, W. H. Acta Metall. 1953, 1 (1), 22.) (c) Structural evolution of the films plotted as
a quasi-2D XRD plot. The intensities are reflected by the corresponding colors. The detectable formation of w-AlN at 1100 C (ta ¼ 1 min) and 900 C
(ta ¼ 120 min) corresponds to the hardness decrease in (a) and the highest strains in (b).
volatility at temperatures above 500 and 700 C, respectively (294). Moreover, stoichiometric MoN and WN preferably crystallize in
the hexagonal structure and are hence difficult to implement in cubic Ti1xyAlxTMyN (295).
Alloying with group VI and V elements such as Zr, Nb, Hf, and Ta is acknowledged for the enhancement of mechanical properties
and wear resistance (296–304), although the impact of the various alloying elements on structure and mechanical properties is not
yet fully understood. Moreover, the combination of experiments with computational studies with respect to thermal stability is
a totally unexplored field. Hence, the goal of this chapter is to yield a comprehensive understanding of the electronic origin of
thermal stability in Ti1xyAlxTMyN and to explore methods to enable a theory-guided materials design instead of the traditional
trial-and-error approach.
4.14.6.1 Modeling of Ti–Al–TM–N Alloys and Comparison to Experiments

4.14.6.1.1 Applied Modeling Methodology for Ti–Al–TM–N Alloys
The use of density functional theory (DFT) modeling in materials science is nowadays widely accepted for investigations of the
electronic structure of condensed matter. Using software packages such as the Vienna Ab Initio Simulation Package (VASP)
(305,306) or Wien2k (17,307) allow for the calculation of alloy specific features, e.g., formation energies (Ef), mixing enthalpies
(Hmix), or the charge transfer between atoms on the lattice (17).
To obtain a computational in-depth insight into materials properties, DFT modeling was used together with the projector
augmented wave pseudopotentials (308), employing the generalized gradient approximation (GGA) as parametrized by Perdew
and Wang (309). Mainly two modifications, namely cubic (fcc, B1-structure, Fm3m space group, NaCl-prototype) and hexagonal
www.iran-mavad.com
(wurtzite, B4-structure, P63mc space group, ZnS-prototype) were considered for the Ti1xyAlxTMyN alloys, and mimicked by
special quasirandom structures (SQS) (310,311). The respective supercells contained 36 atoms (3 3 2) for the cubic and
32 atoms (2 2 2) for the hexagonal structures. Since only one or two atoms at the metallic sublattice (i.e., 18 or 16 sites,
respectively, as the other sublattice is fully occupied by N) were exchanged by TM atoms, compositional steps of 5.56 or 6.25 at%
can be achieved for x and y in the Ti1xyAlxTMyN alloys. Further details on the calculations can be found in Refs. (98,304,312–317).
4.14.6.1.2 Impact of TM-Alloying on the Phase Stability of Ti–Al–N Coatings

The structural transition from cubic to wurtzite in Ti1xyAlxTMyN can be predicted by calculation of Ef as a function of the Al
content, as presented in Figure 24 for TM ¼ Y, Zr, Nb, Hf, and Ta, respectively. The structure with lowest Ef values (i.e., actually
largest negative values) is most likely to be formed during deposition. For clarity, only the data for the ternary Ti1xAlxN and the
quaternaries with y ¼ 11.1 at% (cubic) and 12.5 at% (wurtzite) at the metallic sublattice are shown. In the presented graph, a clear
trend can be found for the impact of TM on the cubic solubility limit.
The formerly discussed crossover of the cubic and wurtzite structure is significantly lowered from x w 0.69 (78) for Ti1xAlxN to
x w 0.56 for the addition of 12.5 at% Y to the metal sublattice (313,318). Additions of 12.5 at% Hf (x w 0.69) (317), Nb (to
x w 0.68) (314), or Zr (x w 0.68) (304) to the metal sublattice have only negligible influence on the extent of the cubic solubility
range of Ti1xyAlxTMyN alloys. The only slightly increased cubic stability range to x w 0.70 (316) of Ti1xAlxN can be observed by
the addition of Ta for the TMN studied here.
The magnitude of Ef is influenced similarly, as highlighted by the arrows in the insert of Figure 24. The addition of group IIIB,
IVB, and VB TM elements also strongly affects cohesive energy, Ecoh, of the quaternary cubic solid solutions, as shown in
Figure 25(a) for TM additions of y ¼ 0 and 11.1 at% at the metal sublattice, respectively. While Y (three valence electrons (VE))
reduces the cohesive forces by significantly changing the electronic structure (158), Hf and Zr (four VE, and isovalent with Ti) have
almost no impact on the extent of Ecoh. In contrast, the group VB elements Nb and Ta (five VE) increase Ecoh over the whole Al
composition range. As the driving force for decomposition strongly increases up to the cubic metastable solubility limit, an Al
content of x ¼ 0.56 was chosen to emphasize the effect of Ti substitution by TM-alloying, see Figure 25(b). The above given trend for
the different effect of the VE configuration on the Ecoh strongly scales with the amount of TM at the metallic sublattice.
Sorting the different TM elements with respect to their extent on Ecoh, a clear trend can be deduced. While Y weakens the bonding,
Zr and Hf exhibit almost no effect on the electronic configuration. Only Nb and Ta significantly strengthen the bonding in cubic
Ti1xyAlxTMyN alloys. The diverse alloying effects can be correlated with changes in the electronic structure, as summarized by the
weighted average of the hybridized and metallic regions of the DOS (compare Section 4.14.2.2). The center of mass of the hybridized
states (COMhyb) gives rise on the covalent part of the bonding, which shifts to a higher energy by Y-alloying but to lower energies by
introduction of Hf, Nb, and Ta. This behavior stems from the available TM d, and in case of Hf and Ta, also f-states. An even more
pronounced effect appears for the near-metallic region, where COMmet refers to the states close to the Fermi level, EF, see Figure 3.
The reduction of the number of VE by Y reduces the sp3d2 hybridization in the t2g-like symmetry, whereas Nb and Ta strengthen the
metallic bonding character by an additional d electron, since COMmet shifts to lower values. The substitution of Ti by isovalent Hf
atoms has thus no significant effect on the metallic bonding behavior. For a detailed description, please see Refs. (313,316,317).
Figure 24 Formation energies, Ef, for Ti1xyAlxTMyN with y ¼ 0 and 0.11 (cub), or y ¼ 0 and 0.125 (wur) as a function of Al content, x. For
a given composition the more stable structure is determined by the lower Ef. The insert qualitatively indicates the shift of the cubic (meta)solubility limit
and the total energy of the system. Data were included from Chen, L.; Holec, D.; Du, Y.; Mayrhofer, P. H. Thin Solid Films 2011, 519 (16), 5503;
Rachbauer, R.; Holec, D.; Lattemann, M.; Hultman, L.; Mayrhofer, P. H. Int. J. Mater. Res. 2011, 102, 735; Mayrhofer, P. H.; Rachbauer, R.; Holec, D.
Scr. Mater. 2010, 63, 807; Rachbauer, R.; Holec, D.; Mayrhofer, P. H. Appl. Phys. Lett. 2010, 97, 151901; Rachbauer, R.; Holec, D.; Mayrhofer, P. H.
Surf. Coat. Technol. 2012, 211, 98; Rachbauer, R.; Blutmager, A.; Holec, D.; Mayrhofer, P. H. Surf. Coat. Technol. 2012, 206, 2667; Moser, M.; Mayrhofer,
P. H. Scr. Mater. 2007, 57 (4), 357 for various TM elements.
www.iran-mavad.com
Figure 25 (a) Cohesive energies, Ecoh, for c-Ti1xyAlxTMyN as a function of Al content, x, for y ¼ 0 and 0.11. (b) The effect of various TM on Ecoh
in the region close to the cubic (meta)solubility limit as function of y. (c) The weighted sum of DOS of the hybridized (COMhyb) and metallic
(COMmet) region (compare Figure 3) indicate a measure for stability of the alloy. Detailed information is included from Rachbauer, R.; Holec, D.;
Lattemann, M.; Hultman, L.; Mayrhofer, P. H. Int. J. Mater. Res. 2011, 102, 735; Mayrhofer, P. H.; Rachbauer, R.; Holec, D. Scr. Mater. 2010, 63, 807;
Rachbauer, R.; Holec, D.; Mayrhofer, P. H. Appl. Phys. Lett. 2010, 97, 151901; Rachbauer, R.; Holec, D.; Mayrhofer, P. H. Surf. Coat. Technol. 2012, 211,
98; Rachbauer, R.; Blutmager, A.; Holec, D.; Mayrhofer, P. H. Surf. Coat. Technol. 2012, 206, 2667.
Due to the fact that in cubic alloys, six nearest nitrogen neighbors surround the metal (compare Section 4.14.2.2), it is more
suggestive to discuss the effect of the various binary TMN in a solid solution with Ti1xAlxN. Therefore, some basic properties of the
respective cubic binaries are schematically summarized in Figure 26 (319). Since the incorporation of TM at the metal sublattice
significantly changes the bonding character and thus also the lattice parameter of the alloy, the cubic lattice parameters, ac, of the
binary TMN related to the reference coating of Ti0.42Al0.58N can facilitate the understanding of stress/strain development with
increasing TM content.
The lattice parameter difference, Dac, of the TMN compared to Ti0.42Al0.58N hence introduces significant strain in the quaternary
Ti1xyAlxTMyN alloys. Consequently the compressive stress, s, of the Ti1xyAlxTMyN films increases steadily with increasing
amount of TM, as shown in Figure 27(a). In addition, all quaternary films exhibit a positive deviation from Vegard’s rule (320).
Moreover, the varying number of VE per formula unit results in a change of the total valence electron concentration (VEC) which
provides a rough estimate for the hardness according to Holleck (321), who observed a hardness maximum at VEC w8.4 for ternary
TM-carbonitrides. By opposing the experimentally observed film hardness as a function of VEC, it is possible to discriminate the
influence of TM-induced lattice strain or compressive stress and the changes of the electronic structure with respect to cohesion, as
presented in Figure 27(b).
Since, Y (three VE) lowers the interatomic bonding strength by substitution of Ti (four VE) at the metallic sublattice, a reduction
of VEC goes along with reduced hardness. Moreover, an sp3 hybridization of the wurtzite phase becomes more favorable for higher
Y contents, in agreement with the experimentally observed structural changes (313,318). On the opposite, the introduction of five-
valent Nb or Ta enhances the interatomic bonding strength, by promoting the interaction of the metal d-states with N p-states. This
behavior favors the sp3d2 hybridization of the covalent region close to EF (313,316), which results in increased cohesive strength of
the quaternary alloy compared to Ti1xAlxN. The solid solution hardening effect of dissolved ZrN and HfN in Ti1xyAlxTMyN can
however almost exclusively be related to Dac, since the incorporation of Zr and Hf exhibit almost no impact on the electronic
structure of the alloy. Details can be found in Refs. (313,317).
4.14.6.1.3 Thermal Stability of Ti–Al–TM–N Coatings

In order to get an estimate for the driving force for isostructural decomposition of Ti1xyAlxTMyN alloys, DFT calculations were
used for the determination of the mixing enthalpy, Hmix, which resembles the development of the energy of formation, Ef (compare
Figure 24), respectively. Since all Hmix values of the quaternary alloys are positive, the total energy of the system will be lowered due
to decomposition into the boundary phases. By examination of Hmix of the individual ternary boundary TMN alloys, it is possible to
predict the stable phases after decomposition, see Figure 28.
All ternary TM1xAlxN exhibit strongly positive Hmix with respect to c-TMN and c-AlN or w-AlN, which scale with the Al content
and reach their maximum around the metastable solubility limit of the respective cubic ternary phase. From this it is possible to
predict that after decomposition, no ternary TM1xAlxN phase exists. Performing the same sort of calculations for the ternary
Ti1zTMzN, using z ¼ y/(1x), the situation changes, as Hmix is significantly lowered with respect to the metastable ternary Ti1xAlxN
for TM-element additions of groups IVB and VB.
www.iran-mavad.com
Figure 26 Schematic summary of cubic binary TMN, including some important properties and their lattice mismatch with respect to Ti0.42Al0.58N.
Data obtained from Predel, B. Landolt-Börnstein – Group IV Physical Chemistry. In Numerical Data and Functional Relationships in Science and Technology,
Cr-Cs – Cu-Zr; Madelung, O., Ed.; Springer: Berlin, Heidelberg, 1994; p 1; Predel, B. Landolt-Börnstein – Group IV Physical Chemistry. In Numerical Data
and Functional Relationships in Science and Technology, Li-Mg – Nd-Zr; Madelung, O., Ed.; Springer: Berlin, Heidelberg, 1997; p 1; Predel, B. Landolt-
Börnstein – Group IV Physical Chemistry. In Numerical Data and Functional Relationships in Science and Technology, Ac-Au – Au-Zr; Madelung, O., Ed.;
Springer: Berlin, Heidelberg, 1991; p 1; Rachbauer, R.; Holec, D.; Lattemann, M.; Hultman, L.; Mayrhofer, P. H. Int. J. Mater. Res. 2011, 102, 735;
Mayrhofer, P. H.; Rachbauer, R.; Holec, D. Scr. Mater. 2010, 63, 807; Rachbauer, R.; Holec, D.; Mayrhofer, P. H. Appl. Phys. Lett. 2010, 97, 151901;
Rachbauer, R.; Holec, D.; Mayrhofer, P. H. Surf. Coat. Technol. 2012, 211, 98; Rachbauer, R.; Blutmager, A.; Holec, D.; Mayrhofer, P. H. Surf. Coat.
Technol. 2012, 206, 2667; Moser, M.; Mayrhofer, P. H. Scr. Mater. 2007, 57 (4), 357; Rovere, F. Theoretical and Experimental Assessment of Cr-Al-Y-N as
Protective Coatings for g-TiAl Based Alloys; Shaker Verlag GmbH: Aachen, 2010.
Figure 27 (a) Biaxial stress of the c-Ti1xyAlxTMyN films as a function of TMN content. (b) Film hardness of c-Ti1xyAlxTMyN films over the valence
electron concentration, VEC. The solid solution hardening effect of Hf-alloying can be attributed mostly to lattice strain, while a VEC increase due to
Nb and Ta substitution of Ti atoms promote the cohesive strength of the alloy, compare Figure 25.
For Nb and Ta additions, a solid solution of Ti1zNbzN and Ti1zTazN is even more stable than the binary TMN throughout the
whole composition range. A negligible driving force for decomposition exists for Ti1zHfzN, while Ti1zZrzN can be regarded as
a border case, since it exhibits positive Hmix but only half the magnitude of Ti1xAlxN. Conversely, Ti1zYzN exhibits an even higher
driving force for decomposition, compared to Ti1xAlxN, suggesting a strong driving force for phase separation into the binary c-TiN
and c-YN (287).
www.iran-mavad.com
Figure 28 (a) Mixing enthalpies, Hmix, of TM1xAlxN (a) and Ti1zTMzN (b) for decomposition into the cubic boundary phases. Details are included
from Rovere, F.; Music, D.; To Baben, M.; Fuss, H.-G.; Mayrhofer, P. H.; Schneider, J. M. J. Phys. D Appl. Phys. 2010, 43 (3), 035302; Holec, D.; Franz, R.;
Mayrhofer, P. H.; Mitterer, C. J. Phys. D Appl. Phys. 2010, 43 (14), 341; Rachbauer, R.; Holec, D.; Lattemann, M.; Hultman, L.; Mayrhofer, P. H. Int. J.
Mater. Res. 2011, 102, 735; Mayrhofer, P. H.; Rachbauer, R.; Holec, D. Scr. Mater. 2010, 63, 807; Rachbauer, R.; Holec, D.; Mayrhofer, P. H. Appl. Phys.
Lett. 2010, 97, 151901; Rachbauer, R.; Holec, D.; Mayrhofer, P. H. Surf. Coat. Technol. 2012, 211, 98; Rachbauer, R.; Blutmager, A.; Holec, D.; Mayrhofer,
P. H. Surf. Coat. Technol. 2012, 206, 2667.
These predictions are extremely helpful in order to determine the stability of the various possible phases forming in the course of
thermal load. As explained in Section 4.14.3.1.1, spinodal decomposition into c-AlN- and c-TiN-rich domains causes age hardening
of the ternary Ti1xAlxN system. Investigations on the structural evolution of c-Ti1xyAlxTMyN thin films as a function of Ta have
proven that the initial stages of decomposition can be related to the spinodal formation of c-AlN-rich and c-Ti1zTMzN-rich
domains, before precipitation of w-AlN occurs and overaging causes decreasing film hardness.
The development of structure and the corresponding volume fractions is schematically shown for ternary c-Ti1xAlxN in
Figure 29(a) and for the quaternary c-Ti1xyAlxTMyN alloys in Figure 29(b) and 29(c). In agreement with ab initio predictions, the
formation of a dual phase structure, consisting of w-AlN and a solid solution of c-Ti1zTMzN, was observed after decomposition
Figure 29 Schematic of the structural evolution for metastable c-Ti1xAlxN as a function of Ta with corresponding phase fractions (a). The formation of
c-AlN- and c-TiN-enriched domains during spinodal decomposition (reproduced from Rachbauer, R.; Massl, S.; Stergar, E.; Holec, D.; Kiener, D.; Keckes,
J.; Patscheider, J.; Stiefel, M.; Leitner, H.; Mayrhofer, P. H. J. Appl. Phys. 2011, 110, 023515; Rachbauer, R.; Stergar, E.; Massl, S.; Moser, M.; Mayrhofer,
P. H. Scr. Mater. 2009, 61, 725) is followed by the precipitation of w-AlN and deteriorating mechanical properties (d) in the final dual phase regime. The
evolution of quaternary c-Ti1xyAlxTMyN alloys is shown for (b) the formation of dual phase structure (w-AlN þ c-Ti1zTMzN) at high temperatures in
the case of group IVB and VB TM elements or (c) the formation of three stable phases after decomposition in the case of Y alloyed Ti1xAlxN.
www.iran-mavad.com
into the thermodynamically stable phases at high temperatures for all c-Ti1xyAlxTMyN alloys up to y w0.1 (see Refs.
(264,287,304,314–317)). The only exception is c-Ti1xyAlxYyN, which further decomposes into w-AlN, c-TiN, and c-YN (287), as
suggested from the interpretation of Hmix (see Figure 29(c)).
The corresponding development of film hardness is shown in Figure 29(d). Although the extent of the hardness increase
becomes less pronounced with increasing TM content in the films, the hardness increase starts at lower Ta but is maintained to
higher Ta (compare Figure 29(a) and 29(b)). Since increasing isostatic pressure is known to promote the isostructural decom-
position of Ti1xAlxN, it is likely that TM-alloying similarly enhances the driving force for spinodal decomposition because of the
increasing compressive stresses in the coatings with an increasing amount of TM at the metallic sublattice (compare Figure 8 in
Section 4.14.3.1.1). Hence, the hardness increases already at lower annealing temperatures. On the other hand, the higher coherency
strain between the isostructurally formed phases, due to an increasing lattice mismatch between the c-AlN-rich and c-Ti(TM)N-rich
domains, implies a retarding force for the precipitation of w-AlN, which requires w25% bigger volume per formula unit than c-AlN.
Further details are given in Refs. (98,142,152,158,312,315–317).
By means of vacuum annealing treatments it is possible to determine, e.g., temperature and mechanisms for the decomposition
route of Ti1xyAlxTMyN under thermal load as discussed above. The observed structural evolution toward the predicted equilib-
rium phases is however only valid for annealing in vacuum or protective gas atmosphere. In order to estimate the thermal stability
of protective coatings, where high-temperature oxidation in ambient air is among the predominant failure mechanisms, additional
annealing treatments were conducted at 850 and 950 C (20 h) under ambient atmosphere. The remaining nitride thickness of the
Ti1xyAlxTMyN films on Al2O3 was determined from SEM cross sections after thermal load and normalized to the as-deposited
thickness, as presented in Figure 30. It can be seen that the ternary Ti1xAlxN suffers from complete oxidation, while all TM-alloyed
films exhibit remaining (nearly unaffected) nitride layers after oxidation at 850 C (Figure 30(a)). Already small TM additions of
3–5 at% at the metallic sublattice result in efficient oxidation protection in the case of TM ¼ Zr, Nb, Hf, and Ta (compare to Refs.
(264,304,316,317)). Similar observations were achieved for thermal exposure at 950 C, where however only the Ti1xyAlxTMyN
films alloyed with Ta provided significantly improved oxidation resistance for little TM contents, whereas Nb and Hf only achieve
satisfying protection for y 8 at% of the metal content (Figure 30(b)).
The growing oxide scale equally consists of a dense Al2O3 layer on top for all Ti1xyAlxTMyN films investigated (compare
Figures 31 and 32). This Al-rich oxide is known to reduce the oxygen indiffusion and only retains the coating lifetime if the porous
Ti-rich oxide underneath does not suffer from mechanical load, which causes spallation of the brittle oxide scale (109). In the case of
the Zr- (304) and Hf-alloyed Ti1xyAlxTMyN films, an alternating layer structure that consists of Al- and Ti-enriched oxides develops
(see Figure 31).
The respective TM remains with Ti, forming a mixed Ti-TM-oxide, as indicated by the EDX-element maps, where a slight oxygen
depletion in the Ti-rich domains compared to the Al-rich scale can be identified. The effective enhancement of the Zr- and
Hf-alloyed films can thus be attributed to the higher activation energy for diffusion of the TM compared to Ti or Al (322). Further
details are included in Ref. (317).
Figure 30 Unaffected nitride layer thicknesses of Ti1xyAlxTMyN films with TM ¼ Hf, Nb, and Ta after annealing in ambient air at (a) 850 C and
(b) 950 C for 20 h, with respect to the as-deposited coating thickness. (c) Electron backscatter images of Ti1xyAlxTMyN fracture cross sections.
From top left to bottom right: the fully oxidized Ti0.42Al0.58N film, and Ti1xyAlxTMyN from (b) with highest TM contents investigated after annealing at
950 C. The oxide scale is referred to as O and the remaining nitride as N, respectively. The best performance was achieved for the Ta-containing
coating where only negligible oxidation occurred.
www.iran-mavad.com
Figure 31 EDX elemental map of a Ti0.35Al0.55Hf0.1N cross-section after annealing at 950 C for 20 h. A self-organized multilayer of Al2O3 and a mixed
Ti–Hf-oxide has formed on top of the unaffected nitride layer.
Figure 32 EDX elemental map of a Ti0.35Al0.57Nb0.08N cross-section after annealing at 950 C for 20 h. A dense Al2O3 top layer and a porous mixed
Ti–Nb-oxide near the interface can clearly be distinguished (compare Figure 30(c)).
On the other hand, the incorporation of Nb and Ta in Ti1xyAlxTMyN exhibits an even more effective protection against
oxidation even after a thermal load of 950 C for 20 h (compare Figures 30 and 32). Like for the TM elements of group IVB (Zr, Hf),
a higher activation energy of Nb or Ta results in a dense Al-rich top layer and Ti(Nb)- or Ti(Ta)-enriched layers below. The different
diffusion kinetics of the species involved also account for the higher porosity at the interface of the Ti–Nb-oxide and the virgin
coating (see Figure 30(c)). In addition, the substitution of Ti3þ ions by Nb5þ or Ta5þ ions effectively slows down the oxygen anion
indiffusion, due to a reduction of oxygen vacancies. Further details can be found in Refs. (302,316,323).
4.14.7 Summary and Outlook
For the industrially employed coating systems based on Ti1xAlxN and Cr1xAlxN, we identified and discussed the mechanisms
relevant for the thermal and chemical stability. State-of-the-art experimental and computational data were utilized to shed light on
www.iran-mavad.com
the decomposition and oxidation pathways. Thermal stability is discussed for Ti1xAlxN-based coating systems, while chemical
stability is analyzed for Cr1xAlxN-based coating systems. Furthermore, the influence of alloying elements such as Y, Nb, Ta, Zr, and
Hf on the phase formation, structure, and mechanical and thermal properties of these ternary Ti1xAlxN and Cr1xAlxN coatings is
discussed.
The experimental studies employing a combination of 3D-APT, HR-TEM, XRD, and nanoindentation enable a comprehensive
understanding of the age-hardening phenomenon of cubic (c) Ti1xAlxN at the atomic scale. Based on these investigations, it has
been necessary to define the most important factors accounting for the thermal stability of c-Ti1xAlxN. It is shown that the
formation of coherency strains between the isostructurally decomposing domains, e.g., c-Ti1xþDAlxDN and c-Ti1xDAlxþDN,
respectively, is crucial for the observed increase of the film hardness with increasing annealing temperature and/or time. The goal to
enhance the thermal stability of the ternary Ti1xAlxN films was addressed by investigating sophisticated TM alloying concepts in
order to influence the wurtzite (w) AlN phase-formation, which is known to result in deteriorated mechanical properties at high
temperatures.
Quantum mechanical calculations are extremely powerful, not only to predict but also to provide explanations for the various
alloying effects of group IIIB to VB transition metals (e.g., Y, Nb, Hf, and Ta). Density functional theory calculations can be
employed to estimate e.g., phase stabilities of the various supersaturated Ti1xyAlxTMyN solid solutions after deposition, but also
to predict the thermodynamically stable products which will form during thermal annealing.
The good agreement between ab initio predicted trends and the observed properties of the Ti1xyAlxTMyN thin films indicates
the strong dependence on the electronic structure in this class of nanostructured materials. Consequently, these studies provide
a significant contribution to the field of theory-guided materials design, as it presents numerous examples for successful imple-
mentation of quantum mechanical calculations to the development of hard coatings with superior high-temperature properties.
Quantum mechanical calculations and experimental data also showed that Cr1xAlxN-based coatings exhibit superior phase
stability as compared to other Al-containing transition metal nitrides. This can be understood due to the small lattice strain together
with a small electronic driving force for decomposition, which are both found to be important factors controlling the phase stability
of Al-containing TMNs. Especially during these studies, N loss was identified as another crucial variable to determine the phase
stability and the decomposition kinetics of such materials. Future stability models must hence include N loss to allow for the
accurate prediction of the phase stability of Al-containing transition metal nitrides.
A cornerstone of this chapter is also the assessment of the potential exploitation of Y-based reactive element effects to improve
the oxidation resistance of TMN coatings, using the model system Cr1xAlxN. Y contents of w1 at% (hence, 2 at% on the metal
sublattice) cause the formation of dense mixed (Al,Cr)2O3 scales that effectively protect the coating up to at least 1000 C. Y
contents exceeding 1 at% are found to be detrimental to the oxidation resistance due to the formation of fast growing, non-
protective oxide scales. Based on the systematic investigations conducted, a chemical composition corresponding to Cr0.30A-
l0.68Y0.02N is identified as most promising in the quaternary materials system Cr1xyAlxYyN. Moreover, exposure to 900 C is
identified as a precondition for the formation of a protective (Al,Cr)2O3 oxide scale. Hence, preliminary heat treatments at 900 C
are necessary to provide excellent oxidation resistance at temperatures well below 900 C.
Future investigations may focus on the interfacial design for the critical transition region between coating and substrate – but
also between various layers or phases – with the goal of developing interlayers that can fulfill diffusion barrier functions and/or
crack-stopping or -branching functions. Thereby coatings can be designed exhibiting increased strengths, thermal stability, and
durability. In this respect, the application of multilayer coatings, featuring different layers, or nanostructured coatings may improve
the overall performance of Ti1xAlxN- and Cr1xAlxN-based protective coatings. A straightforward approach to improve the
mechanical properties of coated components lies within the optimization of deposition parameters. It is envisioned that uniformly
dense coating morphologies without the partially observed film-growth anomalies will suppress microstructural features that
negatively affect the strength of the coatings themselves and coated components in general.
Acknowledgments
We acknowledge the START Project (project Y371) of the Austrian Science Fund (FWF), the Christian Doppler Association
(Christian Doppler Laboratory for Application Oriented Coating Development), the European Commission (project
INNOVATIAL – Innovative processes and materials to synthesize knowledge-based ultra-performance nanostructured PVD
thin-films on gamma titanium aluminides – NMP3-CT-2005-515884), the Deutsche Forschungsgemeinschaft (DFG) within
the projects Schn. 735/14, and the transregional collaborative research center TRR87/1 ‘Pulsed high power plasmas for the
synthesis of nanostructured functional layers’ (SFB-TR 87) for financial support.
References
1. Pliny the elder. Natural History: A Selection; Penguin Classics: Harmondsworth, 2004.
2. Altman, J. Gold Bull. 1979, 12, 75.
3. Nicholson, E. D. Gold Bull. 1979, 12, 161.
4. Sachse, M. Damaszener Stahl: Mythos, Geschichte, Technik, Anwendung, 1. Aufl.; Wirtschaftsverl. NW, Verl. für Neue Wiss: Bremerhaven, 1989
5. Wadsworth, J.; Sherby, O. D. Damascus Steels. Sci. Am. 1985, 252, 94–99.
www.iran-mavad.com
6. Arzt, E. Acta Mater. 1998, 46 (16), 5611.

7. Dehm, G. Prog. Mater. Sci. 2009, 54 (6), 664.
8. Veprek, S.; Argon, A. S.; Zhang, R. F. Philos. Mag. 2010, 90 (31–32), 4104.
9. Greene, J. Handbook of Crystal Growth; Elsevier Science Publishers: Amsterdam, 1993.
10. Knotek, O.; Bohmer, M.; Leyendecker, T. J. Vac. Sci. Technol. A 1986, 4, 2695.
11. Münz, W.-D. J. Vac. Sci. Technol. A 1986, 4, 2717.
12. Holleck, H. Zeitschrift für Werkstofftechnik 1986, 17, 334.
13. Hultman, L. Vacuum 2000, 57 (1), 1.
14. Mitterer, C.; Mayrhofer, P. H.; Musil, J. Vacuum 2003, 71 (1–2), 279.
15. Mayrhofer, P. H.; Mitterer, C.; Hultman, L.; Clemens, H. Prog. Mater. Sci. 2006, 51 (8), 1032.
16. Stueber, M.; Holleck, H.; Leiste, H.; Seemann, K.; Ulrich, S.; Ziebert, C. J. Alloy. Compd. 2009, 483 (1–2), 321.
17. Schwarz, K.; Blaha, P.; Trickey, S. B. Mol. Phys. 2010, 108 (21–23), 3147.
18. Restall, J. E. Metallurgia 1979, 46, 676.
19. Xiao, W.; Jia, S.; Wang, J.; Yang, J.; Wang, L. J. Mater. Res. 2008, 23, 2609.
20. Zhou, H. T.; Zhang, Z. D.; Liu, C. M.; Wang, Q. W. Mater. Sci. Eng. A 2007, 445–446, 1.
21. Zou, H.; Zeng, X.; Zhai, C.; Ding, W. Mater. Sci. Eng. A 2005, 402, 142.
22. Mayrhofer, P. H.; Hörling, A.; Karlsson, L.; Sjölén, J.; Larsson, T.; Mitterer, C.; Hultman, L. Appl. Phys. Lett. 2003, 83 (10), 2049.
23. Münz, W. D.; Gobel, J. Surf. Eng. 1987, 3, 47.
24. PalDey, S.; Deevi, S. C. Mater. Sci. Eng. A 2003, 342 (1–2), 58.
25. Navinsek, B.; Panjan, P.; Urankar, I.; Cvahte, P.; Gorenjak, F. Surf. Coat. Technol. 2001, 142–144, 1148.
26. Mayrhofer, P. H.; Tischler, G.; Mitterer, C. Surf. Coat. Technol. 2001, 142–144, 78.
27. Panjan, P.; Boncina, I.; Bevk, J.; Cékada, M. Surf. Coat. Technol. 2005, 200, 133.
28. Li, W. Z.; Wang, Q. M.; Gong, J.; Sun, C.; Jiang, X. Appl. Surf. Sci. 2009, 255, 8190.
29. Park, K.-C.; Kim, K.-B.; Raaijmakers, I. J. M. M.; Ngan, K. J. Appl. Phys. 1996, 80, 5674.
30. Manaud, J. P.; Poulon, A.; Gomez, S.; Petitcorps, Y. L. Surf. Coat. Technol. 2007, 202, 222.
31. Fox-Rabinovich, G. S.; Veldhuis, S. C.; Dosbaeva, G. K.; Yamamoto, K.; Kovalev, A. I.; Wainstein, D. L.; Gershman, I. S.; Shuster, L. S.; Beake, B. D. J. Appl. Phys. 2008, 103,
083510.
32. Patscheider, J.; Zehnder, T.; Diserens, M. Surf. Coat. Technol. 2001, 146–147, 201.
33. Münz, W. D.; Lewis, D. B.; Hovsepian, P. E.; Schonjahn, C.; Ehiasarian, A.; Smith, I. J. Surf. Eng. 2001, 17, 15.
34. Mayrhofer, P. H.; Mitterer, C.; Clemens, H. Adv. Eng. Mater. 2005, 7 (12), 1071.
35. Holleck, H. In Surface Engineering: Science and Technology I; Kumar, A., Chung, Y. W., Moore, J. J., Smugeresky, J. E., Eds.; The Minerals, Metals and Materials Society,
1999; p 207.
36. Holleck, H. Materialwissenschaft und Werkstofftechnik 1986, 17, 334.
37. Holleck, H. Surf. Coat. Technol. 1990, 43–44, 245.
38. Predel, B. Landolt-Börnstein – Group IV Physical Chemistry. In Numerical Data and Functional Relationships in Science and Technology, Cr-Cs – Cu-Zr; Madelung, O., Ed.;
Springer: Berlin, Heidelberg, 1994; p 1.
39. Predel, B. Landolt-Börnstein – Group IV Physical Chemistry. In Numerical Data and Functional Relationships in Science and Technology, Ga-Gd – Hf-Zr; Madelung, O., Ed.;
40. Predel, B. Landolt-Börnstein – Group IV Physical Chemistry. In Numerical Data and Functional Relationships in Science and Technology, Li-Mg – Nd-Zr; Madelung, O., Ed.;
41. Franke, P.; Neuschütz, D. Numerical Data and Functional Relationships in Science and Technology, Binary Systems. Part 4: Binary Systems from Mn-Mo to Y-Zr. In Landolt-
Börnstein – Group IV Physical Chemistry; Madelung, O., Ed.; Springer: Berlin, 2006; p 1.
42. Baba, K.; Hatada, R.; Udoh, K.; Yasuda, K. Nucl. Instrum. Methods Phys. Res., Sect. B 1997, 127–128, 841.
43. Reiter, A. E.; Derflinger, V. H.; Hanselmann, B.; Bachmann, T.; Sartory, B. Surf. Coat. Technol. 2005, 200 (7), 2114.
44. Hasegawa, H.; Kawate, M.; Suzuki, T. Surf. Coat. Technol. 2005, 200, 2409.
45. Schwarz, K. Crit. Rev. Solid State Mater. Sci. 1987, 13, 211.
46. Colot, J. L. J. Phys. C: Solid State Phys. 1977, 10, 1675.
47. Häglund, J.; Fernández Guillermet, A.; Grimvall, G.; Körling, M. Phys. Rev. B 1993, 48, 11685.
48. Stampfl, C.; Mannstadt, W.; Asahi, R.; Freeman, A. J. Phys. Rev. B 2001, 63, 155106.
49. Rovere, F.; Music, D.; To Baben, M.; Fuss, H.-G.; Mayrhofer, P. H.; Schneider, J. M. J. Phys. D Appl. Phys. 2010, 43 (3), 035302.
50. Rovere, F.; Music, D.; Schneider, J. M.; Mayrhofer, P. H. Acta Mater. 2010, 58, 2708.
51. Braun, R.; Müßener, D.; Moser, M.; Rovere, F.; Mayrhofer, P. H.; Leyens, C. Oxidation Behavior of the g-TiAl Based Alloy Ti-45Al-8Nb Deposited with TiAlYN and CrAlYN. In
Structural Aluminides for Elevated Temperatures; Kim, Y.-W., Morris, D., Yang, R., Leyens, C., Eds.; The Minerals, Metals & Materials Society, 2008; pp 289–296.
52. Braun, R.; Rovere, F.; Mayrhofer, P. H.; Leyens, C. Intermetallics 2010, 18, 479.
53. Fernández Guillermet, A.; Häglund, J.; Grimvall, G. Phys. Rev. B 1992, 45, 11557.
54. Harrison, W. A.; Straub, G. K. Phys. Rev. B 1987, 36, 2695.
55. Holec, D.; Zhou, L.; Rachbauer, R.; Mayrhofer, P. H. Alloy-Based Design of Materials from First Principles: An Application to Functional Hard Coatings. In Density Functional
Theory: Principles, Applications and Analysis; Morin, J., Pelletier, J. M., Eds.; Nova Publishers: New York, 2013. ISBN: 978-1-62417-955-6.
56. Atkins, P. W. Physical Chemistry; Oxford University Press: New York, 1998.
57. Predel, B. Landolt-Börnstein – Group IV Physical Chemistry. In Numerical Data and Functional Relationships in Science and Technology, Ac-Au – Au-Zr; Madelung, O., Ed.;
58. Matzke, H. Defect Diffus. Forum 1992, 83, 111.
59. Shinozuka, K.; Susa, M.; Maruyama, T.; Nagata, K. Defect Diffus. Forum 1997, 143–147, 1237.
60. Lengauer, W. J. Alloy. Compd. 1995, 229, 80.
61. Coad, J. P.; Rickerby, D. S.; Oberlander, B. C. Mater. Sci. Eng. 1985, 74, 93.
62. Nicolet, M. A. Thin Solid Films 1978, 52, 415.
63. Wittmer, M. J. Appl. Phys. 1982, 53, 1007.
64. Kumar, N.; Fissel, M. G.; Pourrezaei, K.; Lee, B.; Douglas, E. C. Thin Solid Films 1987, 153, 287.
65. Östling, M.; Nygren, S.; Petersson, C. S.; Norström, H.; Buchta, R.; Blom, H. O.; Berg, S. Thin Solid Films 1986, 145, 81.
66. McIntyre, D.; Greene, J. E.; Hakansson, G.; Sundgren, J. E.; Munz, W. D. J. Appl. Phys. 1990, 67, 1542.
67. Hasegawa, H.; Suzuki, T. Surf. Coat. Technol. 2004, 188–189, 234.
68. Chim, Y. C.; Ding, X. Z.; Zeng, X. T.; Zhang, S. Thin Solid Films 2009, 517, 4845.
69. Hörling, A.; Hultman, L.; Odén, M.; Sjölén, J.; Karlsson, L. Surf. Coat. Technol. 2005, 191, 384.
www.iran-mavad.com
70. Barshilia, H. C.; Selvakumar, N.; Deepthi, B.; Rajam, K. S. Surf. Coat. Technol. 2006, 201 (6), 2193.
71. Lin, J.; Mishra, B.; Moore, J. J.; Sproul, W. D. Surf. Coat. Technol. 2008, 202, 3272.
72. Schuster, J. C.; Bauer, J. J. Solid State Chem. 1984, 53, 260.
73. Prange, R.; Cremer, R.; Neuschütz, D. Surf. Coat. Technol. 2000, 133–134, 208.
74. Holleck, H. Surf. Coat. Technol. 1988, 36, 151.
75. Mayrhofer, P. H.; Fischer, F. D.; Böhm, H. J.; Mitterer, C.; Schneider, J. M. Acta Mater. 2007, 55 (4), 1441.
76. Xia, Q.; Xia, H.; Ruoff, A. L. J. Appl. Phys. 1993, 73, 8198.
77. Zhou, M.; Makino, Y.; Nose, M.; Nogi, K. Thin Solid Films 1999, 339, 203.
78. Mayrhofer, P. H.; Music, D.; Schneider, J. M. J. Appl. Phys. 2006, 100 (9), 094906.
79. Cremer, R.; Witthaut, M.; Neuschütz, D. Experimental Determination of the Metastable (Ti, Al)N Phase Diagram up to 700 C. In Value-Edition-Metallurgy; Cho, W. D.,
Sohn, H. Y., Eds.; The Minerals & Metals Society: Warrendale, PA, 1998; p 249.
80. Kimura, A.; Hasegawa, H.; Yamada, K.; Suzuki, T. Surf. Coat. Technol. 1999, 120–121, 438.
81. Kutschej, K.; Mayrhofer, P. H.; Kathrein, M.; Polcik, P.; Tessadri, R.; Mitterer, C. Surf. Coat. Technol. 2005, 200 (7), 2358.
82. Hugosson, H. W.; Hogberg, H.; Algren, M.; Rodmar, M.; Selinder, T. I. J. Appl. Phys. 2003, 93, 4505.
83. Mayrhofer, P. H.; Music, D.; Schneider, J. M. Appl. Phys. Lett. 2006, 88 (7), 071922.
84. Zhang, R. F.; Veprek, S. Mater. Sci. Eng. A 2007, 448, 111.
85. Alling, B.; Ruban, A. V.; Karimi, A.; Peil, O. E.; Simak, S. I.; Hultman, L.; Abrikosov, I. A. Phys. Rev. B 2007, 75, 045123.
86. Knotek, O.; Loffler, F.; Scholl, H. J. Surf. Coat. Technol. 1991, 45, 53.
87. Kaindl, R.; Franz, R.; Soldan, J.; Reiter, A.; Polcik, P.; Mitterer, C.; Sartory, B.; Tessadri, R.; O’Sullivan, M. Thin Solid Films 2006, 515, 2197.
88. Kawate, M. J. Vac. Sci. Technol. A 2002, 20, 569.
89. Makino, Y.; Nogi, K. Surf. Coat. Technol. 1998, 98, 1008.
90. Sugishima, A.; Kajioka, H.; Makino, Y. Surf. Coat. Technol. 1997, 97, 590.
91. Zhang, R. F.; Veprek, S. Acta Mater. 2007, 55, 4615.
92. Mayrhofer, P. H.; Music, D.; Reeswinkel, T.; Fuß, H. G.; Schneider, J. M. Acta Mater. 2008, 56, 2469.
93. Alling, B.; Marten, T.; Abrikosov, I. A.; Karimi, A. J. Appl. Phys. 2007, 102, 044314.
94. Bobzin, K.; Lugscheider, E.; Nickel, R.; Immich, P. Materialwissenschaft und Werkstofftechnik 2006, 37, 833.
95. Makino, Y.; Mori, M.; Miyake, S.; Saito, K.; Asami, K. Surf. Coat. Technol. 2005, 193 (1–3), 219.
96. Sheng, S. H.; Zhang, R. F.; Veprek, S. Acta Mater. 2008, 56 (5), 968.
97. Sanjinés, R.; Sandu, C. S.; Lamni, R.; Lévy, F. Surf. Coat. Technol. 2006, 200, 6308.
98. Holec, D.; Rachbauer, R.; Chen, L.; Wang, L.; Luef, D.; Mayrhofer, P. H. Surf. Coat. Technol. 2011, 206, 1698.
99. Selinder, T. I.; Miller, D. J.; Gray, K. E.; Sardela, M. R.; Hultman, L. Vacuum 1995, 46 (12), 1401.
100. Holec, D.; Franz, R.; Mayrhofer, P. H.; Mitterer, C. J. Phys. D Appl. Phys. 2010, 43 (14), 341.
101. Man, B. Y.; Guzman, L.; Miotello, A.; Adami, M. Surf. Coat. Technol. 2004, 180–181, 9.
102. Vaz, F.; Rebouta, L.; Andritschky, M.; Da Silva, M. F.; Soares, J. C. J. Eur. Ceram. Soc. 1997, 17, 1971.
103. Chen, L.; Paulitsch, J.; Du, Y.; Mayrhofer, P. H. Surf. Coat. Technol. 2012, 206, 2954.
104. Barbee, T. W.; Holmes, W. H.; Keith, D. L.; Pyzyna, M. K.; Ilonca, G. Thin Solid Films 1977, 45 (3), 591.
105. Beensh-Marchwicka, G.; Krol-Stepniewska, L.; Posadowski, W. Thin Solid Films 1981, 82 (4), 313.
106. Jehn, H. A.; Hofmann, S.; Rückborn, V. E.; Münz, W. D. J. Vac. Sci. Technol. A 1986, 4 (6), 2701.
107. Knotek, O.; Bosch, W.; Leyendecker, T. In First Int. Conf. on Surface Engineering; Bucklow, I. A., Ed.; Brighton, UK. 1985; pp 197–206.
108. Jehn, H. A.; Hofmann, S.; Münz, W.-D. Thin Solid Films 1987, 153 (1–3), 45.
109. McIntyre, D.; Greene, J. E.; Håkansson, G.; Sundgren, J.-E.; Münz, W.-D. J. Appl. Phys. 1990, 67 (3), 1542.
110. Holleck, H. Surf. Coat. Technol. 1988, 36 (1–2), 151.
111. Cremer, R.; Witthaut, M.; Von Richthofen, A.; Neuschütz, D. Fresenius J. Anal. Chem. 1998, 361 (6–7), 642.
112. Hasegawa, H.; Kimura, A.; Suzuki, T. Surf. Coat. Technol. 2000, 132 (1), 76.
113. Spencer, P. Int. J. Mater. Res. 2001, 92 (10), 1145.
114. Wuhrer, R.; Yeung, W. Y.; Phillips, M. R.; McCredie, G. Thin Solid Films 1996, 290–291, 339.
115. Vaz, F.; Rebouta, L.; Andritschky, M.; Da Silva, M. F.; Soares, J. C. J. Eur. Ceram. Soc. 1997, 17 (15–16), 1971.
116. Vaz, F.; Rebouta, L.; Andritschky, M.; Da Silva, M. F.; Soares, J. C. Surf. Coat. Technol. 1998, 98 (1–3), 912.
117. Huang, J. L.; Shew, B. Y. J. Am. Ceram. Soc. 1999, 82 (3), 696.
118. Hugon, M. C.; Varniere, F.; Letendu, F.; Agius, B.; Vickridge, I.; Kingon, A. I. J. Mater. Res. 2001, 16 (9), 2591.
119. PalDey, S.; Deevi, S. C. Mater. Sci. Eng. A 2003, 361 (1–2), 1.
120. Brizuela, M.; Garcia-Luis, A.; Braceras, I.; Onate, J. I.; Sanchez-Lopez, J. C.; Martinez-Martinez, D.; Lopez-Cartes, C.; Fernandez, A. Surf. Coat. Technol. 2005,
200 (1–4), 192.
121. Klostermann, H.; Fietzke, H.; Modes, T.; Zywitzki, O. Rev. Adv. Mater. Sci. 2007, 15 (1), 33.
122. Barshilia, H. C.; Deepthi, B.; Rajam, K. S.; Bhatti, K. P.; Chaudhary, S. J. Mater. Res. 2008, 23 (5), 1258.
123. Kolozsvari, S.; Pesch, P.; Ziebert, C.; Stueber, M.; Ulrich, S. Plasma Processes Polym. 2009, 6 (1), 146.
124. Franz, R.; Lechthaler, M.; Polzer, C.; Mitterer, C. Surf. Coat. Technol. 2010, 204 (15), 2447.
125. Romero, J.; Gomez, M. A.; Esteve, J.; Montala, F.; Carreras, L.; Grifol, M.; Lousa, A. Thin Solid Films 2006, 515, 113.
126. Vetter, J.; Lugscheider, E.; Guerreiro, S. S. Surf. Coat. Technol. 1998, 98, 1233.
127. Ding, X.-Z.; Tan, A. L. K.; Zeng, X. T.; Wang, C.; Yue, T.; Sun, C. Q. Thin Solid Films 2008, 516, 5716.
128. Ding, X.-Z.; Zeng, X. T. Surf. Coat. Technol. 2005, 200, 1372.
129. Spain, E.; Avelar-Batista, J. C.; Letch, M.; Housden, J.; Lerga, B. Surf. Coat. Technol. 2005, 200, 1507.
130. Kalss, W.; Reiter, A.; Derflinger, V.; Gey, C.; Endrino, J. L. Int. J. Refract. Metals Hard Mater. 2006, 24, 399.
131. Bobzin, K.; Lugscheider, E.; Maes, M.; Gold, P. W.; Loos, J.; Kuhn, M. Surf. Coat. Technol. 2004, 188–189, 649.
132. Lugscheider, E.; Bobzin, K.; Bärwulf, S.; Hornig, T. Surf. Coat. Technol. 2000, 133–134, 540.
133. Lugscheider, E.; Bobzin, K.; Hornig, T.; Maes, M. Thin Solid Films 2002, 420–421, 318.
134. Reiter, A. E.; Brunner, B.; Ante, M.; Rechberger, J. Surf. Coat. Technol. 2006, 200, 5532.
135. Hörling, A.; Hultman, L.; Oden, M.; Sjolen, J.; Karlsson, L. J. Vac. Sci. Technol. A 2002, 20, 1815.
136. Mayrhofer, P. H.; Hultman, L.; Schneider, J. M.; Staron, P.; Clemens, H. Int. J. Mater. Res. 2007, 11, 1054.
137. Kawate, M.; Kimura Hashimoto, A.; Suzuki, T. Surf. Coat. Technol. 2003, 165, 163.
138. Banakh, O.; Schmid, P. E.; Sanjines, R.; Levy, F. Surf. Coat. Technol. 2003, 163–164, 57.
139. Hofmann, S. Thin Solid Films 1990, 193–194, 648.
140. Rovere, F.; Mayrhofer, P. H.; Reinholdt, A.; Mayer, J.; Schneider, J. M. Surf. Coat. Technol. 2008, 202, 5870.
www.iran-mavad.com
141. Kimura, A.; Kawate, M.; Hasegawa, H.; Suzuki, T. Surf. Coat. Technol. 2003, 169–170, 367.
142. Holec, D.; Rovere, F.; Mayrhofer, P. H.; Barna, P. B. Scr. Mater. 2010, 62 (6), 349.
143. Makino, Y. Surf. Coat. Technol. 2005, 193, 185.
144. Mayrhofer, P. H.; Kunc, F.; Musil, J.; Mitterer, C. Thin Solid Films 2002, 415, 151.
145. Mayrhofer, P. H.; Mitterer, C.; Musil, J. Surf. Coat. Technol. 2003, 174–175, 725.
146. Ohring, M. Materials Science of Thin Films – Deposition & Structure; Academic Press: San Diego, 2002.
147. Cahn, J. W. Acta Metall. 1961, 9 (9), 795.
148. Willmann, H.; Mayrhofer, P. H.; Persson, P. O. A.; Reiter, A. E.; Hultman, L.; Mitterer, C. Scr. Mater. 2006, 54, 1847.
149. Rovere, F.; Mayrhofer, P. H. J. Vac. Sci. Technol. A 2008, 26, 29.
150. Willmann, H.; Mayrhofer, P. H.; Hultman, L.; Mitterer, C. J. Mater. Res. 2008, 23 (11), 2880.
151. Gottstein, G. Physikalische Grundlagen der Materialkunde; Springer: Berlin, 2001.
152. Alling, B.; Odén, M.; Hultman, L.; Abrikosov, I. A. Appl. Phys. Lett. 2009, 95 (18), 181906.
153. Porter, A. D.; Easterling, K. E. Phase Transformations in Metals and Alloys; Van Nostrand-Reinhold: New York, 1981.
154. Cahn, J. W.; Haasen, P. Phase Transformations in Materials; VCH: New York, 1991.
155. Alling, B.; Karimi, A.; Hultman, L.; Abrikosov, I. A. Appl. Phys. Lett. 2008, 92 (7), 071903.
156. Christensen, N. E.; Gorczyca, I. Phys. Rev. B 1993, 47 (8), 4307.
157. Rachbauer, R.; Massl, S.; Stergar, E.; Felfer, P.; Mayrhofer, P. H. Surf. Coat. Technol. 2010, 204, 1811.
158. Rachbauer, R.; Massl, S.; Stergar, E.; Holec, D.; Kiener, D.; Keckes, J.; Patscheider, J.; Stiefel, M.; Leitner, H.; Mayrhofer, P. H. J. Appl. Phys. 2011, 110, 023515.
159. Mayrhofer, P. H.; Willmann, H.; Reiter, A. E. Surf. Coat. Technol. 2008, 202, 4935.
160. Mayrhofer, P. H.; Willmann, H.; Hultman, L.; Mitterer, C. J. Phys. D Appl. Phys. 2008, 41, 155316.
161. Thompson, N. G.; Yunovich, M.; Dunmire, D. Corros. Rev. 2007, 25, 247.
162. Darken, L. S.; Gurry, R. W. Physical Chemistry of Metals; McGraw-Hill: New York, 1953.
163. Hindam, H.; Whittle, D. P. Oxid. Met. 1982, 18, 245.
164. Birks, N.; Meier, G. H. Introduction to High Temperature Oxidation of Metals; Edwar Arnold Ltd: London, 1983.
165. Ashby, M. F.; Jones, D. R. H. Engineering Materials 1; Butterworth-Heinemann: Oxford, 1996.
166. Leyens, C. Oxidation and Protection of Titanium Alloys and Titanium Aluminides. In Titanium and Titanium Alloys; Leyens, C., Peters, M., Eds.; Wiley-VCH: Köln, 2003; p 187.
167. Wang, C. C.; Akbar, S. A.; Chen, W.; Patton, V. D. J. Mater. Sci. 1995, 30, 1627.
168. Hagel, W. C. J. Am. Ceram. Soc. 1965, 48, 70.
169. Oishi, Y.; Kingery, W. D. J. Chem. Phys. 1960, 33, 480.
170. Prot, D.; Monty, C. Philos. Mag. A 1996, 73, 899.
171. Reed, D. J.; Wuensch, B. J. J. Am. Ceram. Soc. 1980, 63, 88.
172. Hagel, W. C. Trans. Am. Soc. Metals 1963, 56, 583.
173. Giggins, C. S.; Pettit, F. S. Metall. Trans. 1971, 2, 1071.
174. Lillerud, K. P.; Kofstad, P. J. Electrochem. Soc. 1980, 127, 2397.
175. Kofstad, P.; Lillerud, K. P. J. Electrochem. Soc. 1980, 127, 2410.
176. Kofstad, P.; Lillerud, K. P. Oxid. Met. 1982, 17, 177.
177. Lillerud, K. P.; Kofstad, P. Oxid. Met. 1982, 17, 195.
178. Atkinson, A.; Taylor, R. I.; Hughes, A. E. Philos. Mag. A 1982, 45, 823.
179. Caplan, D.; Sproule, G. I. Oxid. Met. 1975, 9, 459.
180. Stroosnijder, M. F.; Sunderkötter, J. D.; Cristóbal, M. J.; Jenett, H.; Isenbügel, K.; Baker, M. A. Surf. Coat. Technol. 1996, 83, 205.
181. Gulbransen, E. A.; Andrew, K. F. J. Electrochem. Soc. 1957, 104, 334.
182. Stott, F. H.; Wood, G. C.; Stringer, J. Oxid. Met. 1995, 44, 113.
183. Prescott, R.; Graham, M. J. Oxid. Met. 1992, 38, 233.
184. Quadakkers, W. J.; Holzbrecher, H.; Briefs, K. G.; Beske, H. Oxid. Met. 1989, 32, 67.
185. Young, E. W. A.; De Wit, J. H. W. Solid State Ionics 1985, 16, 39.
186. Ben Abderrazik, G.; Moulin, G.; Huntz, A. M.; Young, E. W. A.; De Wit, J. H. W. Solid State Ionics 1987, 22, 285.
187. Reddy, K. P. R.; Smialek, J. L.; Cooper, A. R. Oxid. Met. 1982, 17, 429.
188. Grabke, H. J. Mater. Sci. Forum 1997, 251–254, 149.
189. Stott, F. H. Mater. Sci. Forum 1997, 251–254, 19.
190. Graham, M. J.; Eldrige, J. I.; Mitchell, D. F.; Hussey, R. J. Mater. Sci. Forum 1989, 43, 207.
191. Van Manen, P. A.; Young, E. W. A.; Schalkoord, D.; van der Wekken, C. J.; De Wit, J. H. W. Surf. Interface Anal. 1988, 12, 391.
192. Rybicki, G. C.; Smialek, J. L. Oxid. Met. 1989, 31, 275.
193. Tedmon, C. S. J. J. Electrochem. Soc. 1966, 113, 766.
194. Predel, B. Landolt-Börnstein – Group IV Physical Chemistry. In Numerical Data and Functional Relationships in Science and Technology, Ni-Np – Pt-Zr; Madelung, O., Ed.;
Springer: Berlin Heidelberg, 1998; p 1.
195. Predel, B. Landolt-Börnstein – Group IV Physical Chemistry. In Numerical Data and Functional Relationships in Science and Technology, Dy-Er – Fr-Mo; Madelung, O., Ed.;
Springer: Berlin Heidelberg, 1995; p 1.
196. Stott, F. H. Mater. Sci. Eng. 1987, 87, 267.
197. Briant, C. L.; Mulford, R. A. Metall. Trans. A 1982, 13, 745.
198. Funkenbusch, A. W.; Smeggil, J. G.; Bornstein, N. S. Metall. Trans. A 1985, 16, 1164.
199. Grabke, H. J.; Kurbatov, G.; Schmutzler, H. J. Oxid. Met. 1995, 43, 97.
200. Schmutzler, H. J.; Viefhaus, H.; Grabke, H. J. Surf. Interface Anal. 1992, 18, 581.
201. Grabke, H. J.; Wiemer, D.; Viefhaus, H. Appl. Surf. Sci. 1991, 47, 243.
202. Grabke, H.; Dennert, R.; Wagemann, B. Oxid. Met. 1997, 47, 495.
203. Gibbs, G. B.; Hales, R. Corros. Sci. 1977, 17, 487.
204. Gibbs, G. B. Oxid. Met. 1981, 16, 147.
205. Rommerskirchen, I.; Eltester, B.; Grabke, H. J. Mater. Corros. – Werkstoffe und Korrosion 1996, 47, 646.
206. Kuenzly, J. D.; Douglass, D. L. Oxid. Met. 1974, 8, 139.
207. Kumar, A.; Nasrallah, M.; Douglass, D. L. Oxid. Met. 1974, 8, 227.
208. Hessenbruch, W. Metalle und Legierungen für hohe Temperaturen; Springer: Berlin, 1940.
209. Pfeil, L. B. U.K. Patent 459848, 1937.
210. Chevalier, S.; Larpin, J. P. Mater. Sci. Eng. A 2003, 363, 116.
211. Pint, B. A. Oxid. Met. 1996, 45, 1.
www.iran-mavad.com
212. Ul-Hamid, A. Oxid. Met. 2002, 58, 23.

213. Huntz, A. M. In The Role of Active Elements in the Oxidation of High Temperature Metals and Alloys; Lang, E., Ed.; Elsevier Applied Science, Springer: London and New York,
1989; p 81.
214. Provenzano, V.; Sadananda, K.; Louat, N. P.; Reed, J. R. Surf. Coat. Technol. 1988, 36, 61.
215. Schumann, E. Oxid. Met. 1995, 43, 157.
216. Tawancy, H. M.; Abbas, N. M.; Bennett, A. Surf. Coat. Technol. 1994, 68–69, 10.
217. Grabke, H. J. Mater. Sci. Forum 1999, 294–296, 135.
218. Stringer, J.; Wilcox, B. A.; Jaffee, R. I. Oxid. Met. 1972, 5, 11.
219. Ramanarayanan, T. A.; Raghavan, M.; Petkovic-Luton, R. J. Electrochem. Soc. 1984, 131, 923.
220. Quadakkers, J.; Singheiser, L. Mater. Sci. Forum 2001, 369–372, 77.
221. Jedlinski, J. Oxid. Met. 1993, 39, 55.
222. Burtin, P.; Brunelle, J. P.; Pijolat, M.; Soustelle, M. Appl. Catal. 1987, 34, 225.
223. Rommerskirchen, I.; Kolarik, V. Mater. Corros./Werkstoffe und Korrosion 1996, 47, 625.
224. Burtin, P.; Brunelle, J. P. Appl. Catal. 1987, 34, 239.
225. Besmann, T. M.; Kulkarni, N. S.; Spear, K. E. J. Am. Ceram. Soc. 2006, 89, 638.
226. Bunting, E. N. Bur. Stand. J. Res. 1931, 6, 947.
227. Sitte, W. Investigation of the Miscibility Gap of the System Chromia-Alumina below 1300 C. In Reactivity of Solids; Barret, P., Dufour, L.-C., Eds.; Elsevier: Dijon, 1985; p 451.
228. Braue, W.; Mechnich, P.; Fritscher, K.; Niewolak, L. Surf. Coat. Technol. 2007, 202, 670.
229. Hultman, L.; Hentzell, H. T. G.; Sundgren, J.-E.; Johansson, B.-O.; Helmersson, U. Thin Solid Films 1985, 124 (2), 163.
230. Flink, A.; Andersson, J. M.; Alling, B.; Daniel, R.; Sjölén, J.; Karlsson, L.; Hultman, L. Thin Solid Films 2008, 517 (2), 714.
231. Beckers, M.; Schell, N.; Martins, R. M. S.; Mücklich, A.; Möller, W. Appl. Phys. Lett. 2006, 89 (7), 074101.
232. Beckers, M.; Schell, N.; Martins, R. M. S.; Mücklich, A.; Möller, W.; Hultman, L. J. Appl. Phys. 2006, 99 (3), 034902.
233. Müller, E. W. Zeitschrift fu¨r Physik 1951, 131 (1), 136.
234. Miller, M. K.; Cerezo, A.; Hetherington, M. G.; Smith, G. D. W. Atom Probe Field Ion Microscopy; Oxford University Press: Oxford, 1996.
235. Russell, K. F.; Miller, M. K.; Ulfig, R. M.; Gribb, T. Ultramicroscopy 2007, 107 (9), 750.
236. Miller, M. K. Atom Probe Tomography: Analysis at the Atomic Level; Kluwer Academic/Plenum Publishers: New York, 2000.
237. Miller, M. K. Mater. Charact. 2000, 44 (1–2), 11–27.
238. Cerezo, A.; Abraham, M.; Clifton, P.; Lane, H.; Larson, D. J.; Petford-Long, A. K.; Thuvander, M.; Warren, P. J.; Smith, G. D. W. Micron 2001, 32 (8), 731.
239. Hyde, J. M.; Miller, M. K.; Hetherington, M. G.; Cerezo, A.; Smith, G. D. W.; Elliott, C. M. Acta Metall. Mater. 1995, 43 (9), 3403.
240. Hyde, J. M.; Miller, M. K.; Hetherington, M. G.; Cerezo, A.; Smith, G. D. W.; Elliott, C. M. Acta Metall. Mater. 1995, 43 (9), 3415.
241. Miller, M. K.; Hyde, J. M.; Hetherington, M. G.; Cerezo, A.; Smith, G. D. W.; Elliott, C. M. Acta Metall. Mater. 1995, 43 (9), 3385.
242. Kingham, D. R. Surf. Sci. 1982, 116 (2), 273.
243. Miller, M. K.; Forbes, R. G. Mater. Charact. 2009, 60 (6), 461.
244. Kellogg, G. L. J. Appl. Phys. 1981, 52 (8), 5320.
245. Cerezo, A.; Clifton, P. H.; Galtrey, M. J.; Humphreys, C. J.; Kelly, T. F.; Larson, D. J.; Lozano-Perez, S.; Marquis, E. A.; Oliver, R. A.; Sha, G.; Thompson, K.; Zandbergen, M.;
Alvis, R. L. Mater. Today 2007, 10 (12), 36.
246. Veprek, S.; Veprek-Heijman, M. G. J.; Karvankova, P.; Prochazka, J. Thin Solid Films 2005, 476 (1), 1.
247. Tang, F.; Gault, B.; Ringer, S. P.; Martin, P.; Bendavid, A.; Cairney, J. M. Scr. Mater. 2010, 63 (2), 192.
248. Tang, F.; Gault, B.; Ringer, S. P.; Cairney, J. M. Ultramicroscopy 2010, 110 (7), 836.
249. Gault, B.; Vurpillot, F.; Vella, A.; Gilbert, M.; Menand, A.; Blavette, D.; Deconihout, B. Rev. Sci. Instrum. 2006, 77 (4), 043705.
250. Seidman, D. N. Ann. Rev. Mater. Res. 2007, 37, 127.
251. Blavette, D.; Kassab, T. A.; Cadel, E.; Mackel, A.; Vurpillot, F.; Gilbert, M.; Cojocaru, O.; Deconihout, B. Int. J. Mater. Res. 2008, 99 (5), 454.
252. Shariq, A.; Mutas, S.; Wedderhoff, K.; Klein, C.; Hortenbach, H.; Teichert, S.; Kücher, P.; Gerstl, S. S. A. Ultramicroscopy 2009, 109 (5), 472.
253. Gault, B.; Menand, A.; De Geuser, F.; Deconihout, B.; Danoix, R. Appl. Phys. Lett. 2006, 88 (11).
254. Kelly, T. F.; Larson, D. J.; Thompson, K.; Alvis, R. L.; Bunton, J. H.; Olson, J. D.; Gorman, B. R. Ann. Rev. Mater. Res. 2007, 37, 681.
255. Kelly, T. F.; Miller, M. K. Rev. Sci. Instrum. 2007, 78 (3), 031101.
256. IVAS Software; IMAGO Inc., 2010.
257. Miller, M. K.; Russell, K. F. Ultramicroscopy 2007, 107 (9), 761.
258. Thompson, K.; Lawrence, D.; Larson, D. J.; Olson, J. D.; Kelly, T. F.; Gorman, B. Ultramicroscopy 2007, 107 (2–3), 131.
259. Larson, D. J.; Foord, D. T.; Petford-Long, A. K.; Liew, H.; Blamire, M. G.; Cerezo, A.; Smith, G. D. W. Ultramicroscopy 1999, 79 (1–4), 287.
260. Miller, M. K.; Russell, K. F.; Thompson, G. B. Ultramicroscopy 2005, 102 (4), 287.
261. Miller, M. K.; Russell, K. F.; Thompson, K.; Alvis, R.; Larson, D. J. Microsc. Microanal. 2007, 13 (6), 428.
262. Larson, D. J.; Petford-Long, A. K.; Ma, Y. Q.; Cerezo, A. Acta Mater. 2004, 52 (10), 2847.
263. Thompson, G. B.; Miller, M. K.; Fraser, H. L. Ultramicroscopy 2004, 100 (1–2), 25.
264. Rachbauer, R. Age Hardening in Transition Metal Aluminium Nitride Thin Films Studied at the Atomic Scale. Ph.D. Thesis, Montanuniversität Leoben, Leoben, AT, July 2011.
265. Flink, A.; Beckers, M.; Sjölén, J.; Larsson, T.; Braun, S.; Karlsson, L.; Hultman, L. J. Mater. Res. 2009, 24 (8), 2483.
266. Bunton, J. H.; Olson, J. D.; Lenz, D. R.; Kelly, T. F. Microsc. Microanal. 2007, 13 (6), 418.
267. Williamson, G. K.; Hall, W. H. Acta Metall. 1953, 1 (1), 22.
268. Hellman, O. C.; Vandenbroucke, J. A.; Rüsing, J.; Isheim, D.; Seidman, D. N. Microsc. Microanal. 2000, 6 (5), 437.
269. Langer, J. S.; Bar-On, M.; Miller, H. D. Phys. Rev. A 1975, 11 (4), 1417.
270. Capdevila, C.; Miller, M. K.; Russell, K. F.; Chao, J.; González-Carrasco, J. L. Mater. Sci. Eng. A 2008, 490 (1–2), 277.
271. Rachbauer, R.; Stergar, E.; Massl, S.; Moser, M.; Mayrhofer, P. H. Scr. Mater. 2009, 61, 725.
272. Hellman, O. C.; Du Rivage, J. B.; Seidman, D. N. Ultramicroscopy 2003, 95, 199.
273. Oliver, W. C.; Pharr, G. M. J. Mater. Res. 1992, 7 (6), 1564.
274. Odén, M.; Rogström, L.; Terner, M.; Hedström, P.; Almer, J.; llavsky, J. Appl. Phys. Lett. 2009, 94 (5), 053114.
275. Knutsson, A.; Johansson, M. P.; Persson, P. O.Å.; Hultman, L.; Odén, M. Appl. Phys. Lett. 2008, 93 (14), 143110.
276. Rafaja, D.; Wüstefeld, C.; Baehtz, C.; Klemm, V.; Dopita, M.; Motylenko, M.; Michotte, C.; Kathrein, M. Metall. Mater. Trans. A 2011, 42 (3), 559.
277. Heim, D.; Hochreiter, R. Surf. Coat. Technol. 1998, 98 (1–3), 1553.
278. Holubár, P.; Jílek, M.; Síma, M. Surf. Coat. Technol. 1999, 120–121, 184.
279. Rebholz, C.; Schneider, J. M.; Voevodin, A. A.; Steinebrunner, J.; Charitidis, C.; Logothetidis, S.; Leyland, A.; Matthews, A. Surf. Coat. Technol. 1999, 113 (1–2), 126.
280. Durand-Drouhin, O.; Santana, A. E.; Karimi, A.; Derflinger, V. H.; Schütze, A. Surf. Coat. Technol. 2003, 163–164, 260.
281. Lackner, J. M.; Waldhauser, W.; Ebner, R.; Bakker, R. J.; Schöberl, T.; Major, B. Thin Solid Films 2004, 468 (1–2), 125.
282. Parlinska-Wojtan, M.; Karimi, A.; Coddet, O.; Cselle, T.; Morstein, M. Surf. Coat. Technol. 2004, 188–189 (1–3), 344.
www.iran-mavad.com
283. Stueber, M.; Barna, P. B.; Simmonds, M. C.; Albers, U.; Leiste, H.; Ziebert, C.; Holleck, H.; Kovács, A.; Hovsepian, P.; Gee, I. Thin Solid Films 2005, 493 (1–2), 104.
284. Donohue, L. A.; Smith, I. J.; Münz, W.-D.; Petrov, I.; Greene, J. E. Surf. Coat. Technol. 1997, 94–95, 226.
285. Donohue, L. A.; Lewis, D. B.; Münz, W.-D.; Stack, M. M.; Lyon, S. B.; Wang, H.-W.; Rafaja, D. Vacuum 1999, 55 (2), 109.
286. Pflüger, E.; Schroer, A.; Voumard, P.; Donohue, L.; Münz, W.-D. Surf. Coat. Technol. 1999, 115 (1), 17.
287. Moser, M.; Kiener, D.; Scheu, C.; Mayrhofer, P. H. Materials 2010, 3, 1573.
288. Knotek, O.; Leyendecker, T.; Jungblut, F. Thin Solid Films 1987, 153 (1–3), 83.
289. Kutschej, K.; Mayrhofer, P. H.; Kathrein, M.; Polcik, P.; Mitterer, C. Surf. Coat. Technol. 2004, 188, 358.
290. Mayrhofer, P. H.; Hovsepian, P. E.; Mitterer, C.; Münz, W.-D. Surf. Coat. Technol. 2004, 177–178, 341.
291. Kutschej, K.; Mayrhofer, P. H.; Kathrein, M.; Polcik, P.; Mitterer, C. Surf. Coat. Technol. 2005, 200 (5–6), 1731.
292. Pfeiler, M.; Kutschej, K.; Penoy, M.; Michotte, C.; Mitterer, C.; Kathrein, M. Surf. Coat. Technol. 2007, 202 (4–7), 1050.
293. Pfeiler, M.; Kutschej, K.; Penoy, M.; Michotte, C.; Mitterer, C.; Kathrein, M. Int. J. Refract. Metals Hard Mater. 2009, 27 (2), 502.
294. Gassner, G.; Mayrhofer, P. H.; Kutschej, K.; Mitterer, C.; Kathrein, M. Surf. Coat. Technol. 2006, 201 (6), 3335.
295. Reeswinkel, T.; Sangiovanni, D. G.; Chirita, V.; Hultman, L.; Schneider, J. M. Surf. Coat. Technol. 2011, 205 (20), 4821.
296. Knotek, O.; Böhmer, M.; Leyendecker, T.; Jungblut, F. Mater. Sci. Eng. 1988, 105–106 (2), 481.
297. Donohue, L. A.; Cawley, J.; Brooks, J. S.; Münz, W.-D. Surf. Coat. Technol. 1995, 74–75 (1), 123.
298. Cremer, R.; Neuschütz, D. Surf. Coat. Technol. 2001, 146–147, 229.
299. Kutschej, K.; Fateh, N.; Mayrhofer, P. H.; Kathrein, M.; Polcik, P.; Mitterer, C. Surf. Coat. Technol. 2005, 200 (1–4), 113.
300. Feng, C.; Zhu, S.; Li, M.; Xin, L.; Wang, F. Surf. Coat. Technol. 2008, 202 (14), 3257.
301. Pfeiler, M.; Fontalvo, G. A.; Wagner, J.; Kutschej, K.; Penoy, M.; Michotte, C.; Mitterer, C.; Kathrein, M. Tribol. Lett. 2008, 30 (2), 91.
302. Pfeiler, M.; Scheu, C.; Hutter, H.; Schnöller, J.; Michotte, C.; Mitterer, C.; Kathrein, M. J. Vac. Sci. Technol. A 2009, 27 (3), 554.
303. Braic, V.; Zoita, C. N.; Balaceanu, M.; Kiss, A.; Vladescu, A.; Popescu, A.; Braic, M. Surf. Coat. Technol. 2010, 204 (12–13), 1925.
304. Chen, L.; Holec, D.; Du, Y.; Mayrhofer, P. H. Thin Solid Films 2011, 519 (16), 5503.
305. Kresse, G.; Furthmüller, J. Comput. Mater. Sci. 1996, 6 (1), 15.
306. Kresse, G.; Furthmüller, J. Phys. Rev. B 1996, 54 (16), 11169.
307. Blaha, P.; Sorantin, P. I.; Trickey, S. B. Comput. Phys. Commun. 1990, 59, 399.
308. Kresse, G.; Joubert, D. Phys. Rev. B 1999, 59 (3), 1758.
309. Wang, Y.; Perdew, J. P. Phys. Rev. B 1991, 44 (24), 13298.
310. Wei, S. H.; Ferreira, L. G.; Bernard, J. E.; Zunger, A. Phys. Rev. B 1990, 42 (15), 9622.
311. Holec, D., Ed. Multi-Scale Modelling of III-Nitrides: Selected Topics from Dislocations to the Electronic Structure; VDM Verlag Dr. Müller: Saarbrücken, 2010.
312. Holec, D.; Rachbauer, R.; Kiener, D.; Cherns, P. D.; Costa, P. M. F. J.; McAleese, C.; Mayrhofer, P. H.; Humphreys, C. J. Phys. Rev. B 2011, 83 (16), 165122.
313. Rachbauer, R.; Holec, D.; Lattemann, M.; Hultman, L.; Mayrhofer, P. H. Int. J. Mater. Res. 2011, 102, 735.
314. Mayrhofer, P. H.; Rachbauer, R.; Holec, D. Scr. Mater. 2010, 63, 807.
315. Rachbauer, R.; Holec, D.; Mayrhofer, P. H. Appl. Phys. Lett. 2010, 97, 151901.
316. Rachbauer, R.; Holec, D.; Mayrhofer, P. H. Surf. Coat. Technol. 2012, 211, 98.
317. Rachbauer, R.; Blutmager, A.; Holec, D.; Mayrhofer, P. H. Surf. Coat. Technol. 2012, 206, 2667.
318. Moser, M.; Mayrhofer, P. H. Scr. Mater. 2007, 57 (4), 357.
319. Rovere, F. Theoretical and Experimental Assessment of Cr-Al-Y-N as Protective Coatings for g-TiAl Based Alloys; Shaker Verlag GmbH: Aachen, 2010.
320. Vegard, L. Zeitschrift fu¨r Physik 1921, 5 (1), 17.
321. Holleck, H. J. Vac. Sci. Technol. A 1986, 4 (6), 2661.
322. Feng, C.; Zhu, S.; Li, M.; Xin, L.; Wang, F. Oxid. Met. 2009, 71 (1–2), 63.
323. Leyens, C. In Titanium and Titanium Alloys; Peters, M., Ed.; Wiley-VCH: New York, 2003; pp 207–209.
www.iran-mavad.com
4.15 Advanced Carbon-Based Coatings
E Broitman and L Hultman, Linköping University, Linköping, Sweden

4.15.2 Bonding Configurations in Carbon Films 391
4.15.3 Carbon Nitride Films 392
4.15.3.1 Fullerene-Like Carbon Nitride Films 393
4.15.3.2 Bonding Configuration in Carbon Nitride Films 394
4.15.3.3 The Role of Nitrogen for the Formation of Fullerene-Like Carbon Nitride 397
4.15.3.3.1 Evolution of Curvature 397
4.15.3.3.2 Formation of Interplanar Cross-Linking 398
4.15.3.3.3 Formation of CxNy (x, y 2) Species 398
4.15.3.4 Structure Evolution Mechanisms 400
4.15.4 Phosphorous-Carbide Films 401
4.15.4.1 Synthetic Growth Predictions 402
4.15.4.2 Experimental Data on Fullerene-Like CPx Films 403
4.15.5 Carbon Fluoride Films 405
4.15.5.1 Synthetic Growth Predictions 406
4.15.5.2 Deposition of CFx Films by HiPIMS 407
Acknowledgments 410
Bibliography 410
4.15.1 Introduction
The use of carbon coatings can be traced to prehistoric times, about 32 000 years ago. Ancient peoples decorated walls of protected
caves with a carbon-based paint made from charcoal mixed with spit or animal fat. Other forms of carbon coatings were obtained by
using charcoal from the fire in the form of crayon sticks (1). More recently, another similar application of carbon coatings made of
China ink, invented about 2600 BC by Tien-Tchen, was made by mixing water with the soot produced from the smoke of pines, and
it was used to paint and write with ink sticks (2).
From a scientific point of view, the first research about carbon-based coatings probably was done by Herschel in 1800, when he
studied the optical properties of ‘smoked-glass’ films (3). These coatings, produced on the surface of a glass by the smoke residue of
a candle or other inefficiently burning hydrocarbon, were already used by the end of the eighteenth century for solar observations
with telescopes. Another method of carbon deposition was also reported at the same time by Henry through electric spark
discharges in hydrocarbon gases (4), but he did not propose any practical application. It was almost another century after Henry’s
report when Sawyer and Man patented in 1880 a new method to deposit carbon films by vacuum pyrolysis of hydrocarbon material
to improve the filament life of electric lamps (5). The patent was never used in practice, but this new method would be applied some
years later for other very important applications.
At the end of the nineteenth century, the first industrial-scale use of carbon coatings appeared: carbon-film resistors, consisting of
fine carbon particles dispersed in a resinous binder, were made by Gambrell and Harris in 1897 (6). These resistors were cheap and
easy to produce, but were not very stable. High-stability fixed-carbon resistors of a wide range of values were developed during
World War II for special applications, and were massively used after the war. Smooth ceramic rods were coated with carbon films by
pyrolytic decomposition of methane in a continuous furnace. The obtained resistance value was increased by cutting a spiral track
on the coated surface of the rod. A layer of graphite was applied to the ends of the rods, which were then fitted with metal caps and
lead wires as contacts (7).
In 1954, Bradley proposed the use of carbon films for replica production in transmission electron microscopy (TEM). Films were
deposited using a pointed rodtype carbon arc source (8,9). The method from Bradley was immediately adapted by TEM researchers
in materials science and biology, and Ernest F. Fullam, who was an active researcher in TEM, had the idea to fabricate and sell small
carbon arc evaporators for his colleagues; these became one of the first products of his well-known company. Bradley’s work also
sparked the study of carbon films. Blue and Danielson were the first to study the optical, electrical, and mechanical properties of arc-
evaporated carbon films. They also suggested their practical use as optical attenuators, thermistors, and windows in nuclear physics
studies (10).
Regarding the use of plasmas to deposit carbon films, Henry’s work published in 1797 (4) was rediscovered in 1934 by Stewart,
who noted the deposition of inert and glassy carbonaceous films on the walls of gas discharge tubes containing hydrocarbon vapors
and other carbon-containing vapors (11). In 1953, Schmellenmeir studied plasma-deposited carbon films and found the presence
of ‘diamond-like’ structures (12). In a series of papers published between 1969 and 1973, Aisenberg and Chabot discussed ion beam

www.iran-mavad.com
390 Advanced Carbon-Based Coatings
deposition of elements and compounds using a plasma source where ions of the electrode material (carbon and doped silicon) were
extracted into the deposition chamber (13–15). The energetic carbon ion beam used to deposit carbon films on plastic lenses and
blade edges resulted in a material that resembled the mechanical properties of diamond rather than graphite. The tag ‘diamond-like
carbon’ (DLC) was born, and the research on carbon-based coatings took off. In 1980, Weissmantel proposed the term ‘i-C’ since the
common feature of all known preparation techniques of DLC films at that time seemed to be the supply of activation energy and
momentum by relatively energetic ions (16).
Figure 1 summarizes the number of scientific papers per year that have been published about DLC films in the period
1970–2012. Using the keywords ‘diamond-like’ and ‘carbon films’ in the Google Scholar database, as of 1 April 2013, 14300
publications (including 1700 patents) mention both keywords. Despite their numerous attractive properties, it seems from the
figure that DLC films did not draw much attention throughout the 1970s or even until the mid-1980s, as was also pointed out in
a recent review by Erdermir and Donnet (17). Their view is that this neglect may have originated with the news, at that time, of
crystalline diamond film deposition using low-pressure chemical vapor deposition (CVD). However, inherent difficulties plagued
the large-scale production of diamond as thin films or free-standing large crystals that everybody was dreaming of producing at
reasonable costs (17). Anyway, the data show that over the years, scientific interest in these films has grown tremendously, and
nowadays they can be considered some of the most important tribological coatings in numerous application fields.
At the same time that Aisenberg and Chabot were doing their experiments, Culberton used for the first time a combination of
plasma-enhanced chemical vapor deposition (PECVD) and physical vapor deposition (PVD) to deposit carbon and metal carbide
coatings on electron tube grids to provide a low secondary emission coefficient (18). Ion-plated carbon and other-material (Ti, Cr,
Zr, and Si) carbide coatings from chemical vapor precursors (e.g., TiCl4 and C2H2) were deposited on electron tube grids to decrease
the secondary electron emission. This hybrid process to deposit metal carbide coatings by ion plating is nowadays a common
industrial process.
During the last two decades, nitrogen doping, metal, and other alloying additions have also been developed to improve the
mechanical, optical, and electronic properties of DLC coatings (19). The experimental work regarding the doping of carbon films
with nitrogen was sparked when, in 1989, a theoretical calculation from Liu and Cohen predicted that the hypothetical single crystal
b-C3N4 would have a bulk modulus higher than that of diamond (20,21). The result was misinterpreted by some researchers, who
thought that this carbon nitride material would be ‘harder than diamond.’ So far, all attempts to synthesize this phase resulted in
other crystalline structures or amorphous and short-range-ordered CN allotropes, like the ‘fullerene-like’ (FL) carbon nitride
discovered at Linköping University in 1994 (22).
Metal doping or carbide doping was developed in the mid-1980s to try to improve DLC films by enhancing adhesion, thermal
stability, and toughness. These new tribological films can be considered to have a nanocomposite structure and offer improved
hardness and elastic modulus with a certain ductility. Carbide doped hydrogenated DLCs prepared by PVD are commercially
available, in particular tungsten carbide doped (WC/C) DLC, a chemically inert film with high elasticity and good wear resistance
(19). Another example is Ti–C nanocomposites deposited by PVD methods (23).
The introduction of Si in carbon films has been used to suppress the formation of aromatic structures and to promote the
diamondlike character of the films by forming tetrahedral bonds with hydrogen and CHn groups (24). In terms of the resultant
1000
Number of Publications
800
600
400
200
0
1970 1975 1980 1985 1990 1995 2000 2005 2010
Year
Figure 1 Number of articles on DLC coatings published in the period 1970–2012, determined by searching ‘diamond-like’ and ‘carbon films’ in the
Google Scholar database, 1 April 2013.
www.iran-mavad.com
Advanced Carbon-Based Coatings 391
properties, incorporating silicon into the DLC structure reduces residual internal stress; improves the adhesion to most substrates,
including various metal alloys, steels, and glasses; increases hardness; increases thermal stability; results in low friction coefficients
independent of relative humidity; and decreases the wettability (see (25,26), and references therein).
The incorporation of fluorine and phosphorous in DLC films has also been investigated during the last decade. Fluorine was
found to increase the hydrophobicity of carbon films by lowering the surface energy of the resultant film (27), resembling in some
way an amorphous PFT (Teflon) material. However, when alloying with phosphorous, the carbon-based films show an increase in
electrical conductivity and hardness (28).
In the 1990s, carbon-based coatings enabled the operation of fuel injectors in gasoline and diesel engines, providing necessary
protection against wear, scuffing, and other adverse effects of lubricant-lean friction. Nowadays, carbon-based thin films are
found in most facets of our daily lives, from the protective coatings in the hard disks of computers to the lubricious coating in the
blades of a disposable shaving machine. They are used in the chemical, food, and packaging industries (valves and cutting
hardware); semiconductor processing equipment (wafer handling and mechanical drives of cluster tools); compressors and
pumps; the automotive industry (piston rings, camshafts, and gears); and so on. All the emerging applications of carbon-based
coatings originate in the great variety of crystalline and disordered structures that can be obtained by many different deposition
techniques.
4.15.2 Bonding Configurations in Carbon Films
Besides the well-known diamond and graphite single crystals, carbon exhibits in other forms such as fullerene molecules, nano-
tubes, graphene, nanocrystalline graphite (nc-G), semicrystalline carbon black, and diverse amorphous phases including glassy
carbon (29,30). This variety of allotropic forms is attributable to carbon being the only element in the periodic table known to have
isomers with zero, one, two, or three dimensions. The four valence electrons of carbon ([He] 2s2 2p2) are able to form three
hybridization states enabling three major bond configurations; the sp3 tetrahedral bonds like in diamond, the sp2 planar bonds like
in graphite, and the sp bonds like in alkynes. Figure 2 shows that for sp3-hybridized carbon, all four electrons form s bonds
pointing toward the corners of a tetrahedron, enabling for a strong three-dimensional (3D) network. In a sp2-bonded network, only
three electrons form three planar bonds, while the fourth forms a p-orbital perpendicular to the plane of the three s-bonds. The
electron in the p-orbital can be delocalized to strengthen the bonds by forming a resonance structure of single and double bonds in
an appropriate matrix. The strength of carbon compounds is easily envisioned by the example of diamond, still the hardest material
known, due to the strong carbon–carbon bond, the high 3D coordination, and the small atomic size. However, the overlap of the
p-orbitals for the sp and sp2 carbon–carbon bonds, commonly referred to as double and triple bonds, results in bond energies of 6.3
and 8.7 eV, which are much higher bond energies compared to 3.6 eV for the single bond as in the sp3 hybrid. This gives graphite
and polymers their extraordinary in-plane and chain strength.
The sp3 and sp2 bonds can occur not only in single crystals, like diamond and graphite respectively, but also in amorphous solids
where the atoms are in a random network, with bonding between only a few individual atoms and not in a long-range order
extending over a large number of atoms. Amorphous carbon films containing a mixture of sp2 and sp3 bonds are known under the
generic name of DLC coatings because they display some of the typical properties of diamonds. Following an earlier work from
Jacob and Moller (31), Robertson and Ferrari (29,32,33) proposed a ternary phase diagram for various amorphous carbon-based
coatings according to their sp2, sp3, and H content (Figure 3). The structure phase diagram consists mainly of three regions. The first
one is H-free C coatings along the left axis. The sp2 amorphous carbon (a-C) is typically a coating deposited by pyrolysis of
hydrocarbon polymers or by evaporation. Amorphous coatings with a mixture of sp2 and sp3 bonds, but not containing H, are DLC
coatings typically deposited by sputtering and laser deposition in ultrahigh vacuum (UHV) conditions. When the amount of sp3 is
high, McKenzie has suggested denoting the material as tetrahedral amorphous carbon (ta-C) in order to distinguish it from sp2 a-C
(30). Tetrahedral bonding at levels of approximately 80% can be found in amorphous carbon films deposited by cathodic arc
deposition, ion beam deposition, and laser ablation.
Figure 2 Schematic representation of the three possible types of hybridization for carbon: sp3 (diamond), sp2 (graphite), and sp (alkyne). The elongated
atomic orbitals are the hybrids that form s-type molecular orbitals, while the roundish atomic orbitals are remaining 2p electron(s) that form the
p-type molecular orbitals.
www.iran-mavad.com
Figure 3 Ternary structure phase diagram of bonding in amorphous carbon–hydrogen films.
The second region of the phase diagram is at the bottom right of the figure, where the H content is so high that the material
cannot be deposited in a solid phase. The regions are delimited by C2H2 in the composition on the bottom axis and (CH2)n on the
right axis of the triangle.
The third region, between the other ones, features both amorphous hydrogenated carbon films (a-C:H) and tetrahedral
hydrogenated amorphous carbon (ta-C:H). The a-C:H is typically produced by CVD techniques using hydrocarbon precursors or by
reactive magnetron sputtering of carbon targets in mixtures of Ar and H atmospheres. ta-C:H can be deposited by high-density
plasmas such as electron cyclotron resonance deposition.
It is well established that the key parameters determining the physical properties of carbon coatings are (1) the sp3/sp2 fraction,
(2) the clustering of the sp2 phase, (3) the orientation of the sp2 phase, (4) the cross-sectional nanostructure, and (5) the H content
(29). For nonhydrogenated films, the density, band gap, and mechanical properties (hardness and Young’s modulus) essentially
scale with the fraction of sp3 carbon in the film. That is since the sp3 bonds allow for a higher packing density, and the 3D bonding
gives a higher strength of the material. The band gap increase can be correlated to the lower p-electron density.
The hydrogenated films also can have a high fraction of sp3 bonds, but since the 3D structure may frequently become
terminated by hydrogen, the mechanical strength as well as the density decreases with increasing H content. The band gap,
however, increases with increasing H concentration since the aromatic domains, where the p-electrons can move freely, are
interrupted by hydrogen (33).
An important advantage of DLC over artificially grown diamond films is that they can be deposited at low temperatures
(<200 C), while the CVD process used for diamond deposition requires much higher temperatures. This makes DLC an ideal low-
friction wear-resistant coating on heat-sensitive substrates, such as tool steels (22), ball-bearing steel (34), rubber (35), plastics (36),
and optical storage media (29).
In 2006, the nomenclature in Figure 3 was proposed by the Association of German Engineers as a norm for the classification of
DLC coatings, in order to clarify the existing multiplicity of confusing terms and trade names (37).
This chapter does not feature further structural properties of pure carbon or hydrogenated carbon coatings, mainly because
excellent review papers exist that address both the scientific and application-oriented issues (17,19,30,32,33,38,39). We will instead
concentrate on the microstructural properties of three more recent coatings, which have the possibility to be applied on a large scale:
(1) carbon nitride, (2) phosphorous carbide, and (3) carbon fluoride.
4.15.3 Carbon Nitride Films
The first attempt to produce carbon nitride in the form of a thin film was reported by Cuomo et al. in 1979 (40). The authors used
a radiofrequency (RF) diode-sputtering system to deposit CNx films, which they identified to be a paracyanogenlike material, with
a composition of 49% C, 48% N, and 3% O. In 1984, C. M. Sung predicted in an unpublished patent disclosure letter at the
Diamond Technology Center of Norton Company that C3N4 may be harder than diamond, the hardest material known (41). In
order to estimate the hardness of these hypothetical materials, the author contracted Marvin Cohen to calculate the theoretical
values of their bulk moduli. In 1985, Cohen published an empirical formula relating the bulk modulus of tetrahedrally coordinated
materials to the length and ionicity of their chemical bonds (42). From this relationship, he predicted that a material made of
carbon and nitrogen should exhibit a bulk modulus higher than diamond due to the short length and high covalency of the C–N
bond. Subsequently, Liu and Cohen predicted that b-C3N4, a material isostructural with b-Si3N4, possesses a bulk modulus
approaching that of diamond (20,21). It should be noted here that in none of Cohen’s publications does the term ‘harder than
diamond’ appear. As was pointed out by Teter some years later (43), the shear modulus rather than the bulk modulus is the better
predictor for hardness. Nevertheless, it appears that Cohen’s papers are the beginning of a remarkable interest in the C–N system as
www.iran-mavad.com
200
180
Number of Publications
160
140
120
100
80
60
40
20
0
1990 1995 2000 2005 2010
Year
Figure 4 Estimate of the number of articles on carbon nitride published in the period 1990–2012. Numbers were determined by searching ‘C3N4’and
‘carbon nitride’ in the Google Scholar database, 1 April 2013.
the possible source of a material harder than diamond. This excitement within the scientific community was also reflected by
divulgation articles in the general press (44–48).
Many deposition methods and techniques have been employed in the quest for the elusive superhard crystalline phase b-C3N4
(see, e.g., the reviews from Muhl (49), Wang (50), and Kroke (51)). Figure 4 is an updated analysis of literature found by using only
the keywords ‘C3N4’ and ‘carbon nitride’ in the Google Scholar database; about 1900 publications (including 90 patents) can be
ascribed to this material thus far. The data show that the predictions of Liu and Cohen (20,21,42) sparked off intense theoretical and
experimental activity worldwide with a view to experimentally realize the new C–N superhard material. We can also see that the
activity started to decrease after 1999, a probable consequence of the publication of critical reviews that argued against the
possibility of fabricating a harder-than-diamond b-C3N4 single-crystal film (52). Finally, a new peak started about 2006, probably
related to publications describing new forms of the material like nanowires and FL films.
Since 1990, many investigators have claimed the formation of crystalline carbon nitride. But, to date, no satisfactory evidence has
been presented. Most claims are based on the analysis of electron diffraction patterns, which alone have proven not to be suitable to
identify carbon nitride crystalline phases because the strong peaks of lonsdaleite (hexagonal diamond) do match with the predicted
electron diffraction pattern of hypothetical b-C3N4 (53,54). Also, it has been shown that FeO(OH), which could form during TEM
specimen preparation, matches almost exactly with the predicted patterns of b-C3N4. Furthermore, the elemental compositions in
the obtained crystals were not always reported, or only overall N:C ratios were given, and mostly impurities such as oxygen and
hydrogen or silicon (from the substrate) were not considered or specified to their exact concentrations (53,54).
Difficulties in synthesizing C3N4 films arise primarily from the inherent drive of nitrogen to form strongly covalently bonded N2
with the consequence that as-deposited films are deficient in nitrogen, which readily desorbs (22). In order to structurally incor-
porate the required 57% of nitrogen, the focus was mainly on high-energy deposition techniques to provide for pronounced
nonequilibrium growth conditions. These high energies, however, suppressed any structure evolution by displacing film atoms via
knock-on collisions. Despite these efforts, the nitrogen content in the resulting amorphous CNx material rarely surpasses 30 at%,
and any nitrogen concentration exceeding this limit seems to be due to N2 gas trapped in a porous film material. Alternative
deposition techniques with lower energy per deposited particle exist, such as reactive magnetron sputtering of carbon in a nitrogen-
containing atmosphere. Their typically low ion-to-neutral arrival rate ratio yields film-forming species that carry thermal energies,
which allows for the evolution of structured films, as described next.
Reactive magnetron sputtering with ion assist energies lower than 50 eV was employed by Sjöström et al., which led to the
discovery of the ‘fullerene-like’ allotrope of carbon nitride (FL-CNx) at Linköping University in 1994 (55,56). This nanostructured
CNx (0 < x < 0.25) thin-film material was found to be much more interesting from a materials research and applications point of
view than its amorphous counterpart. By using nitrogen as a reactive gas, a film structure consisting of bent and frequently inter-
secting planes evolves at elevated temperatures, whereas pure carbon films sputtered under the same conditions result in a porous
underdense sootlike material (57).
4.15.3.1 Fullerene-Like Carbon Nitride Films

FL materials are characterized as compound materials with heavily bent and intersecting hexagonal basal planes from structurally
incorporated odd-membered rings. These structures have been recently observed in, for example, C, BN:C, BN, MoS2, TiO, NbS2,
www.iran-mavad.com
Figure 5 Plan view high-resolution transmission electron micrograph of the curved and intersecting FL sheets in CNx for a film grown by reactive
magnetron sputtering from a carbon target in mixed Ar–N2 discharge at 450 C with an N2 fraction of 0.5 and a bias voltage of 25 V. The inset illustrates
the buckling of the graphene sheets by pentagon incorporation obtained from ab-initio calculations. Reproduced from Johansson, A. Charge Dynamics
and Electronic Structure of Pi-Conjugated Systems. Doctoral Thesis No. 783, Linköping Studies in Science and Technology: Linköping, Sweden, 2002.
and WS2 (22), and more recently in CPx (58) and CFx (59). Out of these materials systems, only C, CNx, CPx, and CFx exhibit cross-
linking that enables the material to extend the strength of the covalently bonded two-dimensional (2D) hexagonal network into
three dimensions.
FL-CNx consists of, upon substitution of nitrogen for carbon, bent, intersecting, and cross-linked graphite sheets, as seen in the
high-resolution transmission electron microscopy (HRTEM) plan-view image in Figure 5. This implies that the majority of carbon is
incorporated in a three-coordinated planar sp2-hybridized state. The double bond between sp2-hybridized carbon, as in graphite, is
with a bond energy of 6.3 eV compared to 3.6 eV, which is actually stronger than the single-bond sp3–hybride bond of diamond (56).
However, the overall strength of graphite is limited by its 2D structure of hexagonal planes and, thus, the rather weak interplanar van
der Waals type bonding. Hence, graphite deforms easily by gliding of basal planes. If, however, nitrogen is incorporated during
growth, the energy required to form pentagons is considerably lowered and thus the graphitic sheets curve and buckle (inset Figure 5)
(60). This enables the material to extend the extraordinary strength of a planar sp2-coordinated carbon network in three dimensions.
The assumed nitrogen-induced cross-linkage between the sheets provides strength to FL-CNx by preventing interplanar slip. This
results in an extremely fracture-tough, elastic, and compliant material, which deforms by reversible bond rotation and bond angle
deflection rather than by slip and bond breaking (62). DLC films, for instance, also exhibit properties beneficial for many appli-
cations, such as high hardness due to a high fraction of sp3-carbon bonding and favorable tribological behavior. However, contrary
to FL-CNx, the application of DLC films suffers mostly from the inherently higher brittleness of the material. Depending on the
growth conditions, FL-CNx thin films exhibit a large variation in basal plane curvature and, hence, extension and alignment of
sheets. Thus, due to the variation in microstructure, it denotes a material with a large spread in properties (e.g., elastic modulus,
elasticity, and electrical resistance) (57,63,64), all having in common an extraordinary mechanical strength. This provides for
tailoring of the properties to given applications as, for example, wear-protective coatings.
4.15.3.2 Bonding Configuration in Carbon Nitride Films

FL-CNx is a compound material in which nitrogen is incorporated in a carbon matrix. The effect of impurities such as hydrogen
might also be of importance, as it is known to improve the properties of, for example, DLC films (33). However, the very low
concentrations of impurities such as hydrogen and oxygen in the magnetron-sputtered CNx films should not have any significant
effect upon the film structure. However, Hellgren et al. studied the effect of intentional hydrogen addition to the discharge during
CNx deposition (65). It was shown that hydrogen terminates potential bonding sites for CN precursors, and, thus, the formation of
an FL structure is hindered. Additionally, hydrogen enhances the etching process by the formation of not only volatile C2N2
molecules but also NH3 and volatile CxHy species.
Carbon is an excellent constituent in the synthesis of hard compounds, since it fulfills both the requirements of a small atomic
size and an ability to form strong bonds, as we have seen in Section 4.15.3. Nitrogen, like carbon, predominantly forms covalent
bonds. Its valence electrons ([He] 2s2 2p3) exhibit, despite the additional electron, bond hybridizations similar to those with
carbon. However, the extra valence electron implies some differences. In the case of sp3-hybridized nitrogen, four hybrid orbitals are
formed just like in carbon, while the extra electron forms a nonbonding lone pair with an electron in a hybrid orbital leaving only
three s-bonding orbitals (Figure 6). The bond angle is slightly deflected to 109 instead of 107 compared to diamond due to the
effect of the more spacious lone pair. There are two possibilities for the sp2-hybridized bond configuration. First, the extra electron is
able to form a nonbonding lone pair with an sp-hybrid orbital leaving the atom with two nonlinear s-bonds and the possibility to
participate in p-resonance structures with the electron in the remaining 2p orbital. For the second case, the extra electron forms
www.iran-mavad.com
Figure 6 Schematic of different electron hybridization states for nitrogen; sp3 like in ammonia, and sp2 twofold coordinated like in pyridine and threefold
coordinated as in a substitutional graphite site. The difference in coordination is due to the location of the nonbonding electron pair.
the lone pair with the remaining 2p orbital leaving the atom with three planar s-bonds and no possibility to contribute to
a p-resonance structure, due to the filled, nonbonding 2p orbital. Sp-hybridization of nitrogen is also common, whereas the atom
can form one s- and two p-bonds, while the fifth valence electron forms a lone pair with the second sp-hybrid orbital.
Due to the inherently disordered structure of CNx compounds, it has been difficult to determine the exact bonding structure of
carbon and nitrogen in the material. Thus, complementary spectroscopic techniques have to be used to investigate the local bonding
environment of FL-CNx. Valuable knowledge of the bonding structure can be obtained from X-ray photoelectron spectroscopy
(XPS) (66). The N1s core level peak has typically two strong components (Figure 7), P1 at w400.7 eV and P2 at 398–399 eV, and at
least two smaller ones: P3 positioned at w402 eV and P4 at w399.5 eV. The assignment of these peaks has been widely divergent in
literature. However, strong support can be found for the following peak interpretation by comparing the XPS results of a wide range
of films with other spectroscopic techniques (56):
l P1 (400.7 eV) corresponds to threefold coordinated sp2-hybridized nitrogen bonded to three carbon atoms, substitutional in
a predominantly sp2-coordinated graphite network (see inset of Figure 5), while an sp3-hybridized carbon as a neighbor cannot
be ruled out. N–H bonding originating from reactions with ambient air might also have contributions in this region.
Figure 7 XPS N1s photoelectron spectra of CNx films synthesized at various N2 fractions, a bias voltage of 25 V, and substrate temperatures
from 600 down to 100 C. P1–P4 indicate the 3–4 spectral contributions. The insets illustrate the associated major bonding configurations. Reproduced
from Neidhardt, J.; Hultman, L.; Czigany, Z. Correlated High Resolution Transmission Electron Microscopy and X-Ray Photoelectron Spectroscopy
Studies of Structured CNx (0 < x < 0.25) Thin Solid Films. Carbon 2004, 42 (12–13), 2729.
www.iran-mavad.com
l P2 (398.0–399.0 eV) was initially assigned to nitrogen bonded to an at least partially sp3-hybridized carbon matrix, but has
rather to be seen as nitrogen in a pyridinelike structure (see inset of Figure 5), where a twofold coordinated sp2-hybridized
nitrogen forms two bonds to an otherwise graphitic structure. This bonding configuration may occur along edges of planes or
next to defects, and is essential for eventual cross-linking.
l P3 (402.5 eV) is normally assigned to N–O bonds formed after air exposure. However, recent results from in situ XPS in a UHV
environment show no such correlation. Instead, P3 scales with the intensity of P1, suggesting that it originates from a similar
substitutional graphite structure, where nitrogen donates an electron to the structure leaving it with a virtual positive charge and
thus increasing the binding energy (67).
l P4 (399.5 eV) is a minor contribution and most likely corresponds to sp-hybridized nitrogen in a nitrile structure. N–H bonding
configurations due to contamination or surface reactions may also contribute in this region (67).
In reference (66), it is shown by combining XPS and HRTEM selected-area electron diffraction (SAED) that the N1s core level P1:P2
intensity ratio reflects the structural arrangement of CNx films. A decreasing extension of the FL sheets implies a decreasing number of
nitrogen in a substitutional site (P1) and in turn more nitrogen bonded along edges and defects (P2). Hence, a pronounced FL
microstructure consisting of extended planar domains is characterized by a large P1:P2 ratio, where the peaks are well separated
(z2 eV) and P1 is dominating the spectra. If, on the other hand, the FL structure is disrupted, the P1:P2 ratio is decreasing.
Detailed analysis based on the XPS C1s photoelectron peak is impractical, since it is typically broad and its contributions cannot
be extracted separately. Still, the C1s peak is observed to broaden toward higher binding energies as soon as nitrogen is incorpo-
rated. This is consistent with the above-mentioned distribution in bonding environments, and the broadening is correspondingly
more pronounced at lower growth temperatures (68).
The information extracted from XPS and X-ray absorption near edge structure (XANES) (NEXAFS) is in many aspects
complementary, while the latter assesses the electronic structure of a compound at a much superior resolution. Using this technique,
the bonding structure of CNx, in particular the FL allotrope, was analyzed in detail by Gago et al. (69). As seen in Figure 8, the C1s
spectra consists of two broad major peaks around 284 and 290 eV corresponding to 1s 0 p* and 1s 0 s* transitions, respectively.
The intensity of the p* peak and the position of the s* absorption edge indicate the dominance of sp2 hybrids, as compared to the
reference spectra of graphite and diamond. The states below the graphite peak (G) are attributed to defects in the graphitic network,
while the formation of various CN bonds is revealed by peaks in the 290–294 eV region. The small (C2) peak indicates the existence
of a minor fraction of nitrile-bonding configurations regardless of the substrate temperature. This is not surprising given the large
number of preformed CN species in the deposition flux. The graphitization occurring at higher temperatures can be seen by an
increasing (G) peak. A complete deconvolution and peak assignment for the C1s spectra are, as for XPS, very complex due to the
multitude of bonding configurations.
Similar to XPS, more direct information can be extracted from the N1s spectra, where four distinctively different peaks in the p*
region can be resolved. Nitrogen also seems to be predominantly sp2-hybridized, as indicated by the pronounced p* region. The
N1, N2, and N4 peaks are assigned to nitrogen bonded in a pyridine, nitrile, and graphitelike configuration (comparable to P2, P4,
Figure 8 XANES spectra at the C1s and N1s edge for CNx films grown by reactive magnetron sputtering at an N2 fraction of 0.16, 25 V bias, and
various substrate temperatures. The spectra of graphite and diamond as well as hexagonal and cubic BN are shown as a reference for sp2 and sp3
hybridizations. Reproduced from Gago, R.; Neidhardt, J.; Vinnichenko, M.; Kreissig, U.; Czigany, Z.; Kolitsch, A.; Hultman, L.; Moller, W. Synthesis of
Carbon Nitride Thin Films by Low Energy Ion Beam Assisted Evaporation: On the Mechanisms for Fullerene-Like Microstructure Formation. Thin Solid Films
2005, 483 (1–2), 89–94.
www.iran-mavad.com
and P1 in the XPS spectra), whereas N3 is still under discussion. The development of the N2 peak as a function of the deposition
temperature as analyzed by XANES (see Figure 8) indicates that the nitrile configuration C^N is most pronounced at lower
temperatures. This is due to the fact that the triple bond, present in most of the preformed CN species in the flux, stays intact upon
partial nonaligned incorporation due to a decreased reactivity at lower temperatures. At higher temperature, however, these species
are incorporated along edges of sheets, and thus the electron of the triple bond forms a third bond or a lone pair, as is the case for
a graphite or pyridinelike configuration. In addition, the remaining nitrile groups react to form C2N2, which desorbs by the chemical
desorption process described in this chapter.
In congruence with XPS, the incorporation of nitrogen in a graphitelike configuration (N4) is preferred at elevated temperatures
and lower N2 fractions. This, along with a lower number of nitrogen incorporated in a pyridine (N1) or nitrile (N2) configuration, is
the fingerprint of an evolving FL microstructure, which is even more pronounced at lower N2 fractions. However, these bonding
states are still present, while especially the pyridinelike configuration might play an important role for interplanar cross-linking.
As for the structural signature of CNx allotropes, the lattice image of curved basal planes in Figure 5 is characteristic. While
smaller fragments of FL ordering are also possible, they may project as an amorphous structure. Zsolt et al. assessed the degree of
structuring (FL, graphitic, or plain amorphous) by evaluating peak intensities of the rings in SAED patterns. The SAED patterns from
CNx thin films usually exhibit three rings at w1.2 Å, w2 Å, and w3.5 Å (66). Diffuse SAED rings with a full width at half maximum
(FWHM) of 0.10.2 Å1 and ring diameters of w1.2 Å and w2 Å are typical for amorphous carbon allotropes, whereas the
appearance of a ring with a diameter of w3.5 Å corresponds to graphitic short-range order in the microstructures (i.e., in graphitic or
FL carbon materials) (71).
4.15.3.3 The Role of Nitrogen for the Formation of Fullerene-Like Carbon Nitride
The substitution of nitrogen for carbon in the graphite basal planes affects both the curvature of planes and their reactivity, which is
one of the key processes for the formation of FL-CNx solid films at elevated temperatures. For comparison, pure carbon films grown
under the same conditions result in an amorphous underdense structure consisting of aromatic disordered fragments. It is only
upon introducing even small fractions of nitrogen in the discharge that dense and structured films form. Three major roles were
assigned to nitrogen and will be described in the following subsections.
4.15.3.3.1 Evolution of Curvature

The first role is that nitrogen incorporation induces curvature. This is evident from the fingerprint-like appearance of basal planes of
an FL-CNx film in Figure 5. The highly corrugated and curved basal planes imply a radius of curvature in the nanometer range. This
curvature is induced for all nitrogen fractions of a N2–Ar magnetron sputter discharge, while its degree can be controlled by the N2
fraction (57). Thus, nitrogen is essential for inducing spherical curvature. For pure carbon structures, curvature is known to be due to
the incorporation of odd-member rings in an otherwise hexagonal structure. Twelve pentagons thus yield the spherical appearance
of closed-cage carbon fullerenes, such as C60. Experimental results as well as calculations actually suggest that isostructures to C60
might exist in FL-CNx thin solid films (72). The corresponding so-called nano-onions (Figure 9) form under certain deposition
conditions, such as elevated temperatures (Ts ¼ 450 C) and low N2 fractions (0.1–0.16).
These experimental findings were, furthermore, verified by total energy calculations, which show that nitrogen stabilizes
pentagons in aromatic carbon structures (60). However, the structures are shown to be less stable if more than one carbon is
Figure 9 Spatially resolved electron energy loss spectroscopy showing the CCD counts and relative nitrogen concentration scanned across a CNx nano-
onion as indicated in the respective HRTEM plan-view image. The inset shows the structure of C48N12 that density functional theory (DFT) simulations
suggest may exist in the center of the nano-onion. Reproduced from Hultman, L.; Stafstrom, S.; Czigany, Z.; Neidhardt, J.; Hellgren, N.; Brunell, I.;
Suenaga, K.; Colliex, C. Cross-Linked Nano-Onions of Carbon Nitride in the Solid Phase: Existence of a Novel C48N12 Aza-Fullerene. Phys. Rev. Lett. 2001,
87 (22), 225503.
www.iran-mavad.com
substituted in a ring and/or nitrogen is incorporated next to each other, which from geometrical conditions explains the apparent
saturation in maximum nitrogen concentration of approximately 20–25 at%. In summary, both experiments and theoretical
calculations suggest that nitrogen triggers the evolution of curvature by stabilizing the required pentagons at much lower energies as
compared to pure carbon structures, while the incorporation of more than one nitrogen atom in a carbon ring structure seems to be
energetically less favorable (60).
4.15.3.3.2 Formation of Interplanar Cross-Linking

The second role of nitrogen is in the promotion of interplanar cross-linking. The combination of mechanical properties, such as
elasticity and hardness of FL carbon nitride, is superior to those of most carbon-only materials. The mechanical strength of graphite,
for instance, is limited by its easily activated slip system along the graphite basal planes; however, this weakening interplanar slip
also holds the key for its extremely low friction. Carbon fullerene cages, instead, are known to be extremely resilient and to
withstand severe deformation without any plastic yield via broken carbon–carbon bonds (56). However, the mechanical properties
of crystalline compounds formed from fullerenes – so-called fullerites – still suffer from rather weak intermolecular bonding. It has
also been shown that temperature treatment can initiate intermolecular covalent cross-linking or polymerization, which improves
the mechanical strength (56). The other extreme of hard carbon-only materials is represented by diamond, where the highest degree
of 3D cross-linking is provided by an all-sp3 hybridization. Also, DLC thin films synthesized by various techniques exhibit
a significant fraction of sp3-hybridized carbon that gives the material an outstanding hardness, while still retaining favorable low-
friction properties. However, these superhard materials have only a limited tolerance to deformation, and overloading results in
imminent brittle fracture and, thus, mechanical failure of the substrate-coating system.
FL-CNx compared to common superhard thin-film materials (cubic BN and DLC) has a low-to-moderate resistance to pene-
tration (hardness), as probed by nanoindentation (62). Still, the indentation response surpasses crystalline graphite by far, which
implies the existence of strong interplanar cross-links or interlocking of sheets. The most striking property, however, is its
outstanding resistance to plastic deformation. In fact, FL-CNx deforms predominantly elastically, which combined with its
compliance results in a resilient and fracture-tough material (57,62). Nitrogen plays a crucial role, since pure carbon materials
grown under similar conditions have a sootlike appearance, which is in conjunction with deficient cohesion and strength.
The precise nature of the cross-linking mechanisms is, however, still under discussion. A first explanation was established by total
energy calculations performed in reference (60). The presented results suggest that, upon nitrogen incorporation in a hexagon, it is
for the adjacent carbon atom energetically more favorable to change from an sp2 to an sp3-hybridized state, which could provide for
the required out-of-plane bond. For that reason, nitrogen-doped hetero fullerenes also known as aza-fullerenes (e.g., C59N)
instantaneously dimerize upon synthesis via a carbon–carbon bond next to the substituted nitrogen atom (72). Experimental
results, however, indicate that FL-CNx is predominantly sp2 bonded and the existence of significant numbers of sp3-hybridized
carbon remains to be affirmed (56). In contrast, only a few interplanar cross-links are needed to increase the stiffness significantly,
and any low fraction of sp3-hybridized carbon might pass undetected due to experimental limitations of, for example, nuclear
magnetic resonance (NMR).
Cross-linking at the nitrogen site is, however, rather unlikely. Firstly, it was shown by total energy calculations in reference (60)
that the structure is more stable if two incorporated nitrogen atoms in a graphite fragment are separated by at least two carbon
atoms. Secondly, chemical synthesis of, for example, aza-fullerene dimers indicate that the cross-linking is initiated at the carbon
next to an incorporated nitrogen atom and not at the nitrogen site itself (72).
The high stiffness of FL carbon nitride, in contrast to planar graphite, can also be explained by the 3D interlocking of curved
meandering FL sheets. In addition, purely sp2-based cross-linking mechanisms might also be present, such as branching of sheets
and/or carbon–carbon bond rotation (56,73). For the latter, the electron lone pair in the 2p orbital of substitutionally incorporated
nitrogen or the formation of pentagons disturbs locally the p-resonance structure. Therefore, the double-bond resonance structure
between two neighbor carbon atoms is disrupted, and, thus, the carbon–carbon bond is free to rotate due to a missing 2p–2p orbital
coupling. This might result in an out-of-plane rotation of the subsequently growing structures and hence enables cross-linking
leading to a multitude of possible configurations. This is most likely if one considers also the disruption of the continuous sheet
structure by the incorporation of two-coordinated, pyridinelike bonded nitrogen that is present in significant numbers as suggested
by spectroscopy-based techniques.
Most recent computational studies corroborated the existence of out-of-plane bond deflection and rotation in the presence of
multiatomic precursors (74). Figure 10 exemplifies two of the energetically stable structures exhibiting cross-linking obtained by
simulated (synthetic) growth as developed by Gueorguiev et al. (74).
Another potential sp2-based cross-linking mechanism is the branching of FL sheets. It was shown in Section 4.15.3.2 that
nitrogen is bonded primarily in two different configurations: substitutional at carbon sites, which does not affect continuous planar
growth, or in a pyridine configuration, where one of the three s-bonds is terminated by the electron lone pair. The latter implies the
existence of vacancy defects as depicted in Figure 11(a). Additionally, nitrogen induces pentagon formation, which provides for
spherical curvature at the vacancy site (Figure 11(b)). This presented configuration results in out-of-plane nucleation sites
(Figure 11(c)), where 3D growth can be initiated (Figure 11(d)).
4.15.3.3.3 Formation of CxNy (x, y 2) Species

The third major role of nitrogen is the formation of CxNy (x, y 2) species at the target or cathode, depending on the type of PVD
deposition, to become part of the deposition flux. Carbon nitride thin films are commonly grown by plasma-activated deposition
www.iran-mavad.com
Figure 10 Stable cross-linked structures with a cross-linking bridge (enlarged in the inset). Reproduced from Gueorguiev, G.; Neidhardt, J.; Stafstrom,
S.; Hultman, L. First-Principles Calculations on the Curvature Evolution and Cross-Linkage in Carbon Nitride. Chem. Phys. Lett. 2005, 410 (4–6), 228.
Figure 11 Alternative sp2-based cross-linking mechanisms via a nitrogen-induced vacancy (a), and pentagon formation (b), which results in out-of-
plane bending (c), and finally provides a nucleation site for out-of-plane growth (d). The gray-shaded atoms represent carbon and the black nitrogen;
the hydrogen (white) atoms are terminating the FL fragment. Reproduced from Broitman, E.; Neirdhart, J.; Hultman, L. Fullerene-Like Carbon Nitride:
A New Carbon-Based Tribological Coating. In Tribology of Diamond-Like Carbon Films: Fundamentals and Applications; Donnet, C., Erdemir, A., Eds.;
Springer: New York, 2007; pp 620–653.
techniques, where carbon is prone to react with activated nitrogen originating from the discharge. This reactivity of nitrogen is
considered to be the dominant compound-forming reaction on surfaces for most PVD nitride films, due to the rather low
probability of gas phase reactions under the subatmospheric pressures. However, the experimentally established existence of
a low-mass gaseous compound in the carbon nitride system, namely C2N2 (75), adds numerous degrees of freedom to the
formation mechanisms of FL-CNx films as compared to other nitrides forming from single-atomic precursors. As a consequence,
the deposition flux from a carbon target sputtered in an N2-containing atmosphere consists to a large extent of multiatomic
species. They form via reaction between the carbon atoms in the target and the vast number of nitrogen ions implanted into the
top surface layers during sputtering (75). In turn, C2N2 and fragments thereof are emitted, whereas their number scales with
the nitrogen fraction in the discharge (75). For pulsed laser deposition from carbon in N2, Voevodin et al. showed the formation of
CN species (76).
The presence of nitrogen-containing species in the deposition flux leads to various effects also on the substrate site. One
important process is that they partially react upon adsorption to form volatile gases such as N2 and C2N2, which subsequently
desorb. This results in a film surface etching that is commonly referred to as chemical sputtering. This process is observed for
various CNx deposition techniques, especially those utilizing hyperthermal species to assist the growth (76,77). The extent of the
chemical desorption is determined by the reaction probability to form a volatile gas and, thus, depends on the type and mobility
of the adsorbed species. The overall process can therefore be seen as a preferential etching of less favorably bonded adsorbates
and, thus, results in dense and smooth films. Prevailing CN precursors bonded in an aligned manner along edges of sheets are
www.iran-mavad.com
more resilient to concurrent etching and therefore play an important role for the evolution of FL planes (discussed further in this
chapter). The chemical desorption process also accounts for the apparent maximum of incorporated nitrogen at w25 at%, due to
its higher desorption probability.
4.15.3.4 Structure Evolution Mechanisms

For magnetron sputtering, the presence of nitrogen in the discharge gas leads to complex chemical reactions at the target and
substrate surface (56). The intense high-energy ion bombardment at the target results in sputtered target constituents, whereas most
of the nitrogen ions are implanted during the process. The implanted nitrogen radicals react readily with the carbon atoms in the
target. Thus, the composition of the target surface is altered substantially, especially if one considers the initially rather low sputter
yield of pure C. Hence, predominantly CxNy (x, y 2) molecules are emitted from such a dynamically self-modified target surface,
due to the relatively high strength of the C^N bond, alongside desorbing gases like N2 and C2N2. The extent of these processes scale
with the availability of N2 in the discharge and therefore lead to a strong increase in flux of preformed CN species at higher N2
fractions, while the carbon flux plays only a minor role (75).
Although low ion assist energies (z20–50 eV) are generally beneficial for film growth by enhancing ad-atom mobility at
relatively low substrate temperatures, it has been shown that they are not the dominant structure-determining factor for FL-CNx
films. Instead, the flux and type of preformed CN species carrying only thermal energy are crucial for the evolution of FL structures.
These molecules might be stabilized against desorption in an evolving sheet by multiple bonding on the C and N side upon
aligned incorporation, thus promoting the evolution of planar structures. Gueorguiev et al. has corroborated the energetic
advantage of this process using step-by-step ab-initio approaches in Ref. (78). Furthermore, the formation of C2N2 as volatile gas is
also active on the substrate site and provides for selective etching of misaligned molecules by reaction with adsorbed CN and their
subsequent desorption. This process accounts for the apparent saturation in nitrogen content at approximately 25 at% nitrogen
incorporated into the films, which is in congruence with the most established theoretical stable structures (54,62). The incor-
porated nitrogen seems to originate predominantly from the already-bonded nitrogen in the preformed molecules arriving at the
substrate, while the extensive low-energetic nitrogen ion bombardment from the discharge plasma readily reacts upon adsorption
to form volatile N2. Thus, the evolution of the FL microstructure has to be seen as a delicate interplay of adsorbing preformed
molecules, which are either incorporated in a stabilized and, thus, aligned manner or desorb via the formation of C2N2 if they are
less favorably bonded (56).
These requirements of low to moderate energy input during growth via ion bombardment or substrate heating, and in particular
the existence of a predominant number of preformed molecules in the film-forming flux, are not only inherent to reactive
magnetron sputtering but also can be found for pulsed laser deposition (76), as indicated above, and in arc deposition (79,80). Each
PVD technique can yield FL-CNx films, whereas the structured domains can grow much larger for the magnetron-sputtered films
because of the less intense ion flux. On the other hand, if no preformed species are present in the deposition flux, no structure
evolution is observed as is the case for low-energy ion beam assisted deposition (70). For magnetron sputtering, the growth
parameters’ substrate temperature, N2 fraction, and bias voltage strongly affect the FL structure evolution in terms of extension,
alignment, and cross-linking. All of the mentioned parameters primarily concern the magnitude of the chemical desorption and,
thus, the selectivity incorporation sites either by influencing the ad-atom (molecule) mobility (bias and temperature) or by
changing the number and type of film-forming species (N2 fraction).
The general trends in terms of deposition parameters and resulting structures are summarized in Figure 12. According
to HRTEM (66) and XPS (Figure 5), films grown at higher temperatures (T s > 300 C) and intermediate bias (Vs z 25 V)
are all FL, since both parameters provide sufficient selectivity to incorporate most of the arriving species in an aligned
fashion, regardless of the N 2 fraction. The availability of N2 , however, determines the extension and alignment of the
sheets, since the dominance of preformed species at higher N2 fractions promotes the growth of extended, well-aligned
sheets resulting in an almost graphitelike appearance. At lower N 2 fractions, the flux of carbon atoms becomes relevant
instead and the complexity and total number of the preformed species decrease, which in turn leads to shorter heavily bent
FL fragments. This is an effect of a lower selectivity of incorporation sites via more complex reaction pathways for the
formation of C2N 2.
If the temperature is decreased (T < 300 C), the ad-atom (molecule) mobility is reduced and an increasing number of pre-
formed molecules are incorporated in an out-of-plane orientation due to a decreased efficiency of the chemical desorption process.
This results in the amorphization of the otherwise graphitic nature of the films grown at higher N2 fractions. The smaller number of
species arriving at lower N2 fractions is, however, still effectively incorporated in an aligned fashion into short fragments down to
a deposition temperature of 150 C.
A higher bias or ion energy, exceeding the displacement energy of carbon (E > 30 eV), however, induces an increasing number of
defects into the growing sheets and thus decreases the FL fragment size. This is most pronounced at higher N2 fractions, most likely
due to a high sensitivity of the aligned incorporation process toward energy input (Figure 12).
Most recently, Schmidt et al. reported that FL-CNx thin films can also be grown by reactive high-power pulsed magnetron
puttering (HiPIMS) (81). The HiPIMS process showed higher ion energies and fluxes, particularly for Cþ ions, compared to
conventional (direct current) magnetron sputtering as described in this chapter. The time evolution of the plasma species was
analyzed for HiPIMS and revealed the sequential arrival of working gas ions, ions ejected from the target, and later, during the pulse-
on time, molecular ions, in particular CNþ and C2Nþ.
www.iran-mavad.com
Figure 12 Structure zone diagrams of structures and properties observed for magnetron-sputtered FL-CNx thin films depending on the substrate
temperature and the degree of ion bombardment, comprising the ion-to-neutral ratio and ion energy and the N2 fraction in the discharge. The HRTEM plan-
view inserts should be seen as representatives only. Reproduced from Broitman, E.; Neirdhart, J.; Hultman, L. Fullerene-Like Carbon Nitride: A New
Carbon-Based Tribological Coating. In Tribology of Diamond-Like Carbon Films: Fundamentals and Applications; Donnet, C., Erdemir, A., Eds.; Springer:
New York, 2007; pp 620–653.
4.15.4 Phosphorous-Carbide Films
In 1921, de Mahler reported the synthesis of a new chemical compound containing just carbon and phosphorus (82). The
amorphous, white material was spontaneously inflammable when warmed, yielding phosphoric anhydride and carbon dioxide.
Since that time, most of the work on this has been done in the area of organic chemistry, where the so-called organophosphorus
compounds, organic compounds containing carbon–phosphorus bonds, are used primarily in pest control (83).
The interest of using diamond films for electronic and optoelectronic applications sparked the research on phosphorous-doped
DLC films. The incorporation of up to 1% P into DLC films has been reported to reduce the resistivity by 6–7 orders of magnitude,
www.iran-mavad.com
with no apparent change in the amorphous nature of the carbon films (84). P-doped DLC films using PH3 as a gas on RF plasma
deposition have resulted in films containing w11 at% phosphorous, with an increase of conductivity of nearly five orders of
magnitude (85,86). More recently, Kuo et al. and Pearce et al. studied films deposited with P:C ratios up to 3:1, which resulted in
being hard and amorphous in nature, and with an optical band gap of 2.1–2.6 eV (87,88). Having such high P content, the authors
termed them ‘amorphous carbon phosphide’ instead of ‘doped DLC’ (the given name was, in fact, wrong because a phosphide is
a compound of phosphorus with a less electronegative element, which is not the case for carbon). These findings have encouraged
a number of experimental (89–91) and theoretical (89,92,93) studies of amorphous phosphorus carbide thin films that have
concentrated on aspects of their microstructure, bonding, and electrical properties.
4.15.4.1 Synthetic Growth Predictions

Predictive ab-initio simulations constitute a valuable tool to address structural issues in existing nanostructured materials as well as
to design novel compounds. Recently, the synthetic growth concept (SGC) based on the DFT has been developed for simulations of
film formation during vapor phase deposition. SCG is understood as structural evolution by sequential steps of atomic rear-
rangement where each step is assigned according to the previous relaxed states (78). This enables a variety of precursors for
nanostructured compounds to be systematically addressed for their interaction when they form condensed phases. The structural
patterns thus formed are relevant to the film formation and eventually to the assessment of the materials’ properties. By SGC, the
interplay between bonding at surfaces and interfaces and the properties of the compounds is also addressed.
SGC has been recently used to theoretically study a new class of carbon-based materials that can be unified by the term FL
carbon-based solid compounds alloyed with N, P, and F. The FL-CNx structure evolution and bonding configurations were
successfully described by using SGC, thus contributing to the pioneering knowledge in its growth and characterization (74,78).
Making use of SGC, the existence of FL phosphorus-carbide (FL-CPx) material was also predicted (28,94). C, P, and N exhibit
similarities in valency, but different electronegativity (C atom electronegativity is higher than that of P atoms, yet lower than that
for N). Since P shows a great variety of bonding configurations, the possibilities of stable P fullerenes have also been theoretically
explored. Seifert et al. investigated different P fullerenes, compared them to C fullerenes, and found some pure P cage structures
metastable (95). A systematic classification of film-forming CnPm (1 n, m 3) and Pn (n 4) species was built by geometric
optimizations of different geometries of these small clusters. All comparisons of different possible structures were based on differences
in cohesive energies jDEcohj normalized by the total number of C and P atoms. Figure 13 shows the most stable CnPm and Pn precursor
species. The particular stability of C3P, C2P, and C3P2 clusters, and to a lesser extent CP, is due to their higher cohesive energies per
atom (Ecoh/at), and their smaller number of dangling and P–P bonds with respect to the other CnPm (1 n, m 3) clusters. These may
be formed in film-forming flux by C and P atoms’ interaction, and defined as precursors. Due to the stability of larger CnPm precursors,
they can be seen as defect-introducing agents (cross-linkages and P-segregation) during incorporation in FL-CPx.
By optimizing the model clusters (Figure 14) containing substitutional P atoms at C sites (both internal and peripheral for the
cluster), the energy costs for substitutional P at C sites in a purely hexagonal (graphene) structure of CPx (Figure 14(a)), a cluster
Figure 13 Most stable CnPm and Pn precursor species: (a) C2P; (b) C3P; (c) C3P2; and (d) P4. Reproduced from Furlan, A.; Gueorguiev, G. K.;
Högberg, H.; Stafström, S.; Hultman, L. Fullerene-Like CPx: A First-Principles Study of the Relative Stability of Precursors and Defect Energetics during
Synthetic Growth. Thin Solid Films 2006, 515, 1028–1032.
Figure 14 Optimized CPx model systems representing substitutional P at C sites in (a) a hexagonal (graphene) network; (b) a structure containing
pentagons; (c) structures with a four-membered ring (upper panel: the P atom at an internal site; and lower panel: at a peripheral position); and (d)
a system with a Stone–Wales (SW) defect. Reproduced from Furlan, A.; Gueorguiev, G. K.; Högberg, H.; Stafström, S.; Hultman, L. Fullerene-Like CPx: A
First-Principles Study of the Relative Stability of Precursors and Defect Energetics during Synthetic Growth. Thin Solid Films 2006, 515, 1028–1032.
www.iran-mavad.com
Table 1 Comparison of cohesive energies per atom (Ecoh/at), corresponding to the fully relaxed finite model systems displayed in Figure 14 a
C15P1H8 (C15N1H8) Tetragonal defect

C23P1H12 (C23N1H12) C13P1H8 (C13N1H8) C27P1H14 (C27N1H14)
Structure Hexagonal Two pentagons P(N) at internal site P(N) at external site SW defect
Ecoh/at (eV) 6.30 (6.53) 5.74 (6.27) 5.65 (5.71) 5.57 (5.59) 5.67 (6.12)
DEcoh/at (eV) 0 (0) 0.56 (0.26) 0.65 (0.82) 0.73 (0.94) 0.63 (0.41)
The cohesive energies for the FL-CNx analogous systems (i.e., in which P is substituted for N) are shown in parentheses. The bottom row represents the energy cost for the
corresponding defects relative to the hexagonal structures C23P1H12 (and C23N1H12, respectively) (DEcoh/at). In all these results, the lowest energy positions of the P or the N
atom were used.
a
Reproduced from Furlan, A.; Gueorguiev, G. K.; Ho¨gberg, H.; Stafstro¨m, S.; Hultman, L. Fullerene-Like CPx: A First-Principles Study of the Relative Stability of Precursors and Defect
Energetics during Synthetic Growth. Thin Solid Films 2006, 515, 1028–1032 and Gueorguiev, G.; Neidhardt, J.; Stafstrom, S.; Hultman, L. First-Principles Calculations on the Role of CN
Precursors for the Formation of Fullerene-Like Carbon Nitride. Chem. Phys. Lett. 2005, 401 (1–3), 288.
with pentagon defects (a double-pentagon defect; Figure 14(b)), a cluster containing one tetragon (Figure 14(c)), and a formation
corresponding to the Stone–Wales (SW) defect (Figure 14(d)) were calculated (28). The specific double-pentagon defect shown in
Figure 14(b) was found to be energetically favorable compared to the single-pentagon defect, similar to the case of FL-CNx (78).
Table 1 compares the cohesive energy results for the model clusters in Figure 14, with the data pertaining to the FL-CNx obtained
in Ref. (78) for the structural analogs of those clusters. The calculation shows that the incorporation of P atoms in a graphene
network is energetically more expensive (by 0.2–0.3 eV), depending on if the substitution site is internal or peripheric for the
model cluster and also on the presence of previously incorporated P atoms, than incorporation of an N atom to form CNx. Overall,
the same is also valid for the energy cost of introduction of single and double pentagons, and the SW defect. The higher cost of the
substitutional P at C sites (Table 1) is related to the larger covalent radius and lower electronegativity of the P atoms compared to
those of the N atom. This makes the P atom more difficult to place substitutionally in a grapheme sheet; however, it enhances the
site’s reactivity. The double C–P bond works also as a curvature-enhancing factor. The pentagon defects (single pentagons or
combinations of them) are less favorable in CPx than in CNx, that is, they are by 0.56 eV per atom higher in energy than the
grapheme structure (against only 0.26 eV per atom for CNx). The same is valid to a lesser extent for the SW defects, which is not
surprising since they also contain pentagons. It is remarkable that the energy cost of the tetragons in CPx appears to be, on average,
0.2 eV per atom lower than in CNx (see Table 1, where Ecoh/at for structures containing defects is compared to the Ecoh/at for
graphene systems).
In the case of CPx, the results in Table 1 show that the tetragons are approximately as stable as pentagon defects and/or SW
defects. The predicted stability of the tetragons is an important indication for stronger curved and shorter graphene layers in FL-CPx
than in FL-CNx, which implies also higher density of interlayer cross-linkages (94). Viewed in the context of the P chemistry, the
appearance of tetragons is not surprising, since P is a third-row element with low-energy d-orbitals that allow it to expand its octet
and to form four-membered ring transition states and intermediate structures. This electronic property of P is well-known as
defining its bonding nature, especially in the P-containing ring systems and cyclic compounds that represent an important part of
the organic chemistry of phosphorus (83).
4.15.4.2 Experimental Data on Fullerene-Like CPx Films

Furlan et al. have recently published experimental results that support the predictions about the existence of FL-CPx films (58,96).
They compared CP0.1 and CP0.24 films deposited by DC magnetron sputtering from a target containing 12 at% of phosphorus, at
temperatures of 300 C and 150 C, respectively. Figures 15 and 16 show the respective P 2p spectra from as-deposited film and
after Ar ion sputtering. The spectrum for the as-deposited films shows a high-intensity peak positioned at 133.1 eV, assigned to P–O
bonding, which decreases with Ar sputtering. The peaks at 132.6 eV and 133.8 eV correspond to contributions from P 2p3/2 and 2p1/
2, respectively. Finally, the lower energy peak positioned at 130 eV is assigned to P–C bonding in an FL-CPx compound, and fitted as
contributions P 2p3/2 at 130.2 eV and 2p1/2 at 131.4 eV (58). The proposed deconvolution of the P 2p and P 2s spectra neglects,
however, the possibility of P clusters.
Analysis of the P2p, P2s, and O1s spectra also enabled estimation of the oxidized phosphorus fraction in the CP0.1 and CP0.24
films. For the CP0.1 film, this fraction is 82% of the total P atoms incorporated in the as-deposited state and 18% in the clean-
sputtered state, while the corresponding numbers for the CP0.24 film are 54% and 22%, respectively. Grazing-angle XPS
measurement also revealed for the sputtered samples that oxidized phosphorus is located mainly at the outermost surface region
of the samples. This indicates extremely high sensitivity toward oxidation of the sample surface, which occurs even in UHV
conditions.
The influence of the deposition temperature and sputtering Ar pressure (PAr) on the microstructure was also revealed by HRTEM
imaging and SAED. Figure 17(a) shows that the CP0.1 film deposited at 300 C and PAr ¼ 3 mTorr has essentially amorphous results,
but exhibits short-range ordering with fragments of curved graphene sheets. Although the presence of any larger packages of the FL
structure similar to FL-CNx cannot be ruled out from such HRTEM images due to the possible superposition of nanoscale features, it
is not expected for CPx since the calculations show a strong tendency for P-induced interlinking of the graphene, which breaks the
www.iran-mavad.com
Figure 15 XPS P 2p spectra of CP0.1 film grown at Ts ¼ 300 C and PAr ¼ 3 mTorr: (a) as-deposited; (b) after 20 min sputtering with Ar ions; and
(c) after 60 min sputtering. Reproduced from Furlan, A.; Guerguiev, G.; Högberg, C. Z. H.; Braun, S.; Stafström, S.; Hultman, L. Synthesis of Phosphorous-
Carbide Thin Films by Magnetron Sputtering. Phys. Status Solidi: Rapid Res. Lett. 2008, 2 (4), 191–193.
Figure 16 XPS P 2p spectra of CP0.24 films grown at Ts ¼ 150 C and PAr ¼ 3 mTorr: (a) after 90 min sputtering with Ar ions; and (b) as deposited.
Reproduced from Furlan, A.; Gueorguiev, G.; Czigany, Z.; Darakchieva, V.; Braun, S.; Hogberg, H.; Hultman, L. Structural and Mechanical Properties of
Phosphorus-Carbide Thin Solid Films. Thin Solid Films 2012.
www.iran-mavad.com
Figure 17 HRTEM image and corresponding SAED pattern and its intensity distribution from CPx thin films deposited at Ts ¼ 300 C, Ub ¼ 25 V,
and (a) PAr ¼ 3 mTorr; and (b) PAr ¼ 9 mTorr. Reproduced from Furlan, A.; Guerguiev, G.; Högberg, C. Z. H.; Braun, S.; Stafström, S.; Hultman, L.
Synthesis of Phosphorous-Carbide Thin Films by Magnetron Sputtering. Phys. Status Solidi: Rapid Res. Lett. 2008, 2 (4), 191–193.
continuity of the planes (94). The corresponding SAED pattern shows that the CP0.1 film has broad rings at w1.6, w2.6, and
w5.9 Å. These rings differ from those of other C allotropes, as well as from FL structures like FL-CNx.
In contrast, CPx films deposited at Ts ¼ 300 C and working gas pressure of 9 mTorr show amorphous structure with close to
graphitic short-range ordering. These samples exhibit SAED peaks at 1.26, 2.11, and 3.9 Å (Figure 17(b)). The first two peaks
coincide with those also found in FL-CNx and some other amorphous C allotropes (66). The position of the third peak at w4 Å,
however, differs slightly from the w3.5 Å peak typical for the FL structure known from FL-CNx (66).
4.15.5 Carbon Fluoride Films
The first reported synthesis of a carbon–fluorine compound was in 1836, when the French chemists Dumas and Peligot reported
the synthesis of fluoromethane (CH3F) (97). In 1890, Moissan studied the interaction of fluorine with carbon (98) and reported
erroneously the isolation of carbon tetrafluoride (CF4) from volatile material produced by igniting powdered carbon in fluorine
(99). He reported a boiling point of 15 C, while the correct value is 129 C. In fact, Moissan’s compatriots Lebeau and
Damiens got the credit in 1926 for the synthesis of CF4 (100). In 1934, Ruff and Bretschneider (101) discovered that graphite and
fluorine combined without combustion at about 420 C formed a gray-colored solid. The approximate composition of this solid
was (CFx)n, where x was nearly equal to 1. They called the compound ‘carbon monofluoride’ and found that it was hydrophobic
and had a high electrical resistance. X-ray diffraction studies indicated that the distance between the carbon layer planes had been
expanded from 3.4 Å in graphite to about 8 Å in graphite fluoride (101). Four years later, in 1938, Plunkett discovered acci-
dentally the solid polytetrafluoroethylene (C2F4)n, also known today as PTFE or its trademark name from DuPont, Teflon.
Kinetic Chemicals (a company owned by DuPont and General Motors) patented the new fluorinated plastic in 1941 (102), and
DuPont registered the Teflon trademark in 1945 (103). In 1954 Marc Gregoire, a French engineer, found a way to bond PTFE to
www.iran-mavad.com
aluminum, and the first nonstick cookware was created. He founded in 1956 the Tefal Corporation, and 4 years later the pans
started to be sold in the United States.
The deposition of carbon-fluoride thin films as a conformal surface coating is more recent. In 1963, Logan reported an elec-
trodeposition method to deposit thin PTFE coatings of up to 25 mm thick. The coatings were intended to be used as dry lubricants in
a wide variety of military equipment (104). In 1965, Seyer et al. patented a method to coat razor blades with a carbon-fluoride film
applied as a suspension or dispersion in a volatile liquid medium: the coating thus deposited was then dried, and the blades backed
to fuse the solid fluorocarbon particles and form an adhering thin film (105). The Teflonlike coated razor blade was claimed to glide
more smoothly during shaving than the uncoated one. Other applications in the area of lubrication were studied by NASA when
coatings of about 1 mm thickness were applied to roughened or polished surfaces by mechanically rubbing them over the surfaces at
a constant load (106,107).
The first plasma-deposited carbon-fluoride films were reported at the end of the 1960s. In an article published in 1969, Harrop
et al. describe the frictional properties of PTFE films deposited by RF sputtering (108). The publication mentions a British Provi-
sional Patent application from 1967 to disclose the method, but it seems that the patent was never filled out or was rejected because
there is no information on a patent in their name at the British Patent Office website. In 1972, Murakami et al. (109) discusses the
deposition of fluorinated ethylene propylene (FEP, [CF(CF3)–CF2(CF4)n]m) by vacuum evaporation. He also shows that, in contrast
to FEP, PTFE cannot be thermally evaporated and condensed on the substrate surface because it decomposes below its vaporization
temperature. One year later, Millard et al. deposited fluorocarbon films by introducing perfluorobutene-2 into the afterglow region
of an Ar plasma generated using plates excited at 13.56 MHz at 70 W of power. The films were light yellow in color, adhered strongly
to dry glass substrates, and exhibited liquid contact angles and infrared spectra very similar to those reported on PTFE surfaces (110).
The same year, Morrison and Robertson demonstrated the possibility of depositing PTFE films by RF sputtering of a PTFE target.
Their films were also yellow-colored amorphous substoichiometric films with a higher hardness and lower resistivity than the bulk
material (111).
During the last 15 years, many different approaches have been applied to producing CFx films using PVD and plasma
techniques, but the most common types employed are PECVD, vacuum evaporation, and reactive magnetron sputtering.
A variety of precursors have been tried for producing CFx films by PECVD techniques, like CF4, C2F6, C2F4, C3F6, CHF3, C4F8,
C6F6, and C6H5F (112,113). Vacuum evaporation techniques include thermal evaporation by resistive heating or the use of a laser
(114–116). In the case of magnetron sputtering, targets of pure carbon have been used with reactive gases like in PECVD
techniques, while in other cases the films are directly deposited from PTFE targets in Ar discharges or reactive atmosphere
(59,117–119).
4.15.5.1 Synthetic Growth Predictions

The structural and bonding patterns arising from the incorporation of fluorine atoms in a graphenelike network relevant to the
deposition of CFx films were recently addressed by first-principles calculations (120). The framework adopted for their calcu-
lations was DFT, mostly within its generalized gradient approximation. A systematic study of a variety of CFx (0 < x < 0.16)
model systems (templates) obtained by substitution of C atoms by F atoms in finite graphenelike structures was carried out by
geometry optimizations. Similar model systems have been successfully employed for addressing FL systems (28). In contrast to
what happens in other carbon-based compounds where the dopant element differs less radically from carbon in its bonding
properties (e.g., CNx (74), CPx (94), and CSx (121)), in the case of carbon fluoride, most of the aforementioned structural defects
do not prevail after geometry relaxations due to breakage of bonds and atomic rearrangements. Geometry relaxations of
templates resulted in a set of CFx structure-defining patterns displayed in Figure 18. The diversity of structural patterns identified
in CFx includes large N-member (N ¼ 8–12) rings (Figure 18(a1) and 18(a2)); defects by branching of the model system,
resulting in the formation of carbon chains (Figure 18(b1)) or nanoribbons (Figure 18(b2)); structures defined by a C–F bond
rotation at the edge of a model system (Figure 18(c1) and 18(c2)); and a configuration comprising an F atom belonging to a ring
(Figure 18(d)). The cohesive energy gains associated with these defect patterns vary predominantly in the range of w0.2–0.4 eV/
at, which is comparable with the energy costs of a wide range of defects in C-based nanostructured compounds (120). Different
CFx defect patterns can coexist in close vicinity without merging, thus preserving the system from full amorphization. Both this
result and the particular stability of the large N-member rings favor an ordered FL structure. In contrast, the ‘branching’ pattern
promotes amorphous structure with elements of polymeric structure that are expected to become more pronounced as the F
concentration increases. The simulation results suggest that FL-CFx can be synthesized at F concentrations below w10 at%. As
expected, depending on the particular CFx model system and its selection of defects, the transition from FL to amorphous system
with elements of polymeric structure does not always happen at the same F concentration, but corresponds to a range of F
concentrations, normally between 8 at% and 12.5%. Fluorine is an extremely reactive species, and it is also the most electro-
negative chemical element, so it is not surprising that FL-CFx demands the lowest concentration of the dopant when compared to
other FL compounds such as CNx (up to 30 at% of N), CPx (up to 15 at% of P), and CSx (up to 15–18 at% of S) (120). The
described mechanisms of F incorporation in CFx and the richness of structural patterns in CFx thin films reveal a new compound
with tunable structure (FL, polymeric, or amorphous) achieved by varying its F incorporation rate. The polymeric CFx may lead to
a new material as significant as Teflon. Foreseen applications for CFx films include tuning their mechanical properties and surface
energy by controlling the deposition conditions for, for example, coating low-friction mechanical details, such as rolling and
sliding bearings.
www.iran-mavad.com
Figure 18 Examples of CFx structural patterns arising from the incorporation of F atoms in a graphenelike network: (a1) a large nine-member ring (the
inset represents an inclined view of the same pattern illustrating its curvature); and (a2) a large nine-member ring adjacent to a pure carbon pentagon.
Both structures belong to pattern A. (b1) Branching in two symmetrical chains; and (b2) branching in two symmetrical ribbons of hexagons; both
structures belong to pattern B. (c1) and (c2) Two variants of an ‘uncomplicated’ C–F bond rotation, belonging pattern C; and (d) an F atom included in
a ring, belonging to pattern D. White atoms represent H, while gray and black ones correspond to C and F, respectively. Reproduced from Gueorguiev, G.;
Goyenola, C.; Schmidt, S.; Hultman, L. CF(x): A First-Principles Study of Structural Patterns Arising during Synthetic Growth. Chem. Phys. Lett. 2011, 516
(1–3), 62–67.
4.15.5.2 Deposition of CFx Films by HiPIMS

Recently, Schmidt et al. have shown for the first time that CFx films can also be deposited by HiPIMS (59). HiPIMS offers the
advantage of an increased ionized flux of the target material, due to a high temporal electron density in the plasma. Furthermore, the
ion flux and the ion-to-neutral flux may be controlled to a certain extent by applied bias voltage and pulse energy settings,
respectively. In their experiments, Schmidt et al. used an industrial coating system CC800/9 from CemeCon with a big rectangular
graphite target (440 cm2) and a reactive Ar–CF4 atmosphere of 400 mPa with CF4 partial pressures (pCF4 ) in the range 15–110 mPa.
The cathode was operated in a power-regulated mode with a frequency of 300 Hz, an average power of 1400 W, and the pulse width
kept constant at 200 ms.
Figure 19 shows a set of current–voltage wave forms as a function of CF4 partial pressure. The voltage is not constant throughout
the pulse, but drops from the peak value at the beginning of the pulse when the charge is drawn from the capacitor bank and the
plasma is ignited. Simultaneously, the target current increases to reach the peak value at the time instant that corresponds to the
fastest drop in target voltage. Up to a CF4 partial pressure of 42 mPa, the peak target current is comparable for all applied processes,
but a steady increase is observed at higher partial pressures (cf. Figure 19, inset). This indicates that the amount of charged, sputtered
species is increased, which may be attributed to an increased number of easily ionized species in the target’s vicinity and an increase
in the secondary electron emission yield resulting from the changed sputter gas as well as the reactions taking place at the surface of
the target.
The deposition rate, as shown in Figure 20, first increases steeply as the CF4 partial pressure increases followed by a decrease for
pCF4 > 42 mPa. The simultaneous decrease in deposition rate and increase in target current above pCF4 ¼ 42 mPa indicate a less
effective sputter process. This can be attributed to a combination of a changed target surface chemistry and an increased number of
fractured, light CF4 species that do not equal the momentum transfer of Ar. Another factor influencing the deposition rate, not
coupled to the target current, is the higher probability to form C2F2 species when increasing pCF4 . According to theoretical studies,
C2F2 species play a role as a desorption agent during growth as well as in the removal of F atoms from the substrate surface (59).
As expected, the fluorine content of the films increases as pCF4 increases. The fluorine content shows a peak value of 35 at% at
pCF4 w80 mPa. A further increase in CF4 partial pressure to w110 mPa, however, yields merely a reduction to a value of 30.6 at% of
fluorine within the films. According to theoretical results (59), this reduction of fluorine content is attributed to the higher amount
of fluorine etchants, in particular C2F2 species. In contrast, films with less than 16 at% fluorine could not be obtained from the
applied reactive HiPIMS process with CF4. ERDA furthermore revealed oxygen contents of 1 at% or below and hydrogen contents
below 3 at% in all examined samples. The density of the synthesized CFx films ranged between 1.4 0.1 g cm3 and
1.6 0.2 g cm3, which is lower than that of graphite, and is attributed to the incorporation of fluorine.
Results from XPS and Raman spectroscopy indicate a graphitic nature of CFx films with x 0.23 and a polymeric structure for
films with x 0.26. The bonding in these films comprises C–C, CF, CF2, and very little CF3. CF4 partial pressures above 55 mPa lead
to an even less ordered polymerlike structure, also containing CF4 bonding (59).
www.iran-mavad.com
Figure 19 Recorded target current and voltage wave forms of the deposition processes as a function of CF4 partial pressure. The inset illustrates the
increasing peak target current (I ) with increasing CF4 partial pressure. Reproduced from Schmidt, S.; Greczynski, G.; Goyenola, C.; Gueorguiev, G.;
Czigany, Z.; Jensen, J.; Ivanov, I.; Hultman, L. CFx Thin Films Deposited by High Power Impulse Magnetron Sputtering: Synthesis and Characterization.
Surf. Coat. Technol. 2011, 206, 646–653.
Figure 20 Deposition rate and the fluorine content (F) obtained from elastic recoil detection analysis (ERDA) measurements of CFx thin films as
a function of the CF4 partial pressure. Reproduced from Schmidt, S.; Greczynski, G.; Goyenola, C.; Gueorguiev, G.; Czigany, Z.; Jensen, J.; Ivanov, I.;
Hultman, L. CFx Thin Films Deposited by High Power Impulse Magnetron Sputtering: Synthesis and Characterization. Surf. Coat. Technol. 2011, 206,
646–653.
Based on high-resolution TEM images, the microstructure of the CFx thin films deposited at low substrate temperature is
amorphous: Figure 21(a–c) shows plan views and (d) a cross-section, with corresponding SAED patterns for x ¼ 0.19, 0.23, 0.26,
and 0.35, respectively. No lattice fringes of curved planes could be seen. The SAED patterns, instead, consist of broad diffuse rings
(with FWHM of 0.1–0.2 Å1), characteristic of amorphous materials (71). Specifically, the rings have diameters of w1.15 Å and
w2.15 Å, which are typical for amorphous carbon allotropes (71). Moreover, a broad ring of low intensity was observed between
0.55 Å and 0.75 Å (not visible in Figure 21 due to low intensity). The lack of a ring at w3.5 Å indicates that no graphite or FL short-
range ordering is present in the films, as was otherwise the case for FL-CNx thin films (71).
www.iran-mavad.com
Figure 21 (a–c) Plan view; and (d) cross-sectional HRTEM images with corresponding SAED patterns from CFx films with x ¼ 0.19, x ¼ 0.23, x ¼ 0.26,
and x ¼ 0.35. Reproduced from Schmidt, S.; Greczynski, G.; Goyenola, C.; Gueorguiev, G.; Czigany, Z.; Jensen, J.; Ivanov, I.; Hultman, L. CFx Thin Films
Deposited by High Power Impulse Magnetron Sputtering: Synthesis and Characterization. Surf. Coat. Technol. 2011, 206, 646–653.
Carbon-based coatings have come of age. This review has focused on trends in the development of doped – or rather alloyed –
diamondlike coatings, in particular the new class of FL materials.
We have described unique resilient FL compounds by self-organization of nano-curved sp2-hybridized carbon features, with
tuned mechanical and surface energy properties. They are unique resilient materials consisting of bent and intersecting hexagonal
basal planes, fabricated by the incorporation of odd-member rings. Cross-linking enables the material to extend the strength of the
covalently 2D hexagonal carbon (graphenelike) network into 3D. Presently, the most studied compound that exhibits cross-linking
is carbon nitride. The first magnetron-sputtered FL-CNx films were discovered in Linköping University (122). In FL-CNx (x 0.3),
substituting N for C promotes bending and cross-linkage of layers at a much lower energy cost as compared to C-only. The material
deforms by bond angle deflection and compression of the graphene layers. Concomitantly, the superior strength of the sp2 bonds
inhibits plastic deformation and stores the deformation energy elastically. The resiliency of the FL structure, in combination with the
low friction coefficient of C, is beneficial for tribological applications. Also, alloying can tune the surface energy, to eliminate the
need for lubricating oils or retaining synovial fluids in coated bearings or orthopedic implants (123).
More recently, it was discovered that P in FL-CPx (phosphorous-carbide) has an even stronger tendency of curving graphene than
N due to P’s electronic properties and preference for tetrahedral coordination. Nanoindentation and electron microscopy confirmed
DFT predictions for FL structures with high hardness and extreme elasticity (58). The most recent results for CFx (carbon fluoride)
are promising, offering control over the structure (FL, polymeric, and amorphous) by varying the F content (59). The application
driver is the quest for improved hardness, adhesion, and thermal resistance to replace Teflon-type coatings.
Future FL materials to be developed may be FL-CSx (28), as well as FL-C alloys. The alloying elements should be chosen to
provide bending and cross-linking of the graphene basal planes and to tune dangling bond densities for the design of surface energy
and low friction and wettability. Potential structures and precursor molecules can be identified by ab-initio and molecular dynamics
methods.
Synthesis of FL C-based films is particularly well realized by PVD processes, including reactive magnetron sputtering, high-power
pulsed magnetron sputtering, pulsed laser ablation, and cathodic arc deposition. Elements of chemical sputtering are common to
these processes, including the formation of precursor species that serve as building blocks for the FL basal planes.
For modeling of FL carbon-based compound growth, the so-called synthetic growth concept shows large potential, including its
predictive power in identifying new compounds or for structural design. It is based on first-principles algorithmic structural
optimizations and cohesive energies evaluations (74,78). Synthetic growth modeling of the structural evolution in FL materials has
been demonstrated for the C-N, C-P, and C-F systems. The model employs successive growth events where the starting bonding
www.iran-mavad.com
geometry for each added precursor is defined by the relaxed state of the most favorably bonded precursors at the previous step. It can
be used to consider atomic intercalation and substitution, mixing of sp2 and sp3 networks, structural order and disorder, adsorption
and desorption, dangling bonds, and vacancies and impurities.
Acknowledgments
The authors acknowledge J. -E. Sundgren, G. K. Gueorguiev, H. Sjöström, N. Hellgren, J. Neidhardt, A. Furlan, and S. Schmidt for
valuable discussions. Support has been received from the following organizations: VR, SSF, VINNOVA, ERC, SKF AB, and
CemeCon AG.
Bibliography
1. Guthrie, R. D. The Nature of Paleolithic Art; University of Chicago Press: Chicago, 2005, 30–32.
2. Pines, C. C. The Story of Ink. Am. J. Police Sci. 1931, 2 (4), 290–301.
3. Herschel, W. Investigation of the Powers of the Prismatic Colours to Heat and Illuminate Objects; With Remarks, That Prove the Different Refrangibility of Radiant Heat. To
Which is Added, an Inquiry into the Method of Viewing the Sun Advantageously with Telescopes. Philos. Trans. R. Soc. London 1800, 90, 255–283.
4. Henry, W.; Henry, T. Experiments on Carbonated Hydrogenous Gas; With a View to Determine Whether Carbon be a Simple or a Compound Substance. Philos. Trans. R. Soc.
London 1797, 87, 401–415.
5. Sawyer, W. E.; Man, A. Carbon for Electric Lights. US Patent 229335, June 29, 1880.
6. Gambrell, T. E.; Harris, T. F. Improvements in Electrical Resistances. UK Patent GBD189725412, November 2, 1897.
7. Holmes, T. Fixed High-Stability Carbon Resistors. J. Inst. Electr. Eng. Part IIIA Radiocommunication 1947, 94 (16), 912–914.
8. Bradley, D. E. A High-Resolution Evaporated-Carbon Replica Technique for the Electron Microscope. J. Inst. Met. 1954, 83, 35–38.
9. Bradley, D. E. Evaporated Carbon Films for Use in Electron Microscopy. Br. J. Appl. Phys. 1954, 5, 65–66.
10. Blue, M. D.; Danielson, G. C. Electrical Properties of Arc-Evaporated Carbon Films. J. Appl. Phys. 1957, 28, 583–586.
11. Stewart, R. L. Insulating Films Formed Under Electron and Ion Bombardment. Phys. Rev. 1934, 45, 488–490.
12. Schmellenmeier, H. Die Beeinflussung von festen Oberflachen durch eine ionisierte Gasatmosphare. Experimentelle Technik der Physik 1953, 1, 49–68.
13. Aisenberg, S.; Chabot, R. Study of the Deposition of Single-Crystal Silicon, Silicon Dioxide and Silicon Carbide on Cold-Substrate Silicon. Final Report prepared for NASA
Electronic research Center, Contract No. NAS 12-451, 1969.
14. Aisenberg, S.; Chabot, R. Ion Beam Deposition of Thin Films of Diamondlike Carbon. J. Appl. Phys. 1971, 42 (7), 2953–2958.
15. Aisenberg, S.; Chabot, R. W. Physics of Ion Plating and Ion Beam Deposition. J. Vac. Sci. Technol. 1973, 10 (1), 104–110.
16. Weissmantel, C.; Bewilogua, K.; Dietrich, D.; Erler, H.-J.; Hinneberg, H.-J.; Klose, S.; Nowick, W.; Reisse, G. Structure and Properties of Quasi-Amorphous Films Prepared by
Ion Beam Techniques. Thin Solid Films 1980, 72 (1), 19–32.
17. Erdermir, A.; Donnet, C. Tribology of Diamond-Like Carbon Films: Recent Progress and Future Prospects. J. Appl. Phys. D Appl. Phys. 2006, 39, R311–R327.
18. Culbertson, R. D.; McRae, R. C.; Meyn, H. P. US Patent 3604970, 1971.
19. Hainsworth, S. V.; Uhure, N. J. Diamond like Carbon Coatings for Tribology: Production Techniques, Characterisation Methods and Applications. Int. Mater. Rev. 2007, 52 (3),
153–174.
20. Liu, A. Y.; Cohen, M. L. Prediction of New Low Compressibility Solids. Science 1989, 245, 841–842.
21. Liu, A. Y.; Cohen, M. L. Structural Properties and Electronic Structure of Low-Compressibility Materials: b-Si3N4 and Hypothetical b-C3N4. Phys. Rev. B Condens. Matter
Mater. Phys. 1990, 41 (15), 10727–10734.
22. Broitman, E.; Neirdhart, J.; Hultman, L. Fullerene-Like Carbon Nitride: A New Carbon-Based Tribological Coating. In Tribology of Diamond-Like Carbon Films: Fundamentals
and Applications; Donnet, C., Erdemir, A., Eds.; Springer: New York, 2007; pp 620–653.
23. Vallauri, D.; Atías Adrián, I.; Chrysanthou, A. TiC–TiB2 Composites: A Review of Phase Relationships, Processing and Properties. J. Eur. Ceram. Soc. 2008, 28 (8),
1697–1713.
24. Varma, A.; Palshin, V.; Meletis, E. I. Structure-Property Relationship of Si-DLC Films. Surf. Coat. Technol. 2001, 148 (2–3), 305–314.
25. Berlind, T.; Hellgren, N.; Johansson, M.; Hultman, L. Microstructure, Mechanical Properties, and Wetting Behavior of Si–C–N Thin Films Grown by Reactive Magnet. Surf.
Coat. Technol. 2001, 141 (2–3), 145–155.
26. Damasceno, J. C.; Camargo, S. S.; Freire, F. L.; Carius, R. Deposition of Si-DLC Films with High Hardness, Low Stress and High Deposition Rates. Surf. Coat. Technol. 2000,
133–134, 247–252.
27. Schulz, H.; Leonhardt, M.; Scheibe, H.-J.; Schultrich, B. Ultra Hydrophobic Wetting Behaviour of Amorphous Carbon Films. Surf. Coat. Technol. 2005, 200 (1–4), 1123–1126.
28. Furlan, A.; Gueorguiev, G. K.; Högberg, H.; Stafström, S.; Hultman, L. Fullerene-Like CPx: A First-Principles Study of the Relative Stability of Precursors and Defect Energetics
during Synthetic Growth. Thin Solid Films 2006, 515, 1028–1032.
29. Ferrari, A. C. Diamond-Like Carbon for Magnetic Storage Disks. Surf. Coat. Technol. 2004, 180–181, 190–206.
30. McKenzie, D. Tetrahedral Bonding in Amorphous Carbon. Rep. Prog. Phys. 1996, 59 (12), 1611–1664.
31. Jacob, W.; Moller, W. On the Structure of Thin Hydrocarbon Films. Appl. Phys. Lett. 1993, 63, 1771–1773.
32. Robertson, J. Amorphous Carbon. Adv. Phys. 1986, 35, 317–374.
33. Robertson, J. Diamond-Like Amorphous Carbon. Mater. Sci. Eng., R 2002, 37, 129–281.
34. Vanhulsel, A.; Velasco, F.; Jacobs, R.; Eersels, L.; Havermans, D.; Roberts, E.; Sherrington, I.; Anderson, M.; Gaillard, V. DLC Solid Lubricant Coatings on Ball Bearings for
Space Applications. Tribol. Int. 2007, 40 (7), 1186–1194.
35. Martinez-Martinez, D.; Schenkel, M.; Pei, Y.; Sánchez-López, J.; De Hosson, J. Microstructure and Chemical Bonding of DLC Films Deposited on ACM Rubber. Surf. Coat.
Technol. 2011, 205 (2), S75–S78.
36. Shirakura, A.; Nakaya, M.; Koga, Y.; Kodama, H.; Hasebe, T.; Suzuk, T. Diamond-Like Carbon Films for PET Bottles and Medical Applications. Thin Solid Films 2006, 494
(1–2), 84–91.
37. VDI. Report VDI 2840: Carbon Films – Basic Knowledge, Coating Types and Properties; Association of German Engineers (VDI): Frankfurt, 2006.
38. Erdermir, A.; Donnet, C. Tribology of Diamond, Diamond-Like Carbon, and Related Films. In Modern Tribology Handbook; Bhushan, B., Ed.; CRC Press: New York, 2001.
p. Ch. 24.
39. Spitsyn, B.; Bouilov, L.; Derjaguin, B. Diamond and Diamond-Like Films: Deposition from the Vapour Phase, Structure and Properties. Prog. Cryst. Growth Charact. 1988,
17 (2), 79–170.
40. Cuomo, J. J.; Leary, P. A.; Yu, D.; Reuter, W.; Frisch, M. Reactive Sputtering of Carbon and Carbide Targets in Nitrogen. J. Vac. Sci. Technol. 1979, 16 (2), 299–302.
www.iran-mavad.com
41. Sung, C. M.; Sung, M. Carbon Nitride and Other Speculative Superhard Materials. Mater. Chem. Phys. 1996, 43 (1), 1–28.
42. Cohen, M. L. Calculation of Bulk Moduli of Diamond and Zinc-Blende Solids. Phys. Rev. B Condens. Matter Mater. Phys. 1985, 32, 7988–7991.
43. Teter, D. Computational Alchemy: The Search for New Superhard Materials. MRS Bulletin 1998, 23 (1), 22–27.
44. Carbon Nitride Seen Out-Muscling Diamonds; American Metal Market, 10 June 1992.
45. Diamonds May No Longer be Hardest Substance Carbon Nitride; American Metal Market, 28 July 1993.
46. Harder Than Diamond Is Not Faster Than Light. Discover 1993, 14 (11), 132–133.
47. As Hard as Diamond? Tracking the Elusive Carbon Nitride. Sci. News 11 July 1998.
48. A New Route to a Superhard Material? Sci. News 5 June 1999.
49. Muhl, S.; Mendez, J. M. A Review of the Preparation of Carbon Nitride Films. Diamond Relat. Mater. 1999, 8 (10), 1809–1832.
50. Wang, E. G. Research on Carbon Nitrides. Prog. Mater. Sci. 1997, 41 (5), 241–301.
51. Kroke, E.; Schwarz, M. Novel Group 14 Nitrides. Coord. Chem. Rev. 2004, 248, 493–532.
52. DeVries, R. C. Inventory on Innovative Research: The Case of C3N4. Mater. Res. Innovations 1997, 1 (3), 161–162.
53. Malkow, T. Critical Observations in the Research of Carbon Nitride. Mater. Sci. Eng. A 2001, 302 (2), 311.
54. Matsumoto, S.; Xie, E.; Izumi, F. On the Validity of the Formation of Crystalline Carbon Nitrides, C3N4. Diamond Relat. Mater. 1999, 8 (7), 1175.
55. Sjostrom, H.; Stafstrom, S.; Boman, M.; Sundgren, J.-E. Superhard and Elastic Carbon Nitride Thin Films Having Fullerene-Like Microstructure. Phys. Rev. Lett. 1995,
75 (7), 1336.
56. Neidhardt, J.; Broitman, E.; Hultman, L. Fullerene-Like Carbon Nitride: A New Thin Film Material. In Wide Bandgap Materials and New Developments; Syväjärvi, M.,
Yakimova, R., Eds.; Research Signpost: Singapore, 2006; pp 1–39.
57. Neidhardt, J.; Czigany, Z.; Brunell, I.; Hultman, L. Growth of Fullerene-Like Carbon Nitride Thin Solid Films by Reactive Magnetron Sputtering; Role of Low-Energy Ion Irradiation
in Determining Microstructure and Mechanical Properties. J. Appl. Phys. 2003, 93 (5), 3002.
58. Furlan, A.; Guerguiev, G.; Högberg, C. Z. H.; Braun, S.; Stafström, S.; Hultman, L. Synthesis of Phosphorous-Carbide Thin Films by Magnetron Sputtering. Phys. Status Solidi:
Rapid Res. Lett. 2008, 2 (4), 191–193.
59. Schmidt, S.; Greczynski, G.; Goyenola, C.; Gueorguiev, G.; Czigany, Z.; Jensen, J.; Ivanov, I.; Hultman, L. CFx Thin Films Deposited by High Power Impulse Magnetron
Sputtering: Synthesis and Characterization. Surf. Coat. Technol. 2011, 206, 646–653.
60. Stafstrom, S. Reactivity of Curved and Planar Carbon–Nitride Structures. Appl. Phys. Lett. 2000, 77, 3941–3943.
61. Johansson, A. Charge Dynamics and Electronic Structure of Pi-Conjugated Systems. Doctoral Thesis No. 783, Linköping Studies in Science and Technology: Linköping,
Sweden, 2002.
62. Garcia, I.; -Berasategui, E. G.; Bull, S.; Page, T.; Neidhardt, J.; Hultman, L. How Hard is Fullerene-Like CNx? Some Observations from the Nanoindentation Response of
a Magnetron-Sputtered Coating. Philos. Mag. A 2002, 82 (10 SPEC), 2133.
63. Broitman, E.; Hellgren, N.; Neidhardt, J.; Brunell, I.; Hultman, L. Electrical Properties of Carbon Nitride Thin Films: Role of Morphology and Hydrogen Content. Letters.
J. Electron. Mater. 2002, 31 (9), L11–L15.
64. Neidhardt, J.; Hogberg, H.; Hultman, L. Cryogenic Deposition of Carbon Nitride Thin Solid Films by Reactive Magnetron Sputtering: Suppression of the Chemical Desorption
Processes. Thin Solid Films 2004, 478 (1–2), 34–41.
65. Hellgren, N.; Johansson, M.; Hjorvarsson, B.; Broitman, E.; Ostblom, M.; Liedberg, B.; Hultman, L.; Sundgren, J. Growth, Structure, and Mechanical Properties of CNxHy Films
Deposited by DC Magnetron Sputtering in N2/Ar/H2 Discharges. J. Vac. Sci. Technol. A 2000, 18 (5), 2349.
66. Neidhardt, J.; Hultman, L.; Czigany, Z. Correlated High Resolution Transmission Electron Microscopy and X-Ray Photoelectron Spectroscopy Studies of Structured CNx (0 < x
< 0.25) Thin Solid Films. Carbon 2004, 42 (12–13), 2729.
67. Hultman, L.; Neidhardt, J.; Hellgren, N.; Sjostrom, H.; Sundgren, J.-E. Fullerene-Like Carbon Nitride: A Resilient Coating Material. MRS Bulletin 2003, 28 (3), 194.
68. Hellgren, N.; Nian, L.; Broitman, E.; Serin, V.; Grillo, S.; Twesten, R. P. I.; Colliex, C.; Hultman, L.; Sundgren, J. Thermal Stability of Carbon Nitride Thin Films. J. Mater. Res.
2001, 16 (11), 3188.
69. Gago, R.; Jiménez, I.; Neidhardt, J.; Abendroth, B.; Caretti, I. H. L.; Moller, W. Correlation between Bonding Structure and Microstructure in Fullerene-Like Carbon Nitride Thin
Films. Phys. Rev. B Condens. Matter Mater. Phys. 2005, 71, 125414.
70. Gago, R.; Neidhardt, J.; Vinnichenko, M.; Kreissig, U.; Czigany, Z.; Kolitsch, A.; Hultman, L.; Moller, W. Synthesis of Carbon Nitride Thin Films by Low Energy Ion Beam
Assisted Evaporation: On the Mechanisms for Fullerene-Like Microstructure Formation. Thin Solid Films 2005, 483 (1–2), 89–94.
71. Czigany, Z.; Hultman, L. Interpretation of Electron Diffraction Patterns from Amorphous and Fullerene-Like Carbon Allotropes. Ultramicroscopy 2010, 110 (7), 815–819.
72. Hultman, L.; Stafstrom, S.; Czigany, Z.; Neidhardt, J.; Hellgren, N.; Brunell, I.; Suenaga, K.; Colliex, C. Cross-Linked Nano-Onions of Carbon Nitride in the Solid Phase:
Existence of a Novel C48N12 Aza-Fullerene. Phys. Rev. Lett. 2001, 87 (22), 225503.
73. Gammon, W.; Hoatson, G.; Holloway, B.; Vold, R.; Reilly, A. Bonding in Hard and Elastic Amorphous Carbon Nitride Films Investigated Using 15N, 13C, and 1H NMR
Spectroscopy. Phys. Rev. B Condens. Matter Mater. Phys. 2003, 68 (19), 195401.
74. Gueorguiev, G.; Neidhardt, J.; Stafstrom, S.; Hultman, L. First-Principles Calculations on the Curvature Evolution and Cross-Linkage in Carbon Nitride. Chem. Phys. Lett. 2005,
410 (4–6), 228.
75. Neidhardt, J.; Hultman, L.; Abendroth, B.; Gago, R.; Moller, W. Diagnostics of a N2/Ar Direct Current Magnetron Discharge for Reactive Sputter Deposition of Fullerene-Like
Carbon Nitride Thin Films. J. Appl. Phys. 2003, 94 (11), 7059.
76. Voevodin, A.; Jones, J.; Zabinski, J.; Hultman, L. Plasma Characterization During Laser Ablation of Graphite in Nitrogen for the Growth of Fullerene-Like CNx Films. J. Appl.
Phys. 2002, 92 (2), 724.
77. Hellgren, N.; Johansson, M.; Broitman, E.; Sandstrom, P.; Hultman, L.; Sundgren, J.-E. Effect of Chemical Sputtering on the Growth and Structural Evolution of Magnetron
Sputtered CNx Thin Films. Thin Solid Films 2001, 382 (1–2), 146–152.
78. Gueorguiev, G.; Neidhardt, J.; Stafstrom, S.; Hultman, L. First-Principles Calculations on the Role of CN Precursors for the Formation of Fullerene-Like Carbon Nitride. Chem.
Phys. Lett. 2005, 401 (1–3), 288.
79. Hartmann, J.; Siemroth, P.; Schultrich, B.; Rauschenbach, B. Characterization of Carbon Nitride Produced by High-Current Vacuum Arc Deposition. J. Vac. Sci. Technol. A
1997, 15, 2983–2987.
80. Veisz, B.; Radnoczi, G. Growth Model for Arc-Deposited Fullerene-Like CNx Nanoparticles. Microsc. Res. Tech. 2005, 67, 100–105.
81. Schmidt, S.; Czigány, Z.; Greczynsk G.; Jensen, J.; Hultman, L. Ion Mass Spectrometry Investigations of the Discharge during Reactive High Power Pulsed and Direct Current
Magnetron Sputtering of Carbon in Ar and Ar/N2. J. Appl. Phys. 2012, 112, 013305
82. de Mahler, E. A General Method for the Preparation of Carbides of Metalloids, and the Existence of Carbides of Phosphorus and Arsenic. Bulletin de la Socie´te´ Chimique de
France 1921, 29, 1071–1073.
83. Quin, L. A Guide to Organophosphorus Chemistry; Wiley-Blackwell: New York, 2000.
84. Veerasamy, V.; Amaratunga, G.; Davis, C.; Timbs, A.; Milne, W. I.; McKenzie, D. n-Type Doping of Highly Tetrahedral Diamond-Like Amorphous Carbon. J. Phys. Condens.
Matter 1993, 5, L169–L174.
85. Jones, D.; Stewart, A. Properties of Hydrogenated Amorphous Carbon Films and the Effects of Doping. Philos. Mag. Part B 1982, 46 (5), 423–434.
86. Helmbolda, A.; Hammera, P.; Thielea, J. U.; Rohwera, K.; Meissnera, D. Electrical Conductivity of Amorphous Hydrogenated Carbon. Philos. Mag. Part B 1995, 72 (3),
335–350.
87. Kuo, M.-T.; May, P.; Gunn, A.; Ashfold, M.; Wild, R. Studies of Phosphorus Doped Diamond-Like Carbon Films. Diamond Relat. Mater. 2000, 9, 1222–1227.
www.iran-mavad.com
88. Pearce, S.; May, P.; Wild, R.; Hallam, K.; Heard, P. Deposition and Properties of Amorphous Carbon Phosphide Films. Diamond Relat. Mater. 2002, 11 (3–6), 1041–1046.
89. Claeyssens, F.; Fuge, G.; Allan, N.; May, P.; Pearce, S.; Ashfold, M. Phosphorus Carbide Thin Films: Experiment and Theory. Appl. Phys. A Mater. Sci. Process.
2004, 79 (4–6), 1237–1241.
90. Liu, A.; Wu, H.; Zhu, J.; Han, J.; Niu, L. Evolution of Compressive Stress and Electrical Conductivity of Tetrahedral Amorphous Carbon Films with Phosphorus Incorporation.
Diamond Relat. Mater. 2008, 17 (11), 1927–1932.
91. Broitman, E.; Furlan, A.; Gueorguiev, G.; Czigány, Z.; Högberg, H.; Hultman, L. Structural and Mechanical Properties of CNx and CPx Thin Solid Films. Key Eng. Mater. 2001,
488–489 (9), 581–584.
92. Claeyssensa, F.; Allan, N.; May, P.; Ordejón, P.; Oliva, J. Solid Phosphorus Carbide? Chem. Commun. 2002, 2494–2495.
93. Claeyssensa, F.; Hart, J.; Allan, N.; Oliva, J. Solid Phases of Phosphorus Carbide: An Ab Initio Study. Phys. Rev. B Condens. Matter Mater. Phys. 2009, 79, 134115.
94. Gueorguiev, G.; Furlan, A.; Högberg, H.; Stafström, S.; Hultman, L. First-Principles Calculations on the Structural Evolution of Solid Fullerene-Like CPx. Chem. Phys. Lett. 2006,
426 (4–6), 374–379.
95. Seifert, G.; Heine, T.; Fowler, P. Inorganic Nanotubes and Fullerenes – Structure and Properties of Hypothetical Phosphorus Fullerenes. Eur. Phys. J. D 2001, 16 (1–3),
341–343.
96. Furlan, A.; Gueorguiev, G.; Czigany, Z.; Darakchieva, V.; Braun, S.; Hogberg, H.; Hultman, L. Structural and Mechanical Properties of Phosphorus-Carbide Thin Solid Films.
Thin Solid Films 2012.
97. Dumas, J.; Peligot, E. Nouvelles combinaisons du methylene In: Anales de Chimie et de Physique (Serie 2) 1836, Vol. 61, pp 193–200.
98. Moissan, F. Action du fluor sur les différentes variétés de carbone. Comptes rendus hebdomadaires des seánces de l’Acade´mie des sciences 1890, 110, 276–279.
99. Moissan, F. Sur la preparation et les propriétés du tétrafluorure de carbone. Comptes rendus hebdomadaires des seánces de l’Acade´mie des sciences 1890, 110, 951–954.
100. Lebeau, P.; Damien, A. Sur le tetrafluorure de carrbone. Comptes rendus de l’Acade´mie des sciences 1926, 182, 1340–1342.
101. Ruff, O.; Bretschneider, O.; Ebert, F. The Reaction Products of Various Forms of Carbon with Fluorine. II. Carbon Monofluoride. Zeitschrift für anorganische und allgemeine
Chemie 1934, 217, 1–19.
102. Plunkett, R. J. Tetrafluoroethylene Polymers. US Patent US 2230654, February 4, 1941.
103. DuPont, DuPont Teflon Brand [Online]. Available at: http://www2.dupont.com/Teflon/en_US/index.html (accessed May 13, 2012).
104. Logan, R. Production of Thin Polytetrafluoroethylene Resin (Teflon) Coatings by Electrodeposition Methods; Bureau of Naval Weapons, Department of the Navy, 1963.
105. Seyer, F.; Weidman, C. Razor Blades. US Patent US3203829, August 31, 1965.
106. Fusaro R. L.; Sliney, H. E. Preliminary Investigation of Graphite Fluoride (CFx)n as a Solid Lubricant; NASA TN D-5097, 1969.
107. Fusaro, R. L.; Sliney, H. E. Graphite Fluoride (CFx)n – A New Solid Lubricant. ASLE Trans. 1970, 13 (1), 55–65.
108. Harrop, R.; harrop, P. Friction of Sputtered PTFE Films. Thin Solid Films 1969, 3, 109–117.
109. Murakami, Y.; Shintani, T. Vacuum Deposition of Teflon-FEP. Thin Solid Films 1972, 9 (2), 301–304.
110. Millard, M.; Windle, J.; Pavlath, A. E. Plasma Synthesis of Fluorocarbon Films. J. Appl. Polym. Sci. 1973, 17 (8), 2501–2507.
111. Morrison, D.; Robertson, T. R.F. Sputtering of Plastics. Thin Solid Films 1973, 15 (1), 87–101.
112. Sirghi, L.; Ruiz, A.; Colpo, P.; Rossi, F. Atomic Force Microscopy Indentation of Fluorocarbon Thin Films Fabricated by Plasma Enhanced Chemical Deposition at Low Radio
Frequency Power. Thin Solid Films 2009, 517 (11), 3310–3314.
113. Theil, J. Fluorinated Amorphous Carbon films for Low Permittivity Interlevel Dielectrics. J. Vac. Sci. Technol. B 1999, 17 (6), 2397–2409.
114. Wienecke, M.; Bunescu, M.-C.; Pietrzak, M.; Deistung, K.; Fedtke, P. PTFE Membrane Electrodes with Increased Sensitivity for Gas Sensor Applications. Synth. Met. 2003,
138 (1–2), 165–171.
115. Lippert, T. UV Laser Ablation of Polymers: From Structuring to Thin Film Deposition. Springer Ser. Mater. Sci. 2010, 130, 141–175.
116. Kwong, H.; Wong, M.; Wong, Y.; Wong, K. Superhydrophobicity of Polytetrafluoroethylene Thin Film Fabricated by Pulsed Laser Deposition. Appl. Surf. Sci. 2007, 253 (22),
8841–8845.
117. Yun, Y.; Broitman, E.; Gellman, A. Oxidation of Fluorinated Amorphous Carbon (a-CFx) Films. Langmuir 2010, 26 (2), 908–914.
118. Iwamori, S.; Hasegawa, N.; Uemura, A. Fluorocarbon Polymer Thin Films Prepared by Three Different Types of R.F. Magnetron Sputtering Systems. Surf. Coat. Technol. 2008,
201 (1–2), 59–64.
119. Biederman, H. RF Sputtering of Polymers and Its Potential Application. Vacuum 2000, 59 (2–3), 594–599.
120. Gueorguiev, G.; Goyenola, C.; Schmidt, S.; Hultman, L. CF(x): A First-Principles Study of Structural Patterns Arising during Synthetic Growth. Chem. Phys. Lett. 2011, 516
(1–3), 62–67.
121. Goyenola, C.; Gueorguiev, G.; Stafström, S.; Hultman, L. Fullerene-Like CSx: A First-Principles Study of Synthetic Growth. Chem. Phys. Lett. 2011, 506 (1–3), 86–91.
122. Hellgren, N.; Johansson, M.; Broitman, E.; Hultman, L.; Sundgren, J.-E. Role of Nitrogen in the Formation of Hard and Elastic CNx Thin Films by Reactive Magnetron Sputtering.
Phys. Rev. B Condens. Matter Mater. Phys. 1999, 59 (7), 5162–5169.
123. Broitman, E.; Macdonald, W.; Hellgren, N.; Radnóczi, G.; Czigány, Z.; Wennerberg, A.; Jacobsson, M.; Hultman, L. Carbon Nitride Films on Orthopedic Substrates. Diamond
Relat. Mater. 2000, 9 (12), 1984–1991.
www.iran-mavad.com
4.16 Photocatalytic Coatings
P Pichat, CNRS/Ecole Centrale de Lyon (STMS), Ecully CEDEX, France

4.16.2 Fundamentals of Photocatalysis 414
4.16.2.1 What Is the Basic Phenomenon in Photocatalysis? 414
4.16.2.2 Reactions Involving Directly or Indirectly the Photoproduced Charges 415
4.16.3 Generalities on Photocatalytic Coatings 416
4.16.4 Peculiarities of Photocatalytic Coatings Depending on the Application 418
4.16.4.1 Self-Cleaning Materials 418
4.16.4.2 Air Purification 420
4.16.4.3 Water Purification 420
4.16.5 Peculiarities of Photocatalytic Coatings Depending on the Shape of the Materials 420
4.16.5.1 Flat Materials 420
4.16.5.2 Materials in the Shape of Beads, Rings, and Small Tubes 420
4.16.5.3 Materials in the Shape of Honeycomb Monoliths 421
4.16.5.4 Materials in the Shape of Solid Foams 421
4.16.5.5 Materials in the Shape of Nonwoven Tissues 421
4.16.6 Conclusions 422
References 422
4.16.1 Introduction
Photocatalysis using semiconductors, principally TiO2, is an important research field whose basic phenomena and main results
have been reported in recent reviews (see, e.g., Refs. (1–6)). By contrast, its commercialized applications are still not widespread.
The most developed application is self-cleaning materials (7). Indoor air purifiers based on photocatalysis are also commer-
cialized. However, there is concern about the long-term efficacy of these purifiers because of photocatalyst deactivation that has
been attributed to the photocatalytic formation of amorphous silica from siloxanes often present in indoor air (cf. Figure 1) (8).
Appropriate control of the intermediate products (e.g., toxic low molar mass carbonyls) of the photocatalytic degradation of
initial pollutants also seems to be often ignored (9). Photocatalytic water purification (10) by use of ultraviolet (UV) lamps or
solar light has given rise to various devices mostly, however, at demonstration sites. Other potential applications of photo-
catalysis, that is, organic synthesis, synthesis of fuels, and water splitting, are still at the research stage and therefore are not
considered in this chapter.
Materials such as glass, metals, and polymers can be made self-cleaning by a coating whose composition includes a photo-
catalyst. A photocatalyst can also be incorporated in the top layers of concrete or in paint or mortar used to cover building facades,
Figure 1 The extent of deactivation, induced on a TiO2 (Degussa/Evonik P25) wash-coated microscope slide by exposure to tetramethylsilane (about
1 ppmv) and UV-C irradiation for 0.5–24 h, plotted against the log of the ratio of percent Si to percent Ti as determined by X-ray photoelectron
spectroscopy analysis of the coating. The photocatalytic efficacy was measured for the removal of propanal (100 ppmv) in an air flow. Its decrease was
attributed to amorphous silica formed by photocatalytic oxidation of tetramethylsilane and deposited on TiO2, which gradually hindered access of
molecules to TiO2. Reproduced with permission from Hay, S. O.; Obee, T. N.; Thibaud-Erkey, C. Appl. Catal. B 2010, 99, 435–441.

www.iran-mavad.com
414 Photocatalytic Coatings
which thereby become self-cleaning. As fluidized-bed reactors have not been successful for photocatalytic air purification, the
photocatalyst is coated on materials of various nature and shapes for this application. Coating the photocatalyst on diverse
materials is also most usual for water purification. However, Purifics (www.purifics.com) uses powder TiO2, which is periodically
recovered by use of a ceramic membrane. In the case of the big solar units at demonstration sites in Spain, TiO2 is also suspended in
the water to be treated (11). These cases call attention to the possibility that coating may decrease too much the photocatalytic
efficacy of TiO2.
Nonetheless, most of the current applications of photocatalysis involve coatings. In this chapter, on the basis of an overview of
the fundamentals of photocatalysis, the similarities and specificities of photocatalytic coatings compared with other types of
coatings are presented and discussed. The references are meant to illustrate these points; because of the limitation in their number,
they cannot be exhaustive and apologies are given in advance to the authors whose papers are not cited.
4.16.2 Fundamentals of Photocatalysis

4.16.2.1 What Is the Basic Phenomenon in Photocatalysis?
The term photocatalysis may designate several phenomena that involve photons and a catalyst. Here, the terms heterogeneous
photocatalysis and photocatalysis refer only to cases where the photosensitizer is a semiconductor, which accordingly is at the origin
of the chemical reactions. Moreover, only TiO2 is generally mentioned among the semiconductors because it is by far the photo-
catalyst that is most often used to coat (or be incorporated into) the materials considered in this chapter. The general concepts are
the same for other n-type semiconducting oxides.
Electrons pertaining to an isolated atom occupy discrete energy levels. In a crystal, each of these energy levels is split into as
many energy levels as there are atoms. Consequently, the resulting energy levels are very close and can be regarded as forming
a continuous band of energies. In semiconductors, valence electrons completely fill a band thus called the valence band,
whereas the next higher energy band, termed the conduction band, is empty, at least at 0 K. In a perfect crystal, the energy band
separating the highest level of the valence band from the lowest level of the conduction band is forbidden. Its width is referred
to as the band gap.
Excitation by photons can result in the promotion of electrons from the valence band to the conduction band, provided the
photons have an energy at least equal to that of the band gap (Figure 2). This photoexcitation leaves electron vacancies, also called
holes, in the valence band. For TiO2, photons with wavelengths shorter than about 400 nm (the exact value varying according to the
allotropic form and other characteristics) are required. That corresponds to energies higher than approximately 3.1 eV or
300 kJ mol1. This latter value can easily be compared with the strengths of the bonds that are broken in the course of photocatalytic
reactions. In other words, in heterogeneous photocatalysis, the input of energy is supplied by photons instead of a rise in
temperature as in heterogeneous catalysis.
conduction band e-
e- + O2 O2 -
e--h+
recombination other electron acceptors
(organics, cations)
surface states/imperfections
energy
impurity photons
defect
levels
e--h+pair
formation
electron donors
(organics, H2O/OH-, O2- ,
(organics
etc.)
valence band h+
semiconductor air or water
Figure 2 Schematic band diagram showing the charge transfers corresponding to (1) the photoexcitation of an n-type semiconductor such as TiO2, (2)
the direct and indirect (through bulk and surface energy levels) recombination of the resulting out-of-equilibrium electron–hole pairs, and (3) the
interfacial reactions with electron acceptors or donors. Back charge transfers from radical ions to the semiconductor, e.g., from O2– possibly forming
1
O2, are not shown. Reproduced from Hay, S. O.; Obee, T. N.; Thibaud-Erkey, C. Appl. Catal. B 2010, 99, 435–441.
www.iran-mavad.com
Photocatalytic Coatings 415
4.16.2.2 Reactions Involving Directly or Indirectly the Photoproduced Charges

The conduction band electron (e–CB) can reduce some organic compounds, such as fully halogenated alkanes (e.g., CCl4) and, in
liquid water, cations. The condition is that the redox potentials of these compounds and cations lie below the conduction band
energy.
Above all, e–CB generates the superoxide radical-anion, O2– (Figure 2), from molecular oxygen in the case of n-type semi-
conductors, like TiO2, whose reducing potential of the conduction band permits it. The superoxide radical-anion is poorly active per
se (12). However, it is believed to lead to the formation of OH radicals through further reduction:
O2 þ Hþ / HO2 ; HO2 þ Hþ þ e
CB / H2 O2 ; H2 O2 þ eCB / OH

þ OH;
and additionally, reaction with the H2O2 produced by the second reaction above or by disproportionation
(2O2– þ 2Hþ / H2O2 þ O2):
H2 O2 þ O2 / OH þ OH þ O2 :
The superoxide radical-anion could also react with an organic radical-cation (1,13,14), which itself can be issued from the
reaction of valence band holes (hþ
VB) with an organic compound, P:
P þ hþ þ
VB / P ; P
þ þ O / PO ðe:g:; dioxetaneÞ / oxidized products:
2 2
Indeed, many organic compounds have a redox potential at a higher energy than the valence band edge of TiO2 and other common
semiconductor oxides, thereby allowing these latter reactions to occur, at least from the thermodynamic viewpoint.
The valence band hole (hþ
VB) is also supposed to produce OH radicals via
H2 O þ hþ
VB / H
þ
þ OH; OH þ hþ
VB / OH

where H2O and OH would behave as nucleophilic agents. Note that any photocatalytic coating or film exposed to ambient air is
covered by layers of adsorbed water. However, the existence of these reactions has been questioned because the redox potential of
adsorbed water molecules has been calculated to be at a lower energy level than the top of the TiO2 valence band (15). The
formation of OH radicals via hþ
VB would involve surface oxygen anions – in agreement with the lability of these anions illustrated by
oxygen isotope exchange between TiO2 and gaseous O2 (15,16) – according to the equations
O2 þ hþ
VB / O ; O
þ Hþ / OH
which are equivalent to the overall equation

O2 þ hþ þ
VB þ H / OH:

The OH radical is a very powerful oxidizing agent according to its redox potential, even though its reactivity on surfaces could
differ somewhat from that measured in a fluid phase. Moreover, it is not a nonselective species as written sometimes: in liquid water,
the OH radical reaction rate varies by a factor of about 1000, depending on the molecular structure of the organic compound with
which it reacts. For the photocatalytic transformations of aliphatics, the following reactions have been proposed on the basis of
radiochemistry data (17):
RH þ OH / R þ H2 O; R þ O2 / RO2 ; RO2 þ RH / ROOH þ R ; ROOH / RO þ OH
RCH2 O þ O2 / RCHO þ HO2 ; RCH2 O þ RH / RCH2 OH þ R :

For benzene derivatives, an OH radical would first be added to the ring:
RC6 H5 þ OH / RC6 H5 OH; RC6 H5 OH þ O2 / RC6 H4 OH þ HO2 :
On the other hand, benzene derivatives with electron-donating substituents could also yield cyclohexadienyl radicals, RC6H5OH,
via direct attack by hþ
VB:
RC6 H5 þ hþ þ
VB / RC6 H5 ; RC6 H5
þ þ H O / RC H OH þ Hþ :
2 6 5
Very acidic Rþ could also deprotonate spontaneously and lead to the formation of other products:
R þ / R þ Hþ :
Note that the HO2 radical would often be the leaving species and O2 and H2O are involved as reactants in some of these radical-
based reactions.
Finally, the formation of singlet molecular oxygen, 1O2, on UV-irradiated TiO2 has also been proposed. It would arise from the
reaction of superoxide with hþVB – which is equivalent to back electron transfer from O2 to TiO2. If it does not react with an organic
compound, 1O2 would subsequently de-excite to O2 with an energy release of about 1 eV (or w96.5 kJ mol1) (18,19):
O2 þ hþ 1
VB / O2 / O2 þ energy:
In other words, the photon energy input would be lost from the chemical viewpoint. It is equivalent to the recombination of the
photoproduced charges (vide infra).
www.iran-mavad.com
Possible evidence for 1O2 was drawn from electron spin resonance signals (20–22) for TiO2 in organic solvents or water, with
the usual debate about the attribution of the signals. In the case of an aqueous TiO2 suspension, a very weak phosphorescence at
1270 nm – whose intensity varied as the concentration in superoxide as expected – was attributed to 1O2 (23,24). Furthermore,
the involvement of 1O2 in photocatalytic oxidations was deduced from the acceleration of the phosphorescence decay in the
presence of pyrrole or amino acids or proteins, such as methionine ((2S)-2-amino-4-methylsulfanylbutanoic acid), folic acid,
and collagen (24).
However, regarding the significance of 1O2 in photocatalysis several facts must be considered. First, many of the molecules that
are easily oxidized photocatalytically are not expected to readily react with 1O2. Second, addition of azide ions (25–27), which can
scavenge 1O2, did not affect markedly the photocatalytic oxidations studied, but this scavenger is not selective so that comparisons
with the effect of other scavengers were necessary, with questionable conclusions because relative adsorption of the scavengers
should also be taken into account. Third, the expected increase in 1O2 lifetime on replacing H2O by D2O did not change as much
as expected the photocatalytic degradation rate at least in the case of geosmin, but the isotope effect varied also with the sample of
TiO2 (28). Nevertheless, in spite of some doubts about these arguments, it does not appear that the role of 1O2 in photocatalytic
oxidations may, in general, be major compared with the roles of direct hþ
VB transfer and OH radicals. Anyhow, regarding the
coating technique and coating composition, the aim is to ensure accessibility of O2 to the photocatalyst whatever the pathways
involved.
In summary, the complex chemistry resulting from photoexcitation of a semiconductor shows how the compounds adsorbed
and deposited on the photocatalytic coating can be transformed by reduction or most often oxidation processes. The ‘low’ trans-
formation rates observed can partially be explained by the very low concentrations in pollutants in the cases of purification of air
and water. Back electron transfers can also occur. Radicals can combine, although this process should be very minor because of their
extremely low concentrations. Above all, the out-of-equilibrium e–CB–hþ VB pair can recombine rapidly through radiative and non-
radiative processes. This recombination can occur by either direct electron transfer from the conduction band to the valence band or
most often indirect electron transfer via energy levels located within the band gap (Figure 2). These energy levels correspond to
crystal imperfections due to unavoidable impurities and lattice defects (1–7). This is an inherent and important drawback. It causes
a significant loss of the energy supplied by photon irradiation.
Especially in the case of TiO2, numerous investigations have been performed to increase the photocatalytic efficacy. It has been
found that this efficacy depends on adjustment between the crystal imperfections – which have a detrimental effect on the
recombination rate of the photoproduced charges – the surface area – which in principle allows one to augment the adsorbed
amounts of pollutants and is related to the size of the elementary particles and of the agglomerates – and the ratios of the TiO2
allotropic forms. More sophisticated methods have been extensively explored to increase the efficacy, e.g., the deposition of an
electron-attracting metal and the doping by one or several foreign elements (1–7); until now, none of these modified materials have
been estimated to be economically viable, at least for the most common applications of TiO2 photocatalysis, viz. self-cleaning
materials and air and water purification.
The conclusion to this section is that, for an irradiation with given characteristics, the overall photocatalytic efficacy depends on
both the absorption of useful photons by the coating containing the photocatalyst and the utilization of the resulting e–CB–hþ VB pairs
to effect the photocatalytic transformations. Consequently, any coating should maximize the absorption, and should not impair too
much the use of the photoproduced charges through, for instance, inappropriate thermal treatment and/or addition of unsuitable
components in the coating.
4.16.3 Generalities on Photocatalytic Coatings
Many coating techniques have been utilized to obtain photocatalytic coatings, particularly at the laboratory scale. The choice of the
technique depends, inter alia, on the photocatalytic application and the shape of the material to be coated as is indicated in the
following sections. These techniques include those in which the coating composition containing the photocatalyst is deposited by
dipping, spinning, spraying, painting, flow coating, a doctor blade device, etc. The photocatalyst, most often TiO2, can be in the
form of commercialized particles or it can be prepared by sol–gel methods (29) from a variety of precursors among which titanium
tetraisopropoxide is the most frequently used. Particles and precursors can also be mixed as in the case of the self-cleaning coating of
Figure 3. The numerous variables of the sol–gel methods can be adjusted to change the characteristics of the coating, its adhesion,
and other properties. Experimental design has sometimes been employed for that. A very viscous solvent and/or a substance capable
of creating bonds with the TiO2 precursor have been used to slow TiO2 crystallization and thus to decrease the particles growth rate
in efforts to increase the photocatalytic efficacy (30). Physical vapor deposition, such as thermal evaporation and sputtering
(31–34), and chemical vapor deposition, sometimes plasma assisted (35), have also been employed to prepare photocatalytic
coatings. To obtain a TiO2-containing composite coating, the vaporized Ti can be mixed with another element, e.g., Si. The effects of
the various parameters of these deposition techniques on the photocatalytic efficacy have been investigated.
The coating techniques are not described here since other chapters in this book are devoted to them.
Many of the properties required for photocatalytic coatings do not differ from those required for most other types of coatings.
These similar properties include acceptable cost, durability, and environmental impact (notion of life cycle assessment) according to
each of the applications. It is also self-evident that the coating technique must produce a uniform and reproducible covering of the
supporting material.
www.iran-mavad.com
Figure 3 Scanning electron micrograph of a coating on glass. The coating was obtained by dipping the glass plate, previously covered by a thin SiO2
intermediate layer (appearing as a stratum slightly clearer than the glass in the photograph), in a solution containing porous TiO2 nanoparticles
(>250 m2 g1) and alkoxides of Ti and Si dissolved in ethanol þ 1,2-ethanediol, destined to form a binder. The coated plate was then annealed at
a temperature higher than 623 K (patents WO 97 10185 and 10186). The coating was roughly 45 nm thick and the TiO2 crystalline structure was anatase.
Reproduced with permission from Romeas, V.; Pichat, P.; Guillard, C.; Chopin, T.; Lehaut, C. New J. Chem. 1999, 23, 365.
Other requirements are specific for photocatalytic coatings. First, the photocatalytic efficacy must obviously be as high as
possible. That implies that the photocatalyst must be properly irradiated by the photons needed to move electrons through the
semiconductor band gap (see Section 4.16.2; Figure 2). In other words, the photocatalyst must not be completely embedded by the
other components in the coating that would be susceptible of absorbing these photons. An acceptable photocatalytic efficacy also
implies that the coated photocatalyst remains accessible to O2 and the compounds to be chemically transformed. Finally, the
coating process must preserve as much as possible the initial activity of the photocatalyst particles used or must produce, from
precursors, a photocatalytic phase as active as possible. For that, the choices of the photocatalyst particles and/or the precursor(s),
the thermal treatment if any, and the binder if any, play crucial roles.
In particular, if a thermal treatment is used for better adhesion (vide infra), the support must not release into TiO2 chemical
elements that can decrease the photocatalytic efficacy. For instance, detrimental migrations have been observed for sodium from
glass (36,37), a frequently utilized support. This phenomenon leads one to first cover the support with a layer that cannot be
a source of migrating chemical elements (Figure 4), e.g., for glass, the use of fused silica is common and that of silicon nitride (38)
has been proposed as a better barrier for preventing migration of sodium.
The second requirement specific for photocatalytic coatings corresponds to the necessity of protecting the support and the other
components of the coating if they are photocatalytically degradable. To that end, intermediate layer(s) between the support and the
coating (Figure 4) and special components are used (39). In particular, all polymers are photocatalytically degradable aside from
those formed of C–F bonds only. Consequently, a TiO2-containing coating cannot be applied directly on polymers, except in
extremely rare cases where the replacement of the material before the loss of properties becomes intolerable is affordable and
compatible with the application. The intermediate layer must have sufficient affinity for both the polymer and the photocatalytic
inorganic coating to increase adhesion, while satisfying the other properties required. For instance, the surface of polyethylene
terephthalate was made positive by a layer of another polymer to fix a silica sol, which was subsequently coated with TiO2 (40).
Multiple layers progressing from organic to inorganic character have been proposed (41). The coating process is hence more
complex and presumably more costly. Intermediate layers may, however, be avoided by the use of a TiO2-containing coating
photocatalytic layer
intermediate layer(s) if required
support
Figure 4 Scheme showing the optional presence of intermediate layer(s) between the photocatalytic coating and the support for either (1) preventing
the release of chemical elements from the support into the photocatalytic layer especially if a thermal treatment is necessary during the coating process,
or (2) protecting the material against photocatalytic damage, or (3) providing a surface enabling better adhesion of the photocatalytic coating.
www.iran-mavad.com
including poly(dimethylsiloxane) (PDMS), which can both protect the polymeric support from photocatalytic damage and enhance
adhesion (42). Not surprisingly, the addition of PDMS has, however, an unfavorable effect on the photocatalytic efficacy.
For photocatalytic coatings as for other coatings, adhesion is important for the durability. It is also essential to avoid or minimize
the release of particles that can be a risk for health (43). In fact, both characteristics are related. To achieve acceptable adhesion
without a binder, a thermal treatment is commonly employed. The heating characteristics should be carefully chosen to get an
acceptable trade-off between adhesion and photocatalytic efficacy depending on the application. A pretreatment of the support can
also increase the coating adhesion; for example, a pretreatment of aluminum with phosphates has been proposed (44). The use of
a coating mixture containing both very photocatalytically active TiO2 particles and less active TiO2 particles, which are more able to
adhere strongly on the support, has also been advocated (44).
Among the binders, those that cannot be degraded by photocatalysis are of course favored. Silica, introduced as particles,
colloids, or a precursor, has often been used. Organosilane polymers or purely organic adhesives have also been tried (44). Then, the
subsequent thermal treatment should enable elimination of the organics to obtain a stable coating.
Unfortunately, relevant adhesion tests have not been systematically reported for photocatalytic coatings. Moreover, given the
range of applications and support shapes (vide infra), it is difficult to foresee from the laboratory tests the durability of a given
coating in the field.
4.16.4 Peculiarities of Photocatalytic Coatings Depending on the Application

4.16.4.1 Self-Cleaning Materials
The photocatalytic efficacy of coatings used for self-cleaning materials, although obviously necessary, is not a criterion as essential as
in the case of coatings designed for air or water purification. Clearly, the ‘low’ degradation rates resulting mainly from the
recombination of photoproduced charges (see Section 4.16.2; Figure 2) are more acceptable for self-cleaning, considering the time
scale admissible to obtain the desired result. Moreover, the compounds to be degraded are on the self-cleaning material, whereas
they have to be transferred from the fluid phase to the photocatalytic coating for air or water purification.
By contrast, the durability of the coating – which often involves characteristics opposed to those increasing the photocatalytic
efficacy – is extremely important. For buildings, minimal warranties of 10 years are usual. That puts the emphasis on the coating
adhesion and chemical stability for self-cleaning glasses, polymers, and metals, and on the chemical stability for paints and
cementitious materials.
Also, light-induced hydrophilicity is essential for self-cleaning windows and external mirrors potentially used in automobiles. It
contributes to keep a material visually clean because of the spreading of atmospheric water over the coating. It allows water-soluble
dirt to be flushed away; this is the case for carbonates, nitrates, and sulfates, which are not photocatalytically transformed. For that, the
natural flow rate of water must be higher than the water evaporation rate, which frequently occurs on the vertical surfaces of buildings.
The possible origins of TiO2 light-induced hydrophilicity have been discussed in great detail (3). The most straightforward
interpretation relies on the fact that organic compounds, especially hydrophobic hydrocarbons, adsorbed on TiO2 from ambient air,
hinders the water droplets to spread on the surface (45). Photocatalytic removal of these compounds makes the coating hydrophilic
as illustrated by Figure 5 in the case of n-hexane as a model hydrocarbon. This interpretation is in agreement with the existence of an
induction period to achieve maximum hydrophilicity when self-cleaning materials kept in darkness are exposed to sunlight. This is
also consistent with the need for irradiation with wavelengths at least equal to the photocatalyst band gap to decrease the water
contact angle. Nevertheless, the decreased rate of the water contact angle observed below about 10 was found independent of the
wavelength in the 320–380 nm range at equal irradiance (46); that should not occur if the hydrophilicity and the photocatalytic
cleaning of the TiO2 surface corresponded exactly. Finally, this interpretation based on the presence or absence of an organic film is
also in accordance with the progressive reversibility of hydrophilicity when the material is exposed in the dark to ambient air that
inevitably contains organic compounds prone to be deposited on the material making its surface hydrophobic.
Another interpretation of light-induced hydrophilicity is that water forms films instead of droplets because water molecules are
hydrogen bound to the TiO2 surface OH groups (3). Accordingly, UV irradiation should increase the number of these groups,
whereas darkness would decrease it. However, measurements have shown that UV irradiation of a TiO2 powder exposed to O2 does
not affect the vibrational bands of Ti–OH groups and adsorbed water (45) contrary to earlier measurements (3).
The presence of silica in the coating can improve not only the light-induced hydrophilicity, but also the coating adhesion
(47–50). Unfortunately, silica can decrease the photocatalytic degradation rates because it (1) can shield, at least partially, TiO2
from the reactants and (2) can reduce the mobility of photoproduced charges as a result of both its insulating character and the
possible existence of Si–O–Ti bonds (Figure 6) (51).
For self-cleaning glass, polymers, metals, and paints, the optical properties (such as transmittance, reflectance, gloss, etc.) are
most often a requirement. In this respect, the uniform thickness and homogeneity of the coating are crucial. Keeping the visual
aspect as constant as possible for long periods is often a requisite. For window panes, the loss in transmittance over time must
obviously be minimized. In the case of self-cleaning glass, silica offers the advantage of decreasing the refractive index of the coating
(48). The coating thickness principally depends on the specific optical properties that are required for each material. It ranges from
0.05 mm to a few mm. Given the penetration depth of UV irradiation in TiO2, this thickness range allows total or very significant use
of the UV spectral region of solar light, provided that the useful photons are not substantially and competitively absorbed by the
other components of the coating.
www.iran-mavad.com
Figure 5 The water contact angle evolution during UV irradiation of a rutile (110) single crystal surface in an oxygen atmosphere with n-hexane at the
concentrations indicated shows that the photocatalytic removal of n-hexane is critical to reach complete wetting. Reproduced with permission from
Zubkov, T.; Stahl, D.; Thompson, T. L.; Panayotov, D.; Diwald, O.; Yates, J. T., Jr. J. Phys. Chem. B 2005, 109, 15454.
Additionally, the use of intermediate layer(s) (Figure 4) and components that are not photocatalytically degradable has often to
be taken into account because the goal is making a given material self-cleaning. By contrast, the choice of the supporting material
exists, at least to some extent, for air or water purification.
Concrete and mortar correspond to particular cases of self-cleaning materials because the photocatalyst is either mixed with the
whole material or incorporated into its top layers (52,5). In the case of coarse concrete, a layer made of finer aggregates can be
ln rPA (nm h-1)

0 1 2 3 4 5 6 7 8 9
3
Cleaning efficacy
2.5
1.5
ln D
0.5
-0.5 O2 photosorption capacity

-1
0 1 2 3 4 5 6 7 8 9
ln Psat (Pa)
Figure 6 Log–log plots showing that the relative density of Si–O–Ti bonds with respect to the Si–O–Si bonds (as determined by Fourier transform
infrared spectroscopy-attenuated total reflectance spectrometry) in the coating decreased both the cleaning efficacy and the photoadsorption
capacity for O2. This illustrates both the unfavorable effect of the SiO2 binder and the importance of O2 accessibility to TiO2. The cleaning efficacy
was deduced from measurements of the ratio of the diffuse transmission to the total transmission of the glass onto which hexadecanoic acid (or palmitic
acid ¼ PA) was initially sprayed. The photoadsorption capacity for O2 was derived from the UV irradiation-induced temporal variations in O2
pressure for various initial O2 pressures. The coating was obtained as indicated in the caption to Figure 3. The weight percent of TiO2 in the binder
relative to SiO2 varied from 14 to 75 for the four samples studied, the coating mass being comprised between 0.17 and 0.20 g m2 of glass. Adapted
from Puzenat, E.; Pichat, P. J. Photochem. Photobiol. A: Chem. 2003, 160, 127.
www.iran-mavad.com
applied before spreading the photocatalyst-containing covering. Also, a TiO2-containing cementitious layer with a thickness in the
1–10 mm range can be applied with a roller or a trowel on outside walls. So, strictly speaking, it is not a coating but rather
a covering. The major fraction of incorporated TiO2 is not activated by solar light.
The self-cleaning effect depends on the nature, amounts, and relative grain sizes of the photocatalytic additive and of the cement
constituents. Trials – possibly using experimental design – have to be carried out. About 3 wt% of TiO2 has been reported to be
sufficient to get a satisfactory effect (52,53). Beyond this value, the increased effect may not be worth considering, because of the
relative prices of TiO2 and cement.
Self-cleaning paint is another case where the photocatalytic additive is mixed with the other components and not coated.
Obviously, these components must not be photocatalytically degradable. This restricts the choice to polysiloxanes and silicates;
completely fluorinated organic components can also be used, but that leads to a higher cost (39). As in ordinary paints, TiO2
rendered nonphotocatalytic, that is, pigmentary TiO2, can be one of the inert components in order to enhance whitening (39).
As for cement-based materials, the characteristics and amount of TiO2 must be determined by trials, with the difference that it
is not the cost of TiO2 that usually limits its amount in the paint. The thickness of the applied paint layer(s) is on the order of
0.1 mm.
4.16.4.2 Air Purification

Photocatalytic coatings destined to purify indoor air by use of active devices placed in rooms or in air conditioning ducts must
ensure adsorption and chemical transformation of pollutants as complete as possible for each air pass through the device. On the
other hand, given the volumes of air to be treated relative to the acceptable dimensions of the devices, high flow rates of air are
necessary to achieve satisfactory purification within a suitable time. Three requirements ensue for the photocatalyst, its support, and
the coating technique. First, the photocatalyst surface area must be as large as possible. Second, the photocatalyst support must
permit pressure drops as low as possible; that leads to the use of honeycomb shapes, various types of foams, and interconnected
fibers forming nonwoven tissues (vide infra). Third, the coating must not release particles – possibly, after a conditioning period
following the coating process – in spite of the high flow rates.
In practice, the use of honeycombs or foams leads to a restricted choice of coating methods. The trade-off between the need of
a coating ensuring significant adsorption of pollutants and the need of minimizing the detachment of particles from the coating
implies to adjust the coating thickness and to make appropriate use of a thermal treatment and/or of a binder. Experimental design
is recommended for reducing the number of trials. High surface area of the photocatalyst should be achieved by increasing the
porosity rather than by using smaller particles which are, in principle, easier to be detached from the coating and also potentially
more harmful for the respiratory tract (43).
Walls coated with a photocatalytic layer can also be used to purify indoor air. Double walls have been proposed to permit an
efficacious irradiation of the photocatalytically active wall by UV lamps to which the persons cannot directly be exposed (54).
4.16.4.3 Water Purification

Photocatalytic coatings for water treatment present only a few specificities. The main one refers to the use of photocatalysis – as the
single process or in combination with other processes – to deliver potable water. Because of the health concern about nano- and
even microparticles (43), it is clear that absolutely no release of particles from the coating is permitted. On the other hand, adding
activated carbon to TiO2 has been shown to increase the purifying efficacy (55). This increase is due to the fact that activated carbon,
which is hydrophobic and organophilic, attracts trace organic pollutants, which then could be transferred to TiO2 at the interfaces
between the two solids. Alternatively, a transfer of active species from TiO2 to the pollutants adsorbed on activated carbon has been
postulated. The mixture of activated carbon with TiO2 makes it more difficult to obtain a uniform coating, which both maximizes
the area of interfaces between the two solids and prevents the release of particles into water during the purification process.
4.16.5 Peculiarities of Photocatalytic Coatings Depending on the Shape of the Materials

4.16.5.1 Flat Materials
In the case of glass and metals with a flat and smooth surface, any coating technique can be employed, in principle. The choice
depends on the manufacturer. Clearly, the cost, the reproducibility, and the durability of the coating play essential roles. Obviously,
the use of the coated materials and the size of the coated pieces also intervene; consider, for instance, the difference between float
glass, which will be utilized for self-cleaning window panes, and glass beads for water treatment. For polymers, the choice of the
coating technique is restricted by the material not the shape, because no thermal treatment can of course be involved as
aforementioned.
4.16.5.2 Materials in the Shape of Beads, Rings, and Small Tubes

These kinds of materials are generally used for water treatment and made of glass or polymers. Intuitively, immersion in the coating
mixture is the best choice. Therefore, dip-coating is the coating technique employed most often. If these materials are allowed to
www.iran-mavad.com
float in the water to be treated (56), coating attrition may result from repeated contact between them, emphasizing once again that
adhesion is vital.
4.16.5.3 Materials in the Shape of Honeycomb Monoliths

As aforementioned, honeycomb monoliths are used for indoor air purification. They can be made of metals, ceramics, or polymers.
By analogy with the coating technique of monoliths located in automobile vehicle exhausts, a wash coat containing TiO2 (57) has
been utilized for metallic monoliths introduced into air ducts in various buildings for demonstration. At the laboratory scale,
metallic monoliths were dipped into a previously sonicated aqueous suspension of TiO2. To avoid thicker deposition of TiO2 in the
lower part of the honeycomb cells, a jet of air was forced through the honeycomb alternately from both sides. After drying at room
temperature, the whole deposition process was repeated the number of times necessary to reach the coating optimal thickness for
the photocatalytic efficacy. As no binder and no thermal treatment were used, TiO2 losses were initially observed during the
photocatalytic tests at an air velocity of 0.34 m s1; however, these losses considerably diminished over time (58). In another case,
to solve the problem of TiO2 losses in flowing air, silicates, used as binders, and a precursor of TiO2 were mixed into the ceramic
paste before the extrusion of the monoliths (59). Monoliths made of epoxy resin were obtained in various shapes by stereo-
lithography using a UV laser. They were spin coated with a TiO2 suspension with the purpose of determining the influence of the
shapes on a gas-phase photocatalytic reaction (60).
4.16.5.4 Materials in the Shape of Solid Foams

Like monoliths, they can be made of metals, ceramics, or polymers. More recently, foams in b-SiC have been prepared starting from
polyurethane foams in diverse shapes, which are then transformed using a Si-containing material and subsequent heating in an
argon atmosphere, the shape being preserved (61). Dipping is most often employed to coat foams. The synthesis of a TiC-b-SiC
foam whose thermal oxidation produces a TiO2-b-SiC foam has been reported. The photocatalytic activity was low compared with
that of the TiO2–coated b-SiC foam (62). Indeed, it is usual that oxidation of Ti produces TiO2 with low surface area and crystallinity
that cannot be easily adjusted. But the advantage is that the release of TiO2 particles, which, as aforementioned, can be a health
concern (43), does not occur.
4.16.5.5 Materials in the Shape of Nonwoven Tissues

The use of these types of materials was investigated for both air (58,63–65) and water (66,67) purification. These thin, flexible
tissues are made of interconnected glass fibers (Figure 7). The coating procedure includes a very simple stage of impregnation with
a mixture containing TiO2 and an aqueous colloidal suspension of SiO2 using a ‘sizepress.’ The impregnated tissue is then allowed
to dry at room temperature (68). SiO2 is used to better affix TiO2. However, the adhesion is not completely satisfactory, especially at
Figure 7 Transmission electron micrograph (courtesy of Ahlstrom Research) of a thin nonwoven tissue (fiber dimensions: 8 mm 11 mm)
impregnated by a mixture containing TiO2 nanoparticles and an aqueous colloidal suspension of SiO2 (TiO2/SiO2 mass ratio: 1) using of a ‘size press’ and
allowed to dry at room temperature. The coating mass was 80 g m2 of tissue whose mass was 50–60 g m2 before coating. The photograph
shows that not only the fibers are coated but also the coating forms sheets between the fibers.
www.iran-mavad.com
the beginning of the use in flowing air (58) or water. Moreover, the SiO2 particles very significantly embed the TiO2 particles
(65–67), thereby lowering the photocatalytic activity.
4.16.6 Conclusions
In this chapter, in addition to an overview of the domain of heterogeneous photocatalysis to make the reader aware of the particular
requisites to obtain appropriate photocatalytic coatings, the aspects that depend on the photocatalytic application and on the
nature, shape, and size of the material to be coated have been presented. Apart from difficulties common to many coatings, the main
challenges for photocatalytic coatings are (1) the need for protective intermediate layers in some cases, which increases the
complexity of the coating technique and (2) the trade-off between the efficacy and the durability. Depending upon the case, one of
these two latter properties can be more important, but the other one cannot be ignored. In general, the use of a powder photo-
catalyst to produce the coating may lead to more efficient but less durable coatings, whereas it may be the opposite if precursors of
the photocatalyst are utilized. However, numerous factors can be adjusted, so that this rule-of-thumb conclusion is not always valid.
Aside from some issues about the coating procedures discussed in this chapter, no general guidance can be provided, unfortunately.
One of the reasons is that the relatively rare comparisons were made with coated materials prepared only at the laboratory scale.
Furthermore, most often only the efficacy was thoroughly measured, while the durability was at best roughly evaluated and anyhow
field measurements are required. Even the reported efficacy ranking between various coatings can often be questioned as it depends,
inter alia, on the probe compound (several must be used) and the test conditions. In practice, on the basis of both the abundant
information about photocatalytic coatings available in papers and patents and one’s knowledge of the material, each manufacturer
has to adapt the coating procedure.
References
1. Pichat, P. Photocatalytic Degradation of Pollutants in Water and Air: Basic Concepts and Applications. In Chemical Degradation Methods for Wastes and Pollutants: Envi-
ronmental and Industrial Applications; Tarr, M. A., Ed.; Marcel Dekker, Inc.: New York, Basel, 2003; pp 77–119.
2. Agrios, A. G.; Pichat, P. J. Appl. Electrochem. 2005, 35, 655–663.
3. Fujishima, A.; Zhang, X.; Tryk, D. A. Surf. Sci. Rep. 2008, 63, 515.
4. Zhang, H.; Chen, G.; Bahnemann, D. W. J. Mater. Chem. 2009, 19, 5089–5121.
5. Bak, T.; Norby, T.; Nowotny, J.; Nowotny, N. K.; Sucher, N. Solid State Phenom. 2010, 162, 77.
6. Henderson, M. A. Surf. Sci. Rep. 2011, 66, 185.
7. Pichat, P. Self-cleaning materials based on solar photocatalysis. In New and Future Developments in Catalysis; Suib, I., Ed.; Solar Catalysis, Elsevier: BV, 2003. Vol. 7,
pp 167–190.
8. Hay, S. O.; Obee, T. N.; Thibaud-Erkey, C. Appl. Catal. B 2010, 99, 435–441.
9. Pichat, P. Appl. Catal. B 2010, 99, 428–434.
10. Photocatalysis and Water Purification. From Fundamentals to Recent Applications. In Pichat, P., Ed; Wiley-VCH: Weinheim, 2013.
11. Malato, S.; Fernández-Ibáñez, P.; Maldonado, M. I.; Oller, I.; Polo-López, M. I. Solar photocatalytic pilot plants: Commercially available reactors. In Photocatalysis and Water
Purification. From Fundamentals to Recent Applications; Pichat, P., Ed.; Wiley-VCH: Weinheim, 2013; pp 377–397.
12. Sawyer, D. T.; Valentine, J. S. Acc. Chem. Res. 1981, 14, 393.
13. Cermenati, L.; Pichat, P.; Guillard, C.; Albini, A. J. Phys. Chem. B 1997, 101, 2650.
14. Cermenati, L.; Albini, A.; Pichat, P.; Guillard, C. Res. Chem. Intermed. 2000, 26, 221.
15. Montoya, J. F.; Peral, J.; Salvador, P. Chem. Phys. Chem. 2011, 12, 901.
16. Pichat, P.; Courbon, H.; Enriquez, R.; Tan, T. T. Y.; Amal, R. Res. Chem. Intermed. 2007, 33, 239–250.
17. von Sonntag, C. Free-Radical-Induced DNA Damage and Its Repair: A Chemical Perspective; Springer: Berlin, 2006.
18. Konaka, R.; Kasahara, E.; Dunlap, W. C.; Yamamoto, Y.; Chen, K. C.; Inoue, M. Free Rad. Biol. Med. 1999, 27, 294–300.
19. Janczyk, A.; Krakowska, E.; Stochel, G.; Macyk, W. J. Am. Chem. Soc. 2006, 128, 15574.
20. Barclay, L. R. C.; Basque, M. C.; Vinqvist, M. R. Can. J. Chem. 2003, 81, 457.
21. Konovalova, T. A.; Lawrence, J.; Kispert, L. D. J. Photochem. Photobiol. A 2004, 162, 1–8.
22. Rengifo-Herrera, J. A.; Pierzchala, K.; Sienkiewicz, A.; Forro, L.; Kiwi, J.; Moser, J. E.; Pulgarin, C. J. Phys. Chem. C 2010, 114, 2717.
23. Nosaka, Y.; Daimon, T.; Nosaka, A. Y.; Murakami, Y. Phys. Chem. Chem. Phys. 2004, 6, 2917–2918.
24. Daimon, T.; Hirakawa, T.; Nosaka, Y. Electrochemistry 2008, 76, 136–139.
25. Karunakaran, C.; Senthilvelan, S.; Karuthapandian, S. J. Photochem. Photobiol., A 2005, 172, 207.
26. Xu, Z.; Jing, C.; Li, F.; Meng, X. Environ. Sci. Technol. 2008, 42, 2349.
27. Zheng, S.; Cai, Y.; O’Shea, K. E. J. Photochem. Photobiol., A 2010, 210, 61.
28. Robertson, P. K. J.; Bahnemann, D. W.; Lawton, L. A.; Bellu, E. Appl. Catal. B 2011, 108–109, 1.
29. Jolivet, J.-P. Metal Oxide Chemistry and Synthesis: From Solution to Oxide; Lavoisier: Paris, 2000.
30. Negishi, N.; Matsuzawa, S.; Takeuchi, K.; Pichat, P. Chem. Mater. 2007, 19, 3808.
31. Yamagishi, Makiko; Kuriki, Sina; Song, P. K.; Shigesato, Yuzo. Thin Solid Films 2003, 442, 227–231.
32. Zywitzki, O.; Modes, T.; Sahm, H.; Frach, P.; Goedicke, K.; Glöß, D. Surf. Coat. Technol. 2004, 180–181, 538–543.
33. Kikuchi, Hisashi; Kitano, Masaaki; Takeuchi, Masato; Matsuoka, Masaya; Anpo, Masakazu; Kamat, Prashant V. J. Phys. Chem. B 2006, 110, 5537–5541.
34. Shapovalov, V. I. Glass Phys. Chem. 2010, 36 (2), 121–157.
35. Sarantopoulos, Christos; Gleizes, Alain N.; Maury, Francis Thin Solid Films 2009, 518, 1299–1303.
36. Paz, Y.; Heller, A. J. Mater. Res. 1997, 12, 2759–2766.
37. Ghazzal, M. M.; Chaoui, N.; Aubry, E.; Koch, A.; Robert, D. J. Photochem. Photobiol. A: Chem. 2010, 215, 11–16.
38. Aubry, E.; Ghazzal, M. N.; Demange, V.; Chaoui, N.; Robert, D.; Billard, A. Surf. Coat. Technol. 2007, 201, 7706.
39. Allen, N. S.; Edge, M.; Verran, J.; Stratton, J.; Maltby, J.; Bygott, C. Polym. Degrad. Stab. 2008, 93, 1632.
40. Sanchez, B.; Coronado, J. M.; Candal, R.; Portela, R.; Tejedor, I.; Anderson, M. A.; Tompkins, D.; Lee, T. Appl. Catal. B: Environ. 2006, 66, 295.
www.iran-mavad.com
41. Kasanen, J.; Suvanto, M.; Pakkanen, T. T. J. Appl. Polym. Sci. 2009, 111, 2597.
42. Schmidt, H.; Naumann, M.; Müller, T. S.; Akarsu, M. Thin Solid Films 2006, 502, 132.
43. Pichat, P. J. Adv. Oxid. Technol. 2010, 13, 238 (and refs therein).
44. Paz, Y. Appl. Catal. B 2010, 99, 448–460.
45. Zubkov, T.; Stahl, D.; Thompson, T. L.; Panayotov, D.; Diwald, O.; Yates, J. T., Jr. J. Phys. Chem. B 2005, 109, 15454.
46. Yan, X.; Abe, R.; Ohno, T.; Ohtani, B. Thin Solid Films 2008, 516, 5872.
47. Guan, K. Surf. Coat. Technol. 2005, 191, 155.
48. Zhang, X.; Fujishima, A.; Jin, M.; Emeline, A. V.; Murakami, T. J. Phys. Chem. B 2006, 110, 25142.
49. Allain, E.; Besson, S.; Durand, C.; Moreau, M.; Gacoin, T.; Boilot, J.-P. Adv. Funct. Mater. 2007, 17, 549.
50. Ding, X.; Zhou, S.; Wu, L.; Gu, G.; Yang, J. Surf. Coat. Technol. 2010, 205, 2554.
51. Puzenat, E.; Pichat, P. J. Photochem. Photobiol. A: Chem. 2003, 160, 127.
52. Strini, A.; Cassese, S.; Schiavi, L. Appl. Catal. B 2005, 61, 90.
53. Ruot, B.; Plassais, A.; Olive, F.; Guillot, L.; Bonafous, L.; Quenard, D. Solar Energy 2009, 83, 1794.
54. US Patent 7582256, 2009.
55. Li Puma, G.; Bono, A.; Krishnaiah, D.; Collin, J. G. J. Hazard. Mater. 2008, 157, 209–219. Pichat, P. Water Sci. Technol. in press.
56. Nair, M.; Luo, Z.; Heller, A. Ind. Eng. Chem. Res. 1993, 32, 2318.
57. Hall, R. J.; Bendfeldt, P.; Obee, T. N.; Sangiovanni, J. J. J. Adv. Oxid. Technol. 1998, 3, 243–252.
58. Taranto, J.; Frochot, D.; Pichat, P. Sep. Purif. Technol. 2009, 67, 187–193.
59. Ávila, P.; Sánchez, B.; Cardona, A. I.; Rebollar, M.; Candal, R. Catal. Today 2002, 76 (2–4), 271–278.
60. Furman, M.; Corbel, S.; Wild, G.; Zahraa, O. Chem. Eng. Process 2010, 49, 35–41.
61. Edouard, D.; Lacroix, M.; Pham, Ch.; Mbodjic, M.; Pham, H. C. AIChE J. 2008, 54 (11), 2823–2832.
62. Keller, N.; Keller, V.; Garin, F.; Ledoux, M. J. Mater. Lett. 2004, 58, 970.
63. Taranto, J.; Frochot, D.; Pichat, P. Ind. Eng. Chem. Res. 2007, 46, 7611–7614.
64. Taranto, J.; Frochot, D.; Pichat, P. Ind. Eng. Chem. Res. 2009, 48, 6229–6236.
65. Pichat, P.; Enriquez, R.; Mietton, E. Solid State Phenomena 2010, 162, 41–48.
66. Enriquez, R.; Beaugiraud, B.; Pichat, P. Water Sci. Technol. 2004, 49, 147.
67. Enriquez., R.; Agrios, A. G.; Pichat, P. Catal. Today 2007, 120, 196–202.
68. US Patent 5.965.091, 1999.
www.iran-mavad.com
www.iran-mavad.com
4.17 Biocompatible Coating
Y Qiao and X Liu, Shanghai Institute of Ceramics, Shanghai, PR China

4.17.1.1 Factors Affecting Interfacial Biocompatibility 426
4.17.1.2 Biological Responses Underlying Biocompatibility 426
4.17.2 Hemocompatible Coatings 427
4.17.2.1 Introduction 427
4.17.2.2 Blood–Material Interactions Underlying Hemocompatibility 427
4.17.2.3 Approaches to Hemocompatible Coatings 429
4.17.2.3.1 Passive Coatings 429
4.17.2.3.2 Bioactive Coatings Mimicking Nonthrombogenic Endothelial Cells 431
4.17.2.3.3 Coatings for Promoting Endothelialization 433
4.17.2.4 Conclusions and Outlook 433
4.17.3 Cytocompatible and/or Immunomodulatory Coatings 434
4.17.3.2 Biomaterials in Biological Systems 434
4.17.3.2.1 Cell–Material Interactions 434
4.17.3.2.2 Host Responses to Biomaterials 434
4.17.3.3 Cytocompatible and/or Immunomodulatory Coatings 434
4.17.3.3.1 Surface-Modified Coatings 434
4.17.3.3.2 Anti-Inflammatory Drug-Grafted Biomaterials 435
4.17.3.4 Conclusions 435
4.17.4 Antibacterial Coatings 435
4.17.4.2 Bacterial Responses to Biomaterials 435
4.17.4.3 Approaches to Antibacterial Coatings 436
4.17.4.3.1 Killing Bacteria by Bactericidal Agents 436
4.17.4.3.2 Inhibiting Bacterial Adhesion 437
4.17.4.4 Conclusions and Future Work 438
4.17.5 Summary and Perspectives 438
Acknowledgment 438
References 438
4.17.1 Introduction
In the last century, rapid technological developments in materials science have brought the field of biomedical engineering into
a totally new realm. As one of the most challenging fields in biomedical engineering, biomaterials is highlighted in various medical
applications such as device-based therapies, drug delivery, tissue engineering, and medical imaging (1). Historically, the use of
biomaterials can be traced back to ancient times, where glass eyes, gold-filled teeth, sutures for wound closure, and wooden legs
were applied to repair the human body. From the early part of the twentieth century, the field of biomaterials entered into a period
of explosive growth, and naturally derived materials began to be replaced by synthetic polymers (see also Chapter 1.03, Testing of
Polymeric Materials and Chapter 2.03, Polymeric Materials Characterization and Modeling), ceramics (see also Chapter 1.02,
Techniques for assessing the properties of Ceramic Materials), and metal alloys (see also Chapter 1.08, Testing of Small-Sized
Specimens), which exhibited better biological performance, increased functionality, and more reproducibility than their naturally
derived counterparts (2). This remarkable progress has been witnessed in the development of medical devices such as pacemakers,
cardiac assist devices, external and internal ear repairs, soft-tissue replacements, artificial hip joints, and bone cements. On the basis
of their application, biomaterials are defined as materials that can restore, augment, or replace the natural function of the living
tissues or organs in the body either temporarily or permanently.
Today, the field of biomaterials is highly multidisciplinary and encompasses aspects of materials science, engineering, biology,
chemistry, and medicine. Over the past decades, this field has been rapidly evolving along with advances in molecular biology as
well as the increased exploitation of physicochemical techniques in studying biological phenomena and rational design of
biomedical and functional materials. As a prerequisite to ensure safety and efficacy, biomaterials need to meet basic biocompat-
ibility requirements. In 1980s, biocompatibility was generally defined as ‘the ability of a material to perform with an appropriate
host response in a specific application’ (3); therefore, it is a varying property of a biomaterial depending on its intended medical
application and implanted biological environment. As our understanding of material–biological interaction grows, the concept of

www.iran-mavad.com
426 Biocompatible Coating
biocompatibility began to shift an emphasis from a ‘do no harm’ mission to encouraging positive healing responses (4,5). In other
words of Ratner, biocompatibility can be defined as “the ability of implant surface to interact with cells and liquids of the biological
system and to cause exactly the reactions which analogues body tissue would bring out” (6). In light of the above, the ability to
engineer unique and enhanced biocompatibility into biomaterials has become a subject of intense interest. Since material–
biological interactions are surface-dependent phenomena (7,8), creating biocompatible coatings/films via surface modification
offers a cost-effective solution to satisfy the growing need for the implants with specific biocompatibility but without changing their
bulk physical properties (9).
The objectives of this chapter are to provide a current overview of advances in biocompatible coatings and a basis for interpreting
the biological responses to materials in the context of biocompatibility; illustrate how the material surface properties trigger
different biological reactions at the interfaces by providing examples; and offer insight into the design strategies and principles for
controllable fabrication of biocompatible coatings.
4.17.1.1 Factors Affecting Interfacial Biocompatibility

From the large amount of data available in various publications over the past decade, several determinants have been identified for
the biocompatibility of surface coatings (10), including (1) biological interactions at the interface between material surface and
biological molecules, which is influenced directly by the surface physicochemical features such as surface chemical composition,
wettability and charge; (2) leachable substances such as wear debris, degradation product, released ions, and molecular fragments;
and (3) implant structural integrity influenced by surface geometry, physics, chemistry, and mechanical characteristics.
Apart from these material-related factors, manufacturing, surgical techniques, and patients’ physiological health also
contribute much to implant failure. This chapter describes the influences of material surface-related features on the biocom-
patibility of implants.
4.17.1.2 Biological Responses Underlying Biocompatibility

From a clinical point of view, engineering highly biocompatible coatings requires a thorough understanding of the host responses to
implanted biomaterials. Biomaterial implantation involves surgical trauma, which initiates a cascade of physiological processes
characterized by blood protein adsorption, recruitment of acute and chronic inflammatory cells, formation of granulation tissue,
foreign body reaction, and fibrosis or fibrous encapsulation (11). This important field of research, over the years, has been discussed
by several authors in books and comprehensive review articles (11–17). In this section, we briefly outline up-to-date knowledge on
the biological responses to materials.
When the biomaterials come into contact with biological system, the following reactions were observed at the material–tissue
interface:
1. Blood–Biomaterial Interactions
In vascularized tissue, surgery initially induces a vascular response in intact vessels surrounding the implant site, resulting in the
leakage of fluid, blood proteins, and cells. Then blood coagulation as well as platelet adhesion forms a blood clot, which serves to
reestablish hemostasis and leads to the immediate development of fibrin- and fibronectin-rich provisional matrix for cell adhesion
and migration (18). It is noteworthy that the adsorbed protein layer shows time-dependent conformational changes with respect to
surface characteristics such as wettability, charges, and functional groups (19,20). And the conformational changes may cause
desorption or protein exchange, which has direct influences on the subsequent cellular events.
2. Inflammatory Responses
Following the initial blood–material interactions and provisional matrix formation, acute and chronic inflammation occurs
successively. This phase involves a complex cascade of immune modulators, including inflammatory cells (neutrophils, T and B
lymphocytes, monocytes, macrophages, mast cells, and basophils) and chemical mediators released by resident cells or from the reserves
stored in the granules of platelets. It must be noted that the cellular and molecular interactions at the material–tissue interface are directly
influenced by the continuous presence of nonphagocytosable foreign materials. Therefore, the surface properties of a biomaterial should
be responsible for the variations in the degree and time course of the inflammatory process, which in turn characterize the biocom-
patibility of a material. The biological responses to biomaterials have been reviewed in detail by Anderson et al. (13,15).
3. Bacterial–Biomaterial Interactions
Bacteria and other microorganisms have a natural tendency to adhere to surfaces as a survival mechanism. This process is largely
dependent on surface characteristics of targeted biomaterials and planktonic bacteria or microorganisms. Besides, the kinetics of this
process are affected by environmental factors such as presence of serum proteins and associated flow conditions (21). After
implantation, the initial interaction between bacteria and the adsorbed protein layer is nonspecific through a combination of van
der Waals, gravitational, and Coulombic forces (22). Then bacterial membrane proteins and polysaccharides specifically bind to the
proteins on the device surface. Adherent bacterial species subsequently secrete a protective exopolysaccharide layer (i.e., for
a biofilm) that imparts increased antibiotic resistance to the constituent bacterial cells (23). It is well known that biofilm-associated
www.iran-mavad.com
Biocompatible Coating 427
bacterial infections are exceedingly difficult to treat with conventional antibiotic therapies (24). Thus, reducing bacterial adhesion
during the initial 6 h period is particularly important for preventing device-associated infections.
4.17.2 Hemocompatible Coatings

4.17.2.1 Introduction
The term hemocompatibility defines the tolerance of materials with blood, one aspect of biocompatibility, mainly related to the
thrombogenic and thromboembolic complications induced by biomaterial surfaces. The performance of blood-contacting
biomedical devices (see also Chapter 10.10, Biomedical Applications) such as catheters, blood vessel grafts, vascular stents, artificial
heart valves, circulatory support devices, various extracorporeal tubings, hemodialysis, hemapheresis, and oxygenator membranes is
significantly impaired by these adverse reactions (18,25–28). Although a large number of commercial products are available on the
market, most of them possess an acceptable but inadequate hemocompatibility. Clinically, patients should continuously take
anticoagulation medication to prevent these effects. In recent years, significant efforts have been made to appreciate the funda-
mental aspects of blood–biomaterial interactions (29–31) and molecular mechanisms of surface-induced thrombus formation
(18,32) in order to establish design concepts for hemocompatible biomaterials. And following advances in surface modification
methods, many researchers are currently working on the development of more hemocompatible coatings for blood contact devices
(9,33,34). This section briefly introduces mechanisms of biomaterial-associated thrombosis and the underlying concepts for
hemocompatible coatings engineering.
4.17.2.2 Blood–Material Interactions Underlying Hemocompatibility

When a biomaterial is exposed to the biological system, protein adsorption is the first event in blood–biomaterial interactions,
which takes place immediately. Thrombus or blood clot formation on the surface of a biomaterial is related to the well-known
Vroman effect, which regulates the early time-dependent adsorption, desorption, and exchange behaviors of proteins (35,36).
Consequently, the adsorbed protein layers differ in compositions from plasma proteins and their molecular orientations, which
may activate further physiological processes, including coagulation (37,38), fibrinolytic and complement pathways (39), and
platelet adhesion (40,41), thereby resulting in different degrees of thrombus or blood clot formation (18). These processes interact
intimately and are highly interlinked at many stages (Figure 1 (18)) (18,42). From an ISO standards perspective (ISO 10993-4),
hemocompatibility evaluation covers five categories: thrombosis, coagulation, platelets, hematology, and immunology (comple-
ment and leukocytes) (43). In this section, we only briefly describe the pathways most relevant for biomaterials and attempt to
understand the role of biomaterials in thrombosis.
It is well established that the extent of protein adsorption on substrates is mainly driven by various intermolecular forces
between proteins and biomaterial surfaces (either specifically or nonspecifically), such as ionic bonding, hydrophobic interactions,
and charge–transfer interactions (44,45). These interactions are determined by structural, chemical, and electrical properties of
proteins and biomaterial surfaces, as described in the literature (45). In this way, research into blood coagulation has focused on
interactions with biomaterials that activate the plasma coagulation cascade (see Vogler et al. (37,46,47) for more discussion). This
cascade occurs through a complex series of proteolytic reactions resulting in thrombin generation. Thrombin is a key enzyme
catalyzing the conversion of fibrinogen to fibril units that polymerize to a fibrous matrix of blood clot. There are two interconnected
Figure 1 Overview of blood–material interactions showing the components relevant to thrombosis. With permission from Elsevier Ltd (Gorbet, M. B.;
Sefton, M. V. Biomaterial-Associated Thrombosis: Roles of Coagulation Factors, Complement, Platelets and Leukocytes. Biomaterials 2004, 25 (26),
5681–5703).
www.iran-mavad.com
mechanisms underlying the plasma coagulation cascade (Figure 2 (37)): an intrinsic pathway initiated when blood comes into
contact with subendothelial connective tissues or negatively charged surfaces, and an extrinsic pathway initiated through tissue
factor expression at the site of vascular injury (48). From a biomaterials perspective, the intrinsic pathway is considered to be
responsible for poor hemocompatibility of blood-contacting biomedical devices because it is contact activated by blood–surface
interactions (49). The intrinsic pathway is initiated by the contact activation involving four primary plasma proteins: coagulation
Factors XII and XI, prekallikrein, and high molecular-weight kininogen (HMWK) (37,50,51). It is generally proposed that Factor XII
is activated by adsorption on negatively charged surfaces and undergoes conformational changes that provoke autoactivation
surface
FXII!FXIIa (52). Through complexation of HMWK, FXIIa activates Factor XI to Factor XIa, initiating the intrinsic coagulation
pathway that leads to the formation of thrombin and fibrin. FXIIa also converts prekallikrein into kallikrein, which directly activates
the kinin and fibrinolytic systems.
In addition to the coagulation system, thrombus formation also involves the aggregation of platelets in response to vascular
injury. Platelet adhesion is initially mediated by glycoprotein VI and Ib, the membrane receptors that bind to collagen, and von
Willebrand factor in the subendothelial matrix (53–55), respectively, and is further stabilized by GPIIb/IIIa and GPIa/IIa, which
bind to collagen and fibrinogen/fibrin (53,56), respectively. This indicates that platelets are more prone to adhere to the blood-
contacting surfaces adsorbing these proteins (57,58). Although platelet activation is still not fully understood, the factors identified
to cause activation include platelet adhesion and hydrolysis of protease-activated receptor 1 located on the membrane by thrombin
(59–61).
Apart from coagulation activation, platelet adhesion and activation processes, complement activation may also be an indi-
cator of hemoincompatibility of biomaterials (62–64). The complement system is in the front line in the defense against foreign
substances and microorganisms, and can be stimulated via both the alternative and classical pathways, where the alternative
pathway is most relevant to biomaterials (65). It is initiated by the hydrolysis of C3 and covalent binding of C3b (the cleaved
form of C3) to hydroxyl- or amino-group functionalized surfaces, a process also accompanied by the release of C3a. Furthermore,
C3b binds and activates factor B to produce the active proteolytic enzyme C3 convertase, which cleaves more C3 molecules, thus
amplifying the response. Most of the complement-related responses are likely ascribed to the anaphylatoxins C3a and C5a
(66,67).
Figure 2 Simplified line diagram of the plasma coagulation cascade showing intersection of the intrinsic and extrinsic pathways (many mediators and
cofactors involved in hemostasis are not shown and the interaction with platelets has been ignored for simplicity). With permission from Elsevier Ltd
(Vogler, E. A.; Sledlecki, C. A. Contact Activation of Blood-Plasma Coagulation. Biomaterials 2009, 30 (10), 1857–1869).
www.iran-mavad.com
4.17.2.3 Approaches to Hemocompatible Coatings

Inspiration for the approaches to hemocompatible coatings has been derived from the existing knowledge of blood–material
interactions and thrombogenesis with three key objectives: (1) passivation coatings to prevent or resist nonspecific protein
adsorption, and to avoid undesirable subsequent biological responses; (2) bioactive coatings grafted with anticoagulants and
antiplatelet agents to inhibit the coagulation cascade and prevent the platelet adhesion and activation (see also Chapter 4.18,
Bioactive Coatings); and (3) endothelialization to create a physiologically hemocompatible surface. This section reviews advances
in modifying a material to create optimal interactions with blood, and discusses the challenges that need to be overcome for future
success. Reviews on this subject for a specific application may be found elsewhere (33,68–70).
From the studies mentioned, it is evident that preventing nonspecific protein adsorption, inhibiting blood coagulation and
complement systems, and reducing platelet adhesion may offer opportunities to overcome these adverse responses. And the
following concepts have emerged:
1. Fabricating chemically inert coatings, basically inorganic coatings, to decrease bioreactivity;

2. Passivation of surfaces to minimize the interactions between surfaces and blood proteins or cells;
3. Grafting of drugs to achieve active antithrombotic functions;
4. Imitation of a biomembrane to avoid immune reactions.
4.17.2.3.1 Passive Coatings

It is generally accepted that surface passivation is a widely recognized approach to minimize the interactions of proteins/cells with
surfaces for blood-contacting materials used in antithrombogenic implants, hemodialysis membranes, drug delivery carriers, and
biosensors (69,71,72). According to different modification concepts and routes, these coatings are categorized into two groups:
inorganic passivation coatings and organic passivation coatings.
4.17.2.3.1.1 Inorganic passive coatings

Among hemocompatible surfaces available for blood-contacting medical devices, inorganic coatings have emerged as a practical
solution to minimize biomaterial-associated thrombosis. Examples of such inorganic coatings are currently under development,
including carbon-based coatings (see also Chapter 4.15, Carbon Coatings), metal oxide or nitride, and hydroxyapatite. Generally,
the inorganic coatings offer a good combination of high inertness and mechanical and chemical stability. These advantages may be
important in their ability to prevent surface-induced thrombus formation.
Early investigations of hemocompatible carbon coatings mainly focused on low-temperature isotropic pyrolytic carbon
(LTIC), which has been extensively used as components or coatings for mechanical heart valves or stents since late 1960s. Despite
the clinical success seen with the use of LTIC, inadequate thromboresistance remains a primary complication associated with its
use (73). Hence anticoagulant therapy is required for patients implanted with LTIC-coated devices. In this regard, the alternative
carbon coatings are expected to possess sustained anticoagulant and antithrombotic capacity while retaining excellent durability
and stability. Recently, diamondlike carbon (DLC) films have attracted much attention as antithrombogenic coatings (74–77),
which also exhibit low friction coefficient, high hardness, chemical inertness, and smoothness (78). All these features enable the
application of DLC coatings with prospective potential. Besides, it has been found that the hemocompatibility of DLC coatings
can be tuned by tailoring the surface electrical property, energy, and atomic bond structure of carbon via incorporating different
elements such as Si (79–81), F (82–85), P (86), N (87,88), and their combinations. In an earlier study, tetrahedral amorphous
carbon (ta-C) showed improved anticoagulation compared to LTIC (74). The main reason for this improvement is correlated to
their effective work functions (w1.0 eV for ta-C and w5.5 eV for LTIC), indicating that the charge transfer from the inactive state
of fibrinogen (with a bandgap of 1.8 eV, Figure 3 (89)) to ta-C films is much more difficult than to the surface of LTIC. Thus,
fibrinogen is much easier to be oxidized on LTIC and transforms to fibrin monomer, resulting in thrombus formation. Similarly,
Ogwu et al. found that Si-doped a-C:H coatings (Si-DLC), which lowered the electrical resistivity, effective work function, and
surface energy, lead to reduced platelet aggregation and enhanced hemocompatibility (90). Apart from electrical properties,
surface wettability is also a crucial determinant affecting hemocompatibility. However, available studies show contradicting
tendencies on the effect of surface wettability. It has been found that both hydrophilic phosphorus-doped DLC (P-DLC) films
(86) and hydrophobic fluorine-doped DLC (F-DLC) (84) films can induce a higher ratio of albumin/fibrinogen adsorption and
more strongly suppress platelet adhesion and activation. Chu et al. suggest that the better hemocompatibility of P-DLC coating
may be ascribed to its minimized interactions with proteins, while Hasebe et al. indicate that the hemocompatibility of F-DLC is
not solely determined by wettability but also the chemical composition, interfacial energy, and higher ratio of albumin/
fibrinogen. Another significant factor affecting hemocompatibility relates to the sp2/sp3 fraction. Chu et al. reported that the
hydrogenated DLC films deposited at lower substrate bias delayed the clotting time and the hemocompatibility tended to
improve with increasing sp2/sp3 ratio (75). In a related study, hemocompatibility of La2O3-doped DLC films increased with
increasing of sp2-bonded C content (91). The observations presented in this section and the chapters they are based on are
presented in a more detailed review by Roy et al. (92).
Additionally, ceramic oxides form another versatile subclass of inorganic coatings that have been extensively investigated for
blood-contacting materials. Among ceramics, titanium oxide represents a unique research subject with a high degree of surface
modification. It has been found that the hemocompatibility of titanium oxide is significantly affected by its surface energy and
www.iran-mavad.com
Figure 3 Interaction of fibrinogen with a solid via the charge transfer process, in which “e” represents electrons. With permission from
Elsevier Ltd (Chen, Y. J.; Leng, Y. X.; Tian, X. B.; Wang, L. P.; Huang, N.; Chu, P. K.; Yang, P. Antithrombogenic Investigation of Surface Energy
and Optical Bandgap and Hemocompatibility Mechanism of Ti(Ta)O2 Thin Films. Biomaterials 2002, 23 (12), 2545–2552).
semiconductor properties. As shown by Sharma et al., surface hydrophobicity increased as the thickness of the oxide layer increased,
resulting in a higher albumin/fibrinogen ratio (93). Lin et al. further proved that not only very few of platelets were attached to the
superhydrophobic TiO2 nanotube layers, but also the attached platelets were not activated in this condition (94). This result is
consistent with Fu’s findings (95). It is important to note that fibrinogen has an electronic structure similar to that of a semi-
conductor with a bandgap of 1.8 eV. And fibrinogen electrons transferred from their occupied valence band into the free states of the
materials surface may cause the formation of thrombus. Therefore, after proteins are adsorbed, semiconductor properties of tita-
nium oxide like electronic state play a dominant role in affecting its hemocompatibility. Available evidences suggest that the
electronic structural state of titanium oxide can be tailored physically or chemically, and the nonstoichiometric titanium oxide
(TiO2x), which has n-type semiconductive properties, can inhibit the charge transfer from the valence band of fibrinogen into the
material surface. Previous studies have reported that the rutile titanium oxide exhibited higher hemocompatibility than anatase or
amorphous due to its n-type semiconducting nature, which led to limitation of electron transferring and reduced platelet adhesion
and activation (96,97). In a recent study, La2O3 doping produced pure titanium oxide films, also resulting in improved hemo-
compatibility (98). On the basis of studies described in the literature, n-type semiconductive titanium oxide can be fabricated by
various techniques or doping elements, such as ion beam assisted deposition (99), plasma immersion ion implantation (100, see
also Chapter 4.12, Plasma Sources in Thin Film Deposition and Chapter 4.09, Polymer Deposition and Coatings on Polymers), Ta
doping (89), and magnetron sputtering or F doping (101) via chemical treatment. Chu et al. reported that the Ta-doped titanium
oxide (Ti(Taþ5)O2) showed better hemocompatibility than LTIC in the in vivo tests (89).
4.17.2.3.1.2 Organic passive coatings

Over the last several years, a number of nonfouling coatings have been developed to optimize protein-repellent properties of
biomaterial surfaces, which are frequently grouped into two major categories: synthetic nonfouling materials and inert biomole-
cules. For further details on surface passivation strategies, associated mechanisms, and biological applications, please refer to the
published literature (102,103).
www.iran-mavad.com
According to the broad screening study performed by Whitesides and coworkers, most of the nonfouling materials share
common features such as hydrophilic, electrically neutral, and hydrogen bond acceptors rather than hydrogen bond donors
(104–106), all of which are considered important in resisting protein adsorption and subsequent cell and platelet adhesion (107).
For decades, several types of synthetic materials have been found to exhibit promising nonfouling capabilities, including (1)
hydrophilic molecules (e.g., oligoethylene glycol, hydroxyethyl methacrylate, and polyethylene glycol), (2) zwitterionic molecules
(e.g., phosphobetaine, sulfobetaine, and carboxybetaine), and (3) mixed positively and negatively charged self-assembled
monolayers of equal valence.
One of the most thoroughly investigated hydrophilic molecules is poly(ethylene glycol) (PEG). PEG is a nontoxic, non-
immunogenic, and uncharged polymer that is miscible with aqueous and nonaqueous media (108). In recent years, PEG has been
widely grafted not only to polymers but also to metals and glasses to minimize their interactions with blood components for
providing greater hemocompatibility (109–113). In addition, PEG often acts as a shield to prevent or lower the degree of
hemolysis, which, however, is very significant in the case of nanoparticles, to confer them further longevity in the systematic
circulation (114,115). Excellent reviews on PEGylated materials have been published (116,117). The methods developed to
integrate PEG into biomaterials include physical adsorption (118), self-assembly (119,120), chemical coupling (110), and graft
polymerization (121).
Previous studies have suggested that the protein resistance by PEG chains has been associated with its unique properties such as
the large excluded volume on the surface, steric repulsion between PEG moieties, and the formation of a PEG-hydration layer
(118,122,123). For this reason, many reports have investigated the effects of PEG chain length, density, conformation, and charge at
the membrane surface on protein-resistant efficiency (107,124–128). Extensive theoretical and experimental investigations have
demonstrated that increasing the grafting density and chain length (upper limits appears to be reached at about 20 kDa and
2000 MW, respectively) may result in higher protein-repellent properties of PEGylated surfaces (122,129). Furthermore, grafting
density is known to impart a particular conformation that PEG may adopt in solution. It is well acknowledged that the dense
brushlike PEGylated surface presented higher steric hindrance and more flexibility to shield a wider surface area and offer a higher
hydration capacity than the mushroomlike (low density) or networklike PEGylated surface, thereby more efficiently preventing
protein from reaching the material surface to adsorb (128,130–133). Though not conclusive, the present data suggest that the
nonfouling nature of PEGylated surfaces can be highly regulated by controlling grafting coverage and structures. And due to its
intrinsic properties, PEG modification has also been extended to immobilize bioactive molecules as spacers, keeping molecules at
a distance away from the substrate (134). Although versatile PEG-based materials have been approved by FDA for clinical appli-
cations, disadvantages of PEG such as decomposition in the presence of oxygen or transition metal ions and associated complement
activation promote efforts to develop alternatives (135,136).
Zwitterionic materials, on the other hand, contain both positively and negatively charged groups while maintaining overall
charge neutrality, representing another promising solution for antifouling modification (137–141). The differences between PEG
and zwitterionic molecules are their chemical structures and associated nonfouling mechanisms. Previous experimental and
simulation studies have demonstrated that zwitterionic materials are more strongly hydrated through ionic solvation, as compared
to PEG and other hydrophilic materials, which are hydrated by hydrogen bonding (104,142,143). Due to the key role of surface
hydration in protein-repellent capability, this may explain the recent observations that the zwitterionic surfaces exhibit a nonspecific
protein resistance that is compatible to or better than that of various hydrophilic materials (144). In addition, more recent studies
showed that the nonfouling characteristics of zwitterions can be tuned by adjusting their structures, such as backbones or coun-
terions, the distance between two charged groups, and the type of positively and negatively charged groups (145,146). Mixed-charge
materials, when uniformly distributed at the molecular scale, are equivalent to zwitterionic materials and highly resist to nonspecific
protein adsorption (147,148).
Aside from the synthetic nonfouling materials, albumin, one of the most abundant proteins in blood, is another molecule
frequently used to achieve surface passivation (149–151). The passivation behavior of albumin is associated with significantly lower
platelet adhesion when compared with other serum proteins. Although the direct covalent attachment of functional albumin to
material surfaces has been attempted and results in a significant reduction in platelet adhesion, clinical outcomes have not been able
to establish a clear advantage when albumin-coated grafts were implanted (152). In addition, central issues with albumin
passivation include the possibility of eventual displacement by other proteins in the case of adsorption and potential denaturation.
It should be noteworthy that these synthetic nonfouling coatings not only form excellent hemocompatible surfaces used for
various blood-contacting biomedical devices, but also offer an effective approach to inhibit bacterial adhesion and biofilm
formation, which will be described in later section (see Section 4.17.4.3.2.2).
4.17.2.3.2 Bioactive Coatings Mimicking Nonthrombogenic Endothelial Cells

The function of endothelium (EC) to maintain the balance between blood fluidity and intimal injure-induced arterial thrombosis is
increasingly providing inspiration for fabricating hemocompatible coatings. Therefore, we give a brief overview of the antith-
rombotic mechanism of endothelium, and interested readers are referred to alternative reviews to gather perspective on this subject
(153–157). It is well established that the ability of endothelium to interrupt blood coagulation events depends on producing several
key factors. One of the key factors is heparan sulfate proteoglycans (HSPG) (158,159), which can bind antithrombin III (AT-III),
an inhibitor that neutralizes thrombin (160,161). The other key factor is protein thrombomodulin (TM) (162,163), a cellular
receptor that catalyzes the activation of protein C by thrombin. Activated protein C inhibits Factor Va and Factor VIIa in turn.
The endothelium also synthesizes tissue pathway inhibitor, which binds to both Factor VIIa and Factor Xa (164). This complex
www.iran-mavad.com
subsequently inhibits the activity of tissue factor and thus the inhibition of the extrinsic coagulation pathway. Protein S, a cofactor
for the action of activated protein C, is also expressed by endothelium (165).
In a similar manner, endothelium can counteract platelet adhesion and activation by secreting nitric oxide (NO) (166,167) and
prostacyclin (PGI2) (168). The expression levels of prostacyclin and nitric oxide can be modulated in response to fluid shear stress
and chemical stimuli (169–171). The endothelium surface can also bind to ectonucleotidases that hydrolyze ADP, an agonist of
platelet activation (154,172). In addition, the endothelium surface has a brushlike layer of glycocalyx consisting of negatively
charged proteoglycans with glycosaminoglycan sidechains, which could repel platelets from the endothelium surface and prevent
platelet adhesion. These glycosaminoglycans are expressed by endothelium, including chondroitin/dermatan sulfate proteoglycans,
hyaluronic acid/hyaluronan (HA), and rich HSPG (50–90%) (154).
Apart from synthesizing anticoagulants and antiplatelet agents, endothelial cells also secrete tissue-type plasminogen activator
(t-PA) and urokinase, major initiators of fibrin degradation (173–175). It should be noted that the natural inhibitor of t-PA,
plasminogen activator inhibitor type 1 (PAI-1), is also expressed by endothelial cells (176), while urokinase production is
dependent on the state of EC activation due to inflammatory stimuli.
Both heparin (and heparan sulfate analog) and TM are well known for their anticoagulant function while NO and PGI2 have
long been recognized for their antiplatelet activity. Based on their roles and contribution to the exceptional biocompatibility of the
ECs, numerous attempts have been made to utilize such agents to create more hemocompatible surfaces, which are, in contrast to
nonfouling materials, biologically active surfaces. For general reviews on this subject, readers are referred to previous publications
(69,153). In the following discussions, the authors focus on the current progress in heparin-bound, TM-immobilized, and
NO-releasing coatings.
4.17.2.3.2.1 Heparin-bound coatings

Heparin, which has close structural and functional similarities to heparan sulfate, is a widely used anticoagulant that exerts an
incontrovertible effect on inhibiting thrombin formation indirectly (177,178). Heparin-coated materials were first reported by
Gott et al., and this early report revealed that heparin modification could substantially reduce thrombus formation, both in
vitro and in vivo (179). Recently, extensive heparin-coated materials have been used to manufacture blood-contacting
biomedical devices such as vascular grafts (180), stents (181), heart valves (182), extracorporeal circulation (183), and dialysis
membranes (184).
Current research on heparin modification has been directed toward maximizing its anticoagulant activities and promoting these
heparinized materials from academic research to clinical applications. On the basis of literature reports, heparin has been widely
immobilized or incorporated on a variety of substrates ranging from polymers (e.g., PLA (185), PTFE (186), and Dacron) to metals
(e.g., titanium and its alloy (187), nitinol (188), and stainless steel (189,190)) to prevent thrombus formation and improve
hemocompatibility. However, anticoagulant activities of heparinized surfaces may vary significantly depending on the method of
immobilization (191) and the amount and type of heparin (192). Commonly, immobilization of heparin can be achieved via
physical adsorption, ionic bonding, and covalent linkage, as reviewed by Linhardt et al. (193). This comparative study shows that the
stability of covalently immobilized heparin was found to be higher than those prepared by any other approaches, because the
physically adsorbed or ionically immobilized heparin is highly prone to displacement from the surface by an ion-exchange mech-
anism when exposed to blood. Covalent binding is usually performed through the coupling reaction between activated biomaterial
surface and hydroxyl, carboxyl, or amino groups of heparin. In terms of preserving biological activity of immobilized heparin for
long-term implants, several strategies have been developed to optimize the covalent method. It has been demonstrated that
immobilization of heparin through the use of hydrophilic spacers, like PEG, could effectively enhance the heparin’s bioactivity
compared to its direct immobilization. This enhancement should be ascribed to the reduction in steric hindrance and additional
nonfouling characteristics provided by PEG. Another strategy is to use end-point immobilization, in which the reactive aldehyde
groups introduced at the reducing termini of the heparin chains are directly coupled to primary amino groups of artificial surfaces,
generating a Schiff’s base. Subsequent reduction of the Schiff’s base may lead to stable, covalent bonds. According to previous reports,
stents coated with end-point immobilized heparin exhibited high bioactivity, which could eliminate subacute thrombosis (194,195).
In addition, more efforts on heparin research have been made to develop the cross-linking multilayer, such as heparin/fibro-
nectin (196,197), alginate/heparin (188), collagen/heparin (198), and dexamethasone-heparin systems (199), which benefit from
the cooperative effects of heparin and the other molecules. Although heparin has been used clinically for treatment of thrombosis, it
suffered from its short duration within compartments in vivo that require long-term anticoagulation. A better and promising
alternative is an antithrombin-covalent heparin (ATH) complex, which substantially outperforms standard heparin in terms of both
catalytic and direct AT activity, more efficient interactions with blood components, and the ability to inhibit surface-bound
coagulation factors, Factors XIa, Xa, and IXa (200–202). A recent study by Brash et al. (203) demonstrated that the PEO-ATH
modified gold surface showed higher antithrombin binding and greater proportion of active heparin compared to the heparin-
modified surface. Most importantly, the PEO-ATH modified surface is characterized by a prolonged clotting time of recalcified
plasma.
Despite its wide application in modifying blood-contacting materials, the current limitations of heparinization include
dependence on the local presence of AT-III to exert its activity, inability of the heparin-AT complex to inhibit enzymatically active
and procoagulant fibrin-bound thrombin, and uncertain ‘antiplatelet’ effects (204). More recently, low-molecular-weight heparin
(205), obtained by fractionation or fragmentation of standard heparin, and derivatives of heparin/heparan sulfate (206) have also
become potent alternatives (207).
www.iran-mavad.com
4.17.2.3.2.2 Thrombomodulin-immobilized coatings

Covalent immobilization of TM, an endothelial cell surface glycoprotein that promotes protein C activation, has also provided an
effective strategy to facilitate thrombin inhibition (162). Several studies have investigated the in vivo anticoagulant efficiency of
immobilized TM in reducing thrombogenicity. In an earlier study, human TM (hTM) immobilized hollow fibers have shown
enough nonthrombogenic activity to prolong plasma recalcification times and reduce thrombus formation (208). In a recent study,
hTM has been immobilized onto thin films of poly(octadecene alt maleic) using PEG as spacers (209). The hTM-immobilized
surface not only substantially reduced blood coagulation activation but also significantly suppressed the complement activation,
compared to the PEG surface, which makes TM a privileged candidate for the specific modification of biomaterial surfaces. Besides,
a greater reduction in antithrombogenic effect was observed on a heparin and TM-coimmobilized surface, demonstrating additive
enhancements in antithrombogenic effect (210). However, this modification has not come to clinical application yet.
4.17.2.3.2.3 Nitric oxide-releasing coatings

Nitric oxide (NO), a versatile regulatory agent produced by endothelial cells, has been employed to improve the thrombore-
sistance and hemocompatibility of biomaterial surfaces. The details of these studies are reviewed extensively elsewhere
(211–213). NO is well recognized for protective role within the vascular system, and its continuous release from endothelium
surface effectively prevents platelet activation on the walls of healthy blood vessels (214,215). To mimic the inner surface of
endothelium, various NO donors are incorporated by either doping/dispersing into polymers (216,217) or covalent bonding
onto polymers directly (218). However, it should concern the issues of the release rate of NO, mechanism of NO liberation, and
toxicity of NO donor moiety when released or exposed to the blood (219,220). Among NO donors, diazeniumdiolates
(RR0 N-N2O2) are one typical class as they spontaneously dissociate into NO and parent amines in solutions at physiological pH
and they do not require either an enzyme or a cofactor to generate NO (221,222). In previous work, a markedly reduced pla-
telet adhesion/activation and an improved thromboresistivity were observed on the hydrophobic polymer doped with an
N-diazeniumdiolate compound, while the hydrophilic counterpart was subsequently found to leach from polymer matrix into
blood, which could cause a potential toxicity issue. This toxicity concern can be overcome by anchoring N-diazeniumdiolates
onto other stable polymer matrixes (217). In a recent study, a nitric oxide releasing polymeric coating composed of poly(vinyl
chloride) blended with a lipophilic N-diazeniumdiolate was evaluated in a 4 h rabbit thrombogenicity model using flow
cytometry. This investigation showed that the NO-releasing polymer significantly reduced extracorporeal circulation thrombus
formation compared to polymer control after 4 h blood exposure (223). In addition, incorporation of nitric oxide into the
backbone of a polymer containing cell adhesive peptide may lead to synergistic effects, leading to faster levels of high endo-
thelialization without intimal formation (224,225). By far, extensive efforts have been made to optimizing nonthrombogenic
resistance of NO-releasing polymers (34,218,226,227).
4.17.2.3.3 Coatings for Promoting Endothelialization

As endothelium is intrinsically nonthrombogenic and plays a pivotal role in mediating thrombosis, endothelial seeding is being
explored as a promising strategy to improve graft patency (157,228). Complete endothelial coverage is required to establish
a physical barrier that separates circulating blood from thrombogenic biomaterial surfaces, thus providing nonthrombogenicity. We
give a brief overview of current progress and design principles for this tissue-engineering (see also Chapter 10.09, Tissue Engineering
Application) approach in this section. For in-depth discussion of endothelialization in tissue engineering of blood vessels, the
reader is encouraged to consult several excellent reviews (229–233).
Previous attempts to seed implants with autologous endothelial cells failed because of the lower density and the endothelial
cell retention under in vivo flow. In an earlier in vivo study, vein graft patency was better than in an endothelial-seeded poly-
tetrafluoroethylene (PTFE) graft due to the relatively lower density and smaller coverage (234). One approach to increase the
coverage and retention of endothelium may be to employ extracellular matrix (235) as well as its secreted proteins (236–238), and
peptide sequences (239), which act as ligands for cell proteins. In a previous study, a Gly-Arg-Gly-Asp-Ser-Pro peptide was grafted
on the surface of poly(carbonate urethane) with photoactive 4-benzoylbenzoic acid by UV irradiation, on which increased
retention and enhanced proliferation of endothelial cells were observed (240). And similar observations were also found in
arginine-glycine-aspartic acid (RGD) motif-modified surfaces (241,242). However, the current studies do not support such
a generalization on the RGD-modified surfaces. Besides, RGD is a fibronectin-derived peptide recognized by platelets and
endothelial cells with similar affinity. Alternatively, the lamin-derived Tyr-IIe-Gly-Ser-Arg (YIGSR) peptide is more specific for
endothelial adhesion. Another study of polyurethane grafts containing YIGSR and PEG showed increased endothelial adhesion
and less platelet adhesion (239). However, it is still difficult to achieve full coverage, and this may induce the risk of platelet
adhesion and activation. Recent attempts have been made to overcome these limitations, aided by combining the effects of
endothelial cells and antiplatelet agents, nitric oxide (243).
Aside from endothelial cells, tissue engineering of blood-contacting devices by recruitment of endothelial progenitors and
mesenchymal stem cells can also achieve nonthrombogenic and hemocompatible surfaces (244–246).
4.17.2.4 Conclusions and Outlook

Surface modification of biomaterial surfaces remains an active field of research, with a clear aim to improve the interfacial
hemocompatibility for blood-contacting devices. To achieve this goal, strategies for hemocompatible biomaterials cover a wide
www.iran-mavad.com
range from passivation of surfaces, to functionalization using interacting molecules, to the use of endothelial cells for surface
coverage. In this section, for each application, detailed mechanisms of how surface characteristics affect biological performance have
been elucidated. To satisfy all the technical requirements and biological functions, early benefits of such combination strategies are
already evident. Moreover, interdisciplinary approaches will be critical for designing biomaterial surfaces. Ultimately, advances in
emerging biomedical applications with precisely controlled physical, chemical, and biological properties remain a great oppor-
tunity for future research.
4.17.3 Cytocompatible and/or Immunomodulatory Coatings

The term ‘cytotoxicity’ means to cause toxic effects (death, alterations in cellular membrane permeability, enzymatic inhibition, etc.)
at the cellular level (247). Currently, ‘cytocompatibility’ is assayed as the viability of adhered/spread/proliferating cells, which more
likely reflects the noncytoxicity of biomaterials. Although cytocompatibility is a basic requirement and is critically important in
design of biomaterials for application in tissue engineering, the specific cellular responses on different biomaterial surfaces cannot
be distinguished from this cellular viability assay.
Recently, Williams defined a biomaterial as “.a substance that has been engineered to take a form which, alone or as part of
a complex system, is used to direct, by control of interactions with components of living systems, the course of any therapeutic or
diagnostic procedure” (248). It is well acknowledged that all materials when implanted into living tissue initiate a host response
that reflects the first steps of tissue repair. The advances in our understanding of the healing process and immunology establish the
basis of new design concepts for biomaterials, i.e., making use of this immune response to improve implant integration while
avoiding its perpetuation, leading to chronic inflammation and foreign body reactions and thus loss of the intended function
(5,14). Excellent reviews that discuss immunoengineering are available (67,249).
This section gives a brief overview of the current progress on the investigation of the relationships between surface characteristics
(surface charge, chemistry, topography, hydrophobicity, hydrophilicity, etc.) and cytocompatibility or immunomodulation,
respectively.
4.17.3.2 Biomaterials in Biological Systems

4.17.3.2.1 Cell–Material Interactions
The interaction of cells with their surrounding microenvironment is fundamental to multiple processes, including cell migration,
proliferation, differentiation, and tissue morphogenesis (250). This interaction has been proven to occur through a combination of
biochemical and biophysical signals. It has been reported that specifically engineered surfaces showing selected biofunctional
groups or micrometer-scale patterns have been used to investigate signal interactions in a systematic way (251,252). Besides,
advancements in chemistry, material science, and nanotechnologies greatly improved the possibility of presenting many different
signals.
4.17.3.2.2 Host Responses to Biomaterials

Following the initial blood–material interactions and provisional matrix formation, acute and chronic inflammation occur
successively. This phase involves a complex cascade of immune modulators including inflammatory cells (neutrophils, T and
B lymphocytes, monocytes, macrophages, mast cells, and basophils) and chemical mediators released by resident cells or
from the reserves stored in the granules of platelets. It must be noted that the cellular and molecular interactions at the
material–tissue interface are directly influenced by the continuous presence of nonphagocytosable foreign materials. There-
fore, the surface properties of a biomaterial should be responsible for the variations in the degree and time course of the
inflammatory process, which in turn characterize the biocompatibility of a material. For more detailed discussions, please
refer to Sandra et al. (14).
4.17.3.3 Cytocompatible and/or Immunomodulatory Coatings

Although various strategies are developed to design materials capable of eliciting appropriate immune responses, we only
compare the influences of surface-modified biomaterials or anti-inflammatory drug-grafted biomaterials on the cell behaviors
and immune responses. Sandra et al. have already given a systematic review (14); here we give a brief summary about the
progress.
4.17.3.3.1 Surface-Modified Coatings

Passive modulation of biomaterial surface properties aims at limiting macrophage adhesion, activation, and fusion to foreign body
giant cells (14). It has been reported that the adsorbed protein layer usually provides binding sites for protein-specific receptors on
polymorphonuclear leukocytes, monocytes, and macrophages. It has been found that surface chemistry has profound consequences
on cellular and host responses. A previous in vivo rat cage implant study showed hydrophilic and anionic substrates could effectively
decrease the rates of monocyte/macrophage adhesion and fusion, while adherent cells undergo significant apoptosis on these two
www.iran-mavad.com
surfaces when compared with the polyethylene terephthalate base surface. This study demonstrated that biomaterial-adherent cells
undergo material-dependent apoptosis in vivo, rendering potentially harmful macrophages nonfunctional while the surrounding
environment of the implant remains unaffected (253).
Besides, biomaterial surfaces constructed of well-defined nanotopography are also capable of eliciting the desired cellular
and tissue response (254). It has been reported that a variety of cells such as fibroblasts (255), epithelial cells (256), and
endothelial cells (257) are sensitive to topographic cues on biomaterial surfaces. These cells are found to tend to attach
to elevated groove edges and, in contrast, macrophages has been revealed to be limited on this kind of micron-structure
surfaces (258).
4.17.3.3.2 Anti-Inflammatory Drug-Grafted Biomaterials

Coupling of anti-inflammatory drugs into biomaterial coatings offers a potential strategy to rendering biomaterials immuno-
modulatory (for review, see Kino et al. (259)). They inhibit inflammatory cell activation by abrogating the synthesis of inflam-
matory mediators, including several cytokines and chemokines, prostaglandins, leukotrienes, proteolytic enzymes, free oxygen
radicals, and NO. Simultaneously, they promote resolution of inflammation and of adaptive immune response by enhancing anti-
inflammatory cytokine release and suppressing cellar immunity in favor of humoral immunity and tolerance. More detailed
discussion can be found in Ref. (14).
NO-releasing coatings have evolved as an attractive strategy for durable control of immune responses (260,261). Continuous
and slow liberation of NO results in reduced inflammatory cell recruitment and performance at the implant surface that sustains
even after exhaustion of the NO reservoir (262–264).
4.17.3.4 Conclusions
The materials used in building a medical device must meet stringent functional requirements. Because these requirements are
cytocompatibility concerns, the desired biological activities with specific applications also should be included. As our under-
standing of immunology grows, so will the range of principles for the design of materials used in the immunotherapeutics.
Moreover, the design principles differ in various context.
4.17.4 Antibacterial Coatings

Despite extensive research and development efforts, infections related to biomedical devices and implants (e.g., catheters, heart
valves, and hip and knee implants) remain a major concern for their long-term use. Evidently bacteria can attach to and colonize
upon a broad range of indwelling biomedical devices, resulting in an infection that is difficult to treat. These infections are
associated with prolonged hospitalization, complex revision procedures, implant failure, patient suffering, financial burden, and
even death (265,266). Therefore, there is a growing need to remove bacteria or reduce bacterial colonization by equipping the
surfaces of biomedical devices and implants with chemical or topological features that are unfavorable for bacterial attachment.
Here we review a number of strategies used for the design of antibacterial coatings and discuss associated issues that arise from
using various types of coatings. Additional excellent reviews of the current progress of antibacterial coatings for specific appli-
cations are also available (267–270).
4.17.4.2 Bacterial Responses to Biomaterials

The rational design of antibacterial coatings requires an in-depth understanding of bacterial infection, which often progresses in
a defined order. Upon implantation of a foreign material, a surface protein layer was immediately formed, and this layer has been
shown to be important for bacterial adhesion, colonization, and biofilm formation. The significance of bacterial adhesion on
biomaterial-associated infection should be emphasized, because it is the starting point and an essential step in biofilm formation
and the pathogenesis of these infections. During the last years, extensive studies on initial adhesion events have been performed
theoretically and experimentally (271,272). All these findings suggest that surface topological, chemical, and hydrophobic/
hydrophilic characteristics play an important role in the number and colonization rate of bacteria adhered. For more details
about how these factors influence bacterial adhesion, consult the excellent reviews previously published (22,273). Biofilms are
formed when bacteria come together to constitute a community that is attached to and encased in an exopolysaccharide matrix.
Concerning the pivotal role that biofilm plays in implant-associated infections, the process of biofilm formation has been well
documented (274–276), which will not be detailed here. The biofilm protects adherent bacteria from the host defense system and
bactericidal agents via several proposed mechanisms (23,277). Thus, the biofilms formed on biomaterial surfaces are much more
difficult to eradicate by bactericidal methods than circulating bacteria.
According to the bacterial responses to biomaterial surfaces, we could identify two objectives that are comparatively easier to
combat for preventing infection: (1) bacterial adhesion process influenced by adsorbed protein layer and surface characteristics, and
(2) adherent bacteria without forming exopolysaccharide matrix. In the following section, we give a brief review of strategies
employed to create antibacterial coatings.
www.iran-mavad.com
4.17.4.3 Approaches to Antibacterial Coatings

In light of the above findings, strategies to purposely avoid infection by either killing the adhered microorganisms or inhibiting
bacterial adhesion must be developed. In this section, both of these two strategies are segmented into several categories and briefly
described.
4.17.4.3.1 Killing Bacteria by Bactericidal Agents

It has been reported that active killing of bacteria can be achieved by antimicrobial agents that are released from or immobilized on
the biomaterial surfaces in a controlled manner. In one respect, a great number of organic compounds and inorganic ions (e.g., silver
ions) are known to possess bactericidal activities in solution. Excellent reviews on the development of antibiotic-releasing devices are
well described (278–280). The primary advantage of delivering antibiotics locally is the ability to administrate the high local doses
without exceeding the systemic toxicity level of the drug (271,281). An additional advantage of such coatings is the ease of selection of
antibiotics targeted for specific pathogens. Recent studies showed that various antimicrobial agents that have been used in controlled-
release system include vancomycin, tobramycin, cephalothin, and amoxicillin (282,283); however, by far the most commonly used is
silver. Although such coatings can benefit implants used in revision surgeries, their major limitations such as finite release in case of
permanent implant and local tissue toxicity induced by high level of antimicrobial agents promoted a search for options that are
antibacterial over the lifetime of an implant. Covalent immobilization of antibiotics to biomaterials, unlike elution systems, emerged
as a powerful therapeutic tool that potentially increases the duration of antimicrobial efficacy (for review, see Noreen et al. (284))
(267). Over the last several years, a wide range of antimicrobial agents (quaternary ammonium polymers, chitosan, enzymes, and
peptides) have been immobilized on a variety of materials (metals, plastics, as well as natural and artificial fabrics). In the following
section, organic antibiotics-, silver-, and NO-incorporated antibacterial coatings are presented as representative examples.
4.17.4.3.1.1 Antibiotic-releasing coatings

Antibiotic-releasing system is the most commonly used approach for local delivery of antibiotics to treat biomaterial-associated
infections. However, ideal release kinetics is still a major challenge for further development of delivery systems. Design principles
for this type of system include optimizing the local releasing levels, maintaining sustained release profiles, and improving the
loading efficacy of material carriers. During the last decades, great efforts have been expended to identify the biomaterials which can
deliver antibiotics locally, ranging from inorganic to organic and from natural to synthetic. Detailed reviews on antibiotic-releasing
therapy for biomedical devices are available (304,309).
Generally, materials or biomedical devices chosen for delivery system construction should be biocompatible, which is
a preliminary requirement. Further requirements include selecting appropriate antibiotic according to its antibacterial spec-
trum, specific application, as well as the targeted microorganisms. Owing to their broad antibacterial spectra, aminoglycosides
are the most widely used antibiotics in delivery the most widely used antibiotics in antibiotics-loaded coatings on ceramics
(226,228) (titanium, hydroxyapatite, and calcium phosphate), biodegradable polymers (285), and bioresorbable films (286).
There were several published studies established on the use of these drugs for dental or orthopedic implant coatings. Repre-
sentatively, gentamicin and tobramycin-loaded biomaterials have been frequently investigated and the results revealed that
tobramycin-loaded biomaterials had higher antibacterial activities, less toxicities, and superior elution kinetics than genta-
micin-loaded ones (311-314). However, the high local concentration is the major drawback limiting clinical application of this
type of systems (315-318). In a recent study, 19 of 28 bacterial strains cultured directly from a clinically retrieved gentamicin-
loaded bone cement were gentamicin resistant, raising concerns for the effectiveness of gentamicin-incorporated implants
(284). Besides, other antibiotics with broad antibacterial spectra, for instance, cephalothin, carbenicillin, amoxicillin, and
vancomycin, have also been used in coatings on bone implants (295,296).
4.17.4.3.1.2 Antibiotics-graft coatings

To address the shortcomings of antibiotic-releasing coatings, there has been an increased interest in covalent attachment of
antibiotics to the implant surface to achieve long-lasting antibacterial activity. For example, vancomycin has been successfully
immobilized to titanium and remains bactericidal to Staphylococcus aureus and Staphylococcus epidermidis (268,297). In addition,
Aumsuwan et al. reported that the covalent attachment of penicillin and ampicillin to ePTFE can still maintain their antibacterial
effectiveness against bacteria (298,299). When these drugs were immobilized through a PEG-spacer, the surfaces displayed high
antimicrobial efficiency among their spectrum of activity, indicating that antibiotic mobility is essential to its activity. Despite
long-lasting antimicrobial activity, low incidence of side effects, and nonaccumulation in tissues, the effectiveness of coatings
with classical antibiotics is strongly dependent on the nature of the chosen drug and may be inhibited by antimicrobial resistance
in a relatively short time period (268,300). Therefore, alternatives must be developed. There is a clear need for a broad-spectrum
antimicrobial that prevents bacterial colonization, minimizes the development of bacterial resistance, displays long-term
stability, and has a low cytotoxic profile. Silver-containing coatings have the potential to meet these criteria and therefore
represent a promise for the new generation of antimicrobial surfaces.
4.17.4.3.1.3 Silver-containing coatings

Considerable attention has focused for several decades on the use of silver as an antibacterial agent (301–303). The antibacterial
properties of silver have been traced back to ancient Greeks. Currently, silver-containing agents or materials are widely used in
www.iran-mavad.com
clinical wound dressings (e.g., silver sulfadiazine), bandages for burns and chronic wounds, and the coatings of biomedical
materials (silver- and silver nanoparticle-coated catheters (304)).
A number of chemical forms of silver are known to exhibit antimicrobial activities. The bactericidal effect of silver ions
(Agþ), with its broad spectrum of activity including bacterial, fungal, and viral agents, can be achieved at submicromolar
concentrations. In addition, bulk silver has also been reported to be antimicrobial due to its surface oxide layer and/or release
of silver species. Apart from silver and silver ions, there have been substantial research efforts toward engineering biomaterials
loaded with silver and silver nanoparticles in order to achieve a more prolonged release. Through proteomic analyses, Yuang
et al. have identified a possible mode of action underlying the antibacterial action of the silver nanoparticle, similar to that of
Agþ, by proteomic analysis (305). And the results also demonstrated that the antimicrobial efficacy of the Ag nanoparticle is
much higher than that of Agþ. In a more recent study, Kong et al. synthesized silver nanoparticles-incorporated poly(methyl
methacrylate) nanofibers by radical-mediated dispersion polymerization and investigated antibacterial properties against
Gram-negative (Escherichia coli) and Gram-positive (S. aureus) bacteria; further minimum inhibitory concentration results
showed that the silver-loaded nanofibers have superior antimicrobial efficacy compared with that of silver sulfadiazine and
silver nitrate at the same silver concentration (306). In addition, silver nanoparticles were also synthesized in covalently linked,
layer-by-layer polymeric assemblies from methoxysilane polymers (307), in methoxy polyethylene glycol (308), epoxy polymer
matrix (309), poly(ether urethanes) (310), poly(vinyl alcohol) (311), polytetrafluoroethylene (312), and polyrhodanine
nanofibers (313), as concluded by Krasimir et al. (291). Apart from fibers and polymers, silver has also been ion implanted into
titanium and Ti–Al–Nb alloy to improve their antibacterial activities and wear performance (314, see also Chapter 4.14,
Coatings for Wear Protection). Liu et al. obtained Ag-incorporated titanium by plasma immersion ion implantation technique
and proposed that the antimicrobial activity of silver nanoparticles embedded in titanium may be ascribed to the microgalvanic
effects between metallic silver and titanium (315). Additionally, a titanium–silver hard coating has been deposited on titanium
by physical vapor deposition (316), and a silver-containing hydroxyapatite coating has been produced on titanium by
magnetron sputtering leading to eventually enhanced bactericidal and mechanical properties (317), see also Chapter 4.03,
Magnetron Sputtering.
4.17.4.3.1.4 Quaternary ammonium compound-bound coatings

Inspirited by the pioneering work of Tiller et al., quaternary ammonium compounds (QACs) have attracted considerable interest
over the past several decades (318). Although two main hypotheses have been proposed to explain the antibacterial mechanisms of
QACs, both of them are closely correlated to bacterial membrane distoration, disruption, and plasma leakage, which may be
induced by either long cationic polymer chains (268, (318–319) or ion exchange between the positive charges on the surface and
structurally essential mobile cations within the membrane (268,320). In an elegant experiment, Murata et al. attempted to weigh
both hypotheses (321). It was concluded that the density of cationic surface charges is a key parameter; however, the authors did not
exclude the membrane insertion of alkyl chains as a possible mechanism of action.
The development of various QACs and methods for their applications have been widely reported. Silane-based QACs have
received substantial attention because of the possibility of immobilization onto metal oxides on orthopedic implant surfaces
(321–325). Moreover, the silane group anchors covalently to the substrate and the quaternary ammonium group remains
available for bioactivity. Alternatively, quaternary ammonium thiol derivatives also can be grafted onto metal surfaces (326).
Additionally, numerous methods have been developed to integrate QACs into biomaterial surfaces or biomedical devices
including sol–gel process, via a covalent hydrolyzable ester linkage, atom transfer radical polymerization, plasma polymerization,
and layer-by-layer deposition (327–331).
However, cytotoxicity remains the major concern of QACs related to its clinical use for medical devices (324), which should be
overcome in further explorations.
4.17.4.3.1.5 Nitric oxide-containing coatings

Nitric oxide (NO), which plays a critical role in modulating the natural immune response to infection, now is identified as
a new antibacterial agent. As discussed in literatures, there are multiple mechanisms that can be used to explain the anti-
bacterial activities of nitric oxide including disruption of iron-sulfur clusters in bacterial proteins and inhibition of proteins in
the bacterial respiratory chain or DNA synthesis (332,333). In a recent study, NO-releasing sol–gel polymer coatings for
orthopedic devices showed a significant reduction of adhesion of Pseudomonas aeruginosa, S. aureus, and S. epidermidis (334).
Similar results were also observed by Charville et al. (335). Due to the multiple roles of NO in the biological responses,
polymers incorporating NO-releasing/generating moieties benefit from the combined effects of excellent antibacterial activity
and enhanced hemocompatibility of blood-contacting medical devices (336).
4.17.4.3.2 Inhibiting Bacterial Adhesion

The second strategy to reduce the incidence of infection involves the prevention or reduction of bacterial adhesion to
biomaterials. Careful consideration of the physicochemical factors that can increase or decrease bacterial adhesion to
a biomaterial may be applied in the manufacturing process of the biomaterial. This is normally achieved by surface modifi-
cation of biomaterials, which effects a change in protein adsorption, subsequently affecting both bacterial and cellular inter-
actions. In the following section, biomaterial surface modification strategies that aim to reduce or eliminate bacterial adhesion
are briefly described.
www.iran-mavad.com
4.17.4.3.2.1 Physicochemically modified coatings

Modification of surface physicochemical properties provides a relatively simple and economic way to inhibit bacterial adhesion. It
has been reported that ultraviolet light irradiation is an effective approach to create wettability on titanium dioxide (337).
A previous in vitro experiment demonstrated that UV treatment of Ti6Al4V inhibits bacterial adhesion without compromising the
good response of human bone-forming cells to this alloy (338). Further in vivo studies revealed that UV light pretreatment
of titanium substantially enhances its osteoconductive capacity, possibly in association with UV-catalyzed progressive removal of
hydrocarbons from the TiO2 surface (339).
More recently, plasma treatment has attracted much attention as a potent approach to fabricate antibacterial coatings (for review,
see Jacobs et al. (340)). Several groups have studied plasma created antibacterial surfaces of medical-grade PVC to be applied as
endotracheal tubes (341–345). An oxygen glow discharge has been used to modify small coupons of PVC to prevent the adhesion of
several P. aeruginosa stains (341,342). This treatment made the surface more hydrophilic, and a 57–70% reduction in bacterial
adhesion was observed, which is believed to be attributed to the incorporation of oxygenated functional groups. However, the
disadvantage is its inability to prevent biofilm formation.
Aside from wettability, resistivity is also considered to be correlated with bacterial adhesion for conducting surfaces. It has been
reported that TiNOx coatings can be applied to metals to alter the surface conductivity. A surface resistivity of around 104 mU cm is
found to show the minimal bacterial adhesion onto TiNOx for all strains observed (346), suggesting a new direction for producing
adhesion-resistant surfaces.
4.17.4.3.2.2 Nonfouling coatings

Nonfouling materials that resist nonspecific protein adsorption have shown promise in reducing bacterial adhesion (347–350). Of
the nonfouling surfaces developed, polymer brush coatings have shown the greatest efficacy in reducing bacterial adhesion, with
reductions observed in the orders of magnitude (351). Hydrophilic polymer brushes are highly hydrated in aqueous solution, and it
is hypothesized that the bacteria are hindered in their approach due to the hydrated layer formed with the brushes, which
compresses as a result of the bacterial approach and reduces the motility of the polymer brush. Under these conditions the
approaching bacterium is repelled and bacterial adhesion either does not occur or is reduced. Numerous examples of polymer
brushlike surfaces have been developed and shown in the literature to reduce bacterial adhesion, such as PEG, poly(methacrylic
acid), and zwitterionic materials (352,353). For example, Jiang et al. has investigated the bacterial adhesion and biofilm formation
of a long-chain zwitterionic poly(sulfobetaine methacrylate) (pSBMA) surfaces (354). The results showed that this zwitterionic
pSBMA could dramatically reduce the adhesion and accumulation of Staphylococcus epidermidis and P. aeruginosa. This reduction can
be ascribed to its intrinsically strong hydration via electrostatic interactions.
Though polymer brushes are effective at reducing bacterial adhesion, they do not completely eliminate bacteria from material
surfaces in vitro.
4.17.4.4 Conclusions and Future Work

A wide variety of strategies have been investigated for preparation of antibacterial coatings for biomaterials/devices, however, it is
still years away from practical application. This may be mainly ascribed to the simplified environments used in the in vitro tests and
complexity in real infection situation. Therefore, minimizing the gap between in vitro and in vivo experimental environments is also
a direction to be pursued, which can assist in solving the clinical infection problems.
4.17.5 Summary and Perspectives
Although remarkable progress has been made on biomaterial research, the ideal biomaterial that satisfies all the technical
requirements and biological functions is not available up to now. This is mainly due to the fact that little is known about how
a biomaterial surface has to be designed to exert perfect biocompatibility, considering the interconnection of different stages of
foreign body reactions. This chapter shows that the response of single parameter of the foreign body reactions may be well
controlled by well-defined surface modification. Therefore, further efforts should be devoted to develop multifunctional and
specific biomaterials.
Acknowledgment
The authors would like to acknowledge the joint financial support from the National Basic Research Program of China (973
Program, 2012CB933600) and National Natural Science Foundation of China (31200721) for financial support.
References
1. Peppas, N. A.; Langer, R. New Challenges in Biomaterials. Science 1994, 263 (5154), 1715–1720.
2. Huebsch, N.; Mooney, D. J. Inspiration and Application in the Evolution of Biomaterials. Nature 2009, 462 (7272), 426–432.
www.iran-mavad.com
3. Williams, D. F. Definitions in Biomaterials. In Proceedings of a Consensus Conference of the European Society for Biomaterials, England, Vol. 4; Elsevier: New
York, 1987.
4. Helmus, M. N.; Gibbons, D. F.; Cebon, D. Biocompatibility: Meeting a Key Functional Requirement of Next-Generation Medical Devices. Toxicol. Pathol. 2008, 36 (1),
70–80.
5. Williams, D. F. On the Mechanisms of Biocompatibility. Biomaterials 2008, 29 (20), 2941–2953.
This paper gives an update on understanding of biocompatibility.
6. Ratner, B. D. New Ideas in Biomaterials Science – A Path to Engineered Biomaterials. J. Biomed. Mater. Res. 1993, 27 (7), 837–850.
7. Hench, L. L.; Wilson, J. Surface-Active Biomaterials. Science 1984, 226 (4675), 630–636.
8. Nel, A. E.; Madler, L.; Velegol, D.; Xia, T.; Hoek, E. M. V.; Somasundaran, P.; Klaessig, F.; Castranova, V.; Thompson, M. Understanding Biophysicochemical Interactions
at the Nano-Bio Interface. Nat. Mater. 2009, 8 (7), 543–557.
9. Klee, D.; Hocker, H. Polymers for Biomedical Applications: Improvement of the Interface Compatibility. Adv. Polym. Sci. 1999, 149, 1–57.
10. Spector, M., (chapter 4.401). The Concept of Biocompatibility; In Comprehensive Biomaterials; Ducheyne, P., Healy, K. E., Hutmacher, D. W., Grainger, D. W.,
Kirkpatrick, C. J., Eds.; Elsevier Ltd.: Amsterdam, Boston, 2011; Vol. 4.
11. Anderson, J.; McNally, A. Biocompatibility of Implants: Lymphocyte/Macrophage Interactions. Semin. Immunopathol. 2011, 33 (3), 221–233.
12. Fournier, E.; Passirani, C.; Montero-Menei, C. N.; Benoit, J. P. Biocompatibility of Implantable Synthetic Polymeric Drug Carriers: Focus on Brain Biocompatibility.
Biomaterials 2003, 24 (19), 3311–3331.
13. Anderson, J. M. Biological Responses to Materials. Annu. Rev. Mater. Res. 2001, 31, 81–110.
This is a comprehensive review about biological responses to biomaterials.
14. Franz, S.; Rammelt, S.; Scharnweber, D.; Simon, J. C. Immune Responses to Implants – A Review of the Implications for the Design of Immunomodulatory Biomaterials.
Biomaterials 2011, 32 (28), 6692–6709.
This review presents strategies of triggering appropriate immune response by functional biomaterials and highlights recent approaches of food
biomaterials that mimic the physiological extracellular matrix.
15. Anderson, J. M.; Rodriguez, A.; Chang, D. T. Foreign Body Reaction to Biomaterials. Semin. Immunol. 2008, 20 (2), 86–100.
16. Murphy, H. S., (chapter 3). Inflammation. In Rubin’s Pathology: Clinicopathologic Foundations of Medicine; Rubin, R., Strayer, D. S., Eds; Lippincott Williams & Wilkins,
2008; Vol. 3, pp 37–70.
17. Horbett, T. A., (chapter 4). Host Reactions to Biomaterials and Their Evaluation. In Biomaterials Science: An Introduction to Materials in Medicine; Ratner, B. D.,
Hoffman, A. S., Schoen, F. J., Lemons, J. E., Eds., 2nd ed.; Elsevier Academic Press: San Diego, 2004; pp 293--354.
18. Gorbet, M. B.; Sefton, M. V. Biomaterial-Associated Thrombosis: Roles of Coagulation Factors, Complement, Platelets and Leukocytes. Biomaterials 2004, 25 (26),
5681–5703.
19. Sivaraman, B.; Fears, K. P.; Latour, R. A. Investigation of the Effects of Surface Chemistry and Solution Concentration on the Conformation of Adsorbed Proteins Using an
Improved Circular Dichroism Method. Langmuir 2009, 25 (5), 3050–3056.
20. Roach, P.; Farrar, D.; Perry, C. C. Interpretation of Protein Adsorption: Surface-Induced Conformational Changes. J. Am. Chem. Soc. 2005, 127 (22), 8168–8173.
21. Pascual, A. Pathogenesis of Catheter-Related Infections: Lessons for New Designs. Clin. Microbiol. Infect. 2002, 8 (5), 256–264.
This study is a demonstration about biomedical device-related infection.
22. Katsikogianni, M.; Missirlis, Y. F. Concise Review of Mechanisms of Bacterial Adhesion to Biomaterials and of Techniques Used in Estimating Bacteria-Material Infections.
Eur. Cells Mater. 2004, 8, 37–57.
23. Donlan, R. M.; Costerton, J. W. Biofilms: Survival Mechanisms of Clinically Relevant Microorganisms. Clin. Microbiol. Rev. 2002, 15 (2), 167–193.
24. Martinez, J. L.; Baquero, F. Interactions among Strategies Associated with Bacterial Infection: Pathogenicity, Epidemicity, and Antibiotic Resistance. Clin. Microbiol. Rev.
2002, 15 (4), 647–679.
25. Ratner, B. D. The Catastrophe Revisited: Blood Compatibility in the 21st Century. Biomaterials 2007, 28 (34), 5144–5147.
This paper provides a markedly improved understanding of mechanisms of biocompatibility.
26. Eberhart, R. C.; Clagett, C. P. Catheter Coatings, Blood-Flow, and Biocompatibility. Semin. Hematol. 1991, 28 (4), 42–48.
27. Schaadt, J. Oxygenator Thrombosis: An International Phenomenon. Perfusion 1999, 14 (6), 425–435.
28. Belway, D.; Rubens, F. D. Currently Available Biomaterials for Use in Cardiopulmonary Bypass. Expert Rev. Med. Devices 2006, 3 (3), 345–355.
29. Liu, Z.; Jiao, Y.; Wang, T.; Zhang, Y.; Xue, W. Interactions between Solubilized Polymer Molecules and Blood Components. J. Controlled Release 2012, 160, 11.
30. Deppisch, R.; Storr, M.; Buck, R.; Gohl, H. Blood Material Interactions at the Surfaces of Membranes in Medical Applications. Sep. Purif. Technol. 1998, 14 (1–3), 241–254.
31. Hoffman, A. S. Blood-Biomaterial Interactions: An Overview. In Biomaterials: Interfacial Phenomena and Applications. Advances in Chemistry Series; Cooper, S. L., Peppas, N. A.,
Hoffman, A. S., Ratner, B. D., Eds.; American Chemical Society: Washington, 1982; Vol. 199, pp 3–8.
32. Dee, K. C.; Puleo, D. A.; Bizios, R.., (chapter 4.7). Biomaterials, Devices, and Thrombosis. In An Introduction to Tissue-Biomaterial Interactions; John Wiley & Sons, Inc.: New
York, 2002; pp 81–85.
33. Werner, C.; Maitz, M. F.; Sperling, C. Current Strategies towards Hemocompatible Coatings. J. Mater. Chem. 2007, 17 (32), 3376–3384.
In this paper, progress of hemocompatible coatings was reviewed.
34. Keefer, L. K. Biomaterials – Thwarting Thrombus. Nat. Mater. 2003, 2 (6), 357–358.
35. Noh, H.; Vogler, E. A. Volumetric Interpretation of Protein Adsorption: Competition from Mixtures and the Vroman Effect. Biomaterials 2007, 28 (3), 405–422.
36. Horbett, T. A. The Role of Adsorbed Proteins in Tissue Response to Biomaterials. In Biomaterials Science: An Introduction to Materials in Medicine; Ratner, B. D., Hoffman, A. S.,
Schoen, F. J., Lemons, J. E., Eds., 2nd ed.; Elsevier Academic Press: San Diego, 2004; pp 237–245.
37. Vogler, E. A.; Siedlecki, C. A. Contact Activation of Blood-Plasma Coagulation. Biomaterials 2009, 30 (10), 1857–1869.
This review shows that the consensus biochemical mechanism of contact activation of blood-plasma coagulation that has long served as a rational for poor
hemocompatibility is an inadequate basis for surface engineering of advanced cardiovascular biomaterials.
38. Furie, B.; Furie, B. C. The Molecular-Basis of Blood-Coagulation. Cell 1988, 53 (4), 505–518.
39. Markiewski, M. M.; Nilsson, B.; Ekdahl, K. N.; Mollnes, T. E.; Lambris, J. D. Complement and Coagulation: Strangers or Partners in Crime? Trends Immunol. 2007, 28 (4),
184–192.
40. Granada, J. F.; Price, M. J.; French, P. A.; Steinhubl, S. R.; Cutlip, D. E.; Becker, R. C.; Smyth, S. S.; Dauerman, H. L. Platelet-Mediated Thrombosis and Drug-Eluting Stents.
Circ. Cardiovasc. Interventions 2011, 4 (6), 629–637.
41. Monroe, D.; Hoffman, M.; Roberts, H. R. Platelets and Thrombin Generation. Arterioscler. Thromb. Vasc. Biol. 2002, 22 (9), 1381–1389.
42. Sperling, C.; Fischer, M.; Maitz, M. F.; Werner, C. Blood Coagulation on Biomaterials Requires the Combination of Distinct Activation Processes. Biomaterials 2009, 30 (27),
4447–4456.
43. Seyfert, U. T.; Biehl, V.; Schenk, J. In Vitro Hemocompatibility Testing of Biomaterials According to the ISO 10993-4. Biomol. Eng. 2002, 19 (2–6), 91–96.
This is an FDA standard for blood-contacting biomedical devices.
44. Hlady, V.; Buijs, J. Protein Adsorption on Solid Surfaces. Curr. Opin. Biotech. 1996, 7 (1), 72–77.
45. Dee, K. C.; Puleo, D. A.; Bizios, R. Protein-Surface Interactions. In An Introduction to Tissue-Biomaterial Interactions; John Wiley & Sons, Inc.: New York, 2002; pp
37–52.
www.iran-mavad.com
46. Zhuo, R.; Siedlecki, C. A.; Vogler, E. A. Autoactivation of Blood Factor XII at Hydrophilic and Hydrophobic Surfaces. Biomaterials 2006, 27 (24), 4325–4332.
This paper discusses the role of surface energy in the blood coagulation cascade.
47. Golas, A.; Parhi, P.; Dimachkie, Z. O.; Siedlecki, C. A.; Vogler, E. A. Surface-Energy Dependent Contact Activation of Blood Factor XII. Biomaterials 2010, 31 (6),
1068–1079.
48. Hanson, S. R. Blood Coagulation and Blood Material Interactions. In Biomaterials Science: An Introduction to Materials in Medicine; Ratner, B. D., Hoffman, A. S.,
Schoen, F. J., Lemons, J. E., Eds., 2nd ed.; Elsevier Academic Press: San Diego, 2002; pp 332–337.
49. Dee, K. C.; Puleo, D. A.; Bizios, R. The Coagulation Cascades. In An Introduction to Tissue-Biomaterial Interactions; John Wiley & Sons, Inc.: New York, 2002;
pp 73–79.
50. Schmaier, A. H. Plasma Kallikrein/Kinin System: A Revised Hypothesis for Its Activation and Its Physiologic Contributions. Curr. Opin. Hematol. 2000, 7 (5),
261–265.
51. Davie, E. W.; Fujikawa, K.; Kisiel, W. The Coagulation Cascade – Initiation, Maintenance, and Regulation. Biochemistry 1991, 30 (43), 10363–10370.
52. Zhuo, R.; Siedlecki, C. A.; Vogler, E. A. Competitive-Protein Adsorption in Contact Activation of Blood Factor XII. Biomaterials 2007, 28 (30), 4355–4369.
53. Mackman, N. Triggers, Targets and Treatments for Thrombosis. Nature 2008, 451 (7181), 914–918.
54. Ruggeri, Z. M. Mechanisms Initiating Platelet Thrombus Formation. Thromb. Haemostasis 1997, 78 (1), 611–616.
55. Ruggeri, Z. M. Role of von Willebrand Factor in Platelet Thrombus Formation. Ann. Med. 2000, 32, 2–9.
56. Weiss, H. J.; Hawiger, J.; Ruggeri, Z. M.; Turitto, V. T.; Thiagarajan, P.; Hoffmann, T. Fibrinogen-Independent Platelet-Adhesion and Thrombus Formation on Sub-
endothelium Mediated by Glycoprotein IIb-IIIa Complex at High Shear Rate. J. Clin. Invest. 1989, 83 (1), 288–297.
57. Sivaraman, B.; Latour, R. A. The Relationship between Platelet Adhesion on Surfaces and the Structure versus the Amount of Adsorbed Fibrinogen. Biomaterials 2010,
31 (5), 832–839.
58. Zaidi, T. N.; McIntire, L. V.; Farrell, D. H.; Thiagarajan, P. Adhesion of Platelets to Surface-Bound Fibrinogen under Flow. Blood 1996, 88 (8), 2967–2972.
59. Andersen, H.; Greenberg, D. L.; Fujikawa, K.; Xu, W. F.; Chung, D. W.; Davie, E. W. Protease-Activated Receptor 1 is the Primary Mediator of Thrombin-Stimulated
Platelet Procoagulant Activity. Proc. Natl. Acad. Sci. U. S. A 1999, 96 (20), 11189–11193.
60. Furie, B.; Furie, B. C. Mechanisms of Thrombus Formation. N. Engl. J. Med. 2008, 359 (9), 938–949.
61. Brass, L. F. Thrombin and Platelet Activation. Chest 2003, 124 (3), 18s–25s.
This paper gives a detailed discussion about platelet activation.
62. Ekdahl, K. N.; Lambris, J. D.; Elwing, H.; Ricklin, D.; Nilsson, P. H.; Teramura, Y.; Nicholls, I. A.; Nilsson, B. Innate Immunity Activation on Biomaterial Surfaces: A
Mechanistic Model and Coping Strategies. Adv. Drug Delivery Rev. 2011, 63 (12), 1042–1050.
This review presents models for biomaterial-induced activation processes and describes various strategies.
63. Engberg, A. E.; Rosengren-Holmberg, J. P.; Chen, H.; Nilsson, B.; Lambris, J. D.; Nicholls, I. A.; Ekdahl, K. N. Blood Protein-Polymer Adsorption: Implications for
Understanding Complement-Mediated Hemoincompatibility. J. Biomed. Mater. Res. A 2011, 97A (1), 74–84.
64. Hulander, M.; Lundgren, A.; Berglin, M.; Ohrlander, M.; Lausmaa, J.; Elwing, H. Immune Complement Activation is Attenuated by Surface Nanotopography. Int. J.
Nanomed. 2011, 6, 2653–2666.
65. Nilsson, B.; Ekdahl, K. N.; Mollnes, T. E.; Lambris, J. D. The Role of Complement in Biomaterial-Induced Inflammation. Mol. Immunol. 2007, 44 (1–3), 82–94.
66. Marceau, F.; Hugli, T. E. Effect of C3a and C5a Anaphylatoxins on Guinea-Pig Isolated Blood-Vessels. J. Pharmacol. Exp. Ther. 1984, 230 (3), 749–754.
67. Nilsson, B.; Korsgren, O.; Lambris, J. D.; Ekdahl, K. N. Can Cells and Biomaterials in Therapeutic Medicine be Shielded from Innate Immune Recognition? Trends
Immunol. 2010, 31 (1), 32–38.
This paper discusses the new technologies that are being developed to shield the biomaterial and cell surfaces from recognition by the innate immune
system.
68. de Mel, A.; Cousins, B. G.; Seifalian, A. M. Surface Modification of Biomaterials: A Quest for Blood Compatibility. Int. J. Biomater. 2012, 707863.
69. Li, S.; Henry, J. J. D. Nonthrombogenic Approaches to Cardiovascular Bioengineering. Annu. Rev. Biomed. Eng. 2011, 13, 451–475.
This paper discusses the progress of nonthrombogenic therapeutics and the relevant mechanisms.
70. Schopka, S.; Schmid, T.; Schmid, C.; Lehle, K. Current Strategies in Cardiovascular Biomaterial Functionalization. Materials 2010, 3 (1), 638–655.
71. Chiang, Y. C.; Chang, Y.; Higuchi, A.; Chen, W. Y.; Ruaan, R. C. Sulfobetaine-Grafted Poly(Vinylidene Fluoride) Ultrafiltration Membranes Exhibit Excellent Antifouling
Property. J. Membrane. Sci. 2009, 339 (1–2), 151–159.
72. Langer, R.; Tirrell, D. A. Designing Materials for Biology and Medicine. Nature 2004, 428 (6982), 487–492.
73. Goodman, S. L.; Tweden, K. S.; Albrecht, R. M. Platelet Interaction with Pyrolytic Carbon Heart-Valve Leaflets. J. Biomed. Mater. Res. 1996, 32 (2), 249–258.
74. Yu, L. J.; Wang, X.; Wang, X. H.; Liu, X. H. Haemocompatibility of Tetrahedral Amorphous Carbon Films. Surf. Coat. Technol. 2000, 128, 484–488.
75. Yang, P.; Kwok, S. C. H.; Chu, P. K.; Leng, Y. X.; Chen, J. Y.; Wang, J.; Huang, N. Haemocompatibility of Hydrogenated Amorphous Carbon (a-C: H) Films Synthesized by
Plasma Immersion Ion Implantation-Deposition. Nucl. Instrum. Methods Phys. Res. B 2003, 206, 721–725.
76. Gutensohn, K.; Beythien, C.; Bau, J.; Fenner, T.; Grewe, P.; Koester, R.; Padmanaban, K.; Kuehnl, P. In Vitro Analyses of Diamond-Like Carbon Coated Stents: Reduction
of Metal Ion Release, Platelet Activation, and Thrombogenicity. Thromb. Res. 2000, 99 (6), 577–585.
77. Fedel, M.; Motta, A.; Maniglio, D.; Migliaresi, C. Surface Properties and Blood Compatibility of Commercially Available Diamond-Like Carbon Coatings for Cardiovascular
Devices. J. Biomed. Mater. Res. B 2009, 90B (1), 338–349.
78. Robertson, J. Diamond-Like Amorphous Carbon. Mater. Sci. Eng. R 2002, 37 (4–6), 129–281.
79. Bolz, A.; Schaldach, M. Artificial Heart Valves: Improved Blood Compatibility by PECVD a-SiC: H Coating. Artif. Organs 1990, 14 (4), 260–269.
80. Roy, R. K.; Choi, H. W.; Yi, J. W.; Moon, M. W.; Lee, K. R.; Han, D. K.; Shin, J. H.; Kamijo, A.; Hasebe, T. Hemocompatibility of Surface-Modified, Silicon-Incorporated,
Diamond-Like Carbon Films. Acta Biomater. 2009, 5 (1), 249–256.
81. Roy, R. K.; Choi, H. W.; Park, S. J.; Lee, K. R. Surface Energy of the Plasma Treated Si Incorporated Diamond-Like Carbon Films. Diam. Relat. Mater. 2007, 16 (9),
1732–1738.
82. Saito, T.; Hasebe, T.; Yohena, S.; Matsuoka, Y.; Kamijo, A.; Takahashi, K.; Suzuki, T. Antithrombogenicity of Fluorinated Diamond-Like Carbon Films. Diam. Relat. Mater.
2005, 14 (3–7), 1116–1119.
83. Hasebe, T.; Shimada, A.; Suzuki, T.; Matsuoka, Y.; Saito, T.; Yohena, S.; Kamijo, A.; Shiraga, N.; Higuchi, M.; Kimura, K.; Yoshimura, H.; Kuribayashi, S. Fluorinated
Diamond-Like Carbon as Antithrombogenic Coating for Blood-Contacting Devices. J. Biomed. Mater. Res. A 2006, 76A (1), 86–94.
84. Hasebe, T.; Yohena, S.; Kamijo, A.; Okazaki, Y.; Hotta, A.; Takahashi, K.; Suzuki, T. Fluorine Doping into Diamond-Like Carbon Coatings Inhibits Protein Adsorption and
Platelet Activation. J. Biomed. Mater. Res. A 2007, 83A (4), 1192–1199.
85. Hasebe, T.; Ishimaru, T.; Kamijo, A.; Yoshimoto, Y.; Yoshimura, T.; Yohena, S.; Kodama, H.; Hotta, A.; Takahashi, K.; Suzuki, T. Effects of Surface Roughness on Anti-
Thrombogenicity of Diamond-Like Carbon Films. Diam. Relat. Mater. 2007, 16 (4–7), 1343–1348.
86. Kwok, S. C. H.; Jin, W.; Chu, P. K. Surface Energy, Wettability, and Blood Compatibility Phosphorus Doped Diamond-Like Carbon Films. Diam. Relat. Mater. 2005, 14 (1),
78–85.
87. Kwok, S. C. H.; Yang, P.; Wang, J.; Liu, X. Y.; Chu, P. K. Hemocompatibility of Nitrogen-Doped, Hydrogen-Free Diamond-Like Carbon Prepared by Nitrogen Plasma
Immersion Ion Implantation-Deposition. J. Biomed. Mater. Res. A 2004, 70A (1), 107–114.
88. Maitz, M. F.; Gago, R.; Abendroth, B.; Camero, M.; Caretti, I.; Kreissig, U. Hemocompatibility of Low-Friction Boron-Carbon-Nitrogen Containing Coatings. J. Biomed.
Mater. Res. B 2006, 77B (1), 179–187.
www.iran-mavad.com
89. Chen, J. Y.; Leng, Y. X.; Tian, X. B.; Wang, L. P.; Huang, N.; Chu, P. K.; Yang, P. Antithrombogenic Investigation of Surface Energy and Optical Bandgap and
Hemocompatibility Mechanism of Ti(Taþ5)O-2 Thin Films. Biomaterials 2002, 23 (12), 2545–2552.
90. Okpalugo, T. I. T.; Ogwu, A. A.; Maguire, P. D.; McLaughlin, J. A. D. Platelet Adhesion on Silicon Modified Hydrogenated Amorphous Carbon Films. Biomaterials 2004,
25 (2), 239–245.
91. Zhang, L.; Lv, P.; Huang, Z. Y.; Lin, S. P.; Chen, D. H.; Pan, S. R.; Chen, M. Blood Compatibility of La2O3 Doped Diamond-Like Carbon Films. Diam. Relat. Mater. 2008,
17 (11), 1922–1926.
92. Roy, R. K.; Lee, K. R. Biomedical Applications of Diamond-Like Carbon Coatings: A Review. J. Biomed. Mater. Res. B 2007, 83B (1), 72–84.
93. Sunny, M. C.; Sharma, C. P. Titanium-Protein Interaction: Changes with Oxide Layer Thickness. J. Biomater. Appl. 1991, 6 (1), 89–98.
94. Yang, Y.; Lai, Y. K.; Zhang, Q. Q.; Wu, K.; Zhang, L. H.; Lin, C. J.; Tang, P. F. A Novel Electrochemical Strategy for Improving Blood Compatibility of Titanium-Based
Biomaterials. Colloids Surf. B 2010, 79 (1), 309–313.
95. Sun, T. L.; Tan, H.; Han, D.; Fu, Q.; Jiang, L. No Platelet Can Adhere – Largely Improved Blood Compatibility on Nanostructured Superhydrophobic Surfaces. Small 2005,
1 (10), 959–963.
96. Huang, N.; Chen, Y. R.; Luo, J. M.; Yi, J.; Lu, R.; Xiao, J.; Xue, Z. N.; Liu, X. H. In Vitro Investigation of Blood Compatibility of Ti with Oxide Layers of Rutile Structure.
J. Biomater. Appl. 1994, 8 (4), 404–412.
97. Huang, N.; Yang, P.; Leng, Y. X.; Chen, J. Y.; Sun, H.; Wang, J.; Wang, G. J.; Ding, P. D.; Xi, T. F.; Leng, Y. Hemocompatibility of Titanium Oxide Films. Biomaterials
2003, 24 (13), 2177–2187.
98. Zhang, L.; Chen, D. H.; Wang, K. Q.; Yu, F. M.; Huang, Z. Y.; Pan, S. R. Blood Compatibility Improvement of Titanium Oxide Film Modified by Doping La2O3. J. Mater. Sci.
Mater. Med. 2009, 20 (10), 2019–2023.
99. Zhang, F.; Liu, X. H.; Mao, Y. J.; Huang, N.; Chen, Y.; Zheng, Z. H.; Zhou, Z. Y.; Chen, A. Q.; Jiang, Z. B. Artificial Heart Valves: Improved Hemocompatibility by Titanium
Oxide Coatings Prepared by Ion Beam Assisted Deposition. Surf. Coat. Technol. 1998, 104, 146–150.
100. Tsyganov, I. A.; Maitz, M. F.; Richter, E.; Reuther, H.; Mashina, A. I.; Rustichelli, F. Hemocompatibility of Titanium-Based Coatings Prepared by Metal Plasma Immersion
Ion Implantation and Deposition. Nucl. Instrum. Methods Phys. Res. B 2007, 257, 122–127.
101. Petersson, I. U.; Loberg, J. E. L.; Fredriksson, A. S.; Ahlberg, E. K. Semi-Conducting Properties of Titanium Dioxide Surfaces on Titanium Implants. Biomaterials 2009,
30 (27), 4471–4479.
102. Chen, H.; Yuan, L.; Song, W.; Wu, Z. K.; Li, D. Biocompatible Polymer Materials: Role of Protein-Surface Interactions. Prog. Polym. Sci. 2008, 33 (11),
1059–1087.
103. Emmenegger, C. R.; Brynda, E.; Riedel, T.; Sedlakova, Z.; Houska, M.; Alles, A. B. Interaction of Blood Plasma with Antifouling Surfaces. Langmuir 2009, 25 (11),
6328–6333.
104. Chen, S. F.; Li, L. Y.; Zhao, C.; Zheng, J. Surface Hydration: Principles and Applications toward Low-Fouling/Nonfouling Biomaterials. Polymer 2010, 51 (23),
5283–5293.
This paper discusses the role of surface hydration in providing nonfouling properties to different nonfouling materials.
105. Ostuni, E.; Chapman, R. G.; Holmlin, R. E.; Takayama, S.; Whitesides, G. M. A Survey of Structure-Property Relationships of Surfaces that Resist the Adsorption of
Protein. Langmuir 2001, 17 (18), 5605–5620.
In parallel to experimental studies, this is a modeling investigation to explain protein-repellent properties.
106. Chapman, R. G.; Ostuni, E.; Takayama, S.; Holmlin, R. E.; Yan, L.; Whitesides, G. M. Surveying for Surfaces that Resist the Adsorption of Proteins. J. Am. Chem. Soc.
2000, 122 (34), 8303–8304.
107. Kingshott, P.; Griesser, H. J. Surfaces that Resist Bioadhesion. Curr. Opin. Solid State Mater. Sci. 1999, 4 (4), 403–412.
108. Fuertges, F.; Abuchowski, A. The Clinical Efficacy of Poly(Ethylene Glycol)-Modified Proteins. J. Controlled Release 1990, 11 (1–3), 139–148.
109. Zhang, F.; Kang, E. T.; Neoh, K. G.; Wang, P.; Tan, K. L. Modification of Si(100) Surface by the Grafting of Poly(Ethylene Glycol) for Reduction in Protein Adsorption and
Platelet Adhesion. J. Biomed. Mater. Res. 2001, 56 (3), 324–332.
110. McPherson, T. B.; Shim, H. S.; Park, K. Grafting of PEO to Glass, Nitinol, and Pyrolytic Carbon Surfaces by Gamma Irradiation. J. Biomed. Mater. Res. 1997, 38 (4),
289–302.
111. Balakrishnan, B.; Kumar, D. S.; Yoshida, Y.; Jayakrishnan, A. Chemical Modification of Poly(Vinyl Chloride) Resin Using Poly(Ethylene Glycol) to Improve Blood
Compatibility. Biomaterials 2005, 26 (17), 3495–3502.
112. Shin, H. S.; Park, K.; Kim, J. H.; Kim, J. J.; Han, D. K.; Moon, M. W.; Lee, K. R.; Shin, J. H. Biocompatible PEG Grafting on DLC-Coated Nitinol Alloy for Vascular Stents.
J. Bioact. Compat. Polym. 2009, 24 (4), 316–328.
113. Hansson, K. M.; Tosatti, S.; Isaksson, J.; Wettero, J.; Textor, M.; Lindahl, T. L.; Tengvall, P. Whole Blood Coagulation on Protein Adsorption-Resistant PEG and Peptide
Functionalised PEG-Coated Titanium Surfaces. Biomaterials 2005, 26 (8), 861–872.
114. Gulati, N.; Rastogi, R.; Dinda, A. M.; Saxena, R.; Koul, V. Characterization and Cell Material Interactions of PEGylated PNIPAAM Nanoparticles. Colloids Surf. B 2010,
79 (1), 164–173.
115. Chen, C.; Cheng, Y. C.; Yu, C. H.; Chan, S. W.; Cheung, M. K.; Yu, P. H. F. In Vitro Cytotoxicity, Hemolysis Assay, and Biodegradation Behavior of Biodegradable
Poly(3-Hydroxybutyrate)-Poly(Ethylene Glycol)-Poly(3-Hydroxybutyrate) Nanoparticles as Potential Drug Carriers. J. Biomed. Mater. Res. A 2008, 87A (2),
290–298.
116. Jokerst, J. V.; Lobovkina, T.; Zare, R. N.; Gambhir, S. S. Nanoparticle PEGylation for Imaging and Therapy. Nanomedicine 2011, 6 (4), 715–728.
A comprehensive review about PEG application in biomedical field.
117. Knop, K.; Hoogenboom, R.; Fischer, D.; Schubert, U. S. Poly(Ethylene Glycol) in Drug Delivery: Pros and Cons as Well as Potential Alternatives. Angew. Chem. Int. Ed.
2010, 49 (36), 6288–6308.
118. Claesson, P. Poly(Ethylene Oxide) Surface-Coatings – Relations between Intermolecular Forces, Layer Structure and Protein Repellency. Colloids Surf. A 1993, 77 (2),
109–118.
119. Prime, K. L.; Whitesides, G. M. Adsorption of Proteins onto Surfaces Containing End-Attached Oligo(Ethylene Oxide) – A Model System Using Self-Assembled Monolayers.
J. Am. Chem. Soc. 1993, 115 (23), 10714–10721.
120. Harder, P.; Grunze, M.; Dahint, R.; Whitesides, G. M.; Laibinis, P. E. Molecular Conformation in Oligo(Ethylene Glycol)-Terminated Self-Assembled Monolayers on Gold
and Silver Surfaces Determines Their Ability to Resist Protein Adsorption. J. Phys. Chem. B 1998, 102 (2), 426–436.
121. Han, D. K.; Jeong, S. Y.; Ahn, K. D.; Kim, Y. H.; Min, B. G. Preparation and Surface-Properties of Peo-Sulfonate Grafted Polyurethanes for Enhanced Blood Compatibility.
J. Biomater. Sci. Polym. Ed. 1993, 4 (6), 579–589.
122. Szleifer, I. Protein Adsorption on Surfaces with Grafted Polymers: A Theoretical Approach. Biophys. J. 1997, 72 (2), 595–612.
123. Lee, J. H.; Lee, H. B.; Andrade, J. D. Blood Compatibility of Polyethylene Oxide Surfaces. Prog. Polym. Sci. 1995, 20 (6), 1043–1079.
124. Malmsten, M.; Emoto, K.; Van Alstine, J. M. Effect of Chain Density on Inhibition of Protein Adsorption by Poly(Ethylene Glycol) Based Coatings. J. Colloid Interface Sci.
1998, 202 (2), 507–517.
125. Kingshott, P.; Thissen, H.; Griesser, H. J. Effects of Cloud-Point Grafting, Chain Length, and Density of PEG Layers on Competitive Adsorption of Ocular Proteins.
Biomaterials 2002, 23 (9), 2043–2056.
126. Pasche, S.; Voros, J.; Griesser, H. J.; Spencer, N. D.; Textor, M. Effects of Ionic Strength and Surface Charge on Protein Adsorption at PEGylated Surfaces. J. Phys.
Chem. B 2005, 109 (37), 17545–17552.
www.iran-mavad.com
127. Pei, J.; Hall, H.; Spencer, N. D. The Role of Plasma Proteins in Cell Adhesion to Peg Surface-Density-Gradient-Modified Titanium Oxide. Biomaterials 2011, 32 (34),
8968–8978.
128. Pasche, S.; De Paul, S. M.; Voros, J.; Spencer, N. D.; Textor, M. Poly(L-Lysine)-Graft-Poly(Ethylene Glycol) Assembled Monolayers on Niobium Oxide Surfaces: A
Quantitative Study of the Influence of Polymer Interfacial Architecture on Resistance to Protein Adsorption by ToF-SIMS and In Situ OWLS. Langmuir 2003, 19 (22),
9216–9225.
129. Jeon, S. I.; Lee, J. H.; Andrade, J. D.; Degennes, P. G. Protein Surface Interactions in the Presence of Polyethylene Oxide 1. Simplified Theory. J. Colloid Interface Sci.
1991, 142 (1), 149–158.
130. Pasche, S.; Textor, M.; Meagher, L.; Spencer, N. D.; Griesser, H. J. Relationship between Interfacial Forces Measured by Colloid-Probe Atomic Force Microscopy and
Protein Resistance of Poly(Ethylene Glycol)-Grafted Poly(L-Lysine) Adlayers on Niobia Surfaces. Langmuir 2005, 21 (14), 6508–6520.
131. Halperin, A. Polymer Brushes that Resist Adsorption of Model Proteins: Design Parameters. Langmuir 1999, 15 (7), 2525–2533.
132. Rixman, M. A.; Dean, D.; Ortiz, C. Nanoscale Intermolecular Interactions between Human Serum Albumin and Low Grafting Density Surfaces of Poly(Ethylene Oxide).
Langmuir 2003, 19 (22), 9357–9372.
133. Chang, Y.; Shih, Y. J.; Ko, C. Y.; Jhong, J. F.; Liu, Y. L.; Wei, T. C. Hemocompatibility of Poly(Vinylidene Fluoride) Membrane Grafted with Network-Like and Brush-Like
Antifouling Layer Controlled via Plasma-Induced Surface PEGylation. Langmuir 2011, 27 (9), 5445–5455.
134. Li, D.; Chen, H.; McClung, W. G.; Brash, J. L. Lysine-PEG-Modified Polyurethane as a Fibrinolytic Surface: Effect of Peg Chain Length on Protein Interactions, Platelet
Interactions and Clot Lysis. Acta Biomater. 2009, 5 (6), 1864–1871.
135. Wieland, B.; Lancaster, J. P.; Hoaglund, C. S.; Holota, P.; Tornquist, W. J. Electrochemical and Infrared Spectroscopic Quantitative Determination of the Platinum-Catalyzed
Ethylene Glycol Oxidation Mechanism at Co Adsorption Potentials. Langmuir 1996, 12 (10), 2594–2601.
136. Arima, Y.; Toda, M.; Iwata, H. Complement Activation on Surfaces Modified with Ethylene Glycol Units. Biomaterials 2008, 29 (5), 551–560.
137. Jiang, S. Y.; Cao, Z. Q. Ultralow-Fouling, Functionalizable, and Hydrolyzable Zwitterionic Materials and Their Derivatives for Biological Applications. Adv. Mater. 2010, 22 (9),
920–932.
This work discusses the progress, synthesis method, mechanism and application of zwitterionic materials.
138. Ye, S. H.; Johnson, C. A.; Woolley, J. R.; Murata, H.; Gamble, L. J.; Ishihara, K.; Wagner, W. R. Simple Surface Modification of a Titanium Alloy with Silanated Zwitterionic
Phosphorylcholine or Sulfobetaine Modifiers to Reduce Thrombogenicity. Colloids Surf. B 2010, 79 (2), 357–364.
139. Zhao, J.; Shi, Q. A.; Luan, S. F.; Song, L. J.; Yang, H. W.; Shi, H. C.; Jin, J.; Li, X. L.; Yin, J. H.; Stagnaro, P. Improved Biocompatibility and Antifouling Property of
Polypropylene Non-Woven Fabric Membrane by Surface Grafting Zwitterionic Polymer. J. Membrane. Sci. 2011, 369 (1–2), 5–12.
140. Chang, Y.; Liao, S. C.; Higuchi, A.; Ruaan, R. C.; Chu, C. W.; Chen, W. Y. A. Highly Stable Nonbiofouling Surface with Well-Packed Grafted Zwitterionic Polysulfobetaine for
Plasma Protein Repulsion. Langmuir 2008, 24 (10), 5453–5458.
141. Kuang, J. H.; Messersmith, P. B. Universal Surface-Initiated Polymerization of Antifouling Zwitterionic Brushes Using a Mussel-Mimetic Peptide Initiator. Langmuir 2012,
28 (18), 7258–7266.
142. Hower, J. C.; Bernards, M. T.; Chen, S.; Tsao, H. K.; Sheng, Y. J.; Jiang, S. Hydration of “Nonfouling” Functional Groups. J. Phys. Chem. B 2009, 113 (1), 197–201.
143. Kane, R. S.; Deschatelets, P.; Whitesides, G. M. Kosmotropes Form the Basis of Protein-Resistant Surfaces. Langmuir 2003, 19 (6), 2388–2391.
144. Zhang, Z.; Zhang, M.; Chen, S. F.; Horbetta, T. A.; Ratner, B. D.; Jiang, S. Y. Blood Compatibility of Surfaces with Superlow Protein Adsorption. Biomaterials 2008, 29 (32),
4285–4291.
145. Colak, S.; Tew, G. N. Dual-Functional ROMP-Based Betaines: Effect of Hydrophilicity and Backbone Structure on Nonfouling Properties. Langmuir 2012, 28 (1), 666–675.
146. Colak, S.; Tew, G. N. Amphiphilic Polybetaines: The Effect of Side-Chain Hydrophobicity on Protein Adsorption. Biomacromolecules 2012, 13 (5), 1233–1239.
147. Chang, Y.; Shu, S. H.; Shih, Y. J.; Chu, C. W.; Ruaan, R. C.; Chen, W. Y. Hemocompatible Mixed-Charge Copolymer Brushes of Pseudozwitterionic Surfaces Resistant to
Nonspecific Plasma Protein Fouling. Langmuir 2010, 26 (5), 3522–3530.
148. Chen, S. F.; Jiang, S. Y. A New Avenue To Nonfouling Materials. Adv. Mater. 2008, 20 (2), 335–338.
149. Nakanishi, K.; Sakiyama, T.; Imamura, K. On the Adsorption of Proteins on Solid Surfaces, a Common but Very Complicated Phenomenon. J. Biosci. Bioeng. 2001, 91 (3),
233–244.
150. Park, K.; Mosher, D. F.; Cooper, S. L. Acute Surface-Induced Thrombosis in the Canine Ex Vivo Model – Importance of Protein-Composition of the Initial Monolayer and
Platelet Activation. J. Biomed. Mater. Res. 1986, 20 (5), 589–612.
151. Sweryda-Krawiec, B.; Devaraj, H.; Jacob, G.; Hickman, J. J. A New Interpretation of Serum Albumin Surface Passivation. Langmuir 2004, 20 (6), 2054–2056.
152. Marois, Y.; Chakfe, N.; Guidoin, R.; Duhamel, R. C.; Roy, R.; Marois, M.; King, M. W.; Douville, Y. An Albumin-Coated Polyester Arterial Graft: In Vivo Assessment of
Biocompatibility and Healing Characteristics. Biomaterials 1996, 17 (1), 3–14.
153. Tanzi, M. C. Bioactive Technologies for Hemocompatibility. Expert Rev. Med. Devices 2005, 2 (4), 473–492.
154. Michiels, C. Endothelial Cell Functions. J. Cell. Physiol. 2003, 196 (3), 430–443.
155. van Hinsbergh, V. W. M. The Endothelium: Vascular Control of Haemostasis. Eur. J. Obstet. Gynecol. Reprod. Biol. 2001, 95 (2), 198–201.
156. Tesfamariam, B. Platelet Function in Intravascular Device Implant-Induced Intimal Injury. Cardiovasc. Revasc. Med. 2008, 9 (2), 78–87.
157. Cines, D. B.; Pollak, E. S.; Buck, C. A.; Loscalzo, J.; Zimmerman, G. A.; McEver, R. P.; Pober, J. S.; Wick, T. M.; Konkle, B. A.; Schwartz, B. S.; Barnathan, E. S.;
McCrae, K. R.; Hug, B. A.; Schmidt, A. M.; Stern, D. M. Endothelial Cells in Physiology and in the Pathophysiology of Vascular Disorders. Blood 1998, 91 (10), 3527–3561.
158. Bernfield, M.; Gotte, M.; Park, P. W.; Reizes, O.; Fitzgerald, M. L.; Lincecum, J.; Zako, M. Functions of Cell Surface Heparan Sulfate Proteoglycans. Annu. Rev. Biochem.
1999, 68, 729–777.
This paper gives a brief overview of heparan sulfate and its functions.
159. Whitelock, J. M.; Iozzo, R. V. Heparan Sulfate: A Complex Polymer Charged with Biological Activity. Chem. Rev. 2005, 105 (7), 2745–2764.
160. Opal, S. M.; Kessler, C. M.; Roemisch, J.; Knaub, S. Antithrombin, Heparin, and Heparan Sulfate. Crit. Care Med. 2002, 30 (5), S325–S331.
161. Griffith, M. J. Kinetics of the Heparin-Enhanced Antithrombin-III Thrombin Reaction – Evidence for a Template Model for the Mechanism of Action of Heparin. J. Biol. Chem.
1982, 257 (13), 7360–7365.
162. Dittman, W. A.; Majerus, P. W. Structure and Function of Thrombomodulin – A Natural Anticoagulant. Blood 1990, 75 (2), 329–336.
This paper gives a brief overview of thrombomodulin and its functions.
163. StearnsKurosawa, D. J.; Kurosawa, S.; Mollica, J. S.; Ferrell, G. L.; Esmon, C. T. The Endothelial Cell Protein C Receptor Augments Protein C Activation by the Thrombin-
Thrombomodulin Complex. Proc. Natl. Acad. Sci. U. S. A 1996, 93 (19), 10212–10216.
164. Broze, G. J. Tissue Factor Pathway Inhibitor. Thromb. Haemostasis 1995, 74 (1), 90–93.
165. Hackeng, T. M.; Maurissen, L. F. A.; Castoldi, E.; Rosing, J. Regulation of TFPI Function by Protein S. J. Thromb. Haemostasis 2009, 7, 165–168.
166. Vallance, P.; Chan, N. Endothelial Function and Nitric Oxide: Clinical Relevance. Heart 2001, 85 (3), 342–350.
167. Moncada, S.; Palmer, R. M. J.; Higgs, E. A. Nitric-Oxide – Physiology, Pathophysiology, and Pharmacology. Pharmacol. Rev. 1991, 43 (2), 109–142.
168. Vane, J. R.; Botting, R. M. Pharmacodynamic Profile of Prostacyclin. Am. J. Cardiol. 1995, 75 (3), 3A–10A.
169. Shaul, P. W. Regulation of Endothelial Nitric Oxide Synthase: Location, Location, Location. Annu. Rev. Physiol. 2002, 64, 749–774.
This paper gives a brief overview of nitric oxide and its antiplatelet function.
170. Cannon, R. O. Role of Nitric Oxide in Cardiovascular Disease: Focus on the Endothelium. Clin. Chem. 1998, 44 (8), 1809–1819.
171. Knudsen, H. L.; Frangos, J. A. Role of Cytoskeleton in Shear Stress-Induced Endothelial Nitric Oxide Production. Am. J. Physiol. Heart Circ. Physiol. 1997, 273 (1),
H347–H355.
www.iran-mavad.com
172. Coade, S. B.; Pearson, J. D. Metabolism of Adenine-Nucleotides in Human-Blood. Circ. Res. 1989, 65 (3), 531–537.
173. Pepper, M. S.; Sappino, A. P.; Stocklin, R.; Montesano, R.; Orci, L.; Vassalli, J. D. Upregulation of Urokinase Receptor Expression on Migrating Endothelial Cells. J. Cell Biol.
1993, 122 (3), 673–684.
174. Levin, E. G.; Santell, L.; Osborn, K. G. The Expression of Endothelial Tissue Plasminogen Activator In Vivo: A Function Defined by Vessel Size and Anatomic Location. J. Cell
Sci. 1997, 110, 139–148.
175. Woodhouse, K. A.; Weitz, J. I.; Brash, J. L. Lysis of Surface-Localized Fibrin Clots by Adsorbed Plasminogen in the Presence of Tissue Plasminogen Activator. Biomaterials
1996, 17 (1), 75–77.
176. Stein, C. M.; Brown, N.; Vaughan, D. E.; Lang, C. C.; Wood, A. J. J. Regulation of Local Tissue-Type Plasminogen Activator Release by Endothelium-Dependent and
Endothelium-Independent Agonists in Human Vasculature. J. Am. Coll. Cardiol. 1998, 32 (1), 117–122.
177. Mulloy, B.; Forster, M. J. Conformation and Dynamics of Heparin and Heparan Sulfate. Glycobiology 2000, 10 (11), 1147–1156.
178. Sasisekharan, R.; Venkataraman, G. Heparin and Heparan Sulfate: Biosynthesis, Structure and Function. Curr. Opin. Chem. Biol. 2000, 4 (6), 626–631.
179. Gott, V. L.; Whiffen, J. D.; Dutton, R. C. Heparin Bonding on Colloidal Graphite Surfaces. Science 1963, 142 (3597), 1297.
180. Bosiers, M.; Deloose, K.; Verbist, J.; Schroe, H.; Lauwers, G.; Lansink, W.; Peeters, P. Heparin-Bonded Expanded Polytetrafluoroethylene Vascular Graft for Femoropopliteal
and Femorocrural Bypass Grafting: 1-Year Results. J. Vasc. Surg. 2006, 43 (2), 313–318.
181. Haude, M.; Konorza, T. F. M.; Kalnins, U.; Erglis, A.; Saunamaki, K.; Glogar, H. D.; Grube, E.; Gil, R.; Serra, A.; Richardt, H. G.; Sick, P.; Erbel, R.; Invest, C. T. Heparin-
Coated Stent Placement for the Treatment of Stenoses in Small Coronary Arteries of Symptomatic Patients. Circulation 2003, 107 (9), 1265–1270.
182. Jansen, P.; van Oeveren, W.; Capel, A.; Carpentier, A. In Vitro Haemocompatibility of a Novel Bioprosthetic Total Artificial Heart. Eur. J. Cardiothorac. Surg. 2012, 41 (6),
e166–e172.
183. Tagarakis, G. I.; Tsilimingas, N. B. Heparin-Coated Extracorporeal Circulation Systems in Heart Surgery. Recent Pat. Cardiovasc. Drug Discov. 2009, 4 (3),
177–179.
184. Chanard, J.; Lavaud, S.; Maheut, H.; Kazes, I.; Vitry, F.; Rieu, P. The Clinical Evaluation of Low-Dose Heparin in Haemodialysis: A Prospective Study Using the Heparin-
Coated AN69 ST Membrane. Nephrol. Dial. Transplant. 2008, 23 (6), 2003–2009.
185. Zhu, A. P.; Zhang, M.; Wu, J.; Shen, J. Covalent Immobilization of Chitosan/Heparin Complex with a Photosensitive Hetero-Bifunctional Crosslinking Reagent on PLA
Surface. Biomaterials 2002, 23 (23), 4657–4665.
186. Walluscheck, K. P.; Bierkandt, S.; Brandt, M.; Cremer, J. Infrainguinal ePTFE Vascular Graft with Bioactive Surface Heparin Bonding – First Clinical Results. J. Cardiovasc.
Surg. 2005, 46 (4), 425–430.
187. Yu, S.; Yu, Z. T.; Wang, G.; Dargusch, M. S.; Zhang, M. H. Evaluation of Haemocompatibility of TLM Titanium Alloy with Surface Heparinization. Rare Met. Mater. Eng.
2009, 38 (3), 384–388.
188. Liu, M.; Yue, X. L.; Dai, Z. F.; Xing, L.; Ma, F.; Ren, N. Q. Stabilized Hemocompatible Coating of Nitinol Devices Based on Photo-Cross-Linked Alginate/Heparin Multilayer.
Langmuir 2007, 23 (18), 9378–9385.
189. Yang, Z. L.; Wang, J.; Luo, R. F.; Maitz, M. F.; Jing, F. J.; Sun, H.; Huang, N. The Covalent Immobilization of Heparin to Pulsed-Plasma Polymeric Allylamine Films on
316L Stainless Steel and the Resulting Effects on Hemocompatibility. Biomaterials 2010, 31 (8), 2072–2083.
190. Tan, Q. G.; Ji, J.; Barbosa, M. A.; Fonseca, C.; Shen, J. C. Constructing Thromboresistant Surface on Biomedical Stainless Steel via Layer-by-Layer Deposition
Anticoagulant. Biomaterials 2003, 24 (25), 4699–4705.
191. Wang, A.; Cao, T.; Tang, H.; Liang, X.; Salley, S. O.; Ng, K. Y. In Vitro Haemocompatibility and Stability of Two Types of Heparin-Immobilized Silicon Surfaces. Colloids
Surf. B 2005, 43 (3–4), 245–255.
192. Andersson, J.; Sanchez, J.; Ekdahl, K. N.; Elgue, G.; Nilsson, B.; Larsson, R. Optimal Heparin Surface Concentration and Antithrombin Binding Capacity as Evaluated with
Human Non-Anticoagulated Blood In Vitro. J. Biomed. Mater. Res. A 2003, 67A (2), 458–466.
193. Murugesan, S.; Xie, J.; Linhardt, R. J. Immobilization of Heparin: Approaches and Applications. Curr. Top. Med. Chem. 2008, 8 (2), 80–100.
This paper gives a comprehensive discussion of heparin used as an anticoagulant.
194. Hardhammar, P. A.; vanBeusekom, H. M. M.; Emanuelsson, H. U.; Hofma, S. H.; Albertsson, P. A.; Verdouw, P. D.; Boersma, E.; Serruys, P. W.; VanderGiessen, W. J.
Reduction in Thrombotic Events with Heparin-Coated Palmaz-Schatz Stents in Normal Porcine Coronary Arteries. Circulation 1996, 93 (3), 423–430.
195. Serruys, P. W.; Emanuelsson, H.; vanderGiessen, W.; Lunn, A. C.; Kiemeney, F.; Macaya, C.; Rutsch, W.; Heyndrickx, G.; Suryapranata, H.; Legrand, V.; Goy, J. J.;
Materne, P.; Bonnier, H.; Morice, M. C.; Fajadet, J.; Belardi, J.; Colombo, A.; Garcia, E.; Ruygrok, P.; deJaegere, P.; Morel, M. A. Heparin-Coated Palmaz-Schatz Stents
in Human Coronary Arteries – Early Outcome of the Benestent-II Pilot Study. Circulation 1996, 93 (3), 412–422.
196. Li, G. C.; Zhang, F. M.; Liao, Y. Z.; Yang, P.; Huang, N. Coimmobilization of Heparin/Fibronectin Mixture on Titanium Surfaces and Their Blood Compatibility. Colloids
Surf. B 2010, 81 (1), 255–262.
197. Li, G. C.; Yang, P.; Qin, W.; Maitz, M. F.; Zhou, S.; Huang, N. The Effect of Coimmobilizing Heparin and Fibronectin on Titanium on Hemocompatibility and Endo-
thelialization. Biomaterials 2011, 32 (21), 4691–4703.
198. Chen, J. L.; Chen, C.; Chen, Z. Y.; Chen, J. Y.; Li, Q. L.; Huang, N. Collagen/Heparin Coating on Titanium Surface Improves the Biocompatibility of Titanium Applied as
a Blood-Contacting Biomaterial. J. Biomed. Mater. Res. A 2010, 95A (2), 341–349.
199. Guo, Q. K.; Lu, Z. Q.; Wang, J. Y.; Li, T. In Vivo Evaluation of a Novel Dexamethasone-Heparin-Double-Coated Stent for Inhibition of Artery Restenosis and Thrombosis.
J. Mater. Sci. Mater. Med. 2011, 22 (6), 1615–1623.
200. Holmer, E.; Kurachi, K.; Soderstrom, G. The Molecular-Weight Dependence of the Rate-Enhancing Effect of Heparin on the Inhibition of Thrombin, Factor-Xa, Factor-IXa,
Factor-XIa, Factor-XIIa and Kallikrein by Anti-Thrombin. Biochem. J. 1981, 193 (2), 395–400.
201. Chan, A.; Berry, L.; OBrodovich, H.; Klement, P.; Mitchell, L.; Baranowski, B.; Monagle, P.; Andrew, M. Covalent Antithrombin-Heparin Complexes with High Anticoagulant
Activity – Intravenous, Subcutaneous, and Intratracheal Administration. J. Biol. Chem. 1997, 272 (35), 22111–22117.
202. Berry, L.; Stafford, A.; Fredenburgh, J.; O’Brodovich, H.; Mitchell, L.; Weitz, J.; Andrew, M.; Chan, A. K. C. Investigation of the Anticoagulant Mechanisms of a Covalent
Antithrombin-Heparin Complex. J. Biol. Chem. 1998, 273 (52), 34730–34736.
203. Sask, K. N.; McClung, W. G.; Berry, L. R.; Chan, A. K. C.; Brash, J. L. Immobilization of an Antithrombin-Heparin Complex on Gold: Anticoagulant Properties and Platelet
Interactions. Acta Biomater. 2011, 7 (5), 2029–2034.
This paper discusses the anticoagulant properties of antithrombin-heparin, an alternative of heparin.
204. Weitz, J. I.; Crowther, M. Direct Thrombin Inhibitors. Thromb. Res. 2002, 106 (3), V275–V284.
205. Davenport, A. Review Article: Low-Molecular-Weight Heparin as an Alternative Anticoagulant to Unfractionated Heparin for Routine Outpatient Haemodialysis Treatments.
Nephrology 2009, 14 (5), 455–461.
206. Gandhi, N. S.; Mancera, R. L. Heparin/Heparan Sulphate-Based Drugs. Drug Discov. Today 2010, 15 (23–24), 1058–1069.
207. Mourao, P. A. S.; Pereira, M. S. Searching for Alternatives to Heparin – Sulfated Fucans from Marine Invertebrates. Trends Cardiovasc. Med. 1999, 9 (8), 225–232.
208. Kishida, A.; Akatsuka, Y.; Yanagi, M.; Aikou, T.; Maruyama, I.; Akashi, M. In Vivo and Ex Vivo Evaluation of the Antithrombogenicity of Human Thrombomodulin
Immobilized Biomaterials. ASAIO J. 1995, 41 (3), M369–M374.
209. Sperling, C.; Salchert, K.; Streller, U.; Werner, C. Covalently Immobilized Thrombomodulin Inhibits Coagulation and Complement Activation of Artificial Surfaces In Vitro.
Biomaterials 2004, 25 (21), 5101–5113.
210. Tseng, P. Y.; Rele, S. S.; Sun, X. L.; Chaikof, E. L. Membrane-Mimetic Films Containing Thrombomodulin and Heparin Inhibit Tissue Factor-Induced Thrombin Generation
in a Flow Model. Biomaterials 2006, 27 (12), 2637–2650.
www.iran-mavad.com
211. Bhatt, D. L.; Topol, E. J. Scientific and Therapeutic Advances in Antiplatelet Therapy. Nat. Rev. Drug Discov. 2003, 2 (1), 15–28.
212. de Mel, A.; Murad, F.; Seifalian, A. M. Nitric Oxide: A Guardian for Vascular Grafts? Chem. Rev. 2011, 111 (9), 5742–5767.
213. Reichenbach, G.; Momi, S.; Gresele, P. Nitric Oxide and its Antithrombotic Action in the Cardiovascular System. Curr. Drug Targets Cardiovasc. Haematol. Disord. 2005,
5 (1), 65–74.
214. Cheung, P. Y.; Salas, E.; Etches, P. C.; Phillipos, E.; Schulz, R.; Radomski, M. W. Inhaled Nitric Oxide and Inhibition of Platelet Aggregation in Critically Ill Neonates.
Lancet 1998, 351 (9110), 1181–1182.
215. Radomski, M. W.; Palmer, R. M. J.; Moncada, S. The Role of Nitric-Oxide and CGMP in Platelet-Adhesion to Vascular Endothelium. Biochem. Biophys. Res. Commun.
1987, 148 (3), 1482–1489.
216. Annich, G. M.; Meinhardt, J. P.; Mowery, K. A.; Ashton, B. A.; Merz, S. I.; Hirschl, R. B.; Meyerhoff, M. E.; Bartlett, R. H. Reduced Platelet Activation and Thrombosis in
Extracorporeal Circuits Coated with Nitric Oxide Release Polymers. Crit. Care Med. 2000, 28 (4), 915–920.
217. Schoenfisch, M. H.; Mowery, K. A.; Rader, M. V.; Baliga, N.; Wahr, J. A.; Meyerhoff, M. E. Improving the Thromboresistivity of Chemical Sensors via Nitric Oxide Release:
Fabrication and In Vivo Evaluation of NO-Releasing Oxygen-Sensing Catheters. Anal. Chem. 2000, 72 (6), 1119–1126.
218. Parzuchowski, P. G.; Frost, M. C.; Meyerhoff, M. E. Synthesis and Characterization of Polymethacrylate-Based Nitric Oxide Donors. J. Am. Chem. Soc. 2002, 124 (41),
12182–12191.
219. Wang, P. G.; Xian, M.; Tang, X. P.; Wu, X. J.; Wen, Z.; Cai, T. W.; Janczuk, A. J. Nitric Oxide Donors: Chemical Activities and Biological Applications. Chem. Rev. 2002,
102 (4), 1091–1134.
220. Ignarro, L. J.; Napoli, C.; Loscalzo, J. Nitric Oxide Donors and Cardiovascular Agents Modulating the Bioactivity of Nitric Oxide – An Overview. Circ. Res. 2002, 90 (1),
21–28.
221. Fitzhugh, A. L.; Keefer, L. K. Diazeniumdiolates: Pro- and Antioxidant Applications of the “NONOates”. Free Radicals Biol. Med. 2000, 28 (10), 1463–1469.
222. Smith, D. J.; Chakravarthy, D.; Pulfer, S.; Simmons, M. L.; Hrabie, J. A.; Citro, M. L.; Saavedra, J. E.; Davies, K. M.; Hutsell, T. C.; Mooradian, D. L.; Hanson, S. R.;
Keefer, L. K. Nitric Oxide-Releasing Polymers Containing the N(O)NO.(-) Group. J. Med. Chem. 1996, 39 (5), 1148–1156.
223. Major, T. C.; Brant, D. O.; Reynolds, M. M.; Bartlett, R. H.; Meyerhoff, M. E.; Handa, H.; Annich, G. M. The Attenuation of Platelet and Monocyte Activation in a Rabbit
Model of Extracorporeal Circulation by a Nitric Oxide Releasing Polymer. Biomaterials 2010, 31 (10), 2736–2745.
224. Jun, H. W.; Taite, L. J.; West, J. L. Nitric Oxide-Producing Polyurethanes. Biomacromolecules 2005, 6 (2), 838–844.
225. Taite, L. J.; Yang, P.; Jun, H. W.; West, J. L. Nitric Oxide-Releasing Polyurethane-PEG Copolymer Containing the YIGSR Peptide Promotes Endothelialization with
Decreased Platelet Adhesion. J. Biomed. Mater. Res. B 2008, 84B (1), 108–116.
226. Zhang, H. P.; Annich, G. M.; Miskulin, J.; Stankiewicz, K.; Osterholzer, K.; Merz, S. I.; Bartlett, R. H.; Meyerhoff, M. E. Nitric Oxide-Releasing Fumed Silica Particles:
Synthesis, Characterization, and Biomedical Application. J. Am. Chem. Soc. 2003, 125 (17), 5015–5024.
227. Al-Sa’doni, H.; Ferro, A. S-Nitrosothiols: A Class of Nitric Oxide-Donor Drugs. Clin. Sci. 2000, 98 (5), 507–520.
228. Wu, K. K.; Thiagarajan, P. Role of Endothelium in Thrombosis and Hemostasis. Annu. Rev. Med. 1996, 47, 315–331.
229. Nerem, R. M.; Ensley, A. E. The Tissue Engineering of Blood Vessels and the Heart. Am. J. Transplant. 2004, 4, 36–42.
230. Baguneid, M. S.; Seifalian, A. M.; Salacinski, H. J.; Murray, D.; Hamilton, G.; Walker, M. G. Tissue Engineering of Blood Vessels. Br. J. Surg. 2006, 93 (3),
282–290.
231. Weng, Y. J.; Chen, J. Y.; Tu, Q. F.; Li, Q. L.; Maitz, M. F.; Huang, N. Biomimetic Modification of Metallic Cardiovascular Biomaterials: From Function Mimicking to
Endothelialization In Vivo. Interface Focus 2012, 2 (3), 356–365.
232. de Mel, A.; Jell, G.; Stevens, M. M.; Seifalian, A. M. Biofunctionalization of Biomaterials for Accelerated In Situ Endothelialization: A Review. Biomacromolecules 2008,
9 (11), 2969–2979.
233. McGuigan, A. P.; Sefton, M. V. The Influence of Biomaterials on Endothelial Cell Thrombogenicity. Biomaterials 2007, 28 (16), 2547–2571.
234. Herring, M.; Smith, J.; Dalsing, M.; Glover, J.; Compton, R.; Etchberger, K.; Zollinger, T. Endothelial Seeding of Polytetrafluoroethylene Femoral Popliteal Bypasses – The
Failure of Low-Density Seeding to Improve Patency. J. Vasc. Surg. 1994, 20 (4), 650–655.
235. Kidd, K. R.; Patula, V. B.; Williams, S. K. Accelerated Endothelialization of Interpositional 1-mm Vascular Grafts. J. Surg. Res. 2003, 113 (2), 234–242.
236. Massia, S. P.; Hubbell, J. A. Immobilized Amines and Basic-Amino-Acids as Mimetic Heparin-Binding Domains for Cell-Surface Proteoglycan-Mediated Adhesion. J. Biol.
Chem. 1992, 267 (14), 10133–10141.
237. Zhu, Y. B.; Gao, C. Y.; Liu, Y. X.; Shen, J. C. Endothelial Cell Functions In Vitro Cultured on Poly(L-Lactic Acid) Membranes Modified with Different Methods. J. Biomed.
Mater. Res. A 2004, 69A (3), 436–443.
238. Lu, A.; Sipehia, R. Antithrombotic and Fibrinolytic System of Human Endothelial Cells Seeded on PTFE: The Effects of Surface Modification of PTFE by Ammonia Plasma
Treatment and ECM Protein Coatings. Biomaterials 2001, 22 (11), 1439–1446.
239. Jun, H. W.; West, J. L. Endothelialization of Microporous YIGSR/PEG-Modified Polyurethaneurea. Tissue Eng. 2005, 11 (7–8), 1133–1140.
240. Li, J. H.; Ding, M. M.; Fu, Q.; Tan, H.; Xie, X. Y.; Zhong, Y. P. A Novel Strategy to Graft RGD Peptide on Biomaterials Surfaces for Endothelialization of Small-Diameter
Vascular Grafts and Tissue Engineering Blood Vessel. J. Mater. Sci. Mater. Med. 2008, 19 (7), 2595–2603.
241. Walluscheck, K. P.; Steinhoff, G.; Kelm, S.; Haverich, A. Improved Endothelial Cell Attachment on ePTFE Vascular Grafts Pretreated with Synthetic RGD-Containing
Peptides. Eur. J. Vasc. Endovasc. 1996, 12 (3), 321–330.
242. Chung, T. W.; Lu, Y. F.; Wang, S. S.; Lin, Y. S.; Chu, S. H. Growth of Human Endothelial Cells on Photochemically Grafted Gly-Arg-Gly-Asp (GRGD) Chitosans.
Biomaterials 2002, 23 (24), 4803–4809.
243. Lim, S. H.; Cho, S. W.; Park, J. C.; Jeon, O.; Lim, J. M.; Kim, S. S.; Kim, B. S. Tissue-Engineered Blood Vessels with Endothelial Nitric Oxide Synthase Activity. J. Biomed.
Mater. Res. B 2008, 85B (2), 537–546.
244. Liu, J. Y.; Swartz, D. D.; Peng, H. F.; Gugino, S. F.; Russell, J. A.; Andreadis, S. T. Functional Tissue-Engineered Blood Vessels from Bone Marrow Progenitor Cells.
Cardiovasc. Res. 2007, 75 (3), 618–628.
245. Hoenig, M. R.; Campbell, G. R.; Rolfe, B. E.; Campbell, J. H. Tissue-Engineered Blood Vessels – Alternative to Autologous Grafts? Arterioscler. Thromb. Vasc. Biol. 2005,
25 (6), 1128–1134.
246. Krawiec, J. T.; Vorp, D. A. Adult Stem Cell-Based Tissue Engineered Blood Vessels: A Review. Biomaterials 2012, 33 (12), 3388–3400.
247. Northup, S. J., (chapter 5.2). In Vitro Assessment of Tissue Compatibility. In An Introduction to Tissue-Biomaterial Interactions; Ratner, B. D., Hoffman, A. S.,
Schoen, F. J., Lemons, J. E., Eds., 2nd ed.; John Wiley & Sons, Inc.: New York, 2004; pp 73–79.
248. Williams, D. F. On the Nature of Biomaterials. Biomaterials 2009, 30 (30), 5897–5909.
249. Hubbell, J. A.; Thomas, S. N.; Swartz, M. A. Materials Engineering for Immunomodulation. Nature 2009, 462 (7272), 449–460.
This is a comprehensive review about engineering immunomodulatory materials.
250. Gumbiner, B. M. Cell Adhesion: The Molecular Basis of Tissue Architecture and Morphogenesis. Cell 1996, 84 (3), 345–357.
251. Lehnert, D.; Wehrle-Haller, B.; David, C.; Weiland, U.; Ballestrem, C.; Imhof, B. A.; Bastmeyer, M. Cell Behaviour on Micropatterned Substrata: Limits of Extracellular
Matrix Geometry for Spreading and Adhesion. J. Cell Sci. 2004, 117 (1), 41–52.
252. Flemming, R. G.; Murphy, C. J.; Abrams, G. A.; Goodman, S. L.; Nealey, P. F. Effects of Synthetic Micro- and Nano-Structured Surfaces on Cell Behavior. Biomaterials
1999, 20 (6), 573–588.
253. Brodbeck, W. G.; Patel, J.; Voskerician, G.; Christenson, E.; Shive, M. S.; Nakayama, Y.; Matsuda, T.; Ziats, N. P.; Anderson, J. M. Biomaterial Adherent Macrophage
Apoptosis is Increased by Hydrophilic and Anionic Substrates In Vivo. Proc. Natl. Acad. Sci. U. S. A 2002, 99 (16), 10287–10292.
www.iran-mavad.com
254. Yim, E. K.; Leong, K. W. Significance of Synthetic Nanostructures in Dictating Cellular Response. Nanomedicine 2005, 1 (1), 10–21.
255. Schulte, V. A.; Diez, M.; Moller, M.; Lensen, M. C. Surface Topography Induces Fibroblast Adhesion on Intrinsically Nonadhesive Poly(Ethylene Glycol) Substrates.
Biomacromolecules 2009, 10 (10), 2795–2801.
256. Andersson, A. S.; Backhed, F.; von Euler, A.; Richter-Dahlfors, A.; Sutherland, D.; Kasemo, B. Nanoscale Features Influence Epithelial Cell Morphology and Cytokine
Production. Biomaterials 2003, 24 (20), 3427–3436.
257. Dalby, M. J.; Riehle, M. O.; Johnstone, H.; Affrossman, S.; Curtis, A. S. G. In Vitro Reaction of Endothelial Cells to Polymer Demixed Nanotopography. Biomaterials 2002,
23 (14), 2945–2954.
258. Chen, S. L.; Jones, J. A.; Xu, Y. G.; Low, H. Y.; Anderson, J. M.; Leong, K. W. Characterization of Topographical Effects on Macrophage Behavior in a Foreign Body
Response Model. Biomaterials 2010, 31 (13), 3479–3491.
259. Franchimont, D.; Kino, T.; Galon, J.; Meduri, G. U.; Chrousos, G. Glucocorticoids and Inflammation Revisited: The State of the Art. Neuroimmunomodulation 2002, 10 (5),
247–260.
260. Bogdan, C. Nitric Oxide and the Immune Response. Nat. Immunol. 2001, 2 (10), 907–916.
261. Hetrick, E. M.; Prichard, H. L.; Klitzman, B.; Schoenfisch, M. H. Reduced Foreign Body Response at Nitric Oxide-Releasing Subcutaneous Implants. Biomaterials 2007,
28 (31), 4571–4580.
262. Gifford, R.; Batchelor, M. M.; Lee, Y.; Gokulrangan, G.; Meyerhoff, M. E.; Wilson, G. S. Mediation of In Vivo Glucose Sensor Inflammatory Response via Nitric Oxide
Release. J. Biomed. Mater. Res. A 2005, 75A (4), 755–766.
263. Schwentker, A.; Vodovotz, Y.; Weller, R.; Billiar, T. R. Nitric Oxide and Wound Repair: Role of Cytokines? Nitric Oxide Biol. Chem. 2002, 7 (1), 1–10.
264. Jimenez, J. L.; Gonzalez-Nicolas, J.; Alvarez, S.; Fresno, M.; Munoz-Fernandez, M. A. Regulation of Human Immunodeficiency Virus Type 1 Replication in Human T
Lymphocytes by Nitric Oxide. J. Virol. 2001, 75 (10), 4655–4663.
265. Darouiche, R. O. Current Concepts – Treatment of Infections Associated with Surgical Implants. N. Engl. J. Med. 2004, 350 (14), 1422–1429.
266. Schutzer, S. F.; Harris, W. H. Deep-Wound Infection after Total Hip-Replacement under Contemporary Aseptic Conditions. J. Bone Joint Surg. Am. 1988, 70A (5),
724–727.
267. Vasilev, K.; Cook, J.; Griesser, H. J. Antibacterial Surfaces for Biomedical Devices. Expert Rev. Med. Devices 2009, 6 (5), 553–567.
268. Zhao, L. Z.; Chu, P. K.; Zhang, Y. M.; Wu, Z. F. Antibacterial Coatings on Titanium Implants. J. Biomed. Mater. Res. B 2009, 91B (1), 470–480.
This is a comprehensive review about antibacterial coatings.
269. Ramritu, P.; Halton, K.; Collignon, P.; Cook, D.; Fraenkel, D.; Battistutta, D.; Whitby, M.; Graves, N. A Systematic Review Comparing the Relative Effectiveness of
Antimicrobial-Coated Catheters in Intensive Care Units. Am. J. Infect. Control 2008, 36 (2), 104–117.
270. Coughlan, A.; Boyd, D.; Douglas, C. W. I.; Towler, M. R. Antibacterial Coatings for Medical Devices Based on Glass Polyalkenoate Cement Chemistry. J. Mater. Sci. Mater.
Med. 2008, 19 (12), 3555–3560.
271. Gristina, A. G. Biomaterial-Centered Infection – Microbial Adhesion versus Tissue Integration. Science 1987, 237 (4822), 1588–1595.
272. Gottenbos, B.; Busscher, H. J.; van der Mei, H. C. Pathogenesis and Prevention of Biomaterial Centered Infections. J. Mater. Sci. Mater. Med. 2002, 13 (8), 717–722.
273. Boks, N. P.; Norde, W.; van der Mei, H. C.; Busscher, H. J. Forces Involved in Bacterial Adhesion to Hydrophilic and Hydrophobic Surfaces. Microbiol. SGM 2008, 154,
3122–3133.
274. Donlan, R. M. Biofilm Formation: A Clinically Relevant Microbiological Process. Clin. Infect. Dis. 2001, 33 (8), 1387–1392.
275. Flemming, H. C.; Wingender, J. The Biofilm Matrix. Nat. Rev. Microbiol. 2010, 8 (9), 623–633.
276. Hall-Stoodley, L.; Costerton, J. W.; Stoodley, P. Bacterial Biofilms: From the Natural Environment to Infectious Diseases. Nat. Rev. Microbiol. 2004, 2 (2), 95–108.
277. Lewis, K. Riddle of Biofilm Resistance. Antimicrob. Agents Chemother. 2001, 45 (4), 999–1007.
278. Hetrick, E. M.; Schoenfisch, M. H. Reducing Implant-Related Infections: Active Release Strategies. Chem. Soc. Rev. 2006, 35 (9), 780–789.
279. Zilberman, M.; Elsner, J. J. Antibiotic-Eluting Medical Devices for Various Applications. J. Controlled Release 2008, 130 (3), 202–215.
280. Wu, P.; Grainger, D. W. Drug/Device Combinations for Local Drug Therapies and Infection Prophylaxis. Biomaterials 2006, 27 (11), 2450–2467.
281. Springer, B. D.; Lee, G. C.; Osmon, D.; Haidukewych, G. J.; Hanssen, A. D.; Jacofsky, D. J. Systemic Safety of High-Dose Antibiotic-Loaded Cement Spacers after
Resection of an Infected Total Knee Arthroplasty. Clin. Orthop. Relat. Res. 2004, 427, 47–51.
282. Stigter, M.; Bezemer, J.; de Groot, K.; Layrolle, P. Incorporation of Different Antibiotics into Carbonated Hydroxyapatite Coatings on Titanium Implants, Release and
Antibiotic Efficacy. J. Controlled Release 2004, 99 (1), 127–137.
283. Price, J. S.; Tencer, A. F.; Arm, D. M.; Bohach, G. A. Controlled Release of Antibiotics from Coated Orthopedic Implants. J. Biomed. Mater. Res. 1996, 30 (3), 281–286.
284. Neut, D.; van de Belt, H.; Stokroos, I.; van Horn, J. R.; van der Mei, H. C.; Busscher, H. J. Biomaterial-Associated Infection of Gentamicin-Loaded PMMA Beads in
Orthopaedic Revision Surgery. J. Antimicrob. Chemother. 2001, 47 (6), 885–891.
285. Rossi, S.; Azghani, A. O.; Omri, A. Antimicrobial Efficacy of a New Antibiotic-Loaded Poly(Hydroxybutyric-Co-Hydroxyvaleric Acid) Controlled Release System. J. Anti-
microb. Chemother. 2004, 54 (6), 1013–1018.
286. Aviv, M.; Berdicevsky, I.; Zilberman, M. Gentamicin-Loaded Bioresorbable Films for Prevention of Bacterial Infections Associated with Orthopedic Implants. J. Biomed.
Mater. Res. A 2007, 83A (1), 10–19.
287. Scott, C. P.; Higham, P. A. Antibiotic Bone Cement for the Treatment of Pseudomonas aeruginosa in Joint Arthroplasty: Comparison of Tobramycin and Gentamicin-
Loaded Cements. J. Biomed. Mater. Res. B 2003, 64 (2), 94–98.
288. Scott, C. P.; Higham, P. A.; Dumbleton, J. H. Effectiveness of Bone Cement Containing Tobramycin – An In Vitro Susceptibility Study of 99 Organisms Found in Infected
Joint Arthroplasty. J. Bone Joint Surg. Br. 1999, 81B (3), 440–443.
289. Nelson, C. L.; Griffin, F. M.; Harrison, B. H.; Cooper, R. E. In Vitro Elution Characteristics of Commercially and Noncommercially Prepared Antibiotic PMMA Beads. Clin.
Orthop. Relat. Res. 1992, 284, 303–309.
290. Sterling, G. J.; Crawford, S.; Potter, J. H.; Koerbin, G.; Crawford, R. The Pharmacokinetics of Simplex-Tobramycin Bone Cement. J. Bone Joint Surg. Br. 2003, 85B (5),
646–649.
291. Walenkamp, G. H. I. M. How I Do It – Chronic Osteomyelitis. Acta Orthop. Scand. 1997, 68 (5), 497–506.
292. Heck, D.; Rosenberg, A.; Schinkascani, M.; Garbus, S.; Kiewitt, T. Use of Antibiotic-Impregnated Cement during Hip and Knee Arthroplasty in the United-States.
J. Arthroplasty 1995, 10 (4), 470–475.
293. Kanellakopoulou, K.; Giamarellos-Bourboulis, E. J. Carrier Systems for the Local Delivery of Antibiotics in Bone Infections. Drugs 2000, 59 (6), 1223–1232.
294. Hoff, S. F.; Fitzgerald, R. H.; Kelly, P. J. The Depot Administration of Penicillin-G and Gentamicin in Acrylic Bone-Cement. J. Bone Joint Surg. Am. 1981, 63 (5),
798–804.
295. Stigter, M.; de Groot, K.; Layrolle, P. Incorporation of Tobramycin into Biomimetic Hydroxyapatite Coating on Titanium. Biomaterials 2002, 23 (20), 4143–4153.
296. Radin, S.; Campbell, J. T.; Ducheyne, P.; Cuckler, J. M. Calcium Phosphate Ceramic Coatings as Carriers of Vancomycin. Biomaterials 1997, 18 (11), 777–782.
297. Subramani, K.; Jung, R. E.; Molenberg, A.; Hammerle, C. H. F. Biofilm on Dental Implants: A Review of the Literature. Int. J. Oral Maxillofac. Implants 2009, 24 (4),
616–626.
298. Aumsuwan, N.; Danyus, R. C.; Heinhorst, S.; Urban, M. W. Attachment of Ampicillin to Expanded Poly(Tetrafluoroethylene): Surface Reactions Leading to Inhibition of
Microbial Growth. Biomacromolecules 2008, 9 (7), 1712–1718.
299. Aumsuwan, N.; Heinhorst, S.; Urban, M. W. Antibacterial Surfaces on Expanded Polytetrafluoroethylene; Penicillin Attachment. Biomacromolecules 2007, 8 (2),
713–718.
www.iran-mavad.com
300. Fabiola, Costa; Isabel F., Carvalho; Ronald C., Montelaro; Gomes, P.; Cristina, M.; Martins, L. Covalent Immobilization of Antimicrobial Peptides (AMPs) onto Biomaterial
surfaces. Acta Biomater 2011, 7, 1431.
301. Russell, A. D.; Hugo, W. B. Antimicrobial Activity and Action of Silver. Prog. Med. Chem. 1994, 31, 351–370.
302. Silver, S. Bacterial Silver Resistance: Molecular Biology and Uses and Misuses of Silver Compounds. FEMS Microbiol. Rev. 2003, 27 (2–3), 341–353.
303. Rai, M.; Yadav, A.; Gade, A. Silver Nanoparticles as a New Generation of Antimicrobials. Biotechnol. Adv. 2009, 27 (1), 76–83.
304. Rupp, M. E.; Fitzgerald, T.; Marion, N.; Helget, V.; Puumala, S.; Anderson, J. R.; Fey, P. D. Effect of Silver-Coated Urinary Catheters: Efficacy, Cost-Effectiveness, and
Antimicrobial Resistance. Am. J. Infect. Control 2004, 32 (8), 445–450.
305. Lok, C. N.; Ho, C. M.; Chen, R.; He, Q. Y.; Yu, W. Y.; Sun, H. Z.; Tam, P. K. H.; Chiu, J. F.; Che, C. M. Proteomic Analysis of the Mode of Antibacterial Action of Silver
Nanoparticles. J. Proteome Res. 2006, 5 (4), 916–924.
306. Kong, H.; Jang, J. Antibacterial Properties of Novel Poly(Methyl Methacrylate) Nanofiber Containing Silver Nanoparticles. Langmuir 2008, 24 (5), 2051–2056.
307. Sambhy, V.; Peterson, B. R.; Sen, A. Multifunctional Silane Polymers for Persistent Surface Derivatization and Their Antimicrobial Properties. Langmuir 2008, 24 (14),
7549–7558.
308. Mallick, K.; Witcomb, M. J.; Scurrell, M. S. Self-Assembly of Silver Nanoparticles: Formation of a Thin Silver Film in a Polymer Matrix. Mater. Sci. Eng. C 2006, 26 (1),
87–91.
309. Lu, J.; Moon, K. S.; Wong, C. P. Silver/Polymer Nanocomposite as a High-K Polymer Matrix for Dielectric Composites with Improved Dielectric Performance. J. Mater.
Chem. 2008, 18 (40), 4821–4826.
310. Gray, J. E.; Norton, P. R.; Griffiths, K. Mechanism of Adhesion of Electroless-Deposited Silver on Poly(Ether Urethane). Thin Solid Films 2005, 484 (1–2),
196–207.
311. Galya, T.; Sedlarik, V.; Kuritka, I.; Novotny, R.; Sedlarikova, J.; Saha, P. Antibacterial Poly(Vinyl Alcohol) Film Containing Silver Nanoparticles: Preparation and Char-
acterization. J. Appl. Polym. Sci. 2008, 110 (5), 3178–3185.
312. Zaporojtchenko, V.; Podschun, R.; Schurmann, U.; Kulkarni, A.; Faupel, F. Physico-Chemical and Antimicrobial Properties of Co-Sputtered Ag-Au/PTFE Nanocomposite
Coatings. Nanotechnology 2006, 17 (19), 4904–4908.
313. Voccia, S.; Ignatova, M.; Jerome, R.; Jerome, C. Design of Antibacterial Surfaces by a Combination of Electrochemistry and Controlled Radical Polymerization. Langmuir
2006, 22 (20), 8607–8613.
314. Wan, Y. Z.; Raman, S.; He, F.; Huang, Y. Surface Modification of Medical Metals by Ion Implantation of Silver and Copper. Vacuum 2007, 81 (9), 1114–1118.
315. Cao, H. L.; Liu, X. Y.; Meng, F. H.; Chu, P. K. Biological Actions of Silver Nanoparticles Embedded in Titanium Controlled by Micro-Galvanic Effects. Biomaterials 2011,
32 (3), 693–705.
316. Ewald, A.; Gluckermann, S. K.; Thull, R.; Gbureck, U. Antimicrobial Titanium/Silver PVD Coatings on Titanium. Biomed. Eng. Online 2006, 5.
This paper proposed a bactericidal mechanism for Ag-doped titanium.
317. Chen, W.; Liu, Y.; Courtney, H. S.; Bettenga, M.; Agrawal, C. M.; Bumgardner, J. D.; Ong, J. L. In Vitro Anti-Bacterial and Biological Properties of Magnetron Co-
Sputtered Silver-Containing Hydroxyapatite Coating. Biomaterials 2006, 27 (32), 5512–5517.
318. Tiller, J. C.; Liao, C. J.; Lewis, K.; Klibanov, A. M. Designing Surfaces that Kill Bacteria on Contact. Proc. Natl. Acad. Sci. U. S. A 2001, 98 (11), 5981–5985.
319. Harris, L. G.; Richards, R. G. Staphylococci and Implant Surfaces: A Review. Injury 2006, 37, 3–14.
This review discusses the implant-associated infections induced by Gram-positive staphylococci and gives a summary of the methods used to prevent this
type of infection.
320. Kugler, R.; Bouloussa, O.; Rondelez, F. Evidence of a Charge-Density Threshold for Optimum Efficiency of Biocidal Cationic Surfaces. Microbiol. SGM 2005, 151,
1341–1348.
321. Murata, H.; Koepsel, R. R.; Matyjaszewski, K.; Russell, A. J. Permanent, Non-Leaching Antibacterial Surfaces – 2: How High Density Cationic Surfaces Kill Bacterial Cells.
Biomaterials 2007, 28 (32), 4870–4879.
322. Huang, J. Y.; Koepsel, R. R.; Murata, H.; Wu, W.; Lee, S. B.; Kowalewski, T.; Russell, A. J.; Matyjaszewski, K. Nonleaching Antibacterial Glass Surfaces via “Grafting
Onto”: The Effect of the Number of Quaternary Ammonium Groups on Biocidal Activity. Langmuir 2008, 24 (13), 6785–6795.
323. Bouloussa, O.; Rondelez, F.; Semetey, V. A New, Simple Approach to Confer Permanent Antimicrobial Properties to Hydroxylated Surfaces by Surface Functionalization.
Chem. Commun. 2008, 8, 951–953.
324. Brizzolara, R. A.; Stamper, D. M. The Effect of Covalent Surface Immobilization on the Bactericidal Efficacy of a Quaternary Ammonium Compound. Surf. Interface Anal.
2007, 39 (7), 559–566.
325. Saif, M. J.; Anwar, J.; Munawar, M. A. A Novel Application of Quaternary Ammonium Compounds as Antibacterial Hybrid Coating on Glass Surfaces. Langmuir 2009,
25 (1), 377–379.
326. Thebault, P.; de Givenchy, E. T.; Levy, R.; Vandenberghe, Y.; Guittard, F.; Geribaldi, S. Preparation and Antimicrobial Behaviour of Quaternary Ammonium Thiol Derivatives
Able to be Grafted on Metal Surfaces. Eur. J. Med. Chem. 2009, 44 (2), 717–724.
327. Marini, M.; Bondi, M.; Iseppi, R.; Toselli, M.; Pilati, F. Preparation and Antibacterial Activity of Hybrid Materials Containing Quaternary Ammonium Salts via Sol-Gel
Process. Eur. Polym. J. 2007, 43 (8), 3621–3628.
328. McCubbin, P. J.; Forbes, E.; Gow, M. M.; Gorham, S. D. Covalent Attachment of Quaternary Ammonium Compounds to a Polyethylene Surface via a Hydrolyzable Ester
Linkage: Basis for a Controlled-Release System of Antiseptics from an Inert Surface. J. Appl. Polym. Sci. 2006, 100 (1), 538–545.
329. Huang, J. Y.; Murata, H.; Koepsel, R. R.; Russell, A. J.; Matyjaszewski, K. Antibacterial Polypropylene via Surface-Initiated Atom Transfer Radical Polymerization.
Biomacromolecules 2007, 8 (5), 1396–1399.
330. Jampala, S. N.; Sarmadi, M.; Somers, E. B.; Wong, A. C. L.; Denes, F. S. Plasma-Enhanced Synthesis of Bactericidal Quaternary Ammonium Thin Layers on Stainless
Steel and Cellulose Surfaces. Langmuir 2008, 24 (16), 8583–8591.
331. Li, Z.; Lee, D.; Sheng, X. X.; Cohen, R. E.; Rubner, M. F. Two-Level Antibacterial Coating with Both Release-Killing and Contact-Killing Capabilities. Langmuir 2006,
22 (24), 9820–9823.
332. Mannick., Joan B. Immunoregulatory and Antimicrobial Effects of Nitrogen Oxides. Proc. Am. Thorac. Soc. 2006, 3 (2), 161–165.
333. Fang, F. Mechanisms of Nitric Oxide-Related Antimicrobial Activity. J. Clin. Invest. 1997, 99 (12), 2818–2825.
334. Nablo, B. J.; Rothrock, A. R.; Schoenfisch, M. H. Nitric Oxide-Releasing Sol-Gels as Antibacterial Coatings for Orthopedic Implants. Biomaterials 2005, 26 (8), 917–924.
This paper synthesized a dual-functional coating which showed high bactericidal efficacy.
335. Charville, G. W.; Hetrick, E. M.; Geer, C. B.; Schoenfisch, M. H. Reduced Bacterial Adhesion to Fibrinogen-Coated Substrates via Nitric Oxide Release. Biomaterials 2008,
29 (30), 4039–4044.
336. Frost, M. C.; Reynolds, M. M.; Meyerhoff, M. E. Polymers Incorporating Nitric Oxide Releasing/Generating Substances for Improved Biocompatibility of Blood-Contacting
Medical Devices. Biomaterials 2005, 26 (14), 1685–1693.
337. Watanabe, T.; Nakajima, A.; Wang, R.; Minabe, M.; Koizumi, S.; Fujishima, A.; Hashimoto, K. Photocatalytic Activity and Photoinduced Hydrophilicity of Titanium Dioxide
Coated Glass. Thin Solid Films 1999, 351 (1–2), 260–263.
338. Gallardo-Moreno, A. M.; Pacha-Olivenza, M. A.; Saldana, L.; Perez-Giraldo, C.; Bruque, J. M.; Vilaboa, N.; Gonzalez-Martin, M. L. In Vitro Biocompatibility and Bacterial
Adhesion of Physico-Chemically Modified Ti6Al4V Surface by Means of UV Irradiation. Acta Biomater. 2009, 5 (1), 181–192.
339. Aita, H.; Hori, N.; Takeuchi, M.; Suzuki, T.; Yamada, M.; Anpo, M.; Ogawa, T. The Effect of Ultraviolet Functionalization of Titanium on Integration with Bone. Biomaterials
2009, 30 (6), 1015–1025.
www.iran-mavad.com
340. Oehr, C. Plasma Surface Modification of Polymers for Biomedical Use. Nucl. Instrum. Methods Phys. Res. B 2003, 208, 40–47.
341. Triandafillu, K.; Balazs, D. J.; Aronsson, B. O.; Descouts, P.; Quoc, P. T.; van Delden, C.; Mathieu, H. J.; Harms, H. Adhesion of Pseudomonas Aeruginosa Strains to
Untreated and Oxygen-Plasma Treated Poly(Vinyl Chloride) (PVC) from Endotracheal Intubation Devices. Biomaterials 2003, 24 (8), 1507–1518.
342. Balazs, D. J.; Triandafillu, K.; Chevolot, Y.; Aronsson, B. O.; Harms, H.; Descouts, P.; Mathieu, H. J. Surface Modification of PVC Endotracheal Tubes by Oxygen Glow
Discharge to Reduce Bacterial Adhesion. Surf. Interface Anal. 2003, 35 (3), 301–309.
343. Balazs, D. J.; Triandafillu, K.; Wood, P.; Chevolot, Y.; van Delden, C.; Harms, H.; Hollenstein, C.; Mathieu, H. J. Inhibition of Bacterial Adhesion on PVC Endotracheal
Tubes by RF-Oxygen Glow Discharge, Sodium Hydroxide and Silver Nitrate Treatments. Biomaterials 2004, 25 (11), 2139–2151.
344. Asadinezhad, A.; Novak, I.; Lehocky, M.; Sedlarik, V.; Vesel, A.; Junkar, I.; Saha, P.; Chodak, I. A Physicochemical Approach to Render Antibacterial Surfaces on Plasma-
Treated Medical-Grade PVC: Irgasan Coating. Plasma Process. Polym. 2010, 7 (6), 504–514.
345. Asadinezhad, A.; Novak, I.; Lehocky, M.; Sedlarik, V.; Vesel, A.; Junkar, I.; Saha, P.; Chodak, I. An In Vitro Bacterial Adhesion Assessment of Surface-Modified Medical-
Grade PVC. Colloids Surf. B 2010, 77 (2), 246–256.
346. Koerner, R. J.; Butterworth, L. A.; Mayer, I. V.; Dasbach, R.; Busscher, H. J. Bacterial Adhesion to Titanium-Oxy-Nitride (TiNOx) Coatings with Different Resistivities: A
Novel Approach for the Development of Biomaterials. Biomaterials 2002, 23 (14), 2835–2840.
347. Park, K. D.; Kim, Y. S.; Han, D. K.; Kim, Y. H.; Lee, E. H. B.; Suh, H.; Choi, K. S. Bacterial Adhesion on PEG Modified Polyurethane Surfaces. Biomaterials 1998, 19 (7–9),
851–859.
348. Terada, A.; Yuasa, A.; Tsuneda, S.; Hirata, A.; Katakai, A.; Tamada, M. Elucidation of Dominant Effect on Initial Bacterial Adhesion onto Polymer Surfaces Prepared by
Radiation-Induced Graft Polymerization. Colloids Surf. B 2005, 43 (2), 99–107.
349. Fernandez, I. C. S.; van der Mei, H. C.; Metzger, S.; Grainger, D. W.; Engelsman, A. F.; Nejadnik, M. R.; Busscher, H. J. In Vitro and In Vivo Comparisons of
Staphylococcal Biofilm Formation on a Cross-Linked Poly(Ethylene Glycol)-Based Polymer Coating. Acta Biomater. 2010, 6 (3), 1119–1124.
350. Holmes, P. F.; Currie, E. P. K.; Thies, J. C.; van der Mei, H. C.; Busscher, H. J.; Norde, W. Surface-Modified Nanoparticles as a New, Versatile, and Mechanically Robust
Nonadhesive Coating: Suppression of Protein Adsorption and Bacterial Adhesion. J. Biomed. Mater. Res. A 2009, 91A (3), 824–833.
351. Nejadnik, M. R.; van der Mei, H. C.; Norde, W.; Busscher, H. J. Bacterial Adhesion and Growth on a Polymer Brush-Coating. Biomaterials 2008, 29 (30), 4117–4121.
352. Fernandez, I. C. S.; van der Mei, H. C.; Lochhead, M. J.; Grainger, D. W.; Busscher, H. J. The Inhibition of the Adhesion of Clinically Isolated Bacterial Strains on Multi-
Component Cross-Linked Poly(Ethylene Glycol)-Based Polymer Coatings. Biomaterials 2007, 28 (28), 4105–4112.
353. Maddikeri, R. R.; Tosatti, S.; Schuler, M.; Chessari, S.; Textor, M.; Richards, R. G.; Harris, L. G. Reduced Medical Infection Related Bacterial Strains Adhesion on
Bioactive RGD Modified Titanium Surfaces: A First Step toward Cell Selective Surfaces. J. Biomed. Mater. Res. A 2008, 84A (2), 425–435.
354. Cheng, G.; Zhang, Z.; Chen, S. F.; Bryers, J. D.; Jiang, S. Y. Inhibition of Bacterial Adhesion and Biofilm Formation on Zwitterionic Surfaces. Biomaterials 2007, 28 (29),
4192–4199.
www.iran-mavad.com
www.iran-mavad.com
4.18 Antimicrobial Bioactive Polymer Coatings
R Förch, Max Planck Institute for Polymer Research, Mainz, Germany and Institut für Mikrotechnik Mainz GmbH, Mainz, Germany
L Duque and A Lotz, Max Planck Institute for Polymer Research, Mainz, Germany

4.18.1.1 The Bacterial Problem 449
4.18.2 Current Thin-Film Approaches 451
4.18.2.1 Leaching Systems 451
4.18.2.1.1 Metal Ion Release Systems 451
4.18.2.1.2 Nonmetal Release Systems 454
4.18.2.2 Nonleaching Systems 455
4.18.2.2.1 Quaternary Ammonium Surfaces 455
4.18.2.3 Biologically Derived Solutions 456
4.18.2.3.1 Essential Oils 456
4.18.2.3.2 Antimicrobial Peptides 457
4.18.2.3.3 Phages 458
4.18.3 Summary 458
References 459
4.18.1 Introduction
With the increasing worldwide problem of multiresistant bacteria, new approaches to combat infectious disease are a major priority
in health care research. Bacteria are evolving to be resistant to antibiotics faster than new drugs are being discovered and licensed for
use. The increasing number of reports on bacterial infections that cannot be treated with antibiotics are a consequence of unlimited
global travel. Recent threats involving the New Delhi metalloprotease and the entirely resistant gonorrhea strains clearly demon-
strated the vulnerability of modern society.
During the past 70 years, antimicrobial drugs (e.g., antibiotics) have been successfully used to treat patients with bacterial and
infectious diseases; however, many infectious organisms have adapted to the drugs designed to kill them. Even within a few years of
the drug’s development, new strains of bacteria may evolve to develop resistance to commonly used antibiotics. The first reports of
antimicrobial resistance against penicillin were published only 6 months after its first introduction (1). This has started a race
between bacterial evolution and the development of new antibiotics.
Today, resistance to antibiotics is high among gram-positive and gram-negative bacteria that cause serious infections in humans;
it is reported to reach 25% or more in several European Union (EU) States. Various reports and regular updates are published by
international organizations and governments supported by task forces, including U.S. agencies such as the Centers for Disease
Control and Prevention, Food and Drug Administration (FDA), National Institutes of Health, the Department of Agriculture, the
Department of Defense, and the Department of Veterans Affairs (2–4); the World Health Organization (5); the European Center for
Disease Control and Prevention (ECDC) (6,7); and the most recent action plan of the European commission (COM (2011) 748
action plan) (8) that calls for new solutions towards multidrug-resistant bacteria. Worldwide infections caused by resistant bacteria
are considered to be among the most concerning diseases. Unless effective new antibiotics and materials to deliver these drugs are
developed, infection will be the major threat in world health care in the near future (6,7). Antimicrobial resistance originates from
the overuse of antibiotics in medical care as well as in animal husbandry, enabling a passive uptake of antibiotics along the food
chain. Although awareness of this situation has significantly decreased the prophylactic use of antibiotics in health care, there is still
a lack of new agents and new mechanisms of action against multidrug-resistant bacteria.
The literature shows a tremendous effort in the scientific and medical community to understand the mechanisms involved in
bacterial attachment, the evolution of resistance, advance the development of new antibiotics and antiseptics, and development of
new and alternative strategies to overcome bacterial infection without the use of antibiotics. The approaches that have been
intensively studied over the past few decades include metal release systems (9–13), quaternary ammonium salts (14,15), and
surface-bound or entrapped antimicrobials and antibiotics that are released upon contact with a wound. These include approaches
involving natural products such as antimicrobial peptides (16–18), phage lysins (19), and various other antimicrobials from nature
(20–22). The use of new (often evolutionary) antimicrobials has become very popular and accommodates the general trend away
from conventional antibiotics.
4.18.1.1 The Bacterial Problem

Approximately 90% of bacteria are found attached to solid surface-forming structures (biofilms) that are inaccessible to drugs and
antibiotics. These biofilms represent major problems in industry and in the health care sector. The industrial and medical

www.iran-mavad.com
450 Antimicrobial Bioactive Polymer Coatings
importance of such surface-attached bacterial communities is therefore extremely high and significant efforts have been initiated
during the past 20 years to learn more about bacteria, their processes of attachment and mutation, the evolution of resistance to
antibiotics, and the development of new strategies to counteract these evolutionary processes.
The attachment of bacteria to abiotic surfaces is a complex multistage process that involves preconditioning of the surface
with proteins or small molecules, bacterial movement to the surface, initial attachment mediated by physicochemical properties
of the surface (e.g., charge, roughness, hydrophobicity), and then colonization and potentially biofilm growth, as shown in
Figure 1 (23). Although the function and appearance of biofilms in various environments may be different, all biofilms
originate from the same sequence of events (24–26). When microorganisms and solid surfaces are in an aqueous environment
in which organic matter is present (e.g., seawater, milk, tear fluid, urine, blood, saliva), the solid substrate surfaces will first
become covered with a layer of adsorbed, organic molecules simply because transport and adsorption of molecules to a solid
substrate proceeds faster than that of microorganisms. This is generally called a conditioning film that forms the basis for
attachment.
The second step in biofilm formation is the transport of the microorganisms towards a substratum surface, which follows
mechanisms that may include Brownian motion, gravitation, diffusion, convection, or the intrinsic motility of a microorganism.
Pseudomonas aeruginosa are believed to swim, and then attach to surfaces using flagella. Bacteria then aggregate into microcolonies
via twitching motility – that is, crawling via extension and retraction of hairlike structures. Microbial adhesion, either as single
organisms or as (co)aggregates, may then occur. This process is initially mostly reversible, but becomes irreversible with time and is
mediated by the excretion of exopolymeric substances by the microorganisms (27,28). The excretion of these substances leads to the
formation of a microbially derived conditioning film. When a conditioning film is present, any subsequently adhering microor-
ganism is actually not in direct contact with the substratum surface. The strength of the eventual biofilm then becomes dependent
upon the cohesiveness of the conditioning film, rather than upon its direct interaction with the bare substratum surface. Eventually,
the adhering microorganisms start growing, which is the major factor contributing to the accumulation of a high number of cells on
a substrate surface. The formation of the biofilm is dependent on the continual supply of nutrients, hydrodynamic conditions, and
surface roughness, as well as other probably still unknown factors (29–33).
The most virulent and dangerous bacteria include methicillin-resistant strains of Staphylococcus aureus (skin and wound
infections), Streptococcus pyogenes (pharyngitis, tonsillitis, skin and soft-tissue infections), P. aeruginosa (respiratory infection,
pneumonia, sinusitis, and meningitis), and Klebsiella pneumonia (bacteremia, pneumonia) (20,34–36). Staphylococcus aureus is
a bacterium commonly found on human skin and mucosa. It causes diseases when it enters the body, leading to wound
infections, urinary tract infections, pneumonia, and bloodstream infections. Staphylococcus aureus infection begins with
bacterial adhesion to host tissues. The adhesins involved in attachment are grouped into a single family, called microbial
surface components recognizing adhesive matrix molecules; they include fibronectin, collagen, sialoprotein, elastin-binding
proteins, and clumping factors A and B (37). Another important group of virulence factors is represented by a family of
bacterial proteins with superantigen activity: staphylococcal enterotoxins, toxic shock syndrome toxin-1 (tst), exfoliative toxins
A and B, and other toxins such as a-, b-, c- and d-hemolysins, bicomponent leukocidin LukE-LukD, and the Panton–Valentine
leukocidin (38).
Patients with urinary or intravenous catheters are especially vulnerable to methicillin-resistant strains of S. aureus and general
S. aureus infections. Catheter-related bloodstream infections are caused by S. aureus, Candida albicans, and P. aeruginosa, which are
gram-positive, fungal, and gram-negative pathogens responsible for significant morbidity and mortality in hospitals. Today still,
infection remains a major impediment to the long-term use of many implanted or intravascular devices, such as joint prostheses,
heart valves, vascular catheters, contact lenses, and dentures (39,40). Frequently, failure of such devices stems from bacterial biofilm
buildup, which is extremely resistant to host defense mechanisms and antibiotic treatment. Often, the only solution to an infected
implanted device is its surgical removal. Infection remains to the top risk factor in burn wounds (41).
The current paper aims to review and bring together some of the views and recent developments in the area of materials and
strategies to combat bacterial infections. Although this chapter will not provide an exhaustive view of the current efforts worldwide,
it will provide first insights and direct the reader towards further documentation in these research fields.
Figure 1 Steps in biofilm formation.
www.iran-mavad.com
Antimicrobial Bioactive Polymer Coatings 451
4.18.2 Current Thin-Film Approaches

4.18.2.1 Leaching Systems
4.18.2.1.1 Metal Ion Release Systems
Silver and copper have a long history in medicine, going as far back as ancient Rome and Greece when metal coins were dropped
into water as a disinfectant. In the late nineteenth century, silver nitrate solution was used to eliminate blindness in newborns
caused by postpartum infection. In the late nineteenth century, silver nitrate was already used to treat typhoid and anthrax. In 1964,
silver was first used to treat burn wounds. It was shown that a 0.5% solution of silver nitrate (AgNO3) was the lowest concentration
that remained active against bacteria in vitro but, when tested in vivo, it appeared to have no toxic effects on growing epidermal cells
(42). Four years later, silver sulfadiazine was first introduced in modern medicine (10,43).
The antimicrobial efficacy of metals and their ions has been known and used for centuries, as shown in their use in household
items as well as early medical devices. With the rising concern of antibiotic resistance in common pathogenic bacteria and the
reported increase in hospital-acquired infections, nanocomposite coatings have received increased attention by the scientific
community because of their ease of application and often good conformity to many industrial processes. The literature shows metal
ion release systems based on copper (12,44), silver (45–47), bismuth, mercury, gold, and zinc (48–50) as possible candidates for
antimicrobial coatings. Metal ion release systems are typically nanocomposite coatings often consisting of a polymeric matrix with
metal or metal oxide inclusions of different sizes and densities, ranging from individual particles to high levels of coalescence
(51,52). During their original development, many of these systems were studied for their microelectronic and optoelectronic
properties (44,53).
In particular, silver has been a popular candidate because it has been shown to be effective against most pathogenic bacteria,
including the multiresistant strains of S. aureus and P. aeruginosa that have become a major concern to the health care system over the
past 20 years. Silver has been used for centuries as an antimicrobial agent to reduce and prevent infection (9,54). When antibiotics
appeared in the 1940s, the use of silver diminished; however, it remained one of the most popular agents for wound infections,
especially in patients with burn wounds (55–59). Silver has been regarded as an effective non-resistance-inducing antimicrobial to
prevent device-related infections. For decades, both in vitro and in vivo experiments have been performed to develop potent anti-
microbially acting silver-coated devices (60–62). Strong antimicrobial efficacy is generally claimed, and the industry launched
several products on the market. Some examples include Acticoat, Allevyn, and Algisite Ag, which come with different support
materials such as textiles or foams (63).
The use of silver in medical devices has initiated a whole range of different preparative methods. The silver may be integrated into
the original polymer used for the medical device, blended into the original polymer and extruded in the desired shape, or bound
chemically to surface moieties (56). A particularly effective method has been the coating of conventional biomaterials and medical
devices with polymer-based metal release systems using plasma-assisted deposition. The literature shows the development of
different low-pressure plasma-deposited polymeric matrices with metal inclusions, often in the form of the metal or metal oxide
nanoparticles (46,49,64,65). Many of the processes described involve an asymmetrical setup, allowing for the simultaneous
sputtering or evaporation of the metal during plasma deposition of a suitable biocompatible coating (66,67). This has included
ethylene/CO2 mixtures (46) or hexamethyldisiloxane (68,69) to name but a few. The processes lead to the deposition of plasma
polymer coatings with metal particle inclusions whose size, distribution, and levels of coalescence are highly dependent on the
process conditions, such as gas mixtures and input power used. Figure 2 shows examples of plasma polymer coatings with
Figure 2 Transmission electron micrographs of silver/ethylene/CO2 deposits using different gas mixtures and input powers with an asymmetric
geometry. Data from Körner, E.; Aguirre, M.H.; Fortunato, G.; Ritter, A.; Rühe, J.; Hegemann, D. Plasma Process. Polym. 2010, 7, 619.
www.iran-mavad.com
embedded Ag nanoparticles as deposited in a low-pressure radiofrequency (RF) plasma reactor using an asymmetrical setup with an
Ag electrode. The plasma polymer was deposited using a gas/monomer mixture of CO2/C2H4 leading to a functional hydrocarbon
matrix. Argon was used to sputter Ag atoms from the Ag electrode (46,70,71). The release kinetics were shown to correlate with the
inhibitory effects on the pathogens P. aeruginosa or S. aureus and on mammalian cells that were in contact with these coatings. The
authors identified a unique range of Ag content that provided an effective antibacterial burst release, followed by cytocompatible
conditions soon thereafter.
Other work describes the incorporation of monovalent silver in medical-grade poly(vinyl chloride) (PVC) to investigate the
bactericidal effect on P. aeruginosa adhesion and colonization (72). The surface modification involved the radiofrequency-oxygen
glow discharge prefunctionalization, followed by a two-step wet-treatment in sodium hydroxide and silver nitrate solutions. The
creation of silver salt on the PVC resulted in a surface that completely inhibited bacterial adhesion of four strains of P. aeruginosa and
efficiently prevented colonization over periods of up to 72 h. In the work by Vasilev et al. silver nanoparticles stabilized by a surface
layer of polyvinyl sulfonate were immobilized on surfaces produced by the plasma deposition of allylamine on various substrates.
The authors showed good antimicrobial properties towards the attachment of Staphylococcus epidermidis and prevented biofilm
formation (73).
Although silver has probably been the most well-researched metal for use as an antimicrobial agent, copper and nickel have also
been studied as potential alternatives (49). Plasma immersion ion implantation was, for example, used to produce a copper-loaded
antibacterial surface on polyethylene (12). The surfaces produced showed about 11% Cu relative to carbon that was implanted into
the near-surface region and about 3% copper was estimated to be immediately on the surface. Associated antibacterial assays
indicated that the copper-implanted polyethylene showed antibacterial effects against Escherichia coli and S. aureus with an effec-
tiveness of 96.2 and 86.1%, respectively.
Zinc has also been known and used as an antimicrobial agent, in particular when present as an oxide (11,74). It is today an
important ingredient in different antimicrobial creams and hygiene products, yet its inclusion as an antimicrobial agent in medical
devices is not as well established as silver. Similar to silver and copper, zinc and zinc oxide have received most attention for optical
and electronic applications. Zinc oxide particles have been blended or mixed with a polymer or its precursor to produce a ZnO/
polymer blend requiring a number of different wet chemical steps, including (1) the synthesis of ZnO particles and (2) the synthesis
of the polymer with the nanoparticle inclusions. The bust release of the zinc from the material has been used and described for
orthopedic applications and is a key factor in rendering these nanocomposite materials as antimicrobial (11). Figure 3 shows the
results of some recent work in which chemical vapor deposition-synthesized ZnO nanoparticles were embedded within a network
of a biocompatible poly(N-isopropylacrylamide) (PNIPAAm) hydrogel matrix (75). Surface and cross-sectional investigations of
the composite material showed a uniform distribution of the filler particles within the hydrogel matrix, with tunable loadings by
simply varying the thickness of the thin film.
The stabilization of the nanoparticles is postulated in this work to be a result of ligand exchange between oleic acid and acrylic
acid residues in the polymer backbone. Depending on the nanoparticle loading, a decay of the absorption signal could be observed
with time upon immersion into aqueous solution, suggesting a release of material. These composite films exhibited efficient
antimicrobial activity against E. coli at very low ZnO loadings of 0.74 ng cm2 (1.33 mmol cm3) with bacteriostatic activity for as
little as 0.1 ng cm2 (0.04 mmol cm3). Similar antimicrobial activity has been reported for nanoscopic silver/polymer composites,
promoting nanoscale ZnO as an alternative to nanoparticulate silver systems. Complementary to the antibacterial assay, viability
studies with mammalian cells (NIH/3T3) demonstrated noncytotoxic behavior of the nanocomposite films at bactericidal loading
levels, permitting healthy proliferation for a period of 7 days.
Zinc-loaded polymers have been deposited in a single-step, low-pressure RF plasma process to deposit thin films from zinc acetyl
acetonate (pp-Zn(acac)2). The thin films deposited from this organometallic precursor show a comparable homogeneous distri-
bution of ZnO particles in a size range of 2–10 nm within an oxygenated hydrocarbon plasma polymer film (48), as can be seen in
the micrographs shown in Figure 4. The surrounding polymeric matrix is reported to be an oxygenated hydrocarbon that appears to
Figure 3 Cross-sectional scanning electron microscopy image of a ZnO/PNIPAAm nanocomposite thick film. Reproduced from Schwartz, V.; Thétiot,
F.; Pütz, S.; Ritz, S.; Choritz, L.; Lappas, A.; Förch, R.; Landfester, K.; Jonas, U. Antib. Adv. Funct. Mater. 2012, 22, 2376.
www.iran-mavad.com
Figure 4 Transmission electron microscopy data showing the nanocomposite nature of zinc-loaded plasma polymer films. The films were deposited
from Zn(acac)2. Reproduced from Duque, L.; Förch, R. Plasma Process. Polym. 2011, 8, 444.
be biocompatible. The chemical nature and zinc loading is variable by careful choice of deposition parameters or varying the gas
feed from the pure monomer to either dilution in oxygen, hydrogen, or argon as the feed gas. Infrared spectroscopy and X-ray
photoelectron spectroscopy shows some indications that the zinc is present as the metal and as the metal oxide and that some
of the complexing structure to the polymer network is retained.
Upon immersion in an aqueous solution, the nanocomposite films show a burst release of zinc into the surrounding area (76).
This can easily be seen by monitoring the optical properties of the deposits in air and in solution using optical waveguide mode
spectroscopy (OWS). OWS has been described previously and allows for the determination of both refractive index and thickness
changes occurring as the metal ions leach from the nanocomposite (48,77). Table 1 shows the results of an OWS study comparing
the changes in optical parameters of plasma polymers of acetyl acetonate (acac) and zinc acetyl acetonate (Zn(acac)2) before and
after immersion in deionized water for 24 h. Although the plasma polymer of acetyl acetonate shows no significant change in n and
d, the zinc-containing polymer showed a 15% decrease in d and n within 24 h of immersion, which suggests the loss of the metal
from the polymer matrix. The change in optical parameters of the films can be correlated to a measured decrease in zinc within the
film by inductively coupled plasma optical emission spectroscopy (ICP-OES) data (48,78). The coatings act as a zinc reservoir,
allowing for a fast release of zinc within a few hours of being in contact with aqueous solution. It could be shown that although
a minimum concentration of >1 mg cm2 zinc was required to lead to a log 106 reduction in gram-positive S. aureaus, a fourfold
increase in zinc was required to enable similar effects for gram-negative P. aeruginosa. At the same time, these films were shown to be
cytocompatible towards NIH 3T3 mammalian cells up to a zinc content of about 5 mg cm1 (2).
Both the silver- and the zinc-loaded thin plasma polymer films are reported to act only for a short timeframe of one to a few days
when subjected to an aqueous environment. After that time, no further metal release is observed, even though in both cases it was
shown that the metal has not fully leached from the film. To tailor the release rate of the metal from the polymer matrix, a secondary
top layer that acts to reduce the diffusion rate of the metal out of the polymer has been suggested (78). In an example from the
laboratory of the author, a plasma-polymerized layer of allylamine was used with a different thickness to study the influence on the
Zn-release rate. The release of Zn was measured using ICP-OES. The data are summarized in Figure 5, showing that even a top layer
as thin as 14 nm reduces the Zn-release rate significantly. A 10-fold thickness increase of the ad-layer appears to reduce the Zn-
release rate to a minimum.
4.18.2.1.1.1 Future Perspectives

The extensive use of products containing nanocrystalline silver has raised concerns about the development of resistance, toxicity
towards mammalian cells, and effects on the environment (79,80). Similar opinions hold for the nanoparticle inclusions of other
metals. Several studies in the early 2000s already demonstrated the antiproliferative effects of metal ions in general and speculated
about the effects, whether cells were actually killed or unable to divide or both (49,81). Other reports clearly show no significant
differences between silver release dressings and conventional other dressings (82). In fact, randomized clinical studies show that
Table 1 Composite film optical properties (thicknesses and refractive index) based on box-model simulations of the optical waveguide mode
spectrum. Estimation of the optical parameters for the samples as deposited (d0, n0) and after 24 h immersed in deionized water (d24 h, n0)
d0 d24 h n0 n24 h in H
2O
pp-acac 458.8 1.7 445.1 1.3 1.538 0.004 1.532 0.007

pp-Zn(acac)2 396.7 1.4 345.5 2.1 1.576 0.009 1.551 0.008
acac, acetyl acetonate.
www.iran-mavad.com
Figure 5 Zn-release rate as measured by inductively coupled plasma optical emission spectroscopy for different multilayers of pp-Zn(acac)2 and
ad-layers of pp-allylamine of different thicknesses.
statistically significant results of antimicrobial efficacy of silver-coated medical devices in high-risk populations of patients are rare,
dealt with small numbers of patients only, and are often controversial. However, the published evidence for silver resistance and
toxicity is limited and is often associated with frequent and high levels of silver used (83). Several randomized controlled clinical
studies with 1300 patients revealed no significant differences in infection rates between silver-coated and unmodified catheters
(84,85). As a result of the conflicting evidence but the ever-increasing evolution of resistance by pathogenic bacteria towards the
standard assortment of antibiotics available within health systems worldwide, there is a general trend to refine the use of metal ions.
Commercially available metal ion-containing medical devices (e.g., Ag-containing catheters, Ag-containing wound bandages)
therefore tend to be used most effectively over a short timeframe of 1–3 days while bacterial infection is most imminent.
4.18.2.1.2 Nonmetal Release Systems

Although metals and metal oxides have received so much attention, the literature shows equally abundant studies where other
bacteriocidal, biocidal, and antibiotic additives have been incorporated into different polymeric substrates, medical devices,
particles, and thin films (86,87).
Triclosan and phosphonium salt biocides have been used as additives in polystyrene–divinylbenzene (PS–DVB) beads using
suspension polymerization (88). The release of the biocide out of the PS–DVB beads when immersed in water–ethanol mixtures
and in physiological saline was studied using ultraviolet absorption measurements. The amounts released into solution showed
a strong dependence on the original loading of the beads, such that an increase of biocide incorporation in the PS–DVB beads was
accompanied by a corresponding enhancement of its concentration in liquid mixtures. The rate of release of the biocide from the
beads could be controlled by changes in the cross-linking density of the polymer.
To incorporate antimicrobial agents into the bulk of a biomaterial or a medical device, the agent has been integrated or
blended into the original polymer and is extruded into the desired shape (89,90), or it has been added to the original polymer as
micro- or nanoparticles and is cross-linked with the matrix (91,92). These processes typically have several advantages and
disadvantages. As compared to the coating of medical devices, there is no need to process the medical device after being made, the
quantity of antimicrobial agent in the medical device is easily controlled, and the processes are generally easy to implement on an
industrial scale.
As typical examples, polyurethanes have been successfully loaded with antimicrobials against either gram-negative or gram-
positive bacteria. In a given example, polyurethane-containing methylene diphenylene diisocyanate, poly(tetramethylene oxide),
and the quaternized biocidal agent N,N-bis(2-hydroxyethyl) isonicotinamide was shown to exhibit up to 95% antibacterial activity
against gram-positive S. aureus, but at the same time only modest 10% antibacterial efficacy against gram-negative E. coli (89).
However, by the addition of hypochlorite-activated dimethylhydanto in pendant groups or alkylammonium-functionalized soft
blocks the polyurethanes were reported to be effective towards both gram-positive (S. aureus) and gram-negative (P. aeruginosa and
E. coli) (93,94).
The major limitations of such processes are that not many antimicrobials are stable at extrusion temperatures or resist free radical
chemical initiators. Further, the processes generally call for considerable amounts of antimicrobial agents in order to homoge-
neously fill the bulk of the medical device, and the incorporation of the antimicrobial agent might affect the mechanical and
biological properties of the medical device, or inversely, components of the bulk material may affect the efficacy of the
antimicrobial.
Developments in this area also include plasma surface treatment and plasma-deposited surface coatings. The modification of
polymeric surfaces with antibiotics has been described and reviewed recently (13,95,96). Described are, for example, different low-
pressure processes to incorporate metal ions, antibiotics, and antimicrobials. Different examples are given that demonstrate the
attachment of penicillin, ampicillin, or vancomycin to plasma polymerized maleic anhydride and poly(ethylene glycol) (PEG)-
grafted coatings (97). The immobilization of the antibiotics onto surfaces generally requires several wet chemical steps, which
makes preparative processes often time consuming, cumbersome, and costly. Because, however, most of the antibiotics used have
already been approved by the FDA, the modification of biomaterials and medical devices with these agents might still find
applications and approval by regulatory agencies. However, the use of antibiotics as the antimicrobial agents is also cause for
increasing concern, because they will contribute to the development of resistant pathogens.
www.iran-mavad.com
4.18.2.2 Nonleaching Systems

4.18.2.2.1 Quaternary Ammonium Surfaces
Surfaces that are equipped with cationic groups have been considered as promising antimicrobial solutions since the early work of
Battice et al. in the mid 1980s (98). Quaternary ammonium compounds (QACs) are regarded as nonleaching, contact-killing
antimicrobial materials that are generally easy to obtain, cheap, and not susceptible to degradation. In addition, there appears to be
no evidence that they cause bacterial resistance (95,99). Their antimicrobial efficacy towards both gram-negative and gram-positive
bacteria has been reported and is believed to function by initiating a loss of bacterial cell membrane integrity either as a result of
long polymer chains that penetrate and disrupt the cell membrane function or as a result of inducing ion exchange reactions within
the cell membrane (100,101). As such, the antimicrobial efficacy towards bacteria is shown to be highly effective; however, the
impact on the cell wall of mammalian cells is less well researched and there are serious concerns regarding the in vivo cytotoxicity of
these solutions that still remain to be resolved.
The first examples of nonleaching surfaces capable of killing microorganisms on contact were described by researchers from Dow
Corning Corporation, who found that 3-(trimethoxysilyl)-propyldimethylalkyl ammonium chlorides with alkyl chain lengths
varying from 6 to 22 carbons added to water gave a high degree of control over the growth of algae (98). Since then, a large number
of QACs have been synthesized and immobilized of surfaces either wet chemically, generally via silane chemistry for silicon or
titanium substrates (102), or in dry chemical processes coupled with some wet chemical surface derivatization processes (103–105).
Surfaces equipped with QACs can be fabricated by simple dip-coating or by a ‘painting’ methodology. They can be immobilized
via covalent bonds, or they can be physically immobilized on the surface by noncovalent hydrophobic interactions. For example,
Klibanov et al. prepared polymeric coatings of branched N,N-dodecylmethyl–polyethylene imineon glass surfaces by painting the
surfaces with a solution of the polymer in butanol (102,106). The authors showed that the coating was able to eliminate influenza
virus with a 100% efficiency within minutes and was able to reduce the adhesion of E. coli and S. aureus. Later, a coating method
based on emulsion polymerization using water-insoluble antimicrobial emulsifiers was described by Fuchs and Tiller, and it
showed equally promising results (107).
Covalent bonding of QACs to model surfaces has involved different strategies (108). In one strategy, the polymer surface was
plasma activated to generate radicals, which were then grafted with antibacterial monomers, such as diallyldimethylammonium
chloride (DADMAC) by a radical polymerization. In another strategy, the polymer surface was plasma treated to create oxygen
functionalities, which were then used to couple DADMAC copolymers to the activated surface. Microbiological data prove that the
DADMAC copolymers were able to reduce the attachment of bacteria such as Micrococcus luteus (gram-positive) and E. coli (gram-
negative) by a factor of 105–106. Other plasma-based solutions include amine or pyridine plasma polymers, which are further
modified with hexyl bromide or bromobutane to produce antimicrobial QAC surfaces (104,105).
Atmospheric pressure plasma (APP) is a developing branch within the area of plasma surface treatment and offers alternative
routes for the modification and antimicrobial treatment of implants and medical devices. APP technology coupled with the use of
a liquid aerosol precursor (e.g., containing the antimicrobial species) allows for a reasonably well-controlled free-radical-induced
polymerization reaction (109). APPs have been used to chemically graft highly complex chemical functionalities directly onto
a variety of substrates. Compounds that lack polymerizable groups can be physically dissolved in the liquid precursor and are
reported to become physically entrapped within the resultant plasma polymer, retaining many of their biological and chemical
properties. As a preliminary example of such a reaction, the entrapment of various quaternary ammonium salts were deposited
within glycol and acrylic acid plasma polymers. First results seemed to indicate antifungal properties.
The increasing interest in therapeutic applications of biocompatible polymeric nanoparticles and capsules has extended
the above work to the development of quaternized nanospheres. For example, a biocidal, surfactant-free, and thus non-
leaching colloidal system was developed based on highly functional polyelectrolyte nanoparticles with quaternary ammo-
nium moieties (75,110). The surfactant-free nanoparticles were synthesized by miniemulsion and solvent evaporation of
poly(2-methacryloyloxyethyl)-dodecyldimethylammonium methacrylate (PqDMAEMA) polyelectrolyte, which led to the
formation of angular-shaped polyelectrolyte latex particles of 200–400 nm in size. Zeta potential measurements of the
surfactant-free PqDMAEMA nanoparticles confirmed the presence of surface-oriented quaternary ammonium groups. Self-
assembly of the 3% (w/w, water phase) colloidal PqDMAEMA on adequate substrates lead to engineered surfaces that
exhibited bactericidal properties against S. aureus (MSSA476) for a period of up to 5 days (Figure 6(a)). Short-term inter-
actions of the PqDMAEMA nanoparticles with mammalian cells within 24 h permitted attachment and proliferation, while
moderate cytotoxic effects could be observed after long-term studies of 5 days (Figure 6(b)).
Chitosan is a polycationic naturally occurring polymer that can be an alternative to the nonnatural polycationic polymers as
antimicrobial agents for biomedical devices (111). Chitosan is obtained by the N-deacetylation of chitin, the second most abundant
natural polysaccharide obtained from the skeletal structures of crustaceans and the cell walls of fungi. Antibacterial activity is based
on the type of bacterium and the molecular weight of the chitosan, the degree of deacetylation, and the types of substitutents, which
can be either cationic or easily form cations. Gram-negative bacteria, such as E. coli, have an anionic bacterial surface on which
cationic chitosan derivatives interact electrostatically. Thus, many chitosan composites include cationic components such as
ammonium, pyridinium, or piperazinium substituents to increase their positive charge. Gram-positive bacteria such as S. aureus are
inhibited by the binding of lower molecular weight chitosan derivatives to DNA or RNA.
It is evident that chitosan, with its low molecular weight, is a useful carrier for molecular drugs requiring targeted delivery.
An important field of study is the development of chitosan-based matrices for the controlled release of active compounds in areas
www.iran-mavad.com
Figure 6 (a) Long-term study on the effect of surfaces coated with PqDMAEMA NPs on the growth of S. aureus (MSSA476) presented by the
number of viable cell counts. Control: uncoated surface. Second positive control: surface coated with Plystyren (PS) NPs and stabilized with the biocidal
agent Cetyl Trimethylammonium Chloride (CTAC). (b) Fluorescent intensity signal of stained mouse fibroblasts (NIH/3T3), which were harvested from
nanoparticle-coated surfaces after 24 h and 5 days of incubation. The surfaces were coated with different samples of 3% (w/w) PqDMAEMA NPs and
CTAC-stabilized PS NPs. Reproduced from Schwartz, V.; Thétiot, F.; Pütz, S.; Ritz, S.; Choritz, L.; Lappas, A.; Förch, R.; Landfester, K.; Jonas, U. Antib
Adv. Funct. Mater. 2012, 22, 2376.
such as agriculture and food for the control of viruses, bacteria, and fungi (112–114). Some major work in this area involves the
blending of essential oils and natural antimicrobial products for preservation of food (115–117).
Chitosan nanoparticles and chitosan-based hydrogels show good loading capacity and efficacy and have been used as drug
carriers for antitumor active compounds such as doxorubicin, paclitaxel, docetaxel, and norcantharidin (118,119) or antibiotics
such as vancomycin (112,120). Chitosan has been complexed with metals such as copper, silver, and zinc (52,121,122) and is
established for antimicrobial activity in wound healing (61,117,121) and the antimicrobial coating of implants (123,124). Several
biomedical devices containing chitosan, including wound dressings and devices for artery closure, have been approved by the FDA
(125,126).
4.18.2.3 Biologically Derived Solutions

Finding healing powers in plants is an ancient idea. People throughout the world have long applied poultices and imbibed
infusions of hundreds, if not thousands, of indigenous plants, dating back to prehistory. There is evidence that Neanderthals living
60 000 years ago in present-day Iraq used plants such as hollyhock; these plants are still widely used in ethnomedicine around the
world (127). Since the advent of antibiotics in the 1950s, the use of plant derivatives as antimicrobials has been virtually
nonexistent.
4.18.2.3.1 Essential Oils

The worldwide threat of increasing antibiotic resistance has brought the use of alternative, natural antibacterial treatments, which
include natural antimicrobial substances such as plant essential oils again into the forefront (87). In general, published studies
indicate that plant essential oils are highly effective antiseptics against many species, including methicillin-resistant S. aureus,
vancomycin-resistant Enterococci, resistant isolates of P. aeruginosa, Klebsiella pneumoniae, and others; certain plant essential oils may
increase the effectiveness of antibiotics against drug-resistant bacteria and they can be synergistic with bacteriophages (128,129).
Some of the useful antimicrobial phytochemicals include the following (128,130):
Simple phenols and phenolic acids: Cinnamic and caffeic acids and their derivatives are common representatives of a wide group of
phenylpropane-derived compounds. The common herbs tarragon and thyme both contain caffeic acid, which is effective against
viruses, bacteria, and fungi (127,131,132). Catechol and pyrogallol both are hydroxylated phenols, shown to be toxic to
microorganisms.
Quinones: Quinones are aromatic rings with two ketone substitutions. They provide a source of stable free radicals and are known
to complex irreversibly with nucleophilic amino acids in proteins. Probable targets in the microbial cell are surface-exposed
adhesins, cell wall polypeptides, and membrane-bound enzymes. Quinones may also render substrates unavailable to the
microorganism. Possible toxic effects of quinones still need to be thoroughly examined.
Flavones, flavonoids, and flavonols: They are known to be synthesized by plants in response to microbial infection and have been
found in vitro to be effective antimicrobial substances against a wide array of microorganisms (133). Their activity is probably due to
their ability to complex with extracellular and soluble proteins and to complex with bacterial cell walls.
Tannins: Tannins are found in almost every plant part: bark, wood, leaves, fruits, and roots. Examples include gallic acid. One of
their molecular actions is to complex with proteins through nonspecific forces such as hydrogen bonding and hydrophobic effects,
as well as by covalent bond formation (130).
www.iran-mavad.com
Coumarins: Coumarins are phenolic substances made of fused benzene and a-pyrone rings. They are known for their antith-
rombotic, anti-inflammatory, and vasodilatory activities and show some antiviral effects.
Terpenenes or terpenoids: They are active against bacteria, fungi, viruses, and protozoa. Triterpenoid betulinic acid is just one of
several terpenoids which have been shown to inhibit human immunodeficiency virus. The mechanism of action of terpenes is
thought to involve membrane disruption by the lipophilic compounds.
Alkaloids, lectins, and polypeptides as well as different mixtures of the above constitute still a wide range of other naturally occurring
antimicrobials that have been reported to exhibit antimicrobial properties.
Antimicrobial materials derived from natural products represent one of the last approaches in the discovery of biologically active
agents against biofilms. Many of them are already well established in the food industry to prevent pathogenic contamination
(112,116). Salicylic acid, cinnamaldehyde (134), and extracts from garlic (135,136), cranberries (137), and plants such as Artemisia
absinthium (138) have all shown various degrees of antibiofilm properties against a number of bacteria in various studies. An extract
from the Australian desert plant Eremophila neglecta has been shown to possess antibacterial activity in a survey of candidate plants
that may bear novel antimicrobial compounds (139). Extractions of E. neglecta led to several new isolates that showed antimicrobial
activity against gram-positive bacteria, including S. aureus, S. pyogenes, and S. pneumoniae, for which the minimum inhibitory
concentrations (MICs) and the minimum bactericidal concentrations ranged from 6.5 to 101.6 mM and 12.7–202.9 mM, respec-
tively. The authors report no activity against gram-negative bacteria for any of the isolated compounds.
4.18.2.3.2 Antimicrobial Peptides

Antimicrobial peptides (AMPs) are next-generation antibiotics and are far less susceptible to the development of pathogen resis-
tance compared to conventional antibiotics. They exhibit rapid and broad-spectrum killing profiles, and are they effective at low
concentrations. AMPs are small molecular weight proteins that show a broad spectrum of antimicrobial activity against bacteria,
viruses, and fungi (140,141). They are usually positively charged and have both a hydrophobic and hydrophilic side that makes
them soluble in aqueous environments yet still allowing them to enter lipid-rich membranes. Once within a target microbial
membrane, the peptide kills target cells through diverse mechanisms. Two major groups of AMPs are found in humans: cath-
elicidins and defensins (142). The defensins are cationic nonglycosylated peptides containing six cysteine residues that form three
intramolecular disulfide bridges, resulting in a triple-stranded beta-sheet structure. They are produced in bacteria, insects, plants,
and vertebrates and are involved in the defense mechanisms of a wide range of organisms (143–145). AMPs have been shown to be
important in such diverse functions as angiogenesis, wound healing, and chemotaxis. Other studies suggest that beside their role in
treating infectious diseases, AMPs play a role in diseases as diverse as inflammatory disorders, autoimmunity, and cancer (146).
The antimicrobial activity of most peptides occurs as a result of unique structural characteristics that enable them to disrupt the
microbial membrane while leaving human cell membranes intact. However, AMPs also act on host cells to stimulate cytokine
production, cell migration, proliferation, maturation, and extracellular matrix synthesis. They are nontoxic, inherently showing
a high selectivity and no immune response (147,148). These ancestral molecules are therefore vital components of the immune
system and represent an attractive potential for novel therapeutic approaches. To date, more than 1500 AMPs of different origins
have been reported.
The literature shows many reports on surface immobilization of AMPs to different surfaces. Typically, the antimicrobial agents
are either encapsulated within a gel for temporal release or they are covalently attached to a polymeric scaffold or a surface
(149,150). Examples include the immobilization on steel to prevent biofouling (151), and the immobilization on membranes and
biomaterial surfaces (152–154). Atmospheric plasma treatments, especially dielectric barrier discharge plasma deposition, has been
shown to allow for the deposition of an antibacterial coating on steel (155). Duday et al. described an organosilicon-containing
plasma interlayer that improved the adhesion and stability of the antibacterial peptides and was shown to lead to a log 104
reduction in gram-positive bacteria using ISO 12296 antibacterial test. Steven at al. discussed wet chemical modification of
polyethylene film using chromic/sulphuric acid, then PEGylatin using carbodiimide chemistry to attach omega-amino-alpha-
carboxyl-PEG. The peptide was covalently coupled to the free terminus of the PEG also using carbodiimide coupling (156).
Antimicrobial activity against E. coli showed a log 103 reduction compared to controls.
The broad-spectrum synthetic cationic peptide melimine has been covalently linked to a surface via two different azide linkers.
The authors reported a 40-fold reduction in bacterial adhesion for the Anti Bacterial Adhesion (ABA) linker and a fivefold reduction
for the Free Nitrous Acid (FNA) linker (157). Gabriel et al. showed a method of covalently attaching Cathelin LL37 to titanium
surfaces (158). In this example, the AMP was linked to the surface by a flexible hydrophilic PEG spacer to produce a surface peptide
layer, which was capable of killing bacteria on contact. From the combined data available, it is obvious that the coupling chemistry
used plays a significant role in retaining the functionality of the immobilized AMP. The advantages of AMPs are generally
compromised upon AMP tethering to solid supports (159–161).
The incorporation of AMPs within gels that are able to release their cargo is an alternative to covalent attachment to the surface.
Laverty et al. investigated the inclusion of maximin-4, the ultrashort peptide H-Orn-Orn-Trp-Trp-NH(2), and the lipopeptide
C(12)-Orn-Orn-Trp-Trp-NH(2) using both matrix- and immersion-loaded poly(HEMA) hydrogels to study the antimicrobial
activity (162). The antimicrobial activity was correlated to both the concentration released and the relative antimicrobial
concentrations of each compound against Staphylococcus epidermidis ATCC 35984. Immersion-loaded samples containing C(12)-
Orn-Orn-Trp-Trp-NH(2) showed the lowest bacterial adherence profile for all peptides studied.
Other solutions involve the development of peptide based hydrogels that exhibit inherent antibacterial activity. This includes
biohybrid approaches of integrating the structural and functional features of antimicrobial proteins and peptides with synthetic
www.iran-mavad.com
polymers and the resulting unique properties in these hybrids, such as bioresponsive–bioactive systems (163–165). An example is
the hydrogel scaffold prepared from the self-assembling peptide, MAX1, whose surface exhibits inherent antibacterial activity (150).
Results show that the hydrogel surface is active against gram-positive (S. epidermidis, S. aureus, and S. pyogenes) and gram-negative (K.
pneumoniae and E. coli) bacteria, thus showing broad-spectrum antibacterial activity. Coculture experiments indicate that hydrogel
surfaces show selective toxicity to bacterial versus mammalian cells. Roy et al. described L-tryptophan-based cationic amphiphilic
hydrogelators with varying degrees of hydrophobicity that exhibited remarkable bactericidal activity against wide range of gram-
positive (MIC ¼ 0.1–75 mg ml1) and gram-negative bacteria (MIC ¼ 0.5–5 mg ml1). Antimicrobial efficacy of the amphiphiles
was shown to be influenced by their alkyl chain length. In some cases, they were found to be significantly better than that of
clinically available antibiotics while remaining biocompatible to the mammalian cells (166).
Naturally occurring cationic AMPs and their mimics form a diverse class of antibacterial agents that are currently being validated
in preclinical and clinical settings for the treatment of infections caused by antimicrobial-resistant bacteria (167). Numerous studies
have strongly supported the hypothesis that the physicochemical properties, rather than any specific amino acid sequence, are
responsible for their microbiological activities. Their topology is essential for insertion into and disruption of the cytoplasmic
membrane. The therapeutic use of naturally occurring AMPs is, unfortunately, hindered by high manufacturing costs, poor phar-
macokinetic properties, and low bacteriological efficacy in animal models. To overcome these problems, a variety of novel and
structurally diverse cationic amphiphiles are appearing, which exhibit superior pharmacokinetic properties and reduced in vitro
toxicity while retaining potent antibacterial activity against resistant and nonresistant bacteria.
4.18.2.3.3 Phages
Bacteriophages are viruses that infect bacteria and are estimated to be the most abundant life form on the planet (168,169).
Bacteriophages attach to the host bacterial cell via specific receptors, whose nature precludes whether or not the phage can infect the
bacterial strain. The use of phage therapies in human medicine dates back to the early twentieth century (170), but it was quickly
overlooked with the discovery of broad-spectrum antibiotics. With the emergence of multiple-drug-resistant bacteria and a lack of
new antibiotics, renewed interests has been spurred in this area by both academic and industrial research. There has been an
increasing interest in the use of bacteriophages and their gene products to control bacterial pathogens as alternatives to currently
used antibiotics for food related applications (171–173) and more recently in wound healing (174–176). In particular, in countries
that have complied with the ban on antimicrobial growth promoters in animal feed, there is a need for alternatives. Bacteriophage
gene therapies could eventually be used to target bacterial pathogens to control animal and human diseases (177,178).
The coevolution of phages and bacteria is a perpetual struggle, with both mutating in concert in hopes of one gaining the upper
hand for continued existence. The application of phage-mediated infections of bacteria that form biofilms has shown some
promising results; however, the question of whether new therapies based on phages represent a viable avenue for the treatment of
biofilms remains open. Despite the advantages of bacteriophages as a natural control over biofilms, there are a number of potential
drawbacks. These include mutations of the protein receptor sites on the bacterial cell surface, which could lead to phage resistance.
Also, phage immunogenicity is a concern if it is to be used as a viable therapy in human treatment. The immune system recognizes
the phage as a foreign substance and triggers antigenic responses (169). Furthermore, once phage-mediated cell lysis occurs, the
bacterial cells’ contents are released into the surrounding environments of the organism and the likelihood of high levels of toxin
release is imminent. This can lead to a number of biological consequences, including inflammation and endotoxic shock (169). The
specificity of phage to one particular bacterium is high, such that this will imply the use of multiple phages in a phage cocktail in
order to ensure complete infection of the biofilm community. This inherently means exposure of patients to a multitude of phages,
each with their own side effects.
A final challenge for phage therapies are the technological challenges that still need to be met. Examples include the immo-
bilization or encapsulation of phages on surfaces and how to retain their integrity over time. The literature shows a number of recent
reports on the immobilization of phages on functionalised thin films (179), on silica nanoparticles (180), incorporated in micron-
sized vesicles for antimicrobial creams (181), or tethered to different sensor surfaces for bacterial capture (182–184). Although the
functionality towards a particular bacterial strain has been proven in all cases described, the inherent side issues mentioned above
still need to be addressed.
4.18.3 Summary
The danger associated with the buildup of resistance in bacteria such as S. aureus, P. aeruginosa, S. pyogenes, K. pneumonia, and other
common pathogens to common antimicrobials and antibiotics has spurred a huge effort in establishing new antimicrobials and
new antimicrobial therapies. Although it was realized quite early already that pathogens are able to mutate and build up resistances,
only the past 10–15 years have seen the current high interdisciplinary activity towards finding solutions. Antimicrobial bio-
functional coatings represent just one strategy to combat infection finding use in both the detection and the treatment of resistant
pathogens. In this context, the coatings can act either as the carrier system that delivers the antimicrobial substance to the site of
infection or its inherent structure exhibits antimicrobial properties.
Different variants of antimicrobial coatings include those with the following:
1. Tethered molecules such quaternary ammonium compounds and antimicrobial peptides
www.iran-mavad.com
2. Embedded moieties such as essential oils, metal or metal oxide nanoparticles (silver, copper, and zinc in general), or conven-
tional antibiotics and other drugs that elude from the polymer coating, such as upon contact with a moist wound bed. These
coatings may include delivery vehicles such as hydrogels and nanocontainers or combinations thereof.
Solutions such as quaternary ammonium compounds, antibiotics, and natural antimicrobials tethered to surface coatings as
possible candidates for new therapeutic solutions against multidrug-resistant pathogens and their evolution of resistance have been
evaluated and discussed. Immobilization to the surface is often associated with a marked reduction in efficacy, such that more recent
efforts are directed towards encapsulation an embedding within different release matrices. Each of the solutions currently being
pursued has advantages and disadvantages. With time, individual areas of optimum application for each solution will probably be
identified as research and development proceeds.
References
1. Rammelkamp, C. H.; Maxon, T. Proc. Royal Soc. Exper. Biol. Med. 1942, 51, 386.
2. http://www.cdc.gov/drugresistance/actionplan/aractionplan.pdf, 1999.
3. http://www.cdc.gov/drugresistance/pdf/public-health-action-plan-combat-antimicrobial-resistance.pdf, 2011.
4. http://www.cdc.gov/drugresistance/pdf/PReport-2009-2010-final-june2011.pdf, 2010.
5. http://whqlibdoc.who.int/HQ/2004/WHO_EDM_PAR_2004.7.pdf, 2004.
6. ECDC/EMEA Joint Technical Report. The Bacterial Challenge: Time to React; European Centre for Disease Prevention and Control, 2009. ISBN 978-92-9193-193-4.
7. Annual Epidemiological Report on Communicable Diseases in Europe; European Centre for Disease Prevention and Control (ECDC): Stockholm, Sweden, 2008. ECDC.
Available from: http://ecdc.europa.eu/en/publications/Publications/0812_SUR_Annual_Epidemiological.
8. http://ec.europa.eu/dgs/health_consumer/docs/communication_amr_2011_748_en.pdf.
9. Monteiro, D.; Gorup, L.; Takamiya, A.; Ruvollo-Filho, A.; de Camargo, E.; Barbosa, D. Int. J. Antimicrob. Agents 2009, 34, 103.
10. Brett, D.W. OWM Comments 2006, 1/1/2006.
11. Boyd, D.; Li, H.; Tanner, D. A.; Towler, M. R.; Wall, J. G. J. Mater. Sci. Mater. Med. 2006, 17, 489.
12. Zhang, W.; Zhang, Y.-H.; Ji, J.-H.; Zhao, J.; Yan, Q.; Chu, P. K. Polymer 2006, 47, 7441.
13. Vasilev, K.; Griesser, S. S.; Griesser, H. J. Plasma Process. Polym. 2011, 8, 1010.
14. Hegstad, K.; Langsrud, S.; Lunestad, B. T.; Scheie, A. A.; Sunde, M.; Yazdankhah, S. P. Microb. Drug Resist. 2010, 16, 91.
15. Tischer, M.; Pradel, G.; Ohlsen, K.; Holzgrabe, U. Chem. Med. Chem. 2012, 7, 22.
16. Siedenbiedel, F.; Tiller, J. C. Polymers 2012, 4, 46.
17. Seo, M.-D.; Won, H.-S.; Kim, J.-H.; Mishig-Ochir, T.; Lee, B.-J. Molecules 2012, 17, 12276.
18. Brandenburg, L. O.; Merres, J.; Albrecht, L. J.; Varoga, D.; Pufe, T. Polymers 2012, 4, 539.
19. Coffey, B.; Mills, S.; Coffey, A.; McAuliffe, O.; Ross, R. P. Annu. Rev. Food Sci. Technol. 2010, 1, 449.
20. Singh, G. S.; Pandeya, S. N. In Natural Products in Discovery of Potential and Safer Antibacterial Agents; Singh, G. S., Ed.; Research Signpost: Kerala, India, 2011; p 63.
21. Lohans, C. T.; Vederas, J. C. Int. J. Microbiol. 2012, 2012, 386410.
22. Anakok, O.; Ndi, C.; Barton, M.; Griesser, H.; Semple, S. J. Appl. Microbiol. 2011, 112, 197.
23. Katsikogianni, M.; Missirlis, Y. F. Eur. Cells Mater. 2004, 8, 37.
24. Gristina, A. G. Science 1987, 237, 1588.
25. Escher, A.; Characklis, W. G. In Biofilms; Characklis, W. G., Marshall, K. C., Eds.; John Wiley and Sons: New York, NY, 1990; p 445.
26. Schneider, R. P.; Marshall, K. C. Colloids Surf., B 1994, 2, 387.
27. Dufrene, Y. F.; Boonaert, C. P. J.; Rouxhet, P. G. Colloids Surf., B 1996, 7, 113.
28. Neu, T. R.; Marshall, K. C. J. Biomater. Appl. 1990, 5, 107.
29. Bradshaw, D. J.; Homer, K. A.; Marsh, P. D.; Beighton, D. Microbiology 1994, 140, 3407.
30. Fletcher, M.; Lessmann, J. M.; Loeb, G. I. Biofouling 1991, 4, 129.
31. Rittmann, B. E. In Structure and Function of Biofilms; Characklis, W. G., Wilderer, P. A., Eds.; John Wiley and Sons: Chichester, 1986; p 49.
32. Boulange-Petermann, L.; Rault, J.; Bellon-Fontaine, M. N. Biofouling 1997, 11, 201.
33. Quirynen, M.; Bollen, C. M. J. Clin. Periodontol. 1995, 22, 1.
34. Marraffini, L. A.; Dedent, A. C.; Schneewind, O. Microbiol. Mol. Biol. Rev. 2006, 70, 192.
35. Mazmanian, S. K.; Ton-That, H.; Schneewind, O. Mol. Microbiol. 2001, 40, 1049.
36. Scott, J. R.; Barnett, T. C. Annu. Rev. Microbiol. 2006, 60, 397.
37. Foster, T. J.; Höök, M. Trends Microbiol. 1998, 6, 484.
38. Sauer, P.; Sõla, J.; Stosova, T. J. Med. Microbiol. 2008, 57, 403.
39. Geesey, G. Curr. Opin. Microbiol. 2001, 4, 296.
40. Lejeune, P. Trends Microbiol. 2003, 11, 179.
41. Church, D.; Elsayed, S.; Reid, O.; Winston, B.; Lindsay, R. Clin. Microbiol. Rev. 2006, 19, 403.
42. Moyer, C.; Brentano, L.; Gravens, D.; Margraf, H.; Monafo, W. Arch. Surg. 1965, 90, 812.
43. Fox, C. L.; Rappole, B. W.; Stanford, W. Surg. Gynecol. Obstet. 1969, 14, 168.
44. Oehr, C.; Suhr, H. Appl. Phys. A-Mater. Sci. Process. 1988, 45, 151.
45. Mackova, A.; Perina, V.; Hnatowicz, V.; Biederman, H.; Slavinska, D.; Choukourov, A. Acta Phys. Slovaca 2004, 54, 7.
46. Körner, E.; Aguirre, M. H.; Fortunato, G.; Ritter, A.; Rühe, J.; Hegemann, D. Plasma Process. Polym. 2010, 7, 619.
47. Griesser, H.J.; Hall, H.; Jenkins, T.A.; Griesser, S.S.; Vasilev, K.; Textor, H.M.G.A.M.; Eds. John Wiley & Sons, Inc.: Hoboken, NJ, USA, 2012.
48. Duque, L.; Förch, R. Plasma Process. Polym. 2011, 8, 444.
49. Wataha, J.; Lockwood, P.; Schedle, A. J. Biomed. Mater. Res. 2000, 52, 360.
50. Silver, S.; Phung, L. T. J. Ind. Microbiol. Biotechnol. 2005, 32, 587.
51. Reddy, P. R.; Varaprasad, K.; Reddy, N. N.; Raju, K. M.; Reddy, N. S. J. Appl. Polym. Sci. 2012, 125, 1357.
52. Liu, Y.; Yu, F.; Liu, L. Asian J. Chem. 2012, 24, 2913.
53. Uglov, V. V.; Kuleshov, A. K.; Astashynskaya, M. V.; Anishchik, V. M.; Dub, S. N.; Thiery, F.; Pauleau, Y. Compos. Sci. Technol. 2005, 65, 785.
54. Alexander, J. W. Surg. Infect. 2009, 10, 289.
55. Tian, J.; Wong, K. K. Y.; Ho, C. M.; Lok, C.-N.; Yu, W.-Y.; C., C.-M.; Chiu, J.-F.; Tam, P. K. H. ChemMedChem 2007, 2, 129.
www.iran-mavad.com
56. Kumar, R.; Howdle, S.; Munstedt, H. J. Biomed. Mater. Res., Part B 2005, 75B, 311.
57. Jeong, S. H.; Hwang, Y. H.; Yi, S. C. J. Mat. Sci. 2005, 40, 5413.
58. Guthrie, K. M.; Agarwal, A.; Tackes, D. S.; Johnson, K. W.; Abbott, N. L.; Murphy, C. J.; Czuprynski, C. J.; Kierski, P. R.; Schurr, M. J.; McAnulty, J. F. Ann. Surg. 2012,
256, 371.
59. Hooper, S. J.; Williams, D. W.; Thomas, D. W.; Hill, K. E.; Percival, S. L. Ostomy Wound Management 2012, 58, 16.
60. Lin, Y.-H.; Lin, J.-H.; Wang, S.-H.; Ko, T.-H.; Tseng, G.-C. J. Biomed. Mater. Res., Part B 2012, 100B, 2288.
61. Chakavala, S. R.; Patel, N. G.; Pate, N. V.; Thakkar, V. T.; Patel, K. V.; Gandhi, T. R. J. Pharm. & Bioallied Sci. 2012, 4, S54.
62. Barnea, Y.; Weiss, J.; Gur, E. Ther. Clin. Risk Manage. 2010, 21, 6.
63. Acticoat; http://www.smith-nephew.de/produkte/wundmanagement/antimikrobielle-wundauflagen/72-iodosorb-dressing.html.
64. Oehr, C.; Suhr, H. Appl. Phys. A: Mater, Sci. Process. 1989, 49, 691.
65. Ploux, L.; Mateescu, M.; Anselme, K.; Vasilev, K. J. Nanomater. 2012.
66. Suhr, H.; Etspuler, A.; Feurer, E.; Oehr, C. Plasma Chem. Plasma Process. 1988, 8, 9.
67. Salz, D.; Lamber, R.; Wark, M.; Baalmann, A.; Jaeger, N. Phys. Chem. Chem. Phys. 1999, 1, 4447.
68. Despax, B.; Raynaud, P. Plasma Process. Polym. 2007, 4, 127.
69. Balazs, D. J.; Triandafillu, K.; Sardella, E.; Iacoviello, G.; Favia, P.; d’Agostino, R.; Harms, H.; Mathieu, H. J. In Plasma Processes and Polymers; d’Agostino, R., Favia, P.,
Oehr, C., Wertheimer, M. R., Eds.; 2005.
70. Körner, E.; Rupper, P.; Lübben, J. F.; Ritter, A.; Rühe, J.; Hegemann, D. Surf. Coat. Technol. 2011, 205, 2978.
71. Lischer, S.; Körner, E.; Balazs, D. J.; Shen, D.; Wick, P.; Grieder, K.; Haas, D.; Heuberger, M.; Hegemann, D. J. R. Soc., Interface 2011, 8, 1019.
72. Balazs, D. J.; Triandafillu, K.; Wood, P.; Chevolot, Y.; van Delden, C.; Harms, H.; Hollenstein, C.; Mathieu, H. J. Biomaterials 2004, 25, 2139.
73. Vasilev, K.; Sah, V. R.; Goreham, R. V.; Ndi, C.; Short, R. D.; Griesser, H. J. Nanotechnology 2010, 21.
74. Padmavathy, N.; Vijayaraghavan, R. Sci. Technol. Adv. Mater. 2008, 9.
75. Schwartz, V.; Thétiot, F.; Pütz, S.; Ritz, S.; Choritz, L.; Lappas, A.; Förch, R.; Landfester, K.; Jonas, U. Antib. Adv. Funct. Mater. 2012, 22, 2376.
76. Malzahn, K.; Duque, L.; Ciernak, P.; Wiesenmueller, S.; Bender, K.; Foerch, R. Plasma Process. Polym. 2013, 10, 243.
77. Chu, L. Q.; Förch, R.; Knoll, W. Langmuir 2006, 22, 2822.
78. Lotz, A.; Heller, M.; Dohm, N.; Ciernak, P.; Bender, K.; Jansen, B.; Förch, R. J. Mater. Chem. 2012, 22, 19455.
79. Geffers, C.; Gastmeier, P. Dtsch. Arztebl. Int. 2011, 108, 87.
80. Mikelováa, R.; Adama, V.; Trnkovác, L.; Babulad, P.; Hornae, A.; Kizeka, R. Biomed. Papers 2005, 149, 17.
81. Poon, V. K. M.; Burd, A. Burns 2004, 30, 140.
82. Madeo, M.; Martin, C.; Turner, C.; Kirkby, V.; Thompson, D. Intensive Crit. Care Nurs. 1998, 14, 187.
83. Schierholz, J. M.; Beuth, J.; Pulvere, G.; König, D. P. Antimicrob. Agents Chemother. 1999, 43, 2819.
84. Riley, D. K.; Classen, D. C.; Stevens, L. E.; P., B. J. Am. J. Med. 1995, 98, 349.
85. Bassetti, S.; Hu, J.; D’Agostino, R. B. J.; Sherertz, R. J. Antimicrob. Agents Chemother. 2001, 45, 1535.
86. White, R. J.; Cutting, K.; Kingsley, A. Ostomy Wound Management 2006, 52, 26.
87. Lang, G.; Buchbauer, G. Flavour Frag. J. 2012, 27, 13.
88. Iconomopouloua, S. M.; Andreopouloua, A. K.; Sotoa, A.; Kallitsisa, J. K.; Voyiatzis, G. A. J. Contr. Release 2005, 102, 223.
89. Grapski, J. A.; Cooper, S. L. Biomaterials 2001, 22, 2239.
90. Wach, J. Y.; Bonazzi, S.; Gademann, K. Angew. Chem., Int. Ed. 2008, 47, 7123.
91. Beyth, N.; Houri-Haddad, Y.; Baraness-Hadar, L.; Yudovin-Farber, I.; Domb, A. J.; Weiss, E. I. Biomaterials 2008, 29, 4157.
92. Beyth, N.; Yudovin-Farber, I.; Bahir, R.; Domb, A. J.; Weiss, E. I. Biomaterials 2006, 27, 3995.
93. Makal, U.; Wood, L.; O., D. E.; Wynne, K. J. Biomaterials 2006, 27, 1316.
94. Kurt, P.; Wood, L.; Ohman, D. E.; Wynne, K. J. Langmuir 2007, 23, 4719.
95. Ferreira, L.; Zumbueh, A. J. Mater. Chem. 2011, 21, 8841 2009, 19, 7796.
96. Jose, B.; Antoci, V., Jr.; Zeiger, A. R.; Wickstrom, E.; Hickok, N. J. Chem. Biol. 2005, 12, 1041.
97. Aumsuwan, N.; Danyus, R. C.; Heinhorst, S.; Urban, M. W. Biomacromolecules 2008, 9, 1712.
98. Battice, D. R.; Hales, M. G. J. Cell. Plast. 1985, 21, 332.
99. Milovic, N. M.; Wang, J.; Lewis, K.; Klibanov, A. M. Biotechnol. Bioeng. 2005, 90, 715.
100. Murata, H.; Koepsel, R. R.; Matyjaszewski, K.; Russel, A. J. Biomaterials 2007, 28, 4870.
101. Tiller, J. C.; Liao, C. J.; Lewis, K.; Klibanov, A. M. Proc. Natl. Acad. Sci. U.S.A. 2001, 98, 5981.
102. Klibanov, A. M. J. Mater. Chem. 2007, 17, 2479.
103. Thome, J.; Hollander, A.; Jaeger, W.; Trick, I.; Oehr, C. Surf. Coat. Technol. 2003, 174, 584.
104. Jampala, S. N.; Sarmadi, M.; Somers, E. B.; Wong, A. C. L.; Denes, F. S. Langmuir 2008, 24, 8583.
105. Schofield, W. C. E.; Badyal, J. P. S. ACS Appl. Mat. Interfaces 2009, 1, 2763.
106. Haldar, J.; An, D.; Alvarez de Cienfuegos, L.; Chen, J.; Klibanov, A. M. Proc. Natl. Acad. Sci. U.S.A. 2006, 103, 17667.
107. Fuchs, A. D.; Tiller, J. C. Angew. Chem., Int. Ed. 2006, 45, 6759.
108. Thome, J.; Holländer, A.; Jaeger, W.; Trick, I.; Oehr, C. Surf. Coat. Technol. 2003, 174–175, 584.
109. O’Hare, L. A.; O’Neill, L.; Goodwin, A. J. Surf. Interface Anal. 2006, 38, 1519.
110. Schwartz, V.P.T. Johannes Gutenber-University of Mainz, 2012.
111. Rabea, E. I.; Badawy, M. E.; Stevens, C. V.; Smagghe, G.; Steurbaut, W. Biomacromolecules 2003, 4, 1457.
112. Cota-Arriola, O.; Onofre Cortez-Rocha, M.; Burgos-Hernandez, A.; Marina Ezquerra-Brauer, J.; Plascencia-Jatomea, M. J. Sci. Food Agric. 2013, 93, 1525.
113. Alvarez, M. V.; Ponce, A. G.; Moreira, M. D. R. Lwt-Food Sci. Technol. 2013, 50, 78.
114. Elsabee, M. Z.; Abdou, E. S. Mater. Sci. Eng., C 2013, 33, 1819.
115. Bozic, M.; Gorgieva, S.; Kokol, V. Carbohydr. Polym. 2012, 89, 854.
116. Giner, M. J.; Vegara, S.; Funes, L.; Marti, N.; Saura, D.; Micol, V.; Valero, M. J. Sci. Food Agric. 2012, 92, 1917.
117. Hernandez-Ochoa, L.; Macias-Castaneda, C. A.; Nevarez-Moorillon, G. V.; Salas-Munoz, E.; Sandoval-Salas, F. Cyta-J. Food 2012, 10, 85.
118. Vinsova, J.; Vavrikova, E. Curr. Pharm. Des. 2011, 17, 3596.
119. Soo, P. L.; Cho, J.; Grant, J.; Ho, E.; Picluette-Miller, M.; Allen, C. Eur. J. Pharm. Biopharm. 2008, 69, 149.
120. Chakraborty, S. P.; Sahu, S. K.; Pramanik, P.; Roy, S. Int. J. Pharm. 2012, 436, 659.
121. Brunel, F.; El Gueddari, N. E.; Moerschbacher, B. M. Carbohydr. Polym. 2013, 92, 1348.
122. Niranjan, R.; Koushik, C.; Saravanan, S.; Moorthi, A.; Vairamani, M.; Selvamurugan, N. Int. J. Biol. Macromol. 2013, 54, 24.
123. Bumgardner, J. D.; Wiser, R.; Gerard, P. D.; Bergin, P.; Chestnutt, B.; Marin, M.; Ramsey, V.; Elder, S. H.; Gilbert, J. A. J. Biomater.Sci., Polym. Ed. 2003, 14, 423.
124. Marsich, E.; Travan, A.; Donati, I.; Turco, G.; Kulkova, J.; Moritz, N.; Aro, H. T.; Crosera, M.; Paoletti, S. Acta Biomater. 2013, 9, 5088.
125. Burkatovskaya, M.; Tegos, G. P.; Swietlik, E.; Demidova, T. N.; Hamblin, M. R. Biomaterials 2006, 4157, 27.
126. http://www.fda.gov.
www.iran-mavad.com
127. Stockwell, C.; C., S., Eds.; Arrow Books, Limited; New Ed Edition (1989): London, United Kingdom, 1989.
128. Kon, K. V.; Rai, M. K. Expert Rev. Anti-Infect. Ther. 2012, 10, 775.
129. Al Nomaani, R. S. S.; Hossain, M. A.; Weli, A. M.; Al-Riyami, Q.; Al-Sabahi, J. N. Asian Pac. J. Trop. Biomed. 2013, 3, 353.
130. Cowan, M. M. Clin. Microbiol. Rev. 1999, 12, 564.
131. Wild, R.; Wild, R.E., Eds.; Rodale Press, Inc.: Emmaus, PA, 1994.
132. Duke, J. A. CRC Handbook of Medicinal Herbs; CRC Press, Inc.: Boca Raton, Florida, 1985.
133. Dixon, R. A.; Dey, P. M.; Lamb, C. J. Adv. Enzymol. 1983, 55, 1.
134. Brackman, G.; Defoirdt, T.; Miyamoto, C.; Bossier, P.; Van Calenbergh, S.; Nelis, H.; Coenye, T. BMC Microbiol. 2008, 8, 149.
135. Shuford, J. A.; Steckelberg, J. M.; Patel, R. Antimicrob. Agents Chemother. 2005, 49, 473.
136. Ross, Z. M.; O’Gara, E. A.; Hill, D. J.; Sleightholme, H. V.; Maslim, D. J. Appl. Environ. Microbiol. 2001, 67, 475.
137. Labrecque, J.; Bodet, C.; Chandad, F.; Grenier, D. J. Antimicrob. Chemother. 2006, 58, 439.
138. Cruz-Galvez, A. M.; Gomez-Aldapa, C. A.; Villagomez-Ibarra, J. R.; Chavarria-Hernandez, N.; Rodriguez-Banos, J.; Rangel-Vargas, E.; Castro-Rosas, J. Food Contr. 2013,
32, 289.
139. Ndi, C. P.; Semple, S. J.; Griesser, H. J.; Pyke, S. M.; Barton, M. D. J. Nat. Prod. 2007, 70, 1439.
140. Pinheiro da Silva, F.; Machado, M. C. Peptides 2012, 36, 308.
141. Kwakman, P. H.; te Velde, A. A.; Vandenbroucke-Grauls, C. M.; van Deventer, S. J.; Zaat, S. A. Antimicrob. Agents Chemother. 2006, 50, 3977.
142. Ganz, T. Nat. Rev. Immunol. 2003, 3, 710.
143. Boman, H. G. Immunol. Rev. 2000, 173, 5.
144. Lehrer, R. I.; Ganz, T. Curr. Opin. Immunol. 1999, 11, 23.
145. Izadpanah, A.; Gallo, R. L. J. Am. Acad. Dermatol. 2005, 52, 381.
146. Bai, Y.; Liu, S.; Jiang, P.; Zhou, L.; Li, J.; Tang, C.; Verma, C.; Mu, Y.; Beuerman, R. W.; Pervushin, K. Biochemistry 2009, 48, 7229.
147. Cherkasov, A.; Hilpert, K.; Jenssen, H.; Fjell, C. D.; Waldbrook, M.; Mullaly, S. C.; Volkmer, R.; Hancock, R. E. W. J. ACS Chem. Biol. 2009, 4, 65.
148. Guaní-Guerra, E.; Santos-Mendoza, T.; Lugo-Reyes, S. O.; Terán, L. M. Clin. Immunol. 2010, 135, 1.
149. Vreuls, C.; Zocchi, G.; Thierry, B.; Garitte, G.; Griesser, S. S.; Archambeau, C.; Van de Weerdt, C.; Martial, J.; Griesser, H. J. Mater. Chem. 2010, 20, 8092.
150. Salick, D. A.; Kretsinger, J. K.; Pochan, D. J.; Schneider, J. P. J. Am. Chem. Soc. 2007, 129, 14793.
151. Hequet, A.; Humblot, V.; Berjeaud, J.-M.; Pradier, C.-M. Colloids Surf., B 2011, 84, 301.
152. Costa, F.; Carvalho, I. F.; Montelaro, R. C.; Gomes, P.; Martins, M. C. L. Acta Biomater. 2011, 7, 1431.
153. Hall, K.; Aguilar, M.-I. In DeMol, N. J., Fischer, M. J. E., Eds. 2010, Vol. 627; p 213.
154. Berliner, K.; Hershkovitz, E.; Ronen, Z.; Kasher, R. Biopolymers 2007, 88, 602.
155. Duday, D.; Vreuls, C.; Moreno, M.; Frache, G.; Boscher, N. D.; Zocchi, G.; Archambeau, C.; Van de Weerdt, C.; Martial, J.; Choquet, P. Surf. Coat. Technol. 2013, 218, 152.
156. Steven, M. D.; Hotchkiss, J. H. J. Appl. Polym. Sci. 2008, 110, 2665.
157. Chen, R.; Cole, N.; Willcox, M. D. P.; Park, J.; Rasul, R.; Carter, E.; Kumar, N. Biofouling 2009, 25, 517.
158. Gabriel, M.; Nazmi, K.; Veerman, E. C.; Amerongen, A. V. N.; Zentner, A. Bioconjugate Chem. 2006, 17, 548.
159. Onaizi, S. A.; Leong, S. S. J. Biotechnol. Adv. 2011, 29, 67.
160. Bagheri, M.; Bayermann, M.; Dathe, M. J. Pept. Sci. 2008, 14, 5.
161. Bagheri, M.; Beyermann, M.; Dathe, M. Antimicrob. Agents Chemother. 2009, 53, 1132.
162. Laverty, G.; Gorman, S. P.; Gilmore, B. F. J. Biomed. Mater. Res. A 2012, 100, 1803.
163. Krishna, O. D.; Kiick, K. L. Biopolymers 2010, 94, 32.
164. Fairman, R.; Akerfeldt, K. S. Curr. Opin. Struct. Biol. 2005, 15, 453.
165. Mitra, R. N.; Shome, A.; Paul, P.; Das, P. K. Org. Biomol. Chem. 2009, 7, 94.
166. Roy, S.; Das, P. K. Biotechnol. Bioeng. 2008, 100, 756.
167. Findlay, B.; Zhanel, G. G.; Schweizer, F. Antimicrob. Agents Chemother. 2010, 54, 4049.
168. Richards, J. J.; Melander, C. ChemBioChem 2009, 10, 2287.
169. Donlan, R. M. Trends Microbiol. 2009, 17, 66.
170. Sulakvelidze, A.; Alavidze, Z.; Morris, J. G. Antimicrob. Agents Chemother. 2001, 45, 649.
171. Seal, B. S. Poult. Sci. 2013, 92, 526.
172. Coffey, B.; Mills, S.; Coffey, A.; McAuliffe, O.; Ross, R. P. In Doyle, M. P., Klaenhammer, T. R., Eds. 2010, Vol. 1; p 449.
173. Anany, H.; Chen, W.; Pelton, R.; Griffiths, M. W. Appl. Environ. Microbiol. 2011, 77, 6379.
174. Miao, J.; Pangule, R. C.; Paskaleva, E. E.; Hwang, E. E.; Kane, R. S.; Linhardt, R. J.; Dordick, J. S. Biomaterials 2011, 32, 9557.
175. Schmelcher, M.; Korobova, O.; Schischkova, N.; Kiseleva, N.; Kopylov, P.; Pryamchuk, S.; Donovan, D. M.; Abaev, I. J. Biotechnol. 2012, 162, 289.
176. Markoishvili, K.; Tsitlanadze, G.; Katsarava, R.; Morris, J. G.; Sulakvelidze, A. Int. J. Dermatol. 2002, 41, 453.
177. Deresinski, S. Clin. Infect. Dis. 2009, 48, 1096.
178. Gill, J. J.; Hollyer, T.; Sabour, P. M. Expert Opin. Ther. Pat. 2007, 17, 1341.
179. Applegate, B. M.; Perry, L. L.; Morgan, M. T.; Kothapalli, A. Purdue Research Foundation, 2012.
180. Cademartiri, R.; Anany, H.; Gross, I.; Bhayani, R.; Griffiths, M.; Brook, M. A. Biomaterials 2010, 31, 1904.
181. Murthy, K.; Engelhardt, R. Chr Hansen AS, 2012.
182. Gervals, L.; Gel, M.; Allain, B.; Tolba, M.; Brovko, L.; Zourob, M.; Mandeville, R.; Griffiths, M.; Evoy, S. Sens. Actuators B 2007, 125, 615.
183. Garcia-Aljaro, C.; Cella, L. N.; Shirale, D. J.; Park, M.; Javier Munoz, F.; Yates, M. V.; Mulchandani, A. Biosens. Bioelectron. 2010, 26, 1437.
184. Hosseinidoust, Z.; Van de Ven, T. G. M.; Tufenkji, N. Langmuir 2011, 27, 5472.
www.iran-mavad.com
www.iran-mavad.com
4.19 Thin-Film Diffusion Barriers for Electronic Applications
PF Carcia, DuPont Central Research and Development, Wilmington, DE, USA
4.19.1 Background 463

4.19.1.1 Electronic Applications with Ultrabarrier Requirements 463
4.19.1.1.1 Organic Light-Emitting Diodes (OLEDs) 464
4.19.1.1.2 Organic Solar Cells 465
4.19.2 Definitions and Basic Equations of Permeation 465
4.19.2.1 Steady-State Flow 465
4.19.2.2 Nonsteady-State Flow and Lag-Time 467
4.19.2.3 Quasi-Steady-State Flow (Thin Diffusive Wall) Approximation 467
4.19.2.4 Quasi-Steady-State Flow (General) Approximation 467
4.19.2.5 Analysis of Solar Cell Front Sheet/Back Sheet 468
4.19.3 Permeation Models for a Thin-Film Diffusion Barrier 468
4.19.3.1 Pinhole Model 468
4.19.3.2 Percolation Model for Permeation 469
4.19.4 Methods for Permeation Measurement 470
4.19.4.1 Coulometric Measurement of Water Vapor 470
4.19.4.2 Ca-Method 470
4.19.4.2.1 The Optical Ca-Method 470
4.19.4.2.2 The Electrical Ca-Method 470
4.19.4.3 The Radioactive (Tritium) Method 471
4.19.4.4 Mass Spectroscopy Method 472
4.19.5 Ultrabarrier Coating Strategies 472
4.19.5.1 Overview 472
4.19.5.2 Polymer-Inorganic Multilayer Thin-Film Barrier 473
4.19.5.2.1 Applications for PML Thin-Film Barriers 475
4.19.5.2.2 Manufacturing PML Thin-Film Barrier 477
4.19.5.3 General Comments on Single-Layer Thin-Film Barriers 477
4.19.5.4 Single-Layer Ultrabarriers Grown by ALD 479
4.19.5.4.1 ALD Growth Process Description 479
4.19.5.4.2 ALD Thin Films: General Properties 479
4.19.5.4.3 Ultrabarrier Properties of Single-Layer ALD Al2O3 480
4.19.5.4.4 Corrosion of ALD Al2O3 Barrier Thin Film 487
4.19.5.4.5 ALD/CVD Bilayer Ultrabarrier Thin Film 490
4.19.5.4.6 Applications for Single ALD Al2O3 Ultrabarrier 490
4.19.5.4.7 Manufacturing with ALD 494
Acknowledgments 497
References 497
4.19.1 Background
4.19.1.1 Electronic Applications with Ultrabarrier Requirements
No one can live in a modern society today without daily encounters with barrier materials, which are ubiquitous in food packaging.
Most common is the clear plastic wrap or film that allows us to see the food product while retarding permeation of oxygen and
moisture, thereby increasing product shelf life (1). Typically, the clear plastic wrap reduces infiltration of air and moisture to levels of
1.0–100, in units of cc m2 day1 or g m2 day1. For example, poly(vinylidene chloride), or Saran, reduces oxygen transport to
a rate of 5 cm3-O2 m2 day1. Currently, efforts are being made to increase product shelf life further by improving the barrier
properties (w10–50x) of the clear plastic wrap by coating it with a transparent inorganic thin film (1–4). Nontoxic coatings of SiO2
or Al2O3 are common. Sometimes, however, the barrier coating also needs to be opaque because light can reduce the product shelf
life. As an example, potato chip bags incorporate Al foil or are coated with Al metal, which enhances the barrier properties of the
plastic and is opaque. Generally speaking, food packaging films, with or without an inorganic barrier coating, reduce gas infiltration
into the product to the range 0.1–100, in units of cc or g m2 day1.
Optically clear plastics with a barrier coating also offer packaging opportunities in electronics. Compared to glass, which is
a common substrate for fabricating electronic devices, especially displays, plastic is lighter weight, thinner, bendable, and fracture

www.iran-mavad.com
464 Thin-Film Diffusion Barriers for Electronic Applications
resistant. These attributes make plastic particularly attractive for portable/mobile electronic devices. Most electronic devices,
however, are moisture sensitive and need to be protected from moisture permeation. In this regard, glass has an advantage over
plastic because glass is impermeable to atmospheric gases.
One category of emerging electronic devices is based on organic materials. They include thin-film transistors (5), solar cells (6),
and light-emitting materials (7), which are attractive for full color displays and general-purpose lighting (8). What makes organic
electronic materials so promising is that high performance can be achieved with inexpensive materials and low-cost fabrication. And
for solar cells, in particular, low cost is paramount to be competitive with conventional energy sources. Organic electronic materials
can also be processed at low temperature, making them compatible with low-cost plastic substrates. However, organic electronic
devices are particularly susceptible to degradation by atmospheric gases, and fabricating devices on plastic will require a new,
ultrabarrier technology.
4.19.1.1.1 Organic Light-Emitting Diodes (OLEDs)

We first estimate how good a barrier is needed for an organic light-emitting diode (OLED) to operate for 10 000 h with minimal
degradation in performance. Consider an OLED with generic structure shown in Figure 1. By analogy with a solid-state light-
emitting p–n junction diode, the OLED has an upper electronic level – analogous to bottom of conduction band, the lowest
unoccupied molecular orbital (LUMO) – , and a lower electronic level, analogous to top of the valence band, the highest occupied
molecular orbital (HOMO) state. When electrons are injected into the LUMO level, they combine with holes injected into the
HOMO level, as illustrated schematically in Figure 2. The recombination of a hole and an electron forms an exciton, which decays
with light emission. To inject electrons into the LUMO requires a metal cathode with low work function, typically from Group 1A
or 2A in the Periodic Table. These metals, such as Ca, are highly reactive with oxygen and water vapor, and in a typical device are
overcoated with thick Al. By analogy with Burrows (9), we assume that the cathode consists of 5 nm of Ca metal with a much
thicker Al top coat, and that satisfactory device performance ceases when all the Ca is consumed or oxidized to Ca(OH)2 by
reaction with H2O, that is,
Ca þ 2H2 O / CaðOHÞ2 þ H2 [
For a cathode of 1 cm2 area, the volume of Ca oxidized is VCa ¼ (1 cm2) 5 107 cm ¼ 5 107 cm3. The mass of Ca metal
converted to hydroxide, using 1.54 g cm3 as the density of Ca, is mCa ¼ 7.7 107 g. The corresponding mass of water to oxidize
this mass of Ca is, according to the chemical reaction equation above:
mH2 O ¼ 2MH2 O =MCa mCa ¼ 7:7 107 ð36=40Þ ¼ 7 107 g
Figure 1 Generic OLED structure.
www.iran-mavad.com
Thin-Film Diffusion Barriers for Electronic Applications 465
Figure 2 Electronic diagram of OLED light-emission process.
M here refers to the molecular weights. For this reaction to reach completion in not less than 10 000 h, the device lifetime, requires
that the rate for water permeation, or the water vapor transmission rate (WVTR), through the Al top coat with any other protective
layers be less than:

WVTR ¼ 7 107 g= 10 000 h cm2 ¼ 7 1012 g cm2 h1
Or in more common units of g m2 day1 used in packaging,
WVTR ¼ 7 1012 g cm2 h1 24 h day1 104 cm2 m2 ¼ 1:68 106 g m2 day1
This is between 1 105 and 1 106 lower permeation than is common for food packaging barriers. It is indeed an ultrabarrier!
4.19.1.1.2 Organic Solar Cells

Another example of an electronic device requiring an ultrabarrier protective layer is an organic solar cell. A group at Georgia
Institute of Technology studied the effect of barrier structures and their WVTR on the shelf life of pentacene/C 60-based organic
photovoltaic (OPV) cells (10). They determined that cells stored at 20 C and 50% relative humidity lost 50% efficiency, after
a quantity of water equal to w1.6 g m2 had permeated the solar cell. If, then, we expect OPV cells to last 10 years, or 3650 days;
then this corresponds to
WVTR < 1:6 g m2 =3650 days ¼ 4:4 104 g m2 day1
Further, if OPV cells could be employed, as proposed, in remote hot regions in third world tier countries, to replace kerosene for
lighting and to charge a cell phone, then the solar cell might likely operate closer to 60 C. Assuming an activation energy of
w58 kJ mol1 for WVTR (11), then this would correspond to a value at room temperature that is smaller by 0.057, or
2.5 105 g m2 day1. Indeed, this barrier need also qualifies in the category of an ultrabarrier.
An ultrabarrier is not only needed for OPV, but inorganic solar cells also require low permeation to last 20–30 years in the field
with less than 20% loss of efficiency. Dennis Coyle (12) at GE has studied degradation kinetics of solar cells based on Cu(In,Ga)Se2
(CIGS). By relating the degradation rate, or loss of CIGS cell efficiency, for unprotected solar cells at combinations of temperature
and relative humidity to a model for moisture permeation, they were able to relate lifetime directly to the WVTR. Using available
climate data, they predict the CIGS module lifetime at different locations in the United States versus WVTR of the
module encapsulation (Figure 3). For a 20-year lifetime in Miami, a CIGS module would need encapsulation with
WVTR < 104 g m2 day1, which is much more demanding than what a traditional food packaging barrier can provide. Flexible
CIGS inorganic solar modules are therefore another example of an electronic device that requires ultrabarrier encapsulation, if they
are to have a lifetime comparable to the most common solar cells made with Si.
4.19.2 Definitions and Basic Equations of Permeation
First, we begin our discussion of ultrabarrier approaches with a review of the basic laws governing gas permeation in solids (13–15).
4.19.2.1 Steady-State Flow

Consider an enclosed volume V separated from the surrounding atmosphere by a permeable membrane of thickness L and area A
(Figure 4). Assume that the atmosphere is a gas, for example, water vapor, with a concentration c (g cm3). The flux of gas (J) that
permeates the membrane can be written as

J ¼ Q=At g cm2 s1 [1]
www.iran-mavad.com
Figure 3 Barrier requirements for 20-year lifetime of CIGS solar cells in different climate environments. Reprinted with permission from Coyle, D. J. Prog.
Photovolt. Res. Appl. 2011, http://dx.doi.org/10.1002/pip.1172. Copyright 2011, John Wiley and Sons.
Figure 4 Enclosure of volume V with a permeable membrane of thickness L and area A.
where Q is the quantity of gas passing through area (A) of the membrane during time (t). The transfer of gas into V by diffusion is
proportional to the negative gradient of the concentration of gas across the membrane, as defined by Fick’s first law:
J ¼ Dðvc=vxÞ [2]
D is the diffusion coefficient for the gas in the membrane, and x is the space coordinate normal to the reference plane of area A.
Assume that the concentration in both the atmosphere and the container volume are maintained constant. Then a steady-state
profile in the membrane will be established and

vc=vx ¼ constant ¼ c0a c0v L [3]
Using Henry’s Law c0 ¼ Sp, where S is the solubility coefficient for the gas in the membrane and p is the permeant pressure
outside the membrane, Fick’s first law can be written as (13–15):
J¼ ¼ DSðpa pv Þ L ¼ Pðpa pv Þ L [4]

At
where P ¼ DS is defined as the permeability coefficient of the membrane.
(Alternatively, flux can be expressed in terms of a difference in surface gas concentration, (ca cv) using the partition coefficient
K ¼ c0v =cv ¼ c0a =ca where P ¼ DK.) The WVTR, usually expressed in units of g-H2O m2 day1, is the flux of water vapor through the
membrane, with a particular thickness at a specified temperature and humidity. (Sometimes, WVTR values are normalized to
a specific thickness, for example, 1 mil (25 mm).) To calculate permeability from WVTR requires knowledge of the permeant pressure
gradient, corresponding to the specified conditions of temperature and humidity, and the membrane thickness.
P ¼ WVTRðL=DpÞ [5]
www.iran-mavad.com
4.19.2.2 Nonsteady-State Flow and Lag-Time

Fick’s second law describes the nonsteady-state rate of change of gas concentration at any point x > 0 in the membrane during
diffusion (13–15):

ðvc=vtÞ ¼ D v2 c=vx2 [6]
Consider now the case where the concentration of gas at the inlet of the membrane is constant and is initially (t ¼ 0) free of the
diffusing gas, and the diffusant entering the volume V is continually removed (e.g., with a desiccant), that is, zero concentration.
Subject to these boundary conditions
cðx; 0Þ ¼ 0; cð0; tÞ ¼ ca ; cðL; tÞ ¼ 0
the solution to eqn [6] for the time dependence of the quantity of gas per unit area passing through the membrane in time t is

Dca t Lca 2Lca X
N
ð1Þn Dn2 p2 t
Q¼ 2 exp [7]
L 6 p n¼1 n2 L2
For very long times, the exponential terms become negligible so that

Dca L2
Q¼ t [8]
L 6D
Plotted versus time, Q has a constant slope for long times, corresponding to the steady-state flux (Dca/L) of gas passing through
the membrane and a time axis intercept of L2/6D, which is defined as the lag-time, or the time needed to establish steady-state flow.
(Note that Q in eqn [7] is normalized to a unit area.)
4.19.2.3 Quasi-Steady-State Flow (Thin Diffusive Wall) Approximation

A quasi-steady-state analysis allows one to approximate time-dependent permeation using Fick’s first law. Following the analysis of
Tencer (15), consider an enclosure or package of volume V with no desiccant, separated from its environment by thin wall (L)
diffusion membrane of area A, as in Figure 4. Assume the walls of the enclosure and barrier membrane are nonabsorbing and
impermeable. For the quasi-steady-state condition of maintaining a linear gradient in permeant concentration across the
membrane, then we use Fick’s first law to probe the changes in permeant in the enclosure with time. The flux of permeant across the
barrier membrane is then equal to
ca0 cv0 ðca cv Þ ðca cv Þ
J¼D ¼ DK ¼P [9]
L L L
The rate of change for concentration of permeant inside the enclosure or package with volume V is
dcv JA Pðca cv Þ A
¼ ¼ [10]
dt V L V
Solving this equation for the initial condition, cv ¼ 0 at t ¼ 0 leads to

AP
cv ¼ ca 1 exp t [11]
VL
or
t VL
cv ðtÞ ¼ ca 1 exp where s ¼ [12]
s AP
The concentration of permeant in the enclosure will approach that in the outside environment with a time constant, s, which
depends on the volume of the enclosure and the permeation through the barrier membrane. Further, this equation relates how the
concentration of permeant in the enclosure tracks time-dependent changes in the outside environment.
4.19.2.4 Quasi-Steady-State Flow (General) Approximation

For the more general case of absorbing walls, Tencer’s analysis (15) leads to the time constant
VL þ Vwall LK=2
s¼ [13]
AP
where the barrier membrane is the only absorbing wall, Vwall ¼ AL and P ¼ KD leading to
VL L2
s¼ þ [14]
AP D
www.iran-mavad.com
Therefore for a thin membrane with fast diffusion, the volume of the enclosure and the permeability of the membrane will
dominate the time constant. If the volume of the enclosure is minimized, as in many electronic packaging schemes, and diffusion
through the membrane is slow, then the second term (s ¼ L2/D) will dominate the time-dependent response to the outside
environment.
4.19.2.5 Analysis of Solar Cell Front Sheet/Back Sheet

Kempe (16) has applied this quasi-steady-state approach to analyze the effect of encapsulant materials on the solar cell
lifetime. To prevent damage to solar cells by moisture, cells are encapsulated with a front sheet and a back sheet (16), each of
which consists of a thin-barrier membrane layer attached to a thick (400–500 mm) polymeric encapsulation material –
a common one is ethyl vinyl acetate (EVA) – which is thermally bonded to the solar cell. Of particular interest is how
well the front/back sheet design retards moisture ingress to the solar cell. Assuming that the contribution to s in eqn [14] is
dominated by the volume of the encapsulant and defining WVTR sat ¼ Pca =L, the saturation WVTR for the barrier membrane,
then
LE ca
s¼ where LE is thickness of the encapsulant [15]
WVTR sat
Since the volume V is not empty space but an encapsulant material, such as EVA, with a finite solubility and a saturation
concentration for moisture cEsat, then eqn [12] becomes
0
LE cEsat
cE ðtÞ ¼ cEsat 1 et=s where s0 ¼ [16]
WVTR sat
As determined by Kempe, for a typical barrier membrane with WVTR > 0.1 g m2 day1 and a polymeric encapsulant such as
EVA, equilibration occurs in w1 week. However, for WVTRw104 g m2 day1, the time to reach equilibrium could be w20 years.
And counter to intuition, a polymeric encapsulant with a larger solubility, or saturation moisture content can extend the time to
reach equilibrium.
4.19.3 Permeation Models for a Thin-Film Diffusion Barrier

4.19.3.1 Pinhole Model
In principle, a dense, defect-free, inorganic coating should be impermeable to gases. Therefore, the common explanation for gas
permeation through inorganic thin-film barriers is that they contain pinholes – or areas where there is no coating. Pinholes can form
because the substrate is not initially clean or due to process conditions during deposition of the coating. Rossi and Nulman (17)
have modeled the effect of coating flaws, or pinholes, on permeation. They find that for a single hole of dimension R, where R << L,
the wall thickness of the polymer – in our case the polymer substrate thickness – then the quantity of gas passing through the hole
per second is
Q
¼ 4DRcs g or cc s1 [17]
t
where D is the diffusion coefficient through the polymer, R is the ‘hole’ size or diameter, and cs is the concentration of gas permeant
at the polymer surface, for example, polyethylene terephthalate (PET). If we assume that the permeant is water vapor and that its
concentration at room temperature in the near PET surface equals the saturation concentration, cs ¼ 0.0051 g cm3,
D ¼ 4 109 cm2 s1 for PET, and R ¼ 1 mm, then Q/t ¼ 8.2 1015 g s1. If the density of these pinholes is 100 mm2 or
104 cm2, as reported for metal barrier coatings (18,19) then
Q
¼ WVTR ¼ 8:2 1015 104 ¼ 8:2 1011 g cm2 s1 ¼ 0:07g m2 day1
At
This is about what one expects for a very good packaging barrier. When R for the ‘hole’ is large compared to the PET thickness,
that is, R >> L, which is less relevant to an ultrabarrier, the result in this limit becomes
Q
¼ DpR2 cs =L [18]
t
This is the same as one would calculate from Fick’s first law, assuming the moisture concentration at the far surface of the PET is
maintained close to zero. In Figure 5 we show a plot of WVTR calculated using eqn [17] versus average ‘pinhole’ size and number
density. The trends are that WVTR increases with the number and size of pinhole defects. The more compelling message, however, is
that to achieve ultrabarrier performance – WVTR < 105 g m2 day1 – will require defect levels in the range approaching what is
practiced in the manufacture of integrated circuits (<10 cm2), a formidable challenge.
www.iran-mavad.com
Figure 5 Calculated dependence of WVTR on defect density and size.
4.19.3.2 Percolation Model for Permeation

What if one were able to, in a controlled laboratory environment, fabricate a barrier coating with no defects. Would that film be as
impermeable as a slab of quartz or other inorganic crystalline or glassy materials? This is the question our group recently examined.
After systematically measuring the WVTR of Al2O3 barrier films for a range of thickness and growth temperature, we proposed that
an alternative permeation mechanism in pinhole-free films could involve chemical complexing of Al2O3 with the permeant, which
then hops from site to site, eventually traversing the film (20). Reasoning by analogy with electrical breakdown phenomena in thin-
film dielectrics, we postulated that the permeation involved surface adsorption of the permeant, diffusion and chemical complexing
in the barrier film, aggregation of permeant into a continuous percolating network, followed by an abrupt increase in permeation.
We proposed that the time to permeation ‘breakthrough’ could be written as (21)

H0 g
sb ¼ s0 exp [19]
kT dox
where H0 is the exchange energy for water in aqueous solution (or vapor) with metal-oxide bound water. dox is the thickness of the
oxide barrier film, so that thinner layers reduce the time to permeation ‘breakthrough,’ kT is the usual thermal energy, and s0 and g
are constants. For molecular oxides, H0 is in the range 10–120 kJ mol1 (0.1–1.25 eV), which leads to permeation breakthrough
times of 1–108 s (20). In our model, H0 also depends on the film growth temperature. Growth of Al2O3 by atomic layer deposition
(ALD), which we studied, produces more chemically defective films at low temperature, consequently with more facile moisture
permeation. Figure 6 is a plot of eqn [19], assuming a prefactor of 100 (h) and g ¼ 125 Å. For lifetimes exceeding 1 year, H0 0.2 eV
and thickness > 100 Å.
Figure 6 Calculated time-to-failure for a barrier film, assuming a percolation model for gas permeation.
www.iran-mavad.com
4.19.4 Methods for Permeation Measurement
One of the challenges in developing an ultrabarrier technology is the lack of a commercial instrument with the required sensitivity.
This deficiency is particularly acute for potential manufacturers of an ultrabarrier product, who need to qualify the product,
preferably as it is being produced, but at least soon after it is fabricated, with a high degree of confidence. Unfortunately, the
techniques currently available to measure an ultrabarrier are time intensive, requiring the time integration of the small amount of
permeant or long equilibration time, and they lack the needed sensitivity. This section reviews the methods demonstrated to have
sufficient sensitivity to measure permeation through an ultrabarrier.
4.19.4.1 Coulometric Measurement of Water Vapor (22 )

In this scheme, a wet chamber at a selected temperature and humidity is sealed from a dry chamber by the barrier film to be
measured. Water vapor diffusing through the barrier film mixes with the gas in the dry chamber and is carried to a coulometric
sensor, comprising two electrodes, an anode and a cathode, separated from each other by an electrolyte confined in a porous
separator material. Water entering the sensor is absorbed in the electrolyte and ionized by the constant electrical potential applied to
the cell to 2Hþ and O2, which are, respectively, reduced at the cathode and oxidized at the anode, producing a measurable
electrical signal or current proportional to the charge transfer. This method then directly measures the amount of water vapor
permeating the film electrically. The MOCON Corporation, which sells an instrument, the Aqutran 1, based on this coulometric
method to determine WVTR, advertises a sensitivity of at least 5 104 g m2 day1. A schematic diagram of this instrument is
shown in Figure 7. From experience with this instrument in our lab, repeatable measurements require minimally w24 h of data
acquisition to achieve equilibrium or steady-state conditions.
4.19.4.2 Ca-Method
4.19.4.2.1 The Optical Ca-Method ( 11,23,24)
Because it reacts so readily with water vapor or oxygen, Ca metal can be a very sensitive probe of combined oxygen and water vapor
permeation. In the optical Ca-test, shown schematically in Figure 8, an initially thin (w50 nm) semitransparent Ca metal film is
postulated to gradually and uniformly become more optically transparent, as the metal reacts with permeating oxygen or water
vapor to form CaO or Ca(OH)2, which are both optically transparent. The increase in optical transmission, at a particular wave-
length of light (e.g., 600 nm), can be approximately related to WVTR by the relationship (25)

WVTRz0:55½ðDT=TÞ=Dt g m2 day1 [20]
This expression was derived using optical absorption for Ca at 600 nm. Here, DT/T corresponds to the fractional increase in
optical transmission during a time interval Dt (h) of exposure to high temperature and humidity, or damp heat. A 10% fractional
change in optical transmission in 1000 h of damp heat exposure would therefore correspond to a WVTR ¼ 5.5 105 g m2 day1.
Generally, since gases can permeate along edge-seals in the test structure, a reference structure with an impermeable lid, for example,
glass, determines the practical limit for WVTR measurement. This can vary with the test structure geometry and the integrity of the
edge-seal. But for our test structure, which is sealed with a UV-curable epoxy, the limit can be <105 g m2 day1 at 38 C/85% RH.
4.19.4.2.2 The Electrical Ca-Method ( 26,27)

Oxidation of Ca metal also increases its electrical resistance, which can also be a probe of gas permeation. Figure 9 illustrates a test
structure for this electrical Ca-test. If one assumes that oxidation occurs uniformly through the Ca film thickness, which is also
assumed in the optical Ca-test, then the change in resistance can be related to WVTR. Paetzold et al. (26) derived the following
expression
MðH2 OÞ dð1=RÞ l
WVTR ¼ 2 dCa rCa [21]
MðCaÞ dt w
Figure 7 Schematic diagram of equipment for measuring WVTR using a coulometric sensor.
www.iran-mavad.com
Figure 8 Optical Ca-method for measuring WVTR. Reprinted with permission from Carcia, P. F.; McLean, R. S.; Groner, M. D.; Dameron, A. A.; George,
S. M. J. Appl. Phys. 2009, 106, 023533. Copyright 2009, American Institute of Physics.
Figure 9 Electrical resistance Ca-method for measuring WVTR. Reprinted with permission from Schubert, S.; Klumies, H.; Muller-Meskamp, L.; Leo, K.
Rev. Sci. Instrum. 2011, 82, 094101. Copyright 2011, American Institute of Physics.
dð1=RÞ
which is proportional to the time rate of change of the inverse electrical resistance of the Ca thin film. In this expression,
dt
the factor 2 corresponds to the number of molar equivalents of H2O that react with Ca to form Ca(OH)2, M(H2O), and M(Ca) are
the molar masses of H2O and Ca, respectively, dCa is the mass density of Ca, rCa is the electrical resistivity of the Ca film, and l and w
are the respective length and width of the Ca film. With a glass reference barrier, the authors report measurement sensitivity below
106 g m2 day1 with this electrical Ca-test method. For a 100-nm-thick Ca resistor (l/w ¼ 5), this would correspond to a change in
resistance (DR) < 103 U in 1000 h.
4.19.4.3 The Radioactive (Tritium) Method (28,29 )

Another sensitive scheme for measuring moisture permeation is to add a radioactive tracer material, such as tritiated water (HTO),
to the permeating H2O and then use a sensitive scintillation counter to determine the amount of accumulated HTO that permeates
the barrier. Applying the ratio of HTO/H2O allows calculation of WVTR. In one scheme, the HTO–H2O mixture is captured by a LiCl
salt in a vial suspended on the detection side of the barrier (28). The vial is periodically removed for offline measurement of the
radioactive tracer and replaced with fresh LiCl salt. Figure 10 shows a schematic of such a cell. More specific details are described by
Groner et al. (28). With an Al metal foil as the reference barrier, these authors deduced an ultimate measurement sensitivity of
w1 106 g m2 day1. Additional information about using HTO as a radioactive tracer material to measure WVTR can also be
found in published work by the General Atomics group (29). They quote a theoretical detection limit of 2.4 107 g m2 day1.
Needless to say, only a few labs in the world are equipped to safely handle radioactive materials, so this has limited implementation
of this technique.
Figure 10 Schematic diagram for measuring WVTR by a radioactive tracer method with tritium. Reprinted with permission from Groner, M. D.; George,
S. M.; McLean, R. S.; Carcia, P. F. Appl. Phys. Lett. 2006, 88, 051907. Copyright 2006, American Institute of Physics.
www.iran-mavad.com
Figure 11 Schematic of mass spectroscopy scheme to measure WVTR. Reprinted with permission from Zhang, X. D.; Lewis, J. S.; Wolter, S. D.; Parker,
C. B.; Glass, J. T. J. Vac. Sci. Technol. 2007, A25, 1587. Copyright 2007, American Vacuum Society.
4.19.4.4 Mass Spectroscopy Method (30–33 )

Direct measurement of gas permeation by mass spectroscopy has been studied by several groups. While a quadrupole mass
spectrometer does have high intrinsic sensitivity for residual gas detection, the lowest detectable permeation rate is limited prac-
tically by the background partial pressure due to chamber outgassing, virtual leaks, and leaks around the barrier seal. Zhang et al.
(31) have attempted to minimize these effects with their sequential measurement approach. A schematic of their system is shown in
Figure 11. The stepwise method consisted of (1) initiation of permeation upstream, (2) accumulation of the permeant in the
downstream chamber for a fixed duration, (3) opening a valve that isolates the separately pumped accumulation and detection
chambers for a fixed time, and (4) closing this isolation valve and evacuating the accumulation and the detection chambers to their
initial ultrahigh vacuum (UHV) background pressure. This sequence is repeated multiple times until a steady-state pressure rise is
measured by the mass spectrometer. The pressure rise is then used to calculate the permeation rate through the barrier film. Some of
the other critical steps include baking and pumping out the accumulation chamber to a background <2 109 Torr; prior to
introducing permeant, the upstream chamber is pumped to <5 108 Torr, while the volume around the Viton o-ring seal to the
barrier is differentially pumped to at least 102 Torr. With this method, the detection limits are reported to be
1.9 107 g m2 day1 for water vapor and 2.8 102 cm3 m2 day1 for oxygen.
If the WVTR to be measured is w1 106 g m2 day1 or 1.16 1015 g cm2 s1 since 1 mol equivalent of H2O is 18 g mol1,
this corresponds to 6.4 1017 mol cm2 s1. If the diameter of the barrier under test is 2 inches and the accumulation time is
100 s, then the number of moles of water vapor collected is Dn ¼ 1.3 1013 mol. If the detection volume is 1000 cm3, we can
calculate the corresponding pressure rise using the ideal gas law

DnRT 1:3 1013 ð0:082Þð298Þð760Þ
Dp ¼ ¼ ðTorrÞ ¼ 2:4 109 Torr
V 1
To confidently measure this change in pressure, the residual partial pressure of water should be <1 1010 Torr; that is, the
background pressure in the detection chamber needs to be in the UHV range.
4.19.5 Ultrabarrier Coating Strategies

4.19.5.1 Overview
If we use the pinhole model of Rossi and Nulman, the results of which are shown in Figure 6, then a single-layer, ultrabarrier film
must have a very low density of defects, in the range 1–10 cm2. To put this in perspective, these defect levels begin to approach the
tolerable limit for manufacturing integrated circuits, which is carried out in ultraclean rooms on flat, rigorously cleaned Si wafers,
handled robotically. By contrast, the cleaning of commercial plastic film, the barrier substrate, is much less rigorous, and the surface
www.iran-mavad.com
Figure 12 Illustration of increased gas permeation diffusion path in a bilayer laminated barrier structure. Reprinted with permission from Graff, G. L.;
Williford, R. E.; Burrows, P. E. J. Appl. Phys. 2004, 96, 1840. Copyright 2004, American Institute of Physics.
is not flat, with intrinsic topographical features and residual particulates. Further, whereas Si wafers are typically only as large as
300 mm, and batch processing is in a vacuum environment, for economic reasons an ultrabarrier must be produced in a wide
(0.3–2 m) roll of plastic film by a continuous, high line-speed process from an unwind to a windup roll, which can be either inside
or outside the process chamber. In light of these considerations, the prospect for achieving ultrabarrier performance in a manu-
factured film product with a single-barrier layer seems, at first glance, improbable. Recently, however, reports of single-layer,
ultrabarrier films (11,28,34,35), which are grown by ALD (36), a highly conformal gas-phase deposition method, have begun to
challenge this doubt. Nonetheless, the popular strategy for fabricating an ultrabarrier is based on multilayer thin films (37–39).
The general concept of a multilayer barrier film is to decouple the defects in individual barrier layers with separation layers,
which introduce long diffusion paths between defects and consequently decrease overall gas permeation. This is illustrated sche-
matically in Figure 12. Multilayer barriers can assume different configurations. The barrier layer is, however, always a low-
permeation inorganic thin film, which can be fabricated by a physical or chemical vapor deposition process. Common barrier
materials are Al2O3, SiO2, and SiNx. The multilayer structure can be produced by laminating separate barrier-coated polymer films,
using an adhesive (38). The structure can also be directly built up by alternating inorganic barrier layers (37) with an organic layer
that separates or decouples the inorganic barrier layers, such as a polymer, or it could be a ‘sealing’ (39) layer. The ‘sealing’ layer may
be vacuum deposited or wet coated. A polymer separation layer can be deposited by vacuum evaporation, with or without a cross-
linking step. Because its development is the most mature, the focus in the next section will be on polymer-inorganic multilayer
barriers (37).
4.19.5.2 Polymer-Inorganic Multilayer Thin-Film Barrier

The basic structure consists of alternating layers of organic and inorganic materials, as indicated by the cross-section micrograph in
Figure 13. The purpose of the first organic layer is to planarize the polymer film substrate, covering up submicron, residual
particulates, not removed by conventional plastic film cleaning methods. Next, an inorganic barrier layer is deposited on the organic
layer. Growing the inorganic barrier layer on the smooth organic layer presumably results in fewer defects because of the reduction
of surface-induced defects and asperities. The combination of one organic and one inorganic layer is referred to as a dyad. Typical
thicknesses are w200–500 nm for the organic layer, whose chemistry can be a polyacrylate, and 20–50 nm for the inorganic barrier,
which can be sputtered Al2O3. Increasing the number of dyads is postulated to reduce gas permeation, because the additional
organic layers ‘decouple’ defects in the separated barrier layers, creating long permeation pathways.
Figure 13 Electron microscope image of a PML barrier film consisting of 5 dyads of polymer/inorganic thin-film layers. Reprinted with permission from
J. Lewis, Materials Today 2006, 9, 38. Copyright 2006, Elsevier.
www.iran-mavad.com
However, in their analysis of multilayer barrier structures, Graff et al. (40) showed that the dominant mechanism for permeation
was a long lag-time and not low-equilibrium diffusion. Lag-times were calculated using the general formalism of Ash et al. (41) for
diffusion through multilayer structures. Instead of the physical thickness of the polyacrylate film separation layer, they used a much
longer, average spatial distance between defects in the inorganic barrier layers, which they deduced from measurement of simple
structures, to calculate lag-time. The calculated dependence of lag-time on the number of dyads and the average defect spacing is
shown in Figure 14. And Table 1 compares calculated and reported values of (steady-state) WVTR versus the number of dyads.
Remarkably, the calculated equilibrium WVTR has only a weak dependence on the number of dyads, achieving
w7 103 g m2 day1 for 5 dyads.
Using film properties that they deduced for the individual polyacrylate and the sputtered Al2O3 layers, and an effective defect
spacing of 100 mm, the calculated lag-time for 5 dyads was 1752 h (w73 days). When the average defect spacing increased, which
corresponds to a barrier film with lower WVTR for the same average defect size, a lag-time of several years was possible. It is
important to emphasize that long lag-times derived in this analysis scale with lower defect density (larger defect spacing) for the
barrier. Improving the quality of the inorganic barrier layer has a larger effect on increasing the lag-time than adding layers or dyads.
Vogt et al. (42) at the National Institute of Standards and Technology (NIST) also studied the mechanism for moisture
permeation in polymer multilayers (PMLs), using X-ray and neutron reflectivity measurements. They deduced that moisture
accumulates slowly at the Al2O3/polymer interface. This desiccant effect retards moisture transport through the multilayer barrier,
and interfacial H2O strongly absorbs to the Al2O3. After moisture accumulates, it is very slow to leave. In fact, Vogt et al. deduce
longer lag-times (85 days or w2000 h for a 3-dyad structure) than those calculated by Graff et al. (40) from a diffusion model for
permeation in PMLs.
Before the onset of equilibrium gas permeation, corresponding to the lag-time, the measured WVTR is typically below the
measurement limit of even the most sensitive techniques. For some devices, such as OLEDs, where a 1- to 2-year lifetime may be
adequate, a multilayer with long lag-time may be an option as an ultrabarrier. For flexible solar cells with lifetime expectations ranging
from 5 to 25 years, a barrier whose performance is based on lag-time alone may not be a good choice as an ultrabarrier. In this regard,
one cautionary note about a multilayer barrier relates to the interpretation of accelerated lifetime test results. This testing assumes that
degradation in performance at high temperature and humidity for short testing times predicts performance for longer times at the less
demanding conditions of normal device operation. In other words, the mechanism for loss of performance depends on temperature
and humidity in accord with a predictable rate law. These results may be misleading for a multilayer barrier, since the rate laws are
likely to be different in two distinct regimes – nonequilibrium or lag-time regime, and equilibrium or steady-state WVTR regime.
Complementary to Graff ’s study of lag-time in a multilayer barrier stack, Greener et al. (38) modeled steady-state permeation for
a simple laminate barrier structure and multilayer barrier stacks. The laminate structure consisted of two polymer films, each coated
on one side with a thin inorganic barrier layer, laminated together by a thin adhesive that bonded the barrier-coated sides to each
other. Varying the adhesive thickness (La) relative to the average pinhole size (Lh), the model showed two distinct regimes for WVTR.
Figure 14 Dependence of lag-time on the number of dyad pairs in a PML barrier as a function of defect separation distance. Reprinted with permission
from Graff, G. L.; Williford, R. E.; Burrows, P. E. J. Appl. Phys. 2004, 96, 1840. Copyright 2004, American Institute of Physics.
www.iran-mavad.com
Table 1 Predicted versus measured WVTR (steady-state) values for multilayer barrier coatings on
177.7 mm PET substrates
Numbers of dyads Reported WVTR (g m2 day1) Calculated WVTR (g m2 day1)
0 4.7 6.8
1 0.07 0.036
2 <0.005 0.0181
3 <0.005 0.0121
4 <0.005 0.0091
5 w106 0.0073
Reprinted with permission from Graff, G. L.; Williford, R. E.; Burrows, P. E. J. Appl. Phys. 2004, 96, 1840. Copyright 2004,
American Physical Society.
In Regime I with La < Lh, WVTR decreases for thinner adhesive layer thickness (La). For very thin adhesive (La<0.01 Lh), WVTR was
reduced by several orders of magnitude, as shown in Figure 15. In Regime II with adhesive thickness greater than the average,
modeled pinhole size, the WVTR was relatively independent of La. For a multilayer stack, the trends in WVTR with modeled organic
layer thickness were similar. Improvements in WVTR occurred when the organic layer became thinner than the pinhole size. As
Greener points out, however, for pinhole sizes (0.5–2 mm) experimentally reported in barrier films, very thin adhesive layers are
needed to exploit this barrier improvement effect. Neither Greener, for a series of practical laminate structures with La w 2–10 mm,
nor Graff, for multilayer stacks with the organic layer thickness in the range 0.1–3 mm, report an improvement in WVTR for thinner
organic layers.
4.19.5.2.1 Applications for PML Thin-Film Barriers

4.19.5.2.1.1 Flexible OLEDs
The application of polymer/inorganic barrier layers to protect OLED devices was reported in a joint study by Universal Display
Corporation and the Pacific Northwest National Laboratory (PNNL) (43). The test structure consisted of a flexible PET substrate
175 mm thick on which was coated 5 dyads of a polyacrylate/Al2O3 inorganic barrier film. Green-emitting (514 nm peak wave-
length) OLED devices were fabricated on the barrier-coated PET and sealed to it with a glass lid using a UV-cured epoxy edge-seal.
Accelerated lifetime testing of encapsulated devices was carried out in ambient air with devices driven with a constant current of
2.5 mA cm2, which produced an initial luminance of 425 cd cm2. The luminance was monitored with time and compared to
a glass–glass OLED device, driven with a constant current of 2.6 mA cm2 (600 cd cm2). Figure 16 shows normalized luminance
for the glass–glass and PET-glass devices out to 4000 h of testing. The performance is comparable for the device built on PET with
a barrier, decaying to 50% of its initial luminance in 3800 h, compared to 69% for the device on glass. The slower decay for the
OLED device on glass is attributed to a better edge-seal for the epoxy to glass than to the plastic.
4.19.5.2.1.2 CIGS solar cells

As we have previously indicated, CIGS solar cells are susceptible to loss in efficiency when they are exposed to high temperature and
humidity (damp heat). One of the qualification tests for Si solar cell modules requires less than 5% loss in efficiency after 1000 h
Figure 15 Dependence of WVTR in a bilayer laminated barrier as a function of the lamination adhesive thickness (La), normalized to the defect hole size
(Lh). Reprinted with permission from Greener, J.; Ng, K. C.; Vaeth, K. M.; Smith, T. M. J. Appl. Polymer Sci. 2007, 106, 3534. Copyright 2007, John
Wiley and Sons.
www.iran-mavad.com
Figure 16 Comparison of relative luminance lifetimes in OLED devices protected with glass and a PML barrier (graph). Inset photos are of a single
OLED pixel with plastic barrier at (i) 1200 h and (ii) 3000 h, which show no increase in defect number or size. Reprinted with permission
from Weaver, M. S.; Michalski, L. A.; Rajan, K.; Rothman, M. A.; Silvernail, J. A.; Burrows, P. E.; Graff, G. L.; Gross, M. E.; Martin, P. M.; Hall, M.; Mast, E.;
Bonham, C.; Bennett, W.; Zumhoff, M. Appl. Phys. Lett. 2002, 81, 2929. Copyright 2002, American Institute of Physics.
stored at 85 C/85% RH. Thin-film CIGS manufacturers have adopted an analogous criterion (<10% reduction in efficiency) for
1000 h in damp heat.
The group at PNNL tested the effect of damp heat on the performance of commercial mini-modules of CIGS solar cells,
encapsulated with PML barriers (44–47). The barrier structure consisted of 5 dyads. Although loose surface debris from the module
fabrication was cleaned prior to encapsulation, it was still necessary to first coat the module with a thick first polymer layer and
100 nm Al2O3, because of residual surface irregularities. Subsequent dyads consisted of 100 nm polymer and 50 nm Al2O3.
Figure 17 compares module efficiency stored at 60 C/90% RH for up to 1500 h. When the initial polymer planarizing layer was
4 mm thick, modules maintained their efficiency for the test period. Modules with a planarizing layer only 0.5 mm thick, however,
began to lose efficiency after 500 h, while unencapsulated modules lost efficiency immediately, becoming inactive after 1000 h.
For testing at 85 C/85% RH, the loss in performance was significant, even with thick planarizing layers, as shown in Figure 18.
Modules lost 40% efficiency after 1000 h at 85 C/85% RH. Unencapsulated or ‘bare’ modules fared much worse, with 80% reduced
efficiency in less than 100 h. The PNNL group also tested lifetime at 85 C/85% RH for laboratory CIGS solar cells, fabricated at the
Institute of Energy Conversion (IEC) at the University of Delaware. Those cells had a transparent electrode that was indium–tin
oxide (ITO), instead of less expensive aluminum–zinc oxide (AZO) more commonly used in commercial modules. AZO is much
more moisture sensitive than ITO. The lab solar cells with vacuum evaporated electrodes also are cleaner and have less topography
than commercial solar modules. CIGS solar cells were fabricated at IEC and encapsulated with 5 dyads of a polymer-inorganic
Figure 17 Change in efficiency at 60 C/90% RH for CIGS solar cells protected with 5 dyads of a PML barrier with different thickness of the first polymer
smoothing layer. Reprinted with permission from Olsen, L. C.; Kundu, S. N.; Bonham, C. C.; Gross, M. E. In Proc. 31st IEEE Photovoltaic Specialists
Conference, 2005, p 327. Copyright 2005, Institute of Electrical and Electronics Engineers.
www.iran-mavad.com
Figure 18 Change in efficiency at 85 C/85% RH for CIGS solar cells protected with 5 dyads of a PML barrier with different thickness of the first polymer
smoothing layer. Reprinted with permission from Olsen, L. C.; Gross, M. E.; Kundu, S. N.; Graff, G. L. In Proc. 33rd IEEE Photovoltaics Specialists
Conference, 2008, p 1. Copyright 2008, Institute of Electrical and Electronics Engineers.
multilayer barrier, which provided exceptional protection at 85 C/85% RH for 2500 h, after which cells did experience some loss in
efficiency, attributable to moisture infiltration from the cell edges.
4.19.5.2.2 Manufacturing PML Thin-Film Barrier

The challenge for any ultrabarrier technology will be to achieve consistently low permeability in a manufactured barrier product at
low cost. Figures 19 and 20 show a batch and a continuous roll-to-roll system, which are installed at the PNNL, for depositing
polymer-inorganic multilayer barriers (44,47). The throughput for a roll-to-roll tool is reported to be 2–3 m min1, which is limited
by the sputtering rate for Al2O3. Many factors determine cost for this barrier product, including materials, the number of dyads, and
the deposition rate for the individual processes, to name a few. The final application, for example, a solar cell, pricing will ultimately
determine what is or not an acceptable cost for an ultrabarrier product. As an illustration, if we assume a solar cell manufacturing
process achieves the cost objective of $1 per watt for a 15% efficient solar cell, this corresponds to $150 m2, for solar radiation of
1000 W m2. If 10% of the solar cell cost could be allocated to the barrier layer, then a selling price of $15 m2 for the barrier would
be acceptable. Meeting this price objective with a 5-dyad PML barrier could be challenging.
4.19.5.3 General Comments on Single-Layer Thin-Film Barriers

In Section 4.19.3.1, we showed from the pinhole model of barrier films that ultrabarrier performance requires both a low
concentration (<100 cm2) of defects and a small average defect size (<100 nm). As Graff et al. (40) tabulated, and is repeated here
Figure 19 PNNL inline batch coating system for producing polymer PML barrier thin films. Reprinted with permission from Olsen, L. C.; Gross, M. E.;
Graff, G. L.; Kundu, S. N. Reliability of Photovoltaic Cells, Modules, Components, and System. In Proc. SPIE; Dhere, N. G., Ed., 2008; Vol. 7048, p 1.
Copyright 2008, International Society for Optics and Photonics.
www.iran-mavad.com
Figure 20 PNNL roll-to-roll continuous deposition system for producing PML barrier thin films. Reprinted with permission from Olsen, L. C.; Gross, M.
E.; Graff, G. L.; Kundu, S. N. Reliability of Photovoltaic Cells, Modules, Components, and System. In Proc. SPIE; Dhere, N. G., Ed., 2008; Vol. 7048,
p 1. Copyright 2008, International Society for Optics and Photonics.
Table 2 Tabulation of measured defect sizes and defect densities for single-layer barrier films deposited on polymeric substrates
Defect radius (mm) Defects density (mm2) Coating material Deposition method Substrate
0.6 11–1100 SiO2 PECVD PET

0.6 5–1000 Si3N4 PECVD PET
1.0 25–400 Al Evaporated PET
1.0 100–300 Al Evaporated PET
2–3 200 Al Sputtering PET
0.5–1.4 600 AlOxNy Sputtering PET
0.4 100–1000 Al Evaporated BOPP
0.5 700 AlOxNy Sputtering PET
Reprinted with permission from Graff, G. L.; Williford, R. E.; Burrows, P. E. J. Appl. Phys. 2004, 96, 1840. Copyright 2004, American Physical Society.
in Table 2, this requirement is at odds with experimental results for single-layer barriers films grown by physical vapor deposition
(PVD) and chemical vapor deposition (CVD) methods on polymeric substrates. A defect density of w10 000 cm2 with an average
size 0.5–3 mm is common. The likely source of these film defects is debris and chemical contamination on the polymer substrate
surface. Pinholes in the coating can form because debris and particulates shadow film growth, leaving uncoated local areas of the
substrate; chemical contamination can lead to poor adhesion and subsequent local film delamination, or poorly adhered debris can
fall off the substrate, again leaving an uncoated area. These film defects, scratches or pinholes, can form either during the deposition
process or afterward, during film rewinding, when the coating is pressed against the backside of the polymeric film. Of course,
substrate cleanliness is no less an issue when the barrier is directly deposited on a device. In Section 4.19.5.2.1, we showed that even
with a thick planarizing layer to ‘cover’ substrate defects and irregularities, a 5-dyad multilayer barrier did not adequately protect
a CIGS module in the damp heat test. Therefore, one of the key enablers of an ultrabarrier coating is a clean substrate. Strategies for
polymer substrate cleaning and surface modification are reviewed in Charles Bishop’s book on roll-to-roll barrier coatings (48).
Although it is not possible to circumvent substrate cleaning to produce an ultrabarrier coating, a barrier film grown by atomic
layer deposition (ALD) may be more tolerant of substrate surface irregularities and small (submicron) defects, because it con-
formally (49) coats the substrate. And because of the growing interest in barrier films grown by ALD, including our own extensive
research, single-layer barrier films grown by ALD will be the focus of the following sections (11,24,28,34,35).
www.iran-mavad.com
4.19.5.4 Single-Layer Ultrabarriers Grown by ALD

4.19.5.4.1 ALD Growth Process Description
An overview of ALD can be found in the review article by Professor Steven George (University of Colorado) (36). ALD deposits
a film by sequential, self-limiting surface reactions from gas-phase precursors, or reactants. As an example, Al2O3 can be grown in
a reactor by thermal ALD using trimethyl aluminum (TMA) and water. The process steps are (1) dose the reactor with TMA, which
ideally deposits only a monolayer of TMA on the heated substrate surface; (2) allow adequate purging with an inert gas to remove
excess TMA from the reactor atmosphere; (3) dose with H2O, which reacts with TMA to form a near monolayer of aluminum oxide
on the substrate with release of the gaseous CH4 by-product; and (4) purge the reactor with inert gas to remove excess H2O. Repeat
steps 1–4 as many times as needed to grow the desired thickness of Al2O3. The steps in the ALD process are illustrated schematically
in Figure 21. Because reasonably dense films of Al2O3 can be grown by thermal ALD, even close to room temperature, ALD Al2O3 is
particularly attractive for growing a barrier film on temperature-sensitive substrates and devices. Al2O3 is also optically transparent,
which is a requirement as a coating for a display or the light collection surface of a solar cell. Over most of the temperature range of
interest ALD Al2O3 has an amorphous structure with a featureless microstructure, both desirable structural attributes for a low
permeation coating, because of the absence of crystalline grain boundaries or growth boundaries. Therefore, many of the following
sections will describe the barrier properties of thermally grown ALD Al2O3.
4.19.5.4.2 ALD Thin Films: General Properties

The key attributes of ALD for a barrier coating are that the process steps promote layer-by-layer growth, the coating is conformal,
and low-temperature growth, compatible with plastic substrates and thermally sensitive electronic devices (e.g., OLEDs), is
possible. Of the three basic thin-film growth modes (50) illustrated in Figure 22, Frank-van der Merwe or layer-by-layer growth
best describes the epitaxial growth of a monocrystalline thin film on a lattice-matched crystal template. Generally, epitaxial
growth is carried out at high temperature, which drives surface diffusion of deposited adatoms and atomic organization layer by
layer. The other growth modes, typical of sputtering or evaporation, are characterized by island growth that leads to a faceted
microstructure or columnar film features, along which gas permeation can be facile. What is unique about ALD is that dense,
layer-by-layer growth can occur at low temperature, on a noncrystalline substrate, and the film’s structure can be ‘glassy’ or
amorphous with a featureless microstructure, ideal for a gas permeation barrier or a gate dielectric in a field-effect transistor
device. Often, ALD coatings are described as pinhole free, which is deduced from their superior electrical properties and is the key
reason why thin gate dielectrics grown by ALD have replaced thermally grown SiO2 on Si in the highest performance thin-film
transistors for integrated circuits.
Conformal film growth, illustrated in Figure 23 for ALD, also brings unique advantages to applications as an ultrabarrier. If the
surface to be coated has significant topography, ALD will deposit a continuous layer on that surface. One example is a solar cell. The
metal electrode grid on the solar collection surface can be w1–15 mm thick. Whereas these metal lines would shadow film growth by
sputtering or other PVD or CVD methods, leaving uncoated areas on the solar cell surface, ALD conformally and uniformly coats the
metal and the surface of the solar cell together (51). An ALD coating could also ‘cover up’ surface irregularities on a polymeric
substrate, mitigating the need for a separate thick polymer planarizing layer. Of course, this presumes that an initial substrate
cleaning has removed any large, poorly adhered surface debris and that only submicron surface irregularities and particulates
remain. It may also be possible to transform a barrier film for packaging (WVTR w1 g m2 day1) to an ultrabarrier with a thin ALD
coating that ‘seals’ the original barrier defects or pinholes. This possibility will be discussed later in more detail.
Figure 21 Steps in a single cycle of ALD process to produce Al2O3 from trimethyl aluminum and water. Reprinted from www.Beneq.com with
permission from Beneq Oy.
www.iran-mavad.com
Figure 22 Thin-film growth modes versus surface coverage (Q): (a) Island growth – Vollmer-Weber mode; (b) Layer-by-layer growth – Frank-van der
Merwe mode; (c) Mixed growth – Stranski–Krastanov mode.
Figure 23 Example of conformal growth with 300-nm Al2O3 grown by atomic layer deposition in Si trench structure. Reprinted with permission from
Ritala, M.; Leskela, M.; Dekker, J-P.; Mutsaers, C.; Soininen, P. J.; Skarp, J. Chem. Vap. Deposition 1999, 5, 7. Copyright 1999, John Wiley and Sons.
4.19.5.4.3 Ultrabarrier Properties of Single-Layer ALD Al2O3

To demonstrate the outstanding barrier properties of ALD Al2O3 thin films, we grew 25-nm-thick films in a hot-wall ALD flow
reactor at 120 C using TMA (Aldrich) and water (Fisher HPLC-grade) with intermediate nitrogen gas-purging steps (11). The plastic
substrates were 125-mm-thick heat-stabilized polyethylene naphthalate (PEN) (Teonix Q65, DuPont-Teijin Films). Prior to ALD
deposition, the individual plastic substrates were stream-rinsed with methanol, mildly scrubbed with a soft lens tissue, followed by
rinsing in methanol and water, and finally blow dried with dry N2. We verified with atomic force microscopy (AFM) surface imaging
before and after cleaning that this protocol significantly reduced surface particulates without scratching the surface. These clean
substrates were loaded into the reactor residing in a laminar-flow hood operated at Class 100 cleanroom conditions. The substrate
cleanliness and surface chemistry of the polymer were critical for achieving thin, defect-free films with ultralow permeability.
The WVTR for the Al2O3 ALD-coated PEN was measured using the optical Ca-test method (Section 4.19.4.2.1). The change in
optical transmission after aging the test structure at 38 C/85% RH and 60 C/85% RH for a 25 nm Al2O3 ALD barrier on PEN is
shown in Figure 24. There was a small initial increase in %T followed by a slower, constant steady-state change. When the aging
temperature was increased from 38 to 60 C, the optical transmission abruptly changed to a higher steady-state rate. The initial
change in %T was attributed to the outgassing of residual H2O or air in the epoxy that reacted with the Ca film.
From the steady-state change in %T at 38 C/85% RH, the WVTR was 1.7 105 g m2 day1. At 60 C/85% RH, the WVTR was
6.5 105 g m2 day1. These measurements were both for single Al2O3 ALD barriers. At 38 C/85% RH, we calculated
www.iran-mavad.com
Figure 24 Representative change in optical transmission in Ca-test structure protected with a PEN lid coated with 25-nm-thick ALD Al2O3. Reprinted
with permission from Carcia, P. F.; McLean, R. S.; Reilly, M. H.; Groner, M. D.; George, S. M. Appl. Phys. Lett. 2006, 89, 031915. Copyright 2006,
American Institute of Physics.
a permeability of P ¼ 7.33 1019 g-H2O cm1 s1 atm1. Using an estimate of solubility (40) for sputtered Al2O3 of
S ¼ 0.029 g cm3 atm1, the diffusion coefficient was calculated to be D ¼ P/S ¼ 2.5 1017 cm2 s1. From Graff ’s analysis, this
diffusion coefficient corresponds to a single 10-nm-diameter pinhole per square millimeter of coated barrier film (40).
From the data at 38 C and 60 C, we deduced an apparent activation energy Ea ¼ 52 kJ mol1 for WVTR in a single 25-nm-thick
Al2O3 ALD barrier. Using this activation energy, the WVTR extrapolated to 23 C was 6 106 g m2 day1 for the ALD Al2O3 ALD
film. Using the pinhole model (Section 4.19.3.1) of Rossi and Nulman (17) with one 10 nm defect per square millimeter, we
calculated a WVTR at room temperature of 3.5 106 g m2 day1, in reasonably good agreement with our experimentally
extrapolated value.
As further evidence for the outstanding barrier properties of a single-layer ALD barrier, Figure 25 shows 3-year aging data along
with a schematic of the test structure (21). It consists of a PEN barrier lid coated at 120 C with a 25-nm-thick ALD Al2O3 layer,
which is epoxy-sealed to thin Ca pixels (initially w62.5 nm thick) deposited on glass. Initially, the structure was aged at 38 C/85%
RH and 60 C/85% RH for a total of 2470 h. Corresponding to the steady-state change in optical transmission for aging at
60 C/85% RH, we calculated a WVTR ¼ 3 105 g m2 day1. Subsequently, this test structure was stored in the laboratory
ambient for 3 years. After 3 years, no visible defects formed in the Ca, as illustrated in this figure, and only a uniform increase in
optical transmission occurred. From the increase in optical transmission from w8 to w16%, we calculate that about 10 nm of Ca
had oxidized to Ca(OH)2 in 3 years, corresponding to an overall WVTR w 105 g m2 day1, which is about the same rate that
occurs with an impermeable glass lid; therefore oxidation of Ca can be attributed to edge diffusion of H2O through the epoxy-seal.
This is remarkable, when one considers that thin Ca, if exposed to ambient atmosphere, fully oxidizes to a transparent film in less
than 1 min! A single thin ALD layer does indeed provide long-term ultrabarrier performance.
Figure 25 Cumulative change in optical transmission, corresponding to WVTR w 105 g m2 day1 over 3 years, for a Ca-test structure protected by
a PEN lid coated on one side with 25 nm ALD Al2O3. Reprinted with permission from Carcia, P. F.; McLean, R. S.; Sauer, B. B.; Reilly, M. H. J. Nano.
Sci. Technol. 2011, 11, 7994. Copyright 2011, American Scientific Publishing.
www.iran-mavad.com
Figure 26 Barrier film thickness dependence of WVTR measured at 38 C/85% RH and 60 C/85% RH for a single layer of ALD Al2O3 grown at
125 C on PEN. Reprinted with permission from Carcia, P. F.; McLean, R. S.; Groner, M. D.; Dameron, A. A.; George, S. M. J. Appl. Phys. 2009, 106,
023533. Copyright 2009, American Institute of Physics.
4.19.5.3.3.1 Threshold ultrabarrier thickness

A critical characteristic of thin-film ALD Al2O3 is the minimum or threshold thickness needed for ultrabarrier performance. The
thickness dependence (24) of WVTR for ALD Al2O3 films grown at 125 C on PEN is shown in Figure 26. WVTR was determined
using the optical Ca-method for films stored at 38 C/85% RH and 60 C/85% RH. Reference data for the uncoated PEN and a glass
lid are also included in this figure. A steep reduction in the WVTR is observed for an Al2O3 ALD film with a critical thickness of
w5 nm, corresponding to the onset of ultrabarrier properties. For Al2O3 ALD film thicknesses >10 nm, WVTRs were all reduced to
the limit of the Ca-test of 5 105 g m2 day1 at 38 C/85% RH and 4 104 g m2 day1 at 60 C/85% RH, equivalent to an
impermeable glass lid. For comparison, the Wertheimer group (52) reported a threshold thickness of w8 nm for plasma-enhanced
chemically vapor deposited (PECVD) SiN, although the WVTRw0.3 g m2 day1 for thicker films, measured at 33 C/100% RH, is
much higher than ALD barrier films.
4.19.5.3.3.2 Effect of ALD growth temperature and film thickness

While the barrier properties reported in the previous sections were for ALD films grown 120 or 125 C, a lower temperature is
attractive for low-cost plastic substrates and also for thermally sensitive electronic devices, such as OLEDs. Thinner barrier coatings
would also be cheaper to manufacture.
To probe this, we prepared a matrix of single-layer ALD barrier films grown on PET at three temperatures (50, 75, and 100 C)
and three thicknesses, corresponding to 75, 125, and 250 ALD cycles (20,21). The corresponding deposition rates were 0.85 Å per
cycle at 50 C, 0.86 Å per cycle at 75 C, and 0.926 Å per cycle at 100 C.
Figures 27 and 28 summarize visual and quantitative information after aging at 38 C/85% RH for 1200 h, followed by an
additional 500 h at 60 C/85% RH. In these transmitted-light images, areas where metallic Ca fully oxidized to optically transparent
Ca(OH)2 appear as white spots or defects. The time (in days) when these oxidation defects were first observed is indicated in an
individual cell. For ALD growth conditions, for which the Ca pixels retained a featureless, defect-free metallic appearance (no
apparent defects), and damp heat aging only produced a very thin, uniform surface oxidation layer, we report the calculated WVTR
in that cell.
Ca-testing reveals a unique time-dependent behavior for these ALD-coated PET barrier lids. Until defects abruptly appeared in
the Ca, WVTR calculated from the change in optical transmission with aging in damp heat was indistinguishable from that of an
impermeable glass lid. The time when defects first appear is thus the real measure of robustness of the ALD Al2O3 barrier film. Using
this criterion, we conclude for this temperature regime that thicker ALD films grown at a higher temperature are superior to thinner
films grown at lower temperature. This trend is analogous to our observation of higher electric field breakdown (53) in thicker ALD
Al2O3 films, grown at higher temperature.
The abrupt onset of defects in Ca-tests, which mirrors permeation breakthrough in the ALD Al2O3 barrier, after weeks of aging in
damp heat, suggests an alternative model (20,21) for permeation in ALD not due to pinholes (17) in the Al2O3. As previously
described in Section 4.19.3.2, permeation can be described by a percolation process involving a chemical exchange reaction with
molecular H2O along OH-defect clusters that eventually bridge the thickness of the ALD coating. In fact, Al2O3 films grown at
www.iran-mavad.com
Figure 27 Images of defects in Ca pixels after aging 1200 h at 38 C/85% RH for Al2O3 barrier films grown on PET by ALD at 50, 75, and 100 C,
corresponding to 75, 125, and 250 ALD cycles at each growth temperature. Reprinted with permission from Carcia, P. F.; McLean, R. S.; Sauer, B. B.;
Reilly, M. H. J. Nano. Sci. Technol. 2011, 11, 7994. Copyright 2011, American Scientific Publishing.
Figure 28 Images of defective Ca pixels after additional aging at 60 C/85% RH for Al2O3 barrier films in Figure 27. For cells with no defects, the
calculated WVTR is indicated after 500 h at 60 C/85% RH. Images for 125-cycle films grown at 50 and 75 C are after aging 288 h (12 days) at 60 C/85%
RH. These pixels became fully transparent after 500 h. Reprinted with permission from Carcia, P. F.; McLean, R. S.; Sauer, B. B.; Reilly, M. H.
J. Nano. Sci. Technol. 2011, 11, 7994. Copyright 2011, American Scientific Publishing.
low temperature by ALD have been shown to have an increasing amount of hydroxide incorporation (54,55). As detailed in
Section 4.19.3.2, by analogy with models used to explain electrical breakdown in insulating thin-film oxides (56,57), the onset time
for permeation or permeation breakthrough time (sp), described by eqn [19], is an exponential function of the activation energy H0
for the exchange reaction with molecular water and the inverse characteristic film thickness, g. ALD growth of Al2O3 at lower
temperature is predicted to reduce H0, because of more hydroxide content in those films and consequently more facile diffusion of
water by this exchange mechanism. And permeation breakthrough would occur sooner in thinner barrier films. For an ultrabarrier
film, therefore, the recommendation is ALD growth at as high a temperature as is compatible with the substrate and as thick as is cost effective.
4.19.5.3.3.3 Effect of plasma pretreatment of plastic substrate ( 21)

While ALD process temperature and film thickness are important parameters for obtaining low permeation, the initial surface
condition and cleanliness of the polymer substrate are also critical. Plasma treatment of polymer surfaces prior to PVD or CVD
deposition is a common initial processing step used to improve film nucleation and adhesion (48). Early nucleation of a dense
barrier film could reduce the thickness needed for good performance, which would also be more economical. Therefore the effect of
mild Ar- and O2-etching of a PET substrate on WVTR for subsequently deposited Al2O3 barrier films has been investigated. Prior to
ALD deposition at 100 C, the PET substrate was etched either in 100% Ar or in 100% O2 at a negative 50 V bias for 10 min.
Figure 29 compares WVTR for Ar-etched and unetched PET as a function of the number of ALD cycles, or film thickness (21). In
fact, the moisture permeation was less for thinner Al2O3, when it was grown on Ar-etched PET. This is consistent with better
nucleation or earlier continuous film growth on the Ar-etched PET. As the number of cycles or film thickness increases, moisture
permeation converges for etched and unetched substrates to the measurement limit of the commercial instrument, the MOCON
Aquatran 1. In contrast, O-etching the PET had no effect on the film thickness dependence of WVTR, as shown in Figure 30.
www.iran-mavad.com
Figure 29 Dependence of WVTR on the number of ALD cycles of Al2O3 on PET for no etch and for PET Ar-etched prior to deposition. WVTR was
measured with MOCON Aquatran 1 at 38 C/85% RH (w24 h). Reprinted with permission from Carcia, P. F.; McLean, R. S.; Sauer, B. B.; Reilly, M. H.
J. Nano. Sci. Technol. 2011, 11, 7994. Copyright 2011, American Scientific Publishing.
Figure 30 Dependence of WVTR on the number of ALD cycles of Al2O3 on PET for no etch and for PET O2-etched prior to deposition.
From AFM images of unetched and etched PET surfaces (Figures 31 and 32), Ar-etching does not substantially modify the
original PET surface morphology (21). The only subtle difference, deduced from the phase image, appears to be preferential removal
of the softer amorphous material separating crystalline PET fibrils at the surface. Etching with O2, however, does significantly alter
the surface morphology from crystalline fibril segments to more nodular features (Figure 33) (21). This is consistent with the
chemical alteration of the PET surface with reactive etching with oxygen and only physical changes to the PET surface with
nonreactive Ar-etching.
We also measured the change in surface wetting of PET after plasma treatment in Ar or O2. From measurement of advancing and
receding contact angles for water, we calculated the average Young’s contact angle (58) to be 62.2 for PET, 23.6 after Ar-etching at
50 V for 10 min, and 12.5 after etching in O2 also for 10 min at 50 V. Plasma treatment converted the surface from somewhat
hydrophobic to more hydrophilic. This more hydrophilic surface could promote nucleation and ALD growth by presenting a more
www.iran-mavad.com
Figure 31 Topographical (a) and phase (b) AFM images of PET surface. Reprinted with permission from Carcia, P. F.; McLean, R. S.; Sauer, B. B.; Reilly,
M. H. J. Nano. Sci. Technol. 2011, 11, 7994. Copyright 2011, American Scientific Publishing.
Figure 32 Topographical (a) and phase (b) AFM images of PET surface after plasma etching in argon at 50 V for 10 min. Reprinted with permission from
Carcia, P. F.; McLean, R. S.; Sauer, B. B.; Reilly, M. H. J. Nano. Sci. Technol. 2011, 11, 7994. Copyright 2011, American Scientific Publishing.
www.iran-mavad.com
Figure 33 Topographical (a) and phase (b) AFM images of PET surface after plasma etching in oxygen at 50 V for 10 min. Reprinted with permission
from Carcia, P. F.; McLean, R. S.; Sauer, B. B.; Reilly, M. H. J. Nano. Sci. Technol. 2011, 11, 7994. Copyright 2011, American Scientific Publishing.
reactive surface for TMA adsorption. But somehow the change in surface chemistry with O2-etching must counteract any
improvement in nucleation due to better wetting. Of course, there is still wide latitude in gas composition, for example, reducing the
O2 concentration in the etch gas, to optimize its effect on film nucleation and the resulting WVTR.
4.19.5.3.3.4 Effect of substrate surface chemistry ( 21)

A sometimes overlooked fact of commercial polymer films is that the two surfaces are not chemically identical. For example, one
side of PET is commonly coated with a ‘slip’ layer to reduce the coefficient of friction to facilitate winding and unwinding the film
(48). AFM images (topographical and phase) for this surface are shown in Figure 34. Compared to the uncoated side in Figure 31,
this side with a ‘slip’ coating is featureless, and the coating masks the polymer microstructure underneath, while locally the surface is
rougher (4.2 nm versus 1.25 nm). Since ALD growth is sensitive to surface chemistry, nucleation may also proceed differently on
these surfaces, even though the contact angle for wetting with water is about the same (w62 ). Figure 35 compares WVTR for thin
ALD Al2O3 grown at 100 C on the slip-coated and bare surfaces of the PET. The higher WVTR for the coated surface suggests more
sluggish nucleation. Above 80 cycles, the moisture permeation rates for ALD Al2O3 on the two surfaces again converges to the limit
(<5 104 g m2 day1) of the measurement with the MOCON Aquatran 1.
However, more sensitive, longer time Ca-testing has shown that barrier films grown on the slip-coated side eventually fail,
whereas barriers grown on the bare polymer continue to track the glass control sample after several thousand hours testing in damp
heat. Figure 36 shows optical Ca-test data for barrier films grown on the slip-coated side of PET for 75, 125, and 250 cycles of Al2O3
grown at 100 C. Thinner films fail sooner. In Figure 37 the time-to-failure (sb) is plotted, and the fit to eqn [19] yields an activation
energy H0 ¼ 0.1 eV and a characteristic thickness g ¼ 125 cycles (11.6 nm).
If the slip-coated side of PET is first etched in an Ar-plasma (50 V, 10 min), before depositing ALD Al2O3 of 75, 125, and 250
cycles at 100 C, those barrier films tested at 60 C/85% RH show no tendency to fail for up to 2000 h. Since an ALD process can
simultaneously coat both surfaces of the PET, both sides can have exceptional ultrabarrier properties, if the surfaces are plasma
treated.
Ex situ Fourier transform infrared spectroscopy does differentiate earlier stage ALD growth of Al2O3 on Ar-etched and unetched
bare PET by the amount of incorporated –OH in the Al2O3. In Figure 38 the peak of absorbance at 3450 cm1, corresponding to
bulk –OH incorporated in Al2O3 (59), is plotted versus the number of ALD Al2O3 cycles. For both etched and unetched surfaces, the
peak absorbance increases rapidly for the first few cycles, followed by a slower steady increase. The peak absorbance, or the amount
of incorporated –OH is less in Al2O3 grown on etched PET. This is consistent with a more crystalline PET surface, which has less
capacity for H2O absorption than the amorphous segment, which is preferentially etched by Ar. The incorporation of –OH can
facilitate permeation by an exchange mechanism with H2O, and also accelerate corrosion of Al2O3, as discussed in the next section.
www.iran-mavad.com
Figure 34 Topographical (a) and phase (b) AFM images of treated or ‘slip’-coated PET surface. Reprinted with permission from Carcia, P. F.; McLean, R.
S.; Sauer, B. B.; Reilly, M. H. J. Nano. Sci. Technol. 2011, 11, 7994. Copyright 2011, American Scientific Publishing.
Figure 35 Dependence of WVTR on the number of ALD cycles of Al2O3 on PET side with a ‘slip’ coating and for bare PET side without coating. WVTR
was measured with MOCON Aquatran 1 at 38 C/85% RH (w24 h). Reprinted with permission from Carcia, P. F.; McLean, R. S.; Sauer, B. B.; Reilly,
M. H. J. Nano. Sci. Technol. 2011, 11, 7994. Copyright 2011, American Scientific Publishing.
4.19.5.4.4 Corrosion of ALD Al2O3 Barrier Thin Film

There are several reports in the literature that ALD Al2O3 is not stable when contacted by condensed water (60,61). To better
understand the chemical and physical nature of corrosion of thin-film ALD Al2O3 by condensed H2O, it is important to first point
out that crystalline Al2O3 (corundum) and high content (>99%) alumina ceramics are hydrothermally stable, even at T > 300 C
(62,63). Water does not corrode high-purity, crystalline Al2O3. However, Al2O3 thin films, grown at T ¼ 100 C by ALD, are
www.iran-mavad.com
Figure 36 Optical Ca-test data showing time-to-failure for ALD Al2O3 films grown at 100 C on ‘slip-side’ of PET for Al2O3 thickness corresponding to 75,
125, and 250 cycles.
Figure 37 Fit of failure times, obtained from optical Ca-test in Figure 36, to percolation model (eqn [19]) for permeation.
amorphous with significant H-incorporation (54,55). For growth at 100 C, the H content is w15 at.%, increasing to >20% as the
growth temperature is reduced to 50 C. The chemical composition of those films is therefore more accurately described as an oxy-
hydroxide, AlOx(OH)32x. In fact, this chemical composition is close to hydroxide corrosion products – initially AlO(OH) and then
Al(OH)3 – that form when Al or AlN corrode in water (64,65). Coincident with those corrosion products, we identified by X-ray
diffraction (Figure 39) AlO(OH) and Al(OH)3 crystalline phases, after 50-nm-thick Al2O3 grown on Si at 100 C was aged at
85 C/85% RH for 2 weeks. Therefore, we propose that, when the ALD film composition is an oxy-hydroxide, this lowers the barrier
energy for reaction with water and facilitates conversion of ALD amorphous Al2O3 thin films to hydroxide phases, especially in films
grown at low temperature with significant H content. Furthermore, because these hydroxides are crystalline (albeit poorly crys-
talline), they are likely to be highly permeable due to more facile gas permeation along grain boundaries or along hydroxide chains
(20,21).
www.iran-mavad.com
Figure 38 Dependence of bulk –OH incorporated in thin-film ALD Al2O3, as determined by ex situ AFTIR, on the number of ALD cycles for films grown at
100 C on Ar-etched and bare PET.
Figure 39 X-ray diffraction pattern for Al2O3 50-nm-thick film, grown by ALD at 100 C and subsequently aged for 2 weeks at 85 C/85% RH. Peaks
corresponding to aluminum hydroxide corrosion products are indicated. The unmarked sharp peaks are from the Si substrate. Reprinted with permission
from Carcia, P. F.; McLean, R. S.; Li, Z. G.; Reilly, M. H.; Marshal, W. J. J. Vac. Sci. Technol. 2012, A30, 041515–041521. Copyright 2012, American
Vacuum Society.
Also, if the growth process for Al2O3 has a CVD component (59), it can adversely affect film permeation and stability. CVD can
occur in an ALD process sequence, when the purge times are too short or one of the reactants is difficult to remove, for example, H2O
at low temperature, so that films do not grow by a layer-by-layer process but by point nucleation (66) when both reactants are
present simultaneously. Point nucleation can result in a more faceted microstructure (66), which is consequently more permeable
and chemically reactive. Further, Al2O3 films grown by low-temperature CVD have up to 30% incorporated hydrogen (59), which is
responsible for a greater hydroxide content in films. Both a low ALD process temperature and inadequate purge times can
compromise the stability, or corrosion-resistance of an ALD Al2O3 film.
www.iran-mavad.com
Table 3 WVTR for single-layer ALD Al2O3 and bilayer of ALD Al2O3 on 100-nm-thick PECVD SiNx grown on PEN substrate
Substrate
2 1
d (thickness, nm) ALD Al2O3 PEN WVTR (g m day ) 100 nm SiN/PEN WVTR (g m2 day1)
0 0.72 6.5 103

2.5 0.14 3.0 104
5.0 w3.5 104 <5 105
10.0 <5 105 <5 105
Reprinted with permission from Carcia, P. F.; McLean, R. S.; Groner, M. D.; Dameron, A. A.; George, S. M. J. Appl. Phys. 2009, 106, 023533.
Copyright 2009, American Physical Society.
4.19.5.4.5 ALD/CVD Bilayer Ultrabarrier Thin Film

One of the common criticisms of an ALD process is that it is too slow and therefore expensive. In that regard, high-speed ALD
processes are currently in an advanced stage of development (67). However, it is also possible to combine ALD with higher speed
PVD and CVD processes to produce an ultrabarrier film. This can reduce the need for thicker ALD layers and possibly reduce cost.
The function of a thin ALD overcoat would then be to ‘seal’ pinhole defects in the underlying CVD or PVD barrier film and convert
a low-barrier film to an ultrabarrier.
As an illustration, we find that overcoating a SiNx PECVD layer with thin ALD Al2O3 substantially reduces the overall permeation
of water (24). The SiNx PECVD layers were grown without intentionally heating the substrate, using SiH4 and N2 as reactants in
intensified plasma using electron cyclotron resonance (ECR). Table 3 shows the effect on WVTR of thin ALD Al2O3 layer deposited
at 125 C on 100-nm-thick SiN. Depositing 2.5 nm of Al2O3 on 100 nm SiN reduced WVTR to 3 104 g m2 day1, or a reduction
by 22 from 6.5 103 g m2 day1 for a 100-nm SiN barrier film. A 5-nm-thick film of Al2O3 on 100-nm SiNx reduced WVTR
further to the limit of this Ca-test.
4.19.5.4.6 Applications for Single ALD Al2O3 Ultrabarrier

4.19.5.3.6.1 Encapsulation of CIGS solar cells – direct ALD deposition
As an illustration of the excellent barrier properties of an ALD thin film, we compared damp heat aging at 85 C/85% RH of CIGS
solar cells, which were encapsulated by (1) direct ALD deposition, (2) a plastic PET (125 mm thick) lid, and (3) a glass lid (0.5 mm
thick) (51). The CIGS cells had an active area w1 cm2 and were fabricated at the IEC, the University of Delaware. The ALD film was
55 nm of Al2O3 grown at 120 C from TMA and H2O. Following ALD deposition of the Al2O3 barrier layer, a film of DuPont
Teflon FEP 200 C (fluorinated ethylene propylene, 0.002 inches thick – WVTR ¼ 2.3 g-H2O m2 day1 at 40 C/100% RH) was
attached to the CIGS cell with a UV-curable epoxy. The Teflon FEP film was intended to protect the ALD Al2O3 barrier from
potential corrosion by direct condensation of water vapor. Figure 40 is a schematic of the solar cell encapsulated with the 55-nm-
thick ALD A2O3. Both the PET (WVTR ¼ 4.7 g-H2O m2 day1) and glass lid (negligible WVTR) were also bonded to CIGS solar cells
with the same UV-curable epoxy.
The three CIGS test structures were aged in an environmental chamber at 85 C/85% RH for 1000 h. This test is referred to as the
‘damp heat’ test, and it is one of the critical testing protocols used to qualify solar cell modules. In fact, 70% of all thin-film modules
qualification failures occur in the ‘damp heat’ test, making it the most important indicator of a successful packaging strategy for thin-
film CIGS solar cells. While cells are commonly stored in the dark during ‘damp heat’ aging, and then measured ex situ at 25 C with
standard illumination (AM1.5), these cells were simultaneously illuminated from a lamp adjusted to deliver approximately one sun
(1000 W m2) light intensity during aging in ‘damp heat.’ Combining this ‘damp heat’ test with simultaneous solar irradiation is
even more demanding, unique, and realistic than damp heat in the dark.
Figure 41 compares the efficiencies measured ex situ at standard conditions versus time in the environmental chamber for
the CIGS cell encapsulated with a thin ALD Al2O3 barrier film, the cell protected with the PET lid and the one with a glass lid.
Figure 40 Cross-section representation of a CIGS solar cell coated directly with a 55-nm-thick ALD Al2O3 barrier layer. Reprinted with permission from
Carcia, P. F.; McLean R. S.; Hegedus, S. Sol. Energy Mater. Sol. Cells 2010, 94, 2375. Copyright 2010, Elsevier.
www.iran-mavad.com
Figure 41 Dependence of CIGS solar cell efficiency on aging time at 85 C/85% RH, with simultaneous illumination of 1 sun, for cells protected with an
ALD Al2O3 barrier layer, and a glass or a PET lid. Reprinted with permission from Carcia, P. F.; McLean R. S.; Hegedus, S. Sol. Energy Mater. Sol. Cells 2010,
94, 2375. Copyright 2010, Elsevier.
The CIGS cell protected with a PET plastic lid lost about half its efficiency (12.5 / 6.6%) after aging for 1020 h (42.5 days) at
85 C/85% RH with illumination, whereas the CIGS cell protected with the 55 nm ALD Al2O3 barrier and the cell with a glass
lid showed only a small net change (<3%) in efficiency after aging, that is, relative to their efficiency before encapsulation.
Figure 42 summarizes the effect of aging on each of the cell parameters for the CIGS cell protected with a glass lid and the one
coated with ALD Al2O3. For the cell protected with a PET lid, its parameters, not shown, all degraded continuously with stress
time. This loss in cell performance with a PET lid is very similar to results reported by Schmidt et al. for nonencapsulated CIGS
cells (68).
For the CIGS cell protected with ALD Al2O3, there was no degradation in electrical properties; that is, both the open-circuit
voltage (Voc) and the fill factor (FF) were invariant with aging for ‘damp heat’ and illumination. The small decrease observed in
efficiency was associated with a decrease in photocurrent, that is, the short-circuit current density, Jsc. Subsequent examination of
this cell identified ‘yellowing’ of the epoxy layer as the source of reduced photocurrent. This can easily be remedied by substituting
a traditional encapsulant layer, such as EVA (ethylene vinyl acetate) in place of the epoxy.
While the efficiency of the CIGS cell protected with a glass lid was nearly unchanged with aging in ‘damp heat’, Voc did decrease
continually – 50 mV in 1032 h, indicating some electrical degradation occurred in this solar cell. However, there was a compen-
sating increase in FF, due to a small decrease in the cell parasitic series resistance, so that the overall efficiency remained approx-
imately unchanged.
Figure 42 Comparison of the change in solar cell parameters: Open-circuit voltage (Voc), fill factor (FF), and short-circuit current (Jsc) for CIGS cells
protected with ALD Al2O3 (circles) and a glass lid (squares) with aging at 85 C/85% RH and 1 sun illumination. Reprinted with permission from
Carcia, P. F.; McLean R. S.; Hegedus, S. Sol. Energy Mater. Sol. Cells 2010, 94, 2375. Copyright 2010, Elsevier.
www.iran-mavad.com
In summary, a CIGS cell protected by direct deposition of Al2O3, thermally grown by ALD, performed equally well (<3% change
in efficiency) in damp heat testing as a cell protected with a glass lid. As stated in Section 4.19.5.2.1.2, PML barrier films were equally
effective in protecting similar laboratory CIGS solar cells in damp heat.
4.19.5.3.6.2 Encapsulation of CIGS solar cells – flexible front sheet

A practical alternative to direct ALD deposition, which will require new deposition equipment and additional processing for the
CIGS solar cell manufacturer, is to incorporate the ALD barrier in a flexible, transparent front sheet, which can be laminated to the
solar cell with a traditional encapsulant. Our front sheet consisted of two PEN polymer films (Q65-DuPont-Teijin Films, 125 mm
thick), ALD coated on both sides with 25 nm of Al2O3 and then laminated to one another with a contact adhesive. Subsequently,
this front sheet was laminated at w150 C to six individual, commercial CIGS solar cells, 76 110 mm, which had been fabricated
on a flexible stainless steel substrate, using a thermoplastic encapsulant (0.45 mm thick). As a reference, the same number of cells
were encapsulated with a glass lid and an uncoated (no barrier) PEN laminate front sheet. Figure 43 is a plot of the maximum power
for each group of six cells, after aging at 85 C/85% RH for w1700 h. The results show comparably slow degradation for cells with
ALD barrier on PEN and glass front sheets. In fact, the degradation is just slightly less for the ALD/PEN barriers than the reference
glass front sheets. We attribute this to better bonding of the ALD/PEN barriers to CIGS cells than with glass. Since no edge-seal was
used in these structures, moisture infiltration along the glass–encapsulant interface may have been more facile than along an
encapsulant interface with the ALD/PEN front sheet. However, it is quite evident that CIGS cells protected only with uncoated PEN
deteriorated rapidly compared to ALD/PEN and glass front sheets. The cells protected with ALD/PEN retained w85% of their initial
power output after w1700 h exposed to 85 C/85% RH, whereas cells with an uncoated PEN front sheet lost w75% power in less
than 600 h.
4.19.5.3.6.3 Encapsulation of OLEDs – direct ALD deposition

Ghosh et al. demonstrated that top-emitting OLED devices, encapsulated with an Al2O3 layer, which was deposited by ALD,
exhibited negligible change in emission properties after aging at 85 C/85% RH for 1300 h (60). Top emission is attractive because
the aperture for light output is higher. However, the moisture-sensitive cathode must be thin to be transparent, making it more
susceptible to degradation, and desiccants cannot be used because they are opaque. Figure 44 shows individual layers of the OLED
device and the encapsulation. A total of 180 nm of Al2O3 was deposited by ALD at 130 C from TMA and ozone. A 2-mm-thick layer
of parylene was evaporated on Al2O3 to prevent corrosion by condensed water, and this was capped with a glass cover sheet. By
comparison, sputtered barriers and barriers grown by low-temperature PECVD did not perform as well as the ALD Al2O3
Figure 43 Comparison of relative (%) maximum power (Pmax) for flexible, commercial CIGS solar cells protected with a flexible ALD barrier grown on
PEN, bare PEN (DuPont Control), or glass, versus time aged at 85 C/85% RH.
www.iran-mavad.com
Figure 44 Cross-section representation of OLED device encapsulated with thick ALD Al2O3 barrier layer, parylene coating, and glass lid. Reprinted
with permission from Ghosh, A. P.; Gerenser, L. J.; Jarman, C. M.; Fornalik, J. E. Appl. Phys. Lett. 2005, 86, 223503. Copyright 2005, American Institute of
Physics.
encapsulation. The notable feature of this encapsulation scheme, which still uses a glass lid, is that the transparent ALD barrier
coating eliminated the need for a desiccant in the top emitting OLED.
4.19.5.3.6.4 Encapsulation of OLEDs – flexible barrier lid

In our laboratory, collaborating with DuPont Displays, we tested the effectiveness of a flexible barrier lid for protecting blue-
emitting OLED devices. A test device consisted of four pixels, each w2.25 2.35 mm, fabricated on a glass substrate, 15.5 26 mm.
A photographic image of two pairs of blue-emitting OLED test arrays is shown in Figure 45. The devices were bottom emission, that
is, light exits the transparent conducting anode electrode. The opaque top cathode electrode consisted of a thin evaporated layer of
Figure 45 Photographic image of blue-emitting OLED test array.
www.iran-mavad.com
Figure 46 Change in light output, expressed as normalized photovoltage, for OLEDs with indicated encapsulation versus time.
an alkali metal salt, overcoated with thick Al. The completed devices had either a glass lid with an edge-bead, UV-curable epoxy, or
a PET/barrier plastic lid with either an edge-bead or full coverage epoxy. The barrier structure on PET was the same one used for CIGS
solar cells, consisting of two laminated sheets of PET, each coated by ALD with 25-nm-thick Al2O3.
The comparative lifetime of these OLED devices was evaluated in ambient by tracking the change in light output with time for
operation at high voltage and high current, which accelerates device degradation. Specifically, V ¼ 7.3 V and I ¼ 8.7 mA
(165 mA cm2), corresponding to light output of w8200 cd cm2. It is well known that OLED devices will degrade rapidly, if they
are not protected from atmospheric gases. In this study, one set of OLED devices had a glass lid, while three sets had a PET/barrier
lid. One pair of plastic/barrier lids was first heated (bake) at 110 C for 4 h to drive off any residual water before attaching the lid to
the OLED device with an edge-bead epoxy. Identical pairs of unheated plastic/barrier lids were attached to the devices either with an
edge-bead epoxy or full surface coverage of epoxy.
Figure 46 summarizes the change in light output versus time for the four sets of devices, each consisting of four pixels. The light
output for OLED devices with a glass lid decreased to 50% of its initial value in w500 h. The light output for plastic barrier lids
with full epoxy coverage was reduced to 50% at 440 and 470 h, corresponding to between 88 and 94% of the glass lid lifetime. For
plastic/barrier lids that had been baked and were sealed with an edge-bead epoxy, the lifetime was w80% of the reference glass
lids. However, plastic/barrier lids that were not heated before sealing with an edge-bead epoxy provided poor protection for these
OLED devices, losing 50% of their initial light output in less than 150 h. The comparable performance for plastic/barrier lids,
which had full epoxy coverage, to the glass reference devices is encouraging. The data do suggest that better sealing of the plastic lid
to the OLED devices could help improve their lifetime and enable fully flexible OLED displays, and lighting with unique,
nonplanar design.
4.19.5.4.7 Manufacturing with ALD

In the preceding sections, we presented properties for barrier films grown by ALD, which seems uniquely suited to producing
pinhole-free films with ultrabarrier performance. However, a common criticism of ALD is that it is too slow to be economical in
high-volume manufacturing, where throughput is often measured by the line speed (m h1) to fabricate product. In this context, the
deliberate temporal dose and purge steps of traditional ALD would appear to support this criticism. But a favorable attribute of ALD
gas-phase deposition is its ability to coat a large area, for example, multiple wafers, in a batch mode. In fact, the Planar Corporation
(Finland) has manufactured a monochrome electroluminescent display by batch ALD processes (66). With the exception of the
metal electrodes, all layers in their display are grown by ALD.
Adaptation of the batch mode ALD process to depositing a barrier film on flexible plastic was recently demonstrated by the group
at DuPont (69). PET and PEN films with a width 0.3–0.5 m were wound on a spiral cassette, so that adjacent windings were spaced
a few millimeters apart. This is illustrated in Figure 47. Depending on the cassette inner and outer diameter, and the spacing
between windings, or the pitch, more than 10 m length of film could be coated on both sides in an ALD reactor. At 100 C, a cycle
time of less than 4 s was adequate to deposit ALD Al2O3 with good uniformity on PET.
www.iran-mavad.com
Figure 47 (a) Metal-spiral carrier for polymer film. (b) Plastic film loaded on plastic spiral into an ALD reactor enclosure.
As the number of applications for ALD has expanded from the semiconductor industry (gate dielectric, electronic and magnetic
memory devices) to photovoltaics and sensitive organic electronics, a number of groups worldwide have been developing
approaches for high-throughput, continuous ALD (67). The most common of these – spatial ALD – is illustrated schematically in
Figure 48. Whereas in conventional ALD the precursor dose and purge steps are sequential in time, in spatial ALD the precursors are
physically separated and delivered continuously. In this case, the delivery system or ‘head’ can be stationary, and the substrate moves
through the separated half-reaction zones, or the ‘head’ can be moved over a stationary substrate. Two different schemes for spatial
ALD are illustrated in Figures 49 and 50. In the Lotus Technology approach, a flexible substrate moves continuously in a serpentine
path through isolated precursor and purge zones. In a tool designed by Beneq, a flexible substrate is affixed to a drum that rotates
Figure 48 Schematic representation for (a) conventional time-sequence ALD process and (b) spatial ALD process. The solid black line traces the path of
the substrate through the spatially separated reaction zones. Reprinted with permission from Poodt, P.; Cameron, D. C.; Dickey, E.; George, S. M.;
Kuznetsov, V.; Parson, G. N.; Roozeboom, F.; Sundaram, G.; Vermeer, A. J. Vac. Sci. Technol. 2012, A30, 01802–01811. Copyright 2012, American
Vacuum Society.
www.iran-mavad.com
Figure 49 Spatial ALD process under development at Lotus applied technology (Hillsboro, OR). The flexible substrate follows a serpentine path
repeatedly between ALD reaction zones. Reprinted with permission from Poodt, P.; Cameron, D. C.; Dickey, E.; George, S. M.; Kuznetsov, V.; Parson, G. N.;
Roozeboom, F.; Sundaram, G.; Vermeer, A. J. Vac. Sci. Technol. 2012, A 30, 01802–01811. Copyright 2012, American Vacuum Society.
Figure 50 Spatial ALD process under development by Beneq Oy in Finland. The flexible substrate is attached to a drum that rotates the substrate through
separate reaction zones along the perimeter of the reactor. Reprinted with permission from Poodt, P.; Cameron, D. C.; Dickey, E.; George, S. M.;
Kuznetsov, V.; Parson, G. N.; Roozeboom, F.; Sundaram, G.; Vermeer, A. J. Vac. Sci. Technol. 2012, A 30, 01802–1811. Copyright 2012, American Vacuum
Society.
between physically separated precursor dose and purge zones. For both schemes, film thickness will be determined by the number
of passes through these physically separated reaction zones. With all of this development activity, there is reason to be optimistic
about the availability of high-rate, commercial ALD tools to coat plastic film continuously in the near future.
In this chapter we have shown that electronic applications, especially those based on moisture-sensitive organic materials, will
require diffusion barriers with permeation rates up to 106 times lower than barriers used for food packaging. To address this need,
multilayer barriers have emerged as a leading candidate to address the lifetime requirement for sensitive electronic devices.
However, for some acutely cost-driven applications, such as solar cells, a multilayer barrier may be too expensive. A less expensive
alternative would be a single-layer barrier. But the perfection required in a single barrier layer with ultralow permeation is
formidable. One deposition method that does consistently (on a lab scale) produce a barrier film with ultrabarrier performance is
ALD. And we have extensively reviewed properties of ALD barrier films in this chapter. The challenge for single-layer barriers
fabricated by ALD will be to transfer the technical lab-scale success to manufacturing, while maintaining low cost. Nonetheless,
a practical barrier technology based on ALD will still likely require some redundancy of layers, such as two single-layer barriers
laminated to one another, if only to provide mechanical protection for the ALD film.
www.iran-mavad.com
Acknowledgments
Many of the insights, especially from single-layer barrier films grown by ALD, greatly benefited from the experimental skill and keen
observations of my colleague, R. Scott Mclean. Early in our investigation of ALD, we had the good fortune to collaborate with one of
the world’s leading experts in ALD, Professor Steven George at the University of Colorado. I am also grateful to Dr Yoshi Sensaki,
who had the courage to put polymer sheets in his semiconductor ALD reactor, allowing us to verify our intuition that ALD thin films
have ultralow gas permeation.
References
1. Lange, J.; Wyser, Y. Packag. Technol. Sci. 2003, 16, 149.

2. Plog, S.; Schneider, J.; Walker, M.; Schulz, A.; Stroth, U. Surf. Coat. Technol. 2011, 205, 8165.
3. Leterrier, Y.; Singh, B.; Bouchet, J.; Manson, J. -A. E.; Rochat, G.; Fayet, P. Surf. Coat. Technol. 2009, 203, 3398.
4. Rochat, G.; Leterrier, Y.; Garamzegi, L.; Manson, J.-A. E.; Fayet, P. Surf. Coat. Technol. 2003, 174–175, 1029.
5. Katz, H. E. J. Mater. Chem. 1997, 7, 369.
6. Holger, S.; Krebs, F. C. Sol. Energy Mater. Sol. Cells 2004, 83, 125.
7. Tang, C. W.; VanSlyke, S. A. Appl. Phys. Lett. 1987, 51, 913.
8. Bergh, A.; Craford, G.; Duggal, A.; Haitz, R. Phys. Today 2001, 54, 42.
9. Burrows, P. E.; Graff, G. L.; Gross, M. E.; Martin, P. M.; Hall, M.; Mast, E.; Bohnam, C.; Bennett, W.; Michalski, L.; Weaver, M.; Brown, J. J.; Fogarty, D.; Sapochak, L. S.
In Organic Light-emitting Devices IV; Zakya H. Kafafi., Ed. In Proceedings of SPIE, 2001; Vol. 4105, p 75.
10. Kim, N.; Potscavage, W. J., Jr.; Sundaramoothi, A.; Hendersen, C.; Kippelen, B.; Graham, S. Sol. Energy Mater. Sol. Cells 2012, 101, 140.
11. Carcia, P. F.; McLean, R. S.; Reilly, M. H.; Groner, M. D.; George, S. M. Appl. Phys. Lett. 2006, 89, 031915.
12. Coyle, D. J. Prog. Photovolt. Res. Appl. 2011, http://dx.doi.org/10.1002/pip.1172.
13. Crank, J., Park, G. S., Eds. Diffusion in Polymers; Academic Press: New York, 1968.
14. Crank, J. Mathematics of Diffusion; Oxford: London, 1967.
15. Tencer, M. In Proceedings of the 1994 IEEE 44th Electronic and Technology Conference, Washington, D.C., USA, pp 196–209.
16. Kempe, M. D. Sol. Energy Mater. Sol. Cells 2006, 90, 2720.
17. Rossi, G.; Nulman, M. J. Appl. Phys. 1993, 74, 5471.
18. Jamieson, E. H. H.; Windle, A. H. J. Mater. Sci. 1983, 18, 64.
19. Mercea, P.; Muresan, L.; Mecea, V.; Silipas, D. J. Membr. Sci. 1988, 35, 291.
20. Carcia, P. F.; McLean, R. S.; Reilly, M. H. Appl. Phys. Lett. 2010, 97, 221901.
21. Carcia, P. F.; McLean, R. S.; Sauer, B. B.; Reilly, M. H. J. Nano. Sci. Technol. 2011, 11, 7994.
22. http://www.mocon.com/pdfwebinars/2010/High%20Barrier%20WVTR%20Test%20Methods.pdf (accessed 10 August 2012).
23. Nisato, G; Bouten, P. C. P.; Slikkerveer, P. J.; Bennett, W. D.; Graff, G. L.; Rutherford, N.; Heuser, K. In Proceedings of Asia Display/IDW, 2001, p 1435.
24. Carcia, P. F.; McLean, R. S.; Groner, M. D.; Dameron, A. A.; George, S. M. J. Appl. Phys. 2009, 106, 023533.
25. Carcia, P. F.; McLean, R. S.; Li, Z. G.; Reilly, M. H.; Marshal, W. J. J. Vac. Sci. Technol. 2012, A30, 041515-1.
26. Paetzold, R.; Winnacker, A.; Henseler, D.; Cesari, V.; Heuser, K. Rev. Sci. Instrum. 2003, 74, 5147.
27. Schubert, S.; Klumies, H.; Muller-Meskamp, L.; Leo, K. Rev. Sci. Instrum. 2011, 82, 094101.
28. Groner, M. D.; George, S. M.; McLean, R. S.; Carcia, P. F. Appl. Phys. Lett. 2006, 88, 051907.
29. Dunkel, R.; Bujas, R.; Klein, A.; Horndt, V. Proc. IEEE 2005, 93, 1478.
30. Zhang, X. D.; Lewis, J. S.; Wolter, S. D.; Parker, C. B.; Glass, J. T. J. Vac. Sci. Technol. 2007, A25, 1587.
31. Zhang, X. D.; Lewis, J. S.; Parker, C. B.; Glass, J. T.; Wolter, S. D. J. Vac. Sci. Technol. 2008, A26, 1128.
32. Ranade, A.; D’Souza, N. A.; Wallace, R. M.; Gnade, B. E. Rev. Sci. Instrum. 2005, 76, 013902.
33. Norenberg, H.; Miyamoto, T.; Tsukahara, Y.; Smith, G. D. W.; Briggs, G. A. D. Rev. Sci. Instrum. 1999, 70, 2414.
34. Langereis, E.; Creatore, M.; Heil, S. B. S.; van de Sanden, M. C. M.; Kessels, W. M. M. Appl. Phys. Lett. 2006, 89, 081915.
35. Park, S.-H. K.; Oh, J.; Hwang, C.-S.; Lee, J.-I.; Yang, S.; Chu, H. Y. Electrochem. Solid-State Lett. 2005, 8, H21.
36. George, S. M. Chem. Rev. 2010, 110, 111.
37. Affinito, J. D.; Gross, M. E.; Coronado, C. A.; Graff, G. L.; Greenwell, I. N.; Martin, P. M. Thin Solid Films 1996, 290–291, 63.
38 Greener, J.; Ng, K. C.; Vaeth, K. M.; Smith, T. M. J. Appl. Polymer Sci. 2007, 106, 3534.
39. http://www.tera-barrier.com/technology.html (accessed 13 August 2012).
40. Graff, G. L.; Williford, R. E.; Burrows, P. E. J. Appl. Phys. 2004, 96, 1840.
41. Ash, R.; Barrer, R. M.; Petropoulos, J. H. Brit. J. Appl. Phys. 1963, 14, 854.
42. Vogt, B. D.; Lee, H.-J.; Prabhu, V. M.; DeLongchamp, D. M.; Lin, E. K.; Wu, W. J. Appl. Phys. 2005, 97, 114509-1.
43. Weaver, M. S.; Michalski, L. A.; Rajan, K.; Rothman, M. A.; Silvernail, J. A.; Burrows, P. E.; Graff, G. L.; Gross, M. E.; Martin, P. M.; Hall, M.; Mast, E.; Bonham, C.;
Bennett, W.; Zumhoff, M. Appl. Phys. Lett. 2002, 81, 2929.
44. Olsen, L. C. Subcontract Report NREL/SR-520–4752, Barrier Coatings for Thin Film Solar Cells, March 2010.
45. Olsen, L. C.; Kundu, S. N.; Bonham, C. C.; Gross, M. E. In Proceedings 31st IEEE Photovoltaic Specialists Conference, 2005; p 327.
46. Olsen, L. C.; Gross, M. E.; Kundu, S. N.; Graff, G. L. In Proceedings 33rd IEEE Photovoltaics Specialists Conference, 2008; p 1.
47. Olsen, L. C.; Gross, M. E.; Graff, G. L.; Kundu, S. N. Reliability of Photovoltaic Cells, Modules, Components, and System. In Proceedings SPIE; Dhere, N. G., Ed., 2008; Vol.
7048, p 1.
48. Bishop, C. A. Roll-to-Roll Vacuum Deposition of Barrier Coatings; Scrivener: Salem, MA, 2010.
49. Ritala, M.; Leskela, M.; Dekker, J.-P.; Mutsaers, C.; Soininen, P. J.; Skarp, J. Chem. Vap. Deposition 1999, 5, 7.
50. Ohring, M. The Materials Science of Thin Films; Academic Press: San Diego, 1992; p 197.
51. Carcia, P. F.; McLean, R. S.; Hegedus, S. Sol. Energy Mater. Sol. Cells 2010, 94, 2375.
52. daSilva Sobrinho, A. S.; Latreche, M.; Czeremuszkin, G.; Klemberg-Sapieha, J. E.; Werthheimer, M. R. J. Vac. Sci. Technol. 1998, A 16, 3190.
53. Carcia, P. F. In Transparent Electronics: From Synthesis to Application; Fracchetti, A., Marks, T. J., Eds.; J. Wiley: New York, 2010 (chapter 11).
54. Groner, M. D.; Fabreguette, F. H.; Elam, J. W.; George, S. M. Chem. Mater. 2004, 16, 639.
55. Verlaan, V.; van den Elzen, L. R. J. G.; Dingemans, G.; van de Sanden, M. C. M.; Kessels, W. M. M. Phys. Status Solidi 2010, C7, 976.
56. Sune, J.; Placencia, I.; Barniol, N.; Farres, E.; Martin, F.; Aymerich, X. Thin Solid Films 1990, 185, 347.
57. Stathis, J. H. J. Appl. Phys. 1999, 86, 5757.
www.iran-mavad.com
58. Tadmor, R. Langmuir 2004, 20, 7659.

59. Gleizes, A. N.; Vahlas, C.; Sovar, M.-M.; Samelor, D.; Lafont, M.-C. Chem. Vap. Deposition 2007, 13, 23.
60. Ghosh, A. P.; Gerenser, L. J.; Jarman, C. M.; Fornalik, J. E. Appl. Phys. Lett. 2005, 86, 223503.
61. Abdulagatov, I.; Yan, Y.; Cooper, J. R.; Zhang, Y.; Gibbs, Z. M.; Cavanagh, A. S.; Yang, R. G.; Lee, Y. C.; George, S. M. Appl. Mater. Interfaces 2011, 3, 4593.
62. Yue, X.-M.; Zhang, G.-J.; Watanabe, T.; Tai, W.-P. J. Am. Ceram. Soc. 1999, 82, 2560.
63. Oda, K.; Yoshio, T. J. Am. Ceram. Soc. 1997, 80, 3233.
64. Skrovan, J.; Alfantazi, A.; Troczynski, T. J. Appl. Electrochem. 2009, 39, 1695.
65. Li, J.; Nakamura, M.; Shrai, T.; Matsumaru, K.; Ishizaki, C.; Ishizaki, K. J. Am. Ceram. Soc. 2006, 89, 937.
66. Suntola, T. Thin Solid Films 1992, 216, 84.
67. Poodt, P.; Cameron, D. C.; Dickey, E.; George, S. M.; Kuznetsov, V.; Parson, G. N.; Roozeboom, F.; Sundaram, G.; Vermeer, A. J. Vac. Sci. Technol. 2012, A 30, 010802-
010801.
68. Schmidt, M.; Braunger, D.; Schaffler, R.; Schock, H. W.; Rau, U. Thin Solid Films 2000, 361–362, 283.
69. Leffew, K. W.; Boussaad, S.; Reardon, D.; Tate, N.; Glassmaker, N.; Dean, D.; Nunes, G.; Zhao, C. Q.; Samuels, S. L.; Schneider, J.; Carcia, P.; McLean, R. In Proceeding of
the Photovoltaic Module Reliability Workshop 2011, Golden, Colorado, February 16–17, 2011. http://www1.eere.energy.gov/solar/pdfs/pvmrw2011_33_tf_leffew.pdf
(accessed 14 August 2012).
www.iran-mavad.com
4.20 Diffusion Barrier Layers for Edible Food Packaging
S Desobry and E Arab-Tehrany Université de Lorraine, France

4.20.2 Packaging Films 499
4.20.2.1 Mass Transport through Edible Coating and Food Stability 499
4.20.2.1.1 Partition Coefficient Prediction 501
4.20.2.1.2 Molecular Diffusion 501
4.20.2.2 Water Vapor and Gas Properties of Polymers 501
4.20.2.3 Contaminants 503
4.20.3 Edible Coating 504
4.20.3.1 Bioactive Edible Coating 505
4.20.3.1.1 Antimicrobial Compounds 505
4.20.3.1.2 Antioxidants 507
4.20.3.1.3 Lipids 508
4.20.3.2 Physical Properties of Edible Coating 508
4.20.3.2.1 Water Vapor Permeability (WVP) 508
4.20.3.2.2 Gas Permeability 508
4.20.3.2.3 Mechanical Properties 508
4.20.3.2.4 Antimicrobial Properties 509
4.20.4 Added Value and Limitations of Edible Coatings: Sensory Properties 512
4.20.4.1 Visual Properties 513
4.20.4.2 Odors 513
4.20.4.3 Taste 513
4.20.4.4 Touch 513
4.20.5 Toxicity of Polymeric Films into Contact with Foodstuff 513
4.20.6 Conclusion 515
References 515
4.20.1 Introduction
Packaging materials can be of two types: coating or films. Coating may be defined as a thin layer of material that covers food
surfaces, while packaging films are formed previously and later applied to the product. Both coating and film act as barriers between
the environment and the food product to ensure safety and acceptable shelf-life.
In the present chapter, we will study successively both preservation systems starting with the film and then we will focus on
coatings.
4.20.2 Packaging Films
One of the main criteria for selection of a packaging material is its permeability to gas and water, which assesses the packaging
material’s ability for mass transport. This parameter is important mainly for foods susceptible to oxidation, hydration, or dehy-
dration. Aromas and CO2 permeability are increasingly used to evaluate the packaging’s ability to preserve the food product during
a long shelf-life. Development of modified atmosphere packaging also increases the industrial interest in mastering permeabilities.
Since 1980, many authors have published permeability values of polymers used in food packaging. The data are of great interest,
but unfortunately, the values have shown great variability because the method of measurement greatly influences the permeability
values. Static and dynamic permeability measurements are generally not comparable. Most authors have used permeability
measurement for comparison between various films or coatings with a given method.
4.20.2.1 Mass Transport through Edible Coating and Food Stability

Migration of low-molecular-weight compounds from an active layer (film or coating) to food has been the subject of considerable
attention and concern (1–6). The diffusion process describes molecular mobility through film and coating. The first experiments
were developed by J. A. Nollet, a French physicist (1700–1770) who studied diffusion through a pig’s bladder. A few years later, in
1827, R. J. H. Dutrochet (1776–1847) discovered the osmosis process. In 1833, the chemist T. Graham demonstrated that the

www.iran-mavad.com
500 Diffusion Barrier Layers for Edible Food Packaging
– log D
–5
He H2
–6
diffusants sphéríques
–7 Ne O2 N2 diffusants de torme ellongée
–8
H2O CO2
–9 CH4
Ar
–10 Kr MeOH
–11
MeCl EtOH EtCl
–12
Me2C=O
–13
PrOH
–14 benzene butane
–15 BuCH pentane
hexane
–16 CCl4
SF6
–17
0.20 0.30 0.40 d’ 0.50 0.60
Figure 1 Influence of van der Waals (d 0 ) and molecular shape of molecule diffusivity D through PVC. Reproduced from Barens, A. R.; Hopfenburg, H. B.
Diffusion of Organic Vapors at Low Concentration in Glassy PVC, Polystyrene and PMMA. J. Membr. Sci. 1982, 10, 283.
transport speed of a given gas is inversely proportional to the square root of its molecular mass. He developed a global diffusion
approach from the research presented by the German chemist J. W. Döbereiner, who observed that hydrogen mobility was higher
than air mobility. Thirty years ago, Barens and Hopfenburg (1982) (7) published a reference work on gas mobility through
polymers (Figure 1).
The most frequently used approach was developed by A. E. Fick (1829–1901), a German physiologist who studied the diffusion
process in animal tissues and discovered the linear flow of diffusion (8) through a film without any interactions between the
diffusing substance and the matrix. Von Wroblewski (1879) studied molecular sorption and developed the previous works to
provide a quantitative formulation for the steady-state gas flow in terms of kinetic and equilibrium parameters in a gas–rubber
system (9).
It is now accepted that diffusion exists in any system when the migrant molecular size is low and if the coating network structure
allows mobility. For diffusing substances in most packaging materials, a molecular size above 1000 Da is commonly accepted as the
maximum molecular weight to produce a significant diffusion process in films. With regard to coating materials, it is commonly
accepted that crystalline structure does not allow molecular mobility. Molecular transport occurs through the amorphous phase. So,
small compounds must be used to allow molecular transport, and an amorphous coating network is required. When those two
conditions are reached, movement of active compounds in the coating can be controlled according to the Fickian diffusion process.
Mass transport is described by the diffusivity (DP), as shown in Figure 2.
Transport processes can be divided into three classes, based on the limiting mechanism (10–14):
Class 1: Nonmigrating materials. The very low diffusivity into the coating ensures that no migration occurs.
Class 2: Independent migration not controlled by the food, although the presence of food may accelerate the migration. The
high-diffusion coefficient in the coating and the low affinity between coating and migrant allow the migrant to easily leave the
coating material.
Figure 2 Transient diffusion transport profile in film or coating to contact with food. With e ¼ film thickness; DF ¼ diffusivity in food; DP ¼ diffusivity in
polymeric film; Kp ¼ partition coefficient.
www.iran-mavad.com
Diffusion Barrier Layers for Edible Food Packaging 501
Class 3: Leaching; migration controlled by the food; negligible in the absence of food, significant in its presence. Food interacts
highly with the coating material (interdiffusion), and the low coating/food partition coefficient makes migrant transport very fast.
A migration process is fully described by the kinetics of migrant diffusion in each phase (expressed by the diffusion coefficient,
D) and the chemical equilibrium (expressed by the partition coefficient, K). The diffusion coefficient determines the dynamics of the
migration process. The larger the value of D, the shorter is the time to reach equilibrium. The packaging materials most used are the
polyolefins (mostly polyethylene and polypropylene). Their homogenous structure, partial crystallinity, and great industrial use
make them excellent study materials. Piringer (Germany) and Vitrac (France) developed, respectively, statistical and molecular
approaches to predict diffusivities through polyolefin films. Both approaches are of high interest and must be extended to other
materials. The lack of high-precision data limited the use of these approaches for other less studied materials. So, before going
further with the diffusivity, the partition coefficient (KP) has to be defined.
4.20.2.1.1 Partition Coefficient Prediction

To measure the partition coefficient, the coating/food system is left for several weeks in contact until the diffusing component
concentration in both phases of the system is stabilized. The partition coefficient is then calculated as the ratio of migrant equi-
librium concentration in food (CF,N) to the coating (CP,N).
CF;N
K¼ [1]
CP;N
or
MF;N =VF
K¼ [2]
MP;N =VP
where MF,N and MP,N are the amounts of migrant (g) in the food and coating, respectively, while VF and VP are the volumes of food
and coating (m3), respectively.
Partition coefficient determination is experimentally difficult due to the influence of experimental conditions on K measure-
ments. For the same system, data found in the literature can have large variations. Nevertheless, due to the great number of
measurements by various authors, average values can be found which correctly represent the partition of compounds between
polymers and simulated foods (15–19). The values in real foods are very rare. From these data, several approaches were developed
to predict K. One of the most efficient methods is the Quantitative Structure–Property Relationship (19,20). The first predictive
method was developed by Hildebrand et al. (1936) based on cohesive energy density (21). Paik et al. (1993) improved the
prediction by including the entropic contribution (22). The group contribution is more and more being used to predict partition
coefficients, with a good correlation between molecular structure and K. The UNIFAC models (UNIQUAC Functional group Activity
Coefficient) is the most often used QSAR; it is considered the functional group role and is adjusted to real data by the UNIQUAC
(UNIversal, QUAsi-chemical Activity Coefficient) data.
The Flory–Huggins theory has also been used and been demonstrated as a good approach to predict the activity coefficient of
migrants in polymers and simulated foods (23–25). The liquid/air partition coefficients of migrants were calculated according to the
Phase Ratio Variation method. This method is described and represents a convenient way to determine gas–liquid partition
coefficients for practical purposes, utilizing equilibrium headspace–gas chromatography (19).
4.20.2.1.2 Molecular Diffusion

As previously explained, Fick’s laws are the most popular models for molecular migration in foods, coatings, and packaging
materials, when interactions between the diffusing component and the matrix can be neglected. Diffusion in homogeneous media is
based on the assumption that unidirectional transfer through active coating is proportional to the concentration gradient between
both sides of the coating, as:
vCF ðx; tÞ vCF ðx; tÞ
¼D [3]
vt vx2
where CF(x,t) is the concentration of migrants in the film at position x and time t, and D is the diffusion coefficient of migrants
through the coating (cm2 s1).
4.20.2.2 Water Vapor and Gas Properties of Polymers

To determine the molecular diffusivity of gas and water, the experimental arrangement is usually based on the permeability
measurement. The permeability is preferably determined if a ‘barrier’ property is primarily aimed for. For a long time, many studies
have been carried out on food-packaging permeability (26,27). The permeability P (kg m1 s1 bar1) includes the kinetic and
thermodynamic properties of the packaging system and migrant. P is related to the diffusivity, D (m2 s1), and solubility,
S (kg m3 bar1), corresponding to the migrant concentration under given conditions, experienced by the film:
P ¼ D$S [4]
This commonly used equation is based on the assumption that Fickian and Henry laws apply and traduce that D and S are not
dependent on migrant concentration. The parameter P is measured during the transport stationary phase (28). Diffusion of volatile
www.iran-mavad.com
molecules in a gas, food, or packaging is associated with accumulation of the migrant in material. In the case of gas solubility in
a polymer, some authors propose the use of Henry’s equation:
S¼CHp [5]
with H, Henry’s constant; C, gas concentration in the polymer and p, gas partial pressure in the air.
This equation supposes a linear sorption curve, which is rarely the case due to the dependence of S on migrant concentration and
on interactions between migrant and packaging (29–32). The curve of the gas sorption isotherm is experimentally determined with
a gravimetric method.
Sorption in semicrystalline packaging polymers is considered to take place exclusively in amorphous regions. Indeed, the
crystalline organization (periodic repetition of arrangement of atoms, interchain links, and denser structure) makes it difficult
for gas molecules to reach the sorption sites due to very limited mobility. The only known exception to this relates to poly
(4-methyl-1-pentene), which has a crystal structure of open lower density than the amorphous region (33), and its sorption
capacity is reduced to 30% of that in the amorphous region. In this particular case, it is no longer possible to neglect sorption
(and diffusion) in crystalline regions of the packaging polymer.
The effect of molecular weight on solubility is opposite to that observed for the diffusivity. Solubility increases with molecular
weight. For a container of polyvinyl alcohol (PVOH), Figure 3 shows that a higher migrant melting point led to greater solubility
(34). Knowledge of gas permeability is essential for the application of polymers as oxygen barriers. The important factors that affect
permeability are the relative orientation of the fibers in the matrix and the state of aggregation and dispersion of these fibers in
the film matrix (35). Oxygen permeability data for a number of biobased polymers as well as more traditional packaging materials
are shown in Table 1. Although it can be difficult to compare permeability data obtained by various research groups using
1.E-12 1.E-01 diffusivity
solubility
Diffusion Coefficient
Solubility Coefficient
1.E-02
[mol Nm−1]
[m2 s–1]
1.E-13
1.E-03
1.E-14 1.E-04
te te e e e
a na at at oa
t
et io tyr ler r
ac op bu va ap
yl r yl yl yl
c
eth lp th th th
hy e e e
et
Ethyl Ester
Figure 3 Diffusivity and solubility of the ethyl ester in the PVOH at 25 C. Reproduced from Landois-Garza, J.; Hotchkiss, J. H. Permeation of High-
Barrier Films by Ethyl Esters. In Food and Packaging Interactions; Hotchkiss, J. H., Ed.; American Chemical Society: Washington, 1988; Vol. 365, pp 42–58.
Table 1 Oxygen permeability (1016 cm2 s1 Pa1) of packaging films at 23 C
Film PO2 gas (1016 cm2 s1 Pa1) Reference
Polyester 11.9 (36)

Poly(ethylene terephthalate) (PET) 57 (37)
Polyamide (PA) 11.5 (37)
Poly(ethylene naphthalate (PEN) 5.7 (37)
High-density polyethylene 285 (36)
Low-density polyethylene 1003 (36)
Cellophane 1.34 (36)
Ethylene vinyl alcohol EVOH (70% VOH) 14 (38)
Poly(vinyl alcohol) (PVOH) 0.24 (39)
Poly(hydroxyalkanoate (PHA) 173 (40)
Poly(lactic acid) (PLA) 185 (40)
www.iran-mavad.com
Table 2 Water vapor permeability (1011 g m1 Pa1 s1) of packaging films
Film Temperature ( C) WVP (1011 g m1 Pa1 s1) Reference
Poly(ethylene naphthalate (PEN)) 23 0.00798 (37)

Polyamide (PA) 23 0.11403 (37)
Poly(vinyl alcohol) (PVOH) 23 0.3421 (37)
Ethylene vinyl alcohol (EVOH) 23 0.03421 (37)
Cellophane 25 6.9 (43)
Low-density polyethylene (LDPE) 25 0.19 (43)
Hydroxypropylmethylcellulose (HPMC) 25 10 (43)
Polyethylene terephthalate (PET) 37.8 0.28 (44)
Polystyrene (PS), oriented 37.8 0.42 (44)
Poly(lactic acid) (PLA), oriented 37.8 1.34 (44)
different equipments or procedures, data in Table 4 indicate the oxygen permeability of materials. Cellophane, poly (vinyl alcohol)
PVOH, ethylene vinyl alcohol (EVOH), and polyester films have excellent oxygen barrier properties, with values below 15
(1016 cm2 s1 Pa1) in all cases.
The gas transport behavior of a semicrystalline polymer is often considered in terms of a two-phase model, consisting of an
impermeable crystalline phase dispersed in a permeable amorphous matrix. In general, the drop in permeability of polymers by
crystallization is mostly explained by only two factors: crystallinity and tortuosity. The impermeable crystallites decrease the amount
of amorphous material through which the gas molecules can permeate. The crystallites increase the tortuosity of the transport path.
Tables 1 and 2 present data from the literature on gas and water vapor. As previously explained, literature values vary to a great
extent due to varying measurement equipment and conditions.
Although the gas permeabilities of plastic films are relatively well known, biopolymers have been less studied and so remain
relatively unknown. The oxygen permeability depends on the type of polymer used. The oxygen barrier properties of biopolymers
are often very important compared to those of synthetic materials. Material thickness effects on gas permeability have been studied
by Greener-Donhowe and Fennema (41). This increase in permeability to gases as a function of the relative humidity is due to the
increased water content and consequently higher gas solubility in the film.
Water vapor permeability (WVP) is also an important parameter in food packaging that comprises diffusion and sorption. The
interactions between polymer and water molecules are fundamental to evaluate water permeation through film, which usually
occurs through the hydrophilic part of the film (42). The relation between hydrophilic and hydrophobic portions is an important
factor in determining WVP. The values obtained for WVP properties of several commercial polymers as well as some biomaterials
are shown in Table 2. WVP values of PEN and EVOH are several orders of magnitude lower than the corresponding values for the
different hemicellulose films. Compounds with high hydrogen bonding generate films that are susceptible to water vapor, while
compounds with hydrophobic groups make excellent barriers to moisture (45).
The relative humidity plays a very important role in biomaterial gas permeability, more significantly than that shown for
synthetics. The influence of humidity on gas-barrier properties is typical for films based on biopolymers, mainly due to their polar
nature (46,47). However, the behavior of hydrophilic films of natural origin with temperature and relative humidity is very complex
and does not easily follow the laws of Fick and Arrhenius. For protecting fresh products with a short shelf-life from deterioration
while in storage, the development of high barrier multifunctional polymer packaging materials has been a long-standing issue in the
packaging industry.
In recent years, nanomaterials have attracted increasing attention because of their unique advantages. The application of the
nanocomposite concept has proven to be a promising option for improving mechanical and barrier properties (48,49). Adding
nanoplatelets into a polymer matrix seems to be a promising strategy for fabricating high-barrier polymeric films (50–52). With the
presence of nanoplatelets in a polymer matrix, the barrier properties of polymer-based nanocomposites are enhanced, mainly
contributing to a ‘tortuous path effect’ or ‘nano-barrier wall effect’ for diffusing molecules.
4.20.2.3 Contaminants
The amount of the various components that migrate into food depends on the chemical and physical properties of food and its
coating. Various factors such as migrant concentration, molecular weight, solubility, diffusivity, partition coefficient, time,
temperature, coating, and food compositions and structures (density, crystallinity, chain branching) are the main controlling
factors of migration (53,54). Therefore, the properties of the edible films may be changed when active compounds are incor-
porated, mainly through an interaction with the film matrix, which is usually formed of a polysaccharide or protein system. In the
case of coating of humid food, water has a particular role. Interaction between water and the coating film occurs, and the coating
structure changes with time. Active compound diffusion differs from ideality due to polymer relaxation that occurs above the glass
transition, Tg. This state is called ‘rubbery’ opposite to the other ‘glassy’ state that exists below Tg. When relaxation is faster than
diffusion, active compound transport follows pseudo-Fickian diffusion. The polymer structure evolves faster than the molecule
migrates. If the relaxation process is slower than the diffusion process, non-Fickian diffusion occurs. From Fick’s first law
www.iran-mavad.com
(eqn [3]), authors have modeled mass transport through mono- and multilayer films (55–60). All models developed are listed in
a review (61).
These models were used by various authors to calculate diffusivities in various applied systems (62). The diffusion coef-
ficient (D) was determined from the experimental data using a relationship from Fick’s second law for a plane sheet with the
following assumptions: (1) the initial concentration of migrant is uniform across the coating film, (2) the initial migrant
concentration in the food-simulating liquid is zero, and (3) D is not concentration dependent (generally not true but often
acceptable).
" #
Mt 8 X
N
1 2ðn þ 1Þ2 p2 Dt
¼1 2 exp [6]
MN p n¼0 ð2n þ 1Þ2 L2
where Mt and MN are the mass of active compound released from the film into food at time t and at infinite time, respectively. L is
the thickness of the coating film.
For short times (Mt/MN < 2/3), eqn [6] can be simplified to eqn [7], and apparent diffusivity can be determined with eqn [8].
1
Mt Dt 2
¼4 [7]
MN 4L2 p

SL 2
D¼ p [8]
2
where S is the slope of a plot representing Mt/MN versus t1/2.

To investigate the mechanism of diffusion, the following power law model has proved useful.
Mt
¼ kt n [9]
MN
Where k is a constant that describes the macromolecular network and n is a diffusion exponent attribute of the release mechanism.
According to this law, for n ¼ 0.5, the diffusion transport mechanism follows Fick’s law; for n ¼ 1, the solute transport is directly
proportional to time; for 0.5 < n < 1, anomalous (non-Fickian) transport is the predominant mechanism, and for n > 1, other
transport processes occur.
4.20.3 Edible Coating
Coating may be defined as a thin layer of material that covers the food surface, while packaging films are formed separately and later
applied to the product. Films and coatings that can be eaten as part of the whole product are called edible. The composition of
edible coatings must therefore conform to the regulations that apply to food products (63). Edible coatings have already been used
to preserve a variety of foods, and they serve many purposes, such as improving appearance or texture and attractiveness to the
consumer and may enable new product design. Edible coatings can also be barriers against the migration of moisture (64,65),
colorants, oxygen, or other components, and thus increase the shelf-life of a product. For a good barrier function, the film should be
as thin as possible, without defects. The application method in combination with coating formulation and target product properties
will determine the thickness, permeability, and homogeneity of the films produced (66). Another major benefit of edible coatings,
eaten along with the food, is that they can provide additional nutrients, may enhance sensory characteristics, and may include
quality-enhancing antimicrobials. Interest in the development of biodegradable and edible films is also increasing due to envi-
ronmental concerns and regulations.
The composition of edible films determines the functional properties of the whole product. Edible coatings may be
composed of polysaccharides, proteins, lipids, or a blend of these compounds (67–69). Their presence and abundance deter-
mine the barrier properties of material with regard to water vapor, oxygen, carbon dioxide, and lipid transfer in food systems.
Proteins and polysaccharides provide good mechanical and sensory properties. They are effective barriers to aroma compounds
and gases such as oxygen and carbon dioxide, but due to their hydrophilic nature, they are not effective water vapor barriers
(70,71). Polysaccharides have been largely used because of their ability to form films and their selective permeabilities to O2 and
CO2 (72). These films are excellent oxygen, aroma, and oil barriers and provide strength and structural integrity (73,74). Some
polysaccharides that have been used in coating formulations are starch and pectin (75,76), cellulose (75,77), chitosan
(75,78,79), gums (80–82), and alginate (75,77).
Proteins have received great attention for their ability to form edible films and coatings. They include corn zein, wheat
gluten (WG), soy protein, whey protein, casein, collagen/gelatin, pea protein, rice bran protein, cottonseed protein, peanut
protein, and keratin (83–85). However, some considerations with respect to food intolerances, such as wheat gluten intol-
erance (celiac disease), or milk protein intolerance, allergies, or religious beliefs/banning, should be taken into account when
protein-based films and coatings are used. Like polysaccharide-based films, the protein films are also excellent oxygen, aroma,
www.iran-mavad.com
and oil barriers and provide strength and structural integrity; but they are not effective moisture barriers (70,73,86). Proteins
make good film formers and are produced from renewable resources and degrade more readily than other types of polymeric
material (86).
In contrast to polysaccharides and proteins, lipid-based films, due to their apolar nature, have proved to be good moisture
barriers, but they have less effective mechanical properties and possess undesirable sensory properties (70,71). Some of the lipids
that have been used effectively in coating formulations are beeswax, mineral oil, vegetable oil, surfactants, acetylated mono-
glycerides, carnauba wax, and paraffin wax (74). Chocolate is still the coating material preferred by consumers. Lipids offer limited
oxygen barrier properties, due to the presence of microscopic pores and elevated solubility and diffusivity. Lipid components are,
therefore, incorporated in composite coatings made up of at least two materials. The supporting matrix, if made up of hydrophilic
polymers, may affect film resistance to water vapor transmission (87).
The current approaches to extending the functional and mechanical properties of films include (1) incorporation of hydro-
phobic compounds such as lipids to improve their resistance to water (88), (2) optimization of the interaction between polymers
(protein–protein interactions, charge–charge electrostatic complexes between proteins or polysaccharides) and (3) cross-linking or
functionalization through physical, chemical, or enzymatic treatments (89).
4.20.3.1 Bioactive Edible Coating

The main technical function of edible films and coatings is to offer a protective barrier to moisture, oxygen, flavor, aroma, and/or oil
between the food and the environment. They may also maintain food integrity by providing some mechanical protection. These
protective functions may also be enhanced with the addition of antioxidants, antimicrobials, flavors, nutrients, enzymes, or
microorganisms that prevent microbial growth or lipid oxidation in coated food products. The essential oils (EOs) can be
a promising additive since different pieces of work have constituted the evidence of their effectiveness as antimicrobial and anti-
oxidant compounds (90,91). The active coating may contain nutrients or other bioactive compounds that have a positive effect on
health, especially due to the application of new microencapsulation or nanoencapsulation techniques. Moreover, they can slow
food deterioration by inhibiting the growth of microorganisms, due to their natural intrinsic activity or to the incorporation of
antimicrobial compounds (92).
Edible coatings will see great development in the near future to allow better preservation and nutritional value of food products.
Indeed, at the present time, industrial companies are looking for highly nutritive and healthy foods. For that reason, they add
healthy components such as antioxidants (flavonoids and polyunsaturated fatty acids). These compounds have demonstrated for
years their action against several illnesses, such as Alzheimer’s disease. If these active compounds are added directly in food, edible
coating can be used just to limit oxygen access to sensitive molecules and then extend shelf-life. Another way is to incorporate the
active compounds in the coating and so increase the nutritional value without modifying the food processing. In that case, the
incorporation process is as described in Figure 4. This type of encapsulation in a coating film allows the development of a large
variety of products. The first step is to choose the best active compounds dedicated to the objective of food quality (i.e., antioxidant
(93) and antibacterial (94)). Such compounds can sometimes be functionalized to improve their particular activity. The
compounds are then encapsulated and included in active coating film. The coating is deposited on the food surface by immersion in
an active solution or by spraying the food surface.
Medical labs are also involved in the development of these films to ensure the safety of the food system. The medical aspect
concerns mainly toxicity evaluation and could be compulsory in case of new food systems to fulfill the regulations (additives and
European novel food regulation 258/97/EC). This aspect is not within the scope of the present chapter and so is not developed,
despite its great general interest.
Figure 5 shows a SEM photograph of an edible film, including active capsules.
4.20.3.1.1 Antimicrobial Compounds

Edible films and coatings containing additives used to prevent biological deterioration are called antimicrobials or preservatives.
Edibility, biodegradability, and increased food safety are the main benefits of active edible films. Their environmental friendly
aspects make them alternatives in packaging systems, without the ecological costs of synthetic nonbiodegradable materials. In the
future, they will be able to replace partially or totally conventional synthetic packaging (95). Traditionally, antimicrobial agents are
directly added to foods, but their activity can be inhibited by different food components, decreasing their efficiency. In such cases,
implementation of films or coatings can be more efficient than antimicrobial additives used in foodstuffs, since they can selectively
and gradually migrate from the wrapping compounds to food surface (96). Antimicrobial compounds include some inorganic acids
(carbonates, bicarbonates, etc.) or organic acids and their salts (propionates, sorbates, benzoates, etc.), parabens, chitosan, enzymes,
bacteriocins, polypeptides, and EOs or other natural extracts (97). Organic acids are the most common synthetic antimicrobial
agents and include acetic, benzoic, citric, fumaric, lactic, malic, propionic, sorbic, succinic, and tartaric acid, among others. These
acids typically inhibit the outgrowth of bacterial and fungal cells. Potassium sorbate and sodium benzoate are the two organic acid
salts that are more widely used as antimicrobial food additives. Natural antimicrobial agents include chitosan, polypeptides, and
plant oils, extracts, and spices.
Chitosan inhibits the growth of a wide variety of fungi, yeasts, and bacteria and is largely used for its film-forming property;
chitosan is used to prepare films and coatings (98,99). Chitosan is a deacetylated derivative of chitin, a cellulose-like molecule that is
the second most abundant natural polymer widely distributed in living organisms, such as crustaceans, insects, and fungi. Due to its
www.iran-mavad.com
Molecule
Functionalization Nutritional and
Active Compound Active edible films
Coated food
matrix
Edible coating
Absorption
Nano/microstructure Digestion Transport of active compounds to
targeted cells
Figure 4 Process of active compound inclusion in edible coating and absorption in the body.
Figure 5 Incorporation of active compound in edible film observed by SEM. Reproduced from Imran, M.; Revol-Junelles, A. M.; Rene, N.; Jamshidian,
M.; Akhtar M. J.; Arab-Tehrany, E.; Jacquot, M.; Desobry, S. Microstructure and Physico-chemical Evaluation of Nanoemulsion-based Antimicrobial
Peptides Embedded in Bioactive Packaging Films. Food Hydrocolloids. 2012, 29, 407–419.
nontoxic nature, antibacterial and antioxidative activity, film-forming property, biocompatibility, and biodegradability, chitosan
has attracted much attention as a natural food additive. The antimicrobial action of chitosan is influenced by intrinsic factors such as
the degree of deacetylation and polymerization, the chemical or nutrient composition of the medium, or both, and the environ-
mental conditions (e.g., substrate water activity or moisture) (100). In this sense, the highly deacetylated chitosan has more
antimicrobial character than those with a higher proportion of acetylated amino groups because of their greater solubility and
charge density (101).
Other natural substances such as EOs have been recognized for their antibacterial, antifungal, antiviral, insecticidal, and anti-
oxidant properties (102,103). They are characterized as volatile oils obtained from various plants and are widely used as food
flavorings (104). The use of EOs is an alternative to the use of fungicides because EOs also have antimicrobial properties. Several
studies have shown that incorporation of EOs into chitosan films or coatings may not only enhance the film’s antimicrobial and
www.iran-mavad.com
antioxidant properties but also reduce water vapor permeability and slow lipid oxidation in the product on which the film is coated
(105,106). Furthermore, it has been demonstrated that bioactive compounds such as EOs can be added to such coatings in order to
extend shelf-life, prevent microorganism growth, and preserve the nutritional values of foods (107,108). EOs are now natural
alternatives for combatting food-borne pathogens and normal food decay caused by bacterial and mold growth (109).
The activity of EOs and their active constituents against many microorganisms, including several pathogens, has been widely
studied (110,111). Although all the components of EOs may present activity, some studies try to determine which compounds are
responsible for the major antioxidant or antimicrobial effects. Carvacrol, thymol, and eugenol are, for instance, the main
components responsible for the antioxidant activity of basil and thyme oils (112). The major components of a number of EOs are
presented in Table 3 (113).
4.20.3.1.2 Antioxidants
Antioxidants are compounds that inhibit or prevent the oxidation reaction caused by free radicals, with or without oxidation
enzymes, that cause discoloration or browning of certain fruit and vegetable tissues and rancidity. Oxidation can affect the color or
flavor of meat, fish, mushrooms, fruit, and vegetable products. A wide series of compounds have been added to edible films to
enhance antioxidant properties; among these compounds are EOs from spices and herbs (132–134), alpha-tocopherol (135,136),
and organic acids (136,137). Films with EOs could be used to slow down lipid oxidation in products with a high content of
unsaturated fatty acids. Major components in EOs are phenolic substances, which are responsible for antimicrobial properties.
Many of them are classified as GRAS and act synergistically with other components (138). Polyphenols are multifunctional anti-
oxidants that act as reducing agents, hydrogen-donating antioxidants, and singlet oxygen quenchers (139). The mechanism of
antioxidant action can include suppressing reactive oxygen species formation, either by inhibiting enzymes or by chelating trace
elements involved in free radical production, scavenging reactive species, and upregulating or protecting antioxidant defenses (140).
Edible films may be effective carriers of food additives (141). Baldwin and Wood (142) demonstrated the efficiency of cellulose
derivative films with entrapped additives (a-tocopherol, citric acid, BHA, BHT) in prolonging the shelf-life of pecan kernels. Atares
et al. (143) studied the impact of some additives with proven antioxidant properties (ascorbic acid, citric acid, and ginger oil) on the
effectiveness of hydroxypropylmethylcellulose (HPMC) films and coatings as protectors against lipid oxidation. Table 3 shows
some EOs and the corresponding tests used to quantify antioxidant activity. Mantle et al. (117) determined the antioxidant capacity
of plant oil extracts through three complementary assay procedures of ABTSþ (1) radical generation (quantified colorimetrically),
by a metmyoglobin catalyst/hydrogen peroxide system; (2) inhibition of iodophenol-enhanced chemiluminescence by a horse-
radish peroxidase/perborate/luminol system; and (3) protection of a target enzyme (human brain alanyl aminopeptidase, activity
Table 3 Major components and properties of different essential oils
Common name of EO Major components Antioxidant properties Antifungal properties Antibacterial properties References
Bergamot Limonene ABTS – Pathogenic bacteria (114,115)

Linalool
Cinnamon Trans-cinnamaldehyde ABTS Aspergillus Pathogenic bacteria (116–118)
Fusarium
Penicillium
Coriander Linalool DPPH Saccharomyces Pathogenic bacteria (110,119,120)
Clove Eugenol DPPH Aspergillus Pathogenic bacteria (116,121–123)
Eugenyl acetate
Eucalyptus Eucalyptol Thiobarbituric acid Candida Pathogenic bacteria (110,124–126)
DPPH Saccharomyces
Rhodotorula
Lemon Limonene ABTS Aspergillus Pathogenic bacteria (114,115,117,127)
Valencene Penicillium
Ocimene
Oregano Carvacrol Thiobarbituric acid Botrytis Pathogenic bacteria (128–130)
Thymol DPPH Fusarium
g-Terpinene Clavibacter
p-Cymene Candida
Saccharomyces
Rhodotorula
Rosemary a-pinene ABTS Candida Pathogenic bacteria (117,126,131)
Bomyl acetate DPPH Saccharomyces
Camphor Rhodotorula
1,8-cineole
Camphor
a-pinene
b-pinene
Adapted from Sánchez-González, L.; Vargas, M.; González-Martínez, C.; Chiralt, A.; Cháfer, M. Use of Essential Oils in Bioactive Edible Coatings. J. Food Eng. Rev. 2011, 3, 1–16.
www.iran-mavad.com

quantitated via fluorimetric assay) against oxidative damage by OH or O 2 generated by Co60 g radiolysis. Other authors used
a DPPH assay (144) to determine antioxidant capacity; this method was based on the scavenging of the stable DPPH radical by the
antioxidant.
4.20.3.1.3 Lipids
As mentioned previously, lipid-based coatings have been used for over 800 years. Historically, lipid-based coatings include waxing
fruits as well as coating confectionery products with chocolate (74). Lipid coatings are mainly used for their hydrophobic properties,
which make them good barriers to moisture. Lipid coatings have also been used to reduce respiration, thereby extending shelf-life,
and to improve appearance by generating a shiny product in fruits and vegetables. Lipid-based coatings can be made from a wide
array of lipid substances, including acetylated monoglycerides, natural waxes (such as carnauba wax and beeswax), and surfactants
(74). Because of their hydrophobicity, these compounds were added to natural coatings to limit weight loss due to fruit dehydration
during storage.
Lipid coatings have certain disadvantages such as brittleness, lack of homogeneity, and presence of pinholes and cracks at
the surface of the coating. Composite coatings or multilayer coatings, applied either in the form of an emulsion or in successive
layers (multilayer coating), have been developed to combine good structural and gas-barrier properties of hydrocolloid
coatings with good moisture-barrier characteristics of lipids. The coating method must be optimized to ensure the best barrier
properties.
4.20.3.2 Physical Properties of Edible Coating

4.20.3.2.1 Water Vapor Permeability (WVP)
The moisture-barrier properties of edible films have been studied by many research groups. Edible films and coatings made from
carbohydrates and proteins generally exhibit lower moisture barriers due to their hydrophilicity (145). To improve moisture-barrier
characteristics, hydrophobic compounds such as lipids could be added to hydrophilic materials. Surfactants could also be used to
reduce the surface tension of the solution, improving the wettability and adhesion of the coating (146,147).
Studies have focused on films containing lipids, which are less permeable, or proteins, which are more permeable. Protein films
have applications with high moisture foods, such as (processed) fresh fruits and vegetables. Lipids are one of the most frequently
used materials to decrease the hydrophilicity of polysaccharide-based films/coatings leading to a decrease of water sensitivity, thus
improving properties such as WVP and water solubility. Like other lipids, EOs may also improve the water barrier properties of the
films because of their hydrophobic nature.
The direct measurement of water exchange between the coated food and the environment under controlled conditions makes it
possible to determine in situ if the coating shows the required properties for a specific combination of storage conditions (148).
Thus, the water vapor resistance of coated fruits can be obtained by monitoring the weight loss of samples both at a controlled
temperature and controlled relative humidity conditions.
An indirect method of determining the water vapor permeability of coated fruits is to form a coating by casting and drying on
a plate and measure its WVP by gravimetric techniques (149).
4.20.3.2.2 Gas Permeability

Oxygen and carbon dioxide permeability of coatings is also an important characteristic to consider when respiration or oxidation
reactions could affect product quality.
Edible films with antioxidants could provide continuous protection for foods against oxidation during storage or after opening
while also providing a recyclable package material (150–152). Edible films and coatings can prevent the deterioration of many food
products because they often possess excellent oxygen barrier properties. Thus, the relation between antioxidant addition and any
oxygen permeability change in films is of immediate concern.
Chen and Nussinovitch (153) evaluated the PO2 , and PCO2 of wax films (carnauba wax or shellac) blended with locust bean
gum (LBG) and guar gum (GG) and compared their properties with those of commercial coatings (polyethylene and shellac).
The results showed that the incorporation of GG and LBG in wax coatings decreased the permeabilities to oxygen and carbon
dioxide.
The gas transfer rates depend on both internal and external factors. The internal factors include the species, the cultivar, and the
growth state, while the external factors include atmospheric composition O2, CO2 and ethylene ratios, temperature, and other stress
factors (154). Edible films and coatings act as semipermeable barriers that may be able to preserve food quality. The modified
atmosphere created by the coating generates a physical capture of CO2 inside the fruit and a partial sealing of the pores, reducing the
gaseous exchange.
Table 4 shows examples of some proteins and polysaccharides used for coating purposes with their main functional charac-
teristics. They exhibit a wide variety of different molecular characteristics depending on their biological origin and function (155).
4.20.3.2.3 Mechanical Properties

The mechanical properties of edible films and coatings depend on the type of film-forming material and especially on its
structural cohesion. Cohesion is the result of a polymer’s ability to form strong and/or numerous molecular bonds between
polymeric chains, thus hindering their separation. This ability depends on polymer structure and especially molecular strength,
www.iran-mavad.com
Table 4 Water vapor, oxygen, and carbon dioxide permeability of films based on different polysaccharides, lipids, and proteins
Film WVP 1011 a T ( C) P O2 b

P CO 2 b
T ( C)
Methylcellulose 7.55 25 1.12 64.19 30

Hydroxypropylcellulose 11 – 1.01 62.03 30
Hydroxypropylmethylcellulose 10.5 27 0.01–0.1 – 25
Starch 217 23 137.5 2523.7 20
Chitosan 360 25 0.90 15.33 25
Pectin – – 2.55 40.78 25
Carrageenan 190 25 0.362 – –
Zein 11 21 0.25 1.13 20
Gluten 4.3 23 1.88 46.88 25
Soy 354 25 0.067 – 25
Whey proteins 417 25 0.001–0.01 – 25
Sodium caseinate 42.5 25 0.76 4.56 25
g m1 s1 Pa1.
a
ml mm m2 d1 Pa1.

b
Adapted from Vargas, M.; Pastor, C.; Chiralt, A.; McClements, J.; González-Martinez, C. Recent Advances in Edible Coatings for Fresh and Minimally Processed Fruits. Crit. Rev.
Food Sci. Nutr. 2008, 48, 496–511.
geometry, molecular weight distribution, and position of its lateral groups (67). The mechanical properties of edible coatings are
strongly affected by physical, chemical, and temperature conditions, which influence film stability and flexibility. Moreover, it is
well known that the environmental conditions during production, storage, and use of these materials affect their mechanical
properties (156).
The ability of edible films and coatings to protect food against mechanical damage is usually assessed by determining film
tensile properties: (1) Young’s Modulus, which determines film stiffness as determined by the ratio of pulling force/area to degree-
of-film-stretch, (2) tensile strength (TS), which indicates the pulling force per film cross-sectional area required to break the film,
and (3) elongation at break (E), which gives the degree to which the film can stretch before breaking and is expressed as
a percentage (157).
The lipids are often supported on a polysaccharide matrix to provide a film with good mechanical strength (158). The addition
of substances with plasticizing properties is, in general, required to provide films with good mechanical behavior (flexibility). The
plasticizer most often used is glycerol because of its better stability and compatibility with hydrophilic biopolymeric chains in
comparison to sorbitol, polyethylene glycol (PEG), and sugars (159).
Some recent examples of hydrocolloid use in edible films are presented in Table 5. Mechanical parameters are included in the
table for comparison. The table mentions the most usual film-making materials and their most relevant characteristics.
4.20.3.2.4 Antimicrobial Properties

Additives used to prevent biological deterioration are called antimicrobials or preservatives. As they are allowed for food contact
applications, this category comprises natural or synthetic compounds with known and minimal toxicological effects on mammals
and the environment. Studies on antimicrobial, barrier, and mechanical properties should be combined to correctly predict the
behavior of the antimicrobial edible film (46,160). Properties of these films are strongly influenced by the type and concentration of
the antimicrobial compound as well as the nature of the film matrix. Numerous studies have been performed to establish the
effectiveness of antimicrobials in films and coatings. The election of measurement method depends on assay purpose, nature of
antimicrobial and characteristics of target microorganisms, and other factors.
Table 5 Hydrocolloids used in edible film elaboration and film mechanical properties
Mechanical properties
Hydrocolloid Concentration in the film-forming solution Plasticizer Tensile strength (MPa) Strain at break (%)
Cellulose ethers HPC 3% w/w – 20–43 26–41

HPMC 1–6% w/w PEG 3–86 1.9–60.9
Chitosan 1% w/w – 7–12 11–17
1.5% w/w – 18–106 5–20
Proteins Sodium caseinate 4% w/w Sorbitol 2–77 2–130
Soy protein isolate 10% w/w Glycerol 4.7–10.7 –
Blends Whey protein/HPMC Glycerol 4–61 16–112
Starch/chitosan Glycerol 0.36–19.7 61–152
HPC: Hydroxypropylcellulose; HPMC: Hydroxypropylmethylcellulose.
Adapted from Campos, C. A.; Gerschenson, L. N.; Flores, S. K. Development of Edible Films and Coatings with Antimicrobial Activity. Food Bioprocess. Technol. 2011, 4,
849–875.
www.iran-mavad.com
Table 6 Antimicrobial activity of edible polysaccharide-based and protein-based composite films containing antimicrobial agents
Antimicrobial Antimicrobial
Film matrix agent Concentration Target pathogen Pathogen inoculation activitya
HPMCb Chitosan 0.1% Aspergillus niger 104 spores per Petri þ

dish
HPMC Nisin 5 104 IU ml1c Listeria monocytogenes, 70 ml per well þ
Staphylococcus
aureus
HPMC Nisin 104, 105 IU ml1 Micrococcus luteus 270 0.1% of the strain in –
nutritive broth (v/v)
HPMC Nisin 250 mg ml1 Kocuria rhizophila 0.1% of the strain in þ
nutritive broth (v/v)
HPMC-lipid Sodium 2.0% Penicillium digitatum, 103, 104 spores þ
bicarbonate Penicillium italicum per ml
MCd – – Rhodotorula rubra, Penicillium 0.1 ml of suspension –
notatum
MC Natamycin 1.5 mg/10 g film A. niger 104 spores per ml –
solution P. roquefortii 106 spores per ml þ
1.0 mg/10 g film
solution
Chitosan-MC Potassium sorbate, 4.0% R. rubra, P. notatum 0.1 ml of suspension þ
Sodium benzoate
Chitosan-HPMC – – L. monocytogenes 300 cells þ
Chitosan – – R. rubra, P. notatum 0.1 ml of suspension –
Chitosan Nisin 250 mg ml1 A. niger, K. rhizophila 103, 104 spores per ml
ATCC 9341
Chitosan – – A. niger 102 spores per Petri dish þ
Pea starch Grape seed extracts 1.0% L. monocytogenes, S. aureus, 106 CFU ml1 þ
Entercoccus faecium,
E. faecakis, Brochothrix
thermosphacta
Sodium Garlic oil 0.2% S. aureus, B. cereus 105, 106 CFU ml1 þ
alginate
WPIe Lactoferrin 0.1 g g1 Penicillium comune 105 spores –
WPI Lactoperoxidase 0.5% L. monocytogenes 103 CFU ml1b þ
Sodium caseinate Sodium lactate 40.0% L. monocytogenes 102 CFU cm2 þ
a
Antimicrobial activity: ‘þ’ ¼ inhibition; ‘’¼no inhibition.
b
HPMC ¼ hydroxypropylmethylcellulose.
c
IU ¼ international units; 1 mg corresponds to 40 IU.
d
MC ¼ methylcellulose.
e
WPI ¼ whey protein isolated.
Adapted from Valencia-Chamorro, S. A.; Palou., L.; del Río, M. A.; Pérez-Gago, M. B., Antimicrobial Edible Films and Coatings for Fresh and Minimally Processed Fruits and
Vegetables: A review. Crit. Rev. Food Sci. Nutr. 2011, 51, 872–900.
To select an antimicrobial molecule, effectiveness against the target microorganism and possible interactions among the anti-
microbial, the film-forming biopolymer, and other food components present must be considered. These interactions can modify
antimicrobial activity and film characteristics, the key factors in the development of antimicrobial films and coatings.
The antimicrobial activity of different polysaccharide and protein edible films containing antimicrobials is presented in Table 6.
Experimental conditions such as film composition, antimicrobial agent and concentration, strain, and concentration of target
microorganism and analytical method affect film antimicrobial activity. For this reason, the antimicrobial activity of the films is
reported in the tables as inhibition (þ) or no inhibition () of the target pathogenic microorganism with no dependence on the
magnitude of the inhibition, as concluded by the authors, according to their experimental conditions.
In a recent work, edible coatings were prepared with entrapped nisin (94). To study the release kinetic of nisin, an original
fluorescent labeling was done. Figure 5 summarizes the release of the labeled nisin into food and avoids microbial growth.
Figure 6(a) shows the initial system in which nisin is directly added into food without any coating. In this case, nisin action against
spoilage is of short efficiency due to rapid denaturating of free nisin in food.
Figure 6(b) presents nisin migration from a coating to food. The main interest of the coating is that nisin is stable in coating and
diffuses to food. Its slow release into food allows longer nisin efficiency against microorganisms. And finally, in Figure 6(c),
encapsulation slows nisin release and makes coating even more efficient in slowing microorganism growth for an extended time.
In Figure 7, the initial amounts of fluorescently labeled nisin mix in four different films were 0.137 mg nisin cm2. In general,
the release of a bioactive compound from a biopolymeric coating occurs in two stages. In the primary step, water molecules from
www.iran-mavad.com
Figure 6 Controlled release of active compounds from surface coating: (a) nisin directly included in the food; (b) coating with free nisin; (c) coating with
encapsulated nisin. Reproduced from Muhammad Imran, Amira Klouj, Anne-Marie Revol-Junelles, Stéphane Desobry. Nisin Controlled Release from
HPMC, Sodium Caseinate, Poly-Lactic Acid and Chitosan for Active Packaging Applications. J. Microbiol. Meth. 2013. In press.
food penetrate/diffuse into the coating. With hydration, mobility increases, allowing the active agent to diffuse from the film into
the food until the thermodynamic equilibrium between the two phases is reached. Thus, the nisin release kinetics rely on the chain
of events: (1) food penetration/diffusion; (2) biomacromolecular network relaxation; and (3) diffusion of active compound
through the polymeric array. Nisin release kinetics can be described by means of Fick’s Second Law (eqn [6]), intended for a plane
sheet with constant boundary conditions and uniform initial concentration.
Nisin release kinetics for different biopolymers are reported in plot of fractional nisin migration (mass released) with time at
4 and 40 C (Figure 7). Labeled nisin release behaviors differ comparing the four matrices. As expected, the amount of nisin released
in the outer solution increases until equilibrium is reached between coating and simulated food.
Higher amounts of nisin were released from active HPMC films at all times and temperatures compared to other films.
Concerning active HPMC films, 40 and 55% of nisin were released from the films at 4 and 40 C temperature, respectively,
until equilibrium was reached after 160 h. In the case of chitosan films, nisin diffusion was restricted at both temperatures,
www.iran-mavad.com
Figure 7 Diffusion of labeled nisin from biopolymeric films into water–ethanol (10:90 v/v) simulated food. Influence of temperature (4 C C and 40 C
-) and biopolymer type ((a) Hydroxypropylmethylcellulose; (b) Chitosan; (c) Sodium caseinate) on release kinetics.
and only <20% of the initial amount (M0) in films were released into the aqueous-ethanol simulated food. The diffusion of
an antimicrobial compound in one or between several phases depends on the characteristics of the active compound,
solubility, partition coefficients, temperature, physicochemical properties of biopolymers, and active agent–biopolymer
interactions. Among these interactions, the temperature and additive–biopolymer interaction strongly influences the
respective kinetic release. The glass transition temperature of plasticized sodium caseinate films at lower water activity occurs
around 40 C, enabling the characteristic transition to a flexible/rubbery state, which may explain the higher release of nisin
from caseinate at 40 C.
4.20.4 Added Value and Limitations of Edible Coatings: Sensory Properties
The coating layer limits mass transfer from food to the environment (water, aromas) and from the environment to foods
(oxygen, odors, and microorganisms). As previously shown, this coating can also be active, releasing molecules of interest, such
www.iran-mavad.com
as antioxidants or antibacterials. A lot of industrial projects concern conception and demonstration of interest of such new films
for food preservation and safety. Nevertheless, some limitations of edible films can be easily reached and are generally linked to
their intrinsic quality. First, we will consider the sensory properties and weaknesses of edible coatings and then food safety
aspects.
Sensory properties are essential for consumer acceptance. The sensory evaluation concerns the use of four out of our five senses:
visual aspect, smelling, taste, and touch.
4.20.4.1 Visual Properties

Coating can improve the visual presentation of food products or, contrarily, it can decrease the attractiveness. This is one of
the most important limits of edible film use on the industrial scale. Indeed, the objective is that the coating films are
invisible, so their thickness is generally reduced to a few micrometers (5–20 mm). With this low thickness, barrier efficiency
against gas and water vapor is low. This limitation strongly reduces interest in such coatings, except when the coating
formulation makes it water resistant and there is low oxygen permeability. Use of formulated fats is the preferred way to
reach these objectives.
Another visual impact is the color of the coating layer. The objective of most coating producers is to reach transparency or at least
to obtain the same color as the product surface to make it very discreet. The main case where color is not a problem is chocolate, for
which the visible brown color is an attractive part of the food product. This particular attractiveness offers new developments for
food coating because young consumers are always attracted by new products with unexpected colors. Formulators and marketers are
working on this high potential aspect that could give edible coatings new added values.
4.20.4.2 Odors
As previously explained, the coating is expected to be as discreet as possible, or it must give added value to the product. This is
particularly true for odors. Food technologists are looking for odorless coatings, and if this cannot be achieved, they add
perfumes or natural volatile compounds to cover the undesired odor. Generally, cheap compounds such as menthol, flowers, or
fruits aromas are added, but in recent years a strong interest has arisen in EOs that possess the double interest of attractive odor
and additional antibacterial properties. The main technological lock now is to stabilize the volatile compound in the coating to
get a slow compound release during the whole shelf-life. Several research teams are focused on this targeted research. Encap-
sulation of the compounds and inclusion in the film is at present the preferred method of volatile compound stabilization.
Nevertheless, much work remains to be done to manage a fully controlled compound release, particularly for intermediate
moisture and humid foods.
4.20.4.3 Taste
Taste is even more difficult to solve because tongue and tooth are very sensitive to taste and texture changes from uncoated to coated
food with edible film coating. To make it easily acceptable, the coating must be as thin as possible. A few micrometers are not
possible for consumers to detect, but a thicker layer will be detected. One way to avoid this is to make the layer texture resemble as
closely as possible the food surface texture. In that case, the consumer cannot separate the edible film texture from the food surface
texture.
The other way is to develop a completely different coating texture and taste and so transform it into a new type of food that adds
value to the product. For example, consumers can appreciate a crusty coating even if the food product is soft. A lot of edible films
have been developed already (Chitosan, HPMC, Casein, Whey proteins, Starch.) to reach this aim.
4.20.4.4 Touch
Before eating, the consumer often evaluates the coated food products with the fingers or with a fork to feel the food texture. This
phase is also of prime importance for acceptability. Penetration strength and mechanical resistance of the coating have to be in
accordance with the product’s overall texture. If the coating is too soft, it will disturb the consumer and make him reject the product
on the basis that the end of shelf-life is reached. If the coating is too strong, the consumer will associate that with surface dehy-
dration and will also reject the product. Developments are still required to fully control the coating texture, by using new edible
polymers and adding plasticizers to make the coating softer and more deformable.
4.20.5 Toxicity of Polymeric Films into Contact with Foodstuff
Edible film and coating toxicity is the main concern of food technologists. Food contact material and food regulations are moving
very rapidly, facing the large number of innovations coming to industry from research labs. In 2004 and 2009 updated versions of
Directive 89/109 were published (Reglement 1935/2004 and 450/2009) to envisage the case of active film and its strong inter-
actions with food. The main requirement for food-packaging ingredients is that they must be listed in the positive list of
www.iran-mavad.com
components accepted for food contact, a purity criterion is added in the regulation and efficiency of active packaging film has to be
demonstrated. Directive 10/2011/EC is the most recent version of the regulation specifically dedicated to plastic materials (directive
2002/72 and following ones). For the first time, biopolymer films in contact with foodstuffs are regulated with regard to food safety.
If a packaging material includes active compounds, the role of the active substance has to be demonstrated, the molecule reference
number has to be given, and the conditions of use have to be clearly defined (maximum amount, the way to add it in the coating
material, and so forth).
The case of edible film is more restrictive than that of food contact packaging. This is simply because the coating will be
fully eaten. In that case, all compounds present in the coating are considered as food ingredients. Therefore, they all have to
be present in the positive list of food ingredients to make sure that the maximum concentration is respected. For this
reason, the amount of active compound is often limited and reduces the coating’s active function. The key data involve the
maximum amount of active compound (antibacterial such as nisin, for example) that can be incorporated in the coating.
Despite this limited amount, adding active compounds to the film is of prime interest for several reasons: high concen-
tration of the active compound at the food surface, stabilization and controlled release of the compound, and concen-
tration strictly limited to the required action (no overeating of the active compound by consumers and no waste of money
for industrials).
The efficiency of an active compound is generally concentration dependent. The more active molecules you add in food, the
more benefit you have. In Figure 8, one can see that for a given quantity of active molecule added in food formulation, the highest
surface efficiency is reached through coating. Even if the molecule diffuses during storage, the positive action of the compound at the
surface is more efficient after food coating than after direct inclusion into food.
As shown in Figure 9, even in the case of surface deposition on uncoated food, the process is more efficient in coated food due
to two parameters: too fast release and diffusion without coating and active compound protection into the coating (against
oxygen, acids and so on). So, after several days or weeks of storage the surface concentration will decrease at a higher rate without
coating than with coating. The surface action decreases with active component dilution into food. For example, Imran et al. (2011)
studied the activity of nisin on listeria monocytogenes at the food surface (nutritive agar-simulated food) (94). The efficiency of
coated nisin was much higher when a coating was used. The more the nisin is protected, the longer it acts against bacteria. In
Figure 10, the control is the system without any antibacterial treatment, and Listeria develops very rapidly. The coating alone
(HPMC) has no significant effect. The coating with nisin (HPMC þ Nis) has a significant effect and avoids development of
bacteria.
Figure 8 Active compound distribution in food without (a) and with edible film (b) at initial time (c) during storage.
Figure 9 Active compound distribution in food with surface pulverization (a, b) and edible film (c, d) at initial time and during storage. Reproduced
from Imran, M.; Revol-Junelles, A. M.; Rene, N.; Jamshidian, M.; Akhtar M. J.; Arab-Tehrany, E.; Jacquot, M.; Desobry, S. Microstructure and Physico-
chemical Evaluation of Nanoemulsion-based Antimicrobial Peptides Embedded in Bioactive Packaging Films. Food Hydrocolloids. 2012, 29, 407–419.
www.iran-mavad.com
Figure 10 Inhibition of listeria growth with active coating (Hydroxypropylmethylcellulose, HPMC) and nisin. (H ¼ HPMC, L ¼ nisin-free Liposomes,
N ¼ Nisin, E ¼ Encapsulated nisin in liposomes).
4.20.6 Conclusion
To conclude, edible films and coatings are the future of the barrier material in food packaging and coatings. The biological layers
allow reducing packaging impact on the environment and making the food product safer. Research works are nevertheless required
to produce stable coatings and film from biosourced materials and to entrap active compounds with fully controlled delivery.
References
1. Kwappong, O. Y.; Hotchkiss, J. H. Comparative Sorption of Aroma Compounds by Polyethylene and Ionomer Food Contact Plastics. J. Food Sci. 1987, 52, 761–763.
2. Manheim, C. H.; Miltz, J.; Letzter, A. Interactions between Polyethylene Laminated Carbons and Aseptically Packed Citrus Juices. J. Food Sci. 1987, 53, 737–740.
3. Shimoda, M.; Ikegami, T.; Osajima, Y. Sorption of Flavor Compounds in Aqueous Solution into Polyethylene Film. J. Sci. Food Agric. 1988, 42, 157–163.
4. Hirose, K.; Harte, B. R.; Giacin, J. R.; Miltz, J.; Stine, C. Sorption of D Limonene by Sealant Films and Effect of Mechanical Properties. In Food and Packaging Interactions;
Hotchkiss, J. H., Ed.; American Chemical Society: Washington, 1988; pp 28–41.
5. Halek, G. W.; Meyers, M. A. Comparative Sorption of Citrus Flavor Compounds by Low Density Polyethylene. Packag. Technol. Sci. 1989, 2, 141–146.
6. Giacin, J. R. Factors Affecting Permeation, Sorption and Migration Processes in Package – Product Systems. In Food and Packaging Material Chemical Interactions;
Ackermann, P., Jagerstad, M., Ohlsson, T., Eds.; Cambridge Royal Society of Chemistry, 1995; pp 12–22.
7. Barens, A. R.; Hopfenburg, H. B. Diffusion of Organic Vapors at Low Concentration in Glassy PVC, Polystyrene and PMMA. J. Membr. Sci. 1982, 10, 283.
8. Fick, A. Ueber Diffusion. Annalen der Physik 1855, 170, 59–86.
9. Wroblewski, S. V. Ueber das Dispersionsgesetz. Ann. Phys. (Lpz) 1879, 8, 29.
10. Briston, J. H.; Katan, L. L. Plastic in Contact with Food; Food Trade Press: London, 1974.
11. Shepherd, M. J. The Trace Contamination of Foods by Migrations from Plastic Packaging – A Review. Food Chem. 1982, 8, 129–145.
12. Downes, T. W. Practical and Theoretical Considerations in Migrations. In Food Product-Package Compatibility; Gray, J. I., Harte, B. R., Miltz, J., Eds.; Technomics: Lancaster, PA,
1987; pp 44–58.
13. Kim-Kang, H. Volatiles in Packaging Materials. Food Sci. Nutr. 1990, 29, 255–271.
14. Gavara, R.; Hernandez, R. J.; Giacin, J. Methods to Determine Partition, Coefficient of Organic Compounds in Water/Polystyrene System. J. Food Sci. 1996, 61 (5), 947–952.
15. Hernandez-Munoz, P.; Catala, R.; Gavara, R. Food Aroma Partition between Packaging Material and Food Simulant. Food Addit. Contam. 2001, 18 (7), 673–682.
16. Kinigakis, P.; Miltz, J.; Gilbert, S. G. Partition of VSM in Plasticized PVC/Food Simulant System. J. Food Process. Preserv. 1987, 11, 247–255.
17. Pushpa, S. A.; Goonetilleke, P.; Billingham, N. C. Solubility of Antioxidants in Rubber. Rubber Chem. Technol. 1996, 69, 885–896.
18. Arab Tehrany, E.; Desobry, S. Partition Coefficient in Food/Packaging Systems. Food Addit. Contam. 2004, 21 (12), 1186–1202.
19. Arab Tehrany, E.; Mouawad, C.; Desobry, S. Determination of Partition Coefficient of Migrants in Food Simulants by the PRV Method. Food Chem. 2007, 105 (4), 1571–1577.
20. Arab Tehrany, E.; Fournier, F.; Desobry, S. Simple Method to Calculate Partition Coefficient of Migrants in Food Stimulant/Polymer System. J. Food Eng. 2006, 77 (1), 135–139.
21. Hildebrand, J. H.; Scott, R. L. The Solubility of Non-Electrolytes; Reinhold: New York, 1936.
22. Paik, J. S.; Tigani, M. A. Application of Regular Solutions Theory in Predicting Equilibrium Sorption of Flavor Compounds by Packaging Polymers. J. Agric. Food Chem. 1993, 41,
806–808.
23. Gillet, G.; Vitrac, O.; Desobry, S. Development of Decision Tools to Assess Migration from Plastic Materials in Contact with Food. Food Addit. Contam. 2009, 26 (12),
1556–1573.
24. Gillet, G.; Vitrac, O.; Desobry, S. Prediction of Solute Partition Coefficients between Polyolefins and Alcohols Using a Generalized Flory-Huggins Approach. Ind. Eng. Chem. Res.
2009, 48 (11), 5285–5301.
25. Gillet, G.; Vitrac, O.; Desobry, S. Prediction of Partition Coefficients of Plastic Additives between Packaging Materials and Food Simulants. Ind. Eng. Chem. Res. 2010, 49 (16),
7263–7280.
26. Chao, R. R.; Rizvi, S. S. H. Oxygen and Water Vapor Transport through Polymeric Film. In Food and Packaging Interactions; Hotchkiss, J. H., Ed.; London, 1987, pp 217–242.
27. Biquet, B.; Labuza, T. P. Evaluation of the Moisture Permeability Characteristics of Chocolate Films as an Edible Moisture Barrier. J. Food Sci. 1988, 53, 989–998.
28. McCall, D. W.; Douglas, D. C.; Blyler, L. L.; Johnson, G. E.; Jelinski, L. W.; Bair, H. E. Solubility and Diffusion in Low-Density Polyethylene. Macromolecules 1984, 17,
1644–1649.
29. Stannet, V. The Transport of Gases in Synthetic Polymeric Membranes: An Historical Perspective. J. Membr. Sci. 1978, 3, 97–115.
30. Pace, R. J.; Datyner, A. Statistical Mechanical Model of Sorption and Diffusion of Simple Penetrants in Polymers. J. Polym. Sci. 1980, 18, 1103–1124.
31. Fleming, G. K.; Koros, W. J. Membrane Based Gas Separation. J. Membr. Sci. 1993, 83 (1), 1–80.
www.iran-mavad.com
32. Koros, W. J.; Chan, A. H.; Paul, D. R. Sorption and Transport of Various Gases in Polycarbonate. J. Membr. Sci. 1997, 2, 165–190.
33. Bicerano, J. Statistical Thermodynamic Modeling of the Transport of Penetrants through Polymers: Sorption of Carbon Dioxide. Comput. Polym. Sci. 1992, 2, 177–201.
34. Landois-Garza, J.; Hotchkiss, J. H. Permeation of High-Barrier Films by Ethyl Esters. In Food and Packaging Interactions, Hotchkiss, J. H., Ed.; American Chemical Society:
Washington, 1988; Vol. 365, 42–58.
35. Bharadwaj, R. K. Modeling the Barrier Properties of Polymer-Layered Silicate Nanocomposites. Macromolecules 2001, 34 (26), 9189–9192.
36. Perez, C.; Desobry, S. Protein-Based Films for Food Packaging. Int. Rev. Food Sci. Technol. 2002, 1 (10), 101–107.
37. Drieskens, M.; Peeters, R.; Mullens, J.; Franco, D.; Lemstra, P. J.; Hristova-Bogaerds, D. G. Structure versus Properties Relationship of Poly(Lactic Acid). I. Effect of Crystallinity
on Barrier Properties. J. Polym. Sci. B. Polym. Phys. 2009, 47, 2247–2258.
38. Salame, M. Barrier Polymers. In Wiley Encyclopedia of Packaging Technology; Bakker, M., Ed.; John Willey: New York, 1986; pp 48–54.
39. Grondahl, M.; Eriksson, L.; Gatenholm, P. Material Properties of Plasticized Hardwood Xylans for Potential Application as Oxygen Barrier Films. Biomacromolecules 2004, 5 (4),
1528–1535.
40. van Tuil, R.; Fowler, P.; Lawther, M.; Weber, C. J. In Biobased Packaging Materials for the Food Industries status and Perspectives; Weber, C. J., Ed.; Department of Dairy and
Food Science: Denmark, Frederiksberg, 2000; pp 13–44.
41. Greener-Donhowe, I.; Fennema, O. Edible Films and Coating: Characteristics, Formation, Definitions, and Testing Methods. In Edible Coatings and Preservation: Theory and
Practice; Krachta, J. M., Baldwin, E. A., Nisperos-Carriedo, M., Eds.; Elsevier Applied Science Publishing Co.: London, 1994; pp 371–394.
42. Sangaj, N. S.; Malshe, V. C. Permeability of Polymers in Protective Organic Coatings. Prog. Org. Coat. 2004, 50 (1), 28–39.
43. Peroval, C.; Debeaufort, F.; Despre, D.; Voilley, A. J. Edible Arabinoxylan Films. I. Effects of Lipid Type on Water Vapor Permeability, Film Structure, and Other Physical
Characteristics. J. Agric. Food Chem. 2002, 50, 3977–3983.
44. Auras, R.; Singh, S. P.; Singh, J. Performance Evaluation of PLA against Existing PET and PS Containers. J. Test. Eval. 2006, 34 (6), 530–536.
45. Souza, B. W. S.; Cerqueira, M. A.; Casariego, A.; Lima, A. M. P.; Teixeira, J. A.; Vicente, A. A. Effect of Moderate Electric Fields in the Permeation Properties of Chitosan
Coatings. Food Hydrocoll. 2009, 23 (8), 2110–2115.
46. McHugh, T. H.; Krochta, J. M. Sorbitol- vs Glycerol-Plasticized Whey Protein Edible Films: Integrated Oxygen Permeability and Tensile Property Evaluation. J. Agric. Food Chem.
1994, 42, 841–845.
47. Gontard, N.; Thibault, R.; Cuq, B.; Guilbert, S. Influence of Relative Humidity and Film Composition on Oxygen and Carbon Dioxide Permeabilities of Edible Films. J. Agric. Food
Chem. 1996, 44, 1064–1069.
48. Avella, M.; De Vlieger, J. J.; Errico, M. E.; Fischer, S.; Vacca, P.; Volpe, M. G. Biodegradable Starch/Clay Nanocomposite Films for Food Packaging Applications. Food Chem.
2005, 93, 467–474.
49. Zhao, L.; Li, F.; Chen, G.; Fang, Y.; An, X.; Zheng, Y.; Xin, Z.; Zhang, M.; Yang, Y.; Hu, Q. Effect of Nanocomposite-Based Packaging on Preservation Quality of Green Tea. Int. J.
Food Sci. Technol. 2012, 47, 572–578.
50. Nawani, P.; Burger, C.; Rong, L.; Chu, B.; Hsiao, B. S.; Tsou, A. H.; Weng, W. Characterization of Nanoclay Orientation in Polymer Nanocomposite Film by Small-Angle X-Ray
Scattering. Polymer 2010, 51, 5255–5266.
51. Maji, P. K.; Das, N. K.; Bhowmick, A. K. Preparation and Properties of Polyurethane Nanocomposites of Novel Architecture as Advanced Barrier Materials. Polymer 2010, 51,
1100–1110.
52. Dunkerley, E.; Schmidt, D. Effects of Composition, Orientation and Temperature on the O2 Permeability of Model Polymer/Clay Nanocomposites. Macromolecules 2010, 43,
10536–10544.
53. Chang, S.; Guttman, C. M.; Sanchez, I. C.; Smith, S. E. Theoretical and Computational Aspects of Migration of Package Components to Food. In Food and Packaging Interactions,
Hotchkiss, J. H., Ed.; American Chemical Society: Washington, 1988; Vol. 365, 106–117.
54. Gilbert, J. R.; Miltz, J.; Giacin, J. R. Transport Considerations of Potential Migrants from Packaging Materials. J. Food Process. Preserv. 1980, 4, 27–49.
55. Jost, W. Diffusion in Solids, Liquids, Gases. In Physical Chemistry, Loebl, E. M., Ed.; Academic Press: New York, 1960; Vol. 1.
56. Stannet, V. Simple Gases. In Diffusion in Polymers; Crank, J., Park, J., Eds.; Academic Press: New York, 1968; pp 45–56.
57. Skelland, A. Diffusional Mass Transfer; Wiley Interscience: New York, 1974.
58. Sherwood, T. K.; Pigford, R. L.; Wilke, C. R. Mass Transfer; McGraw Hill: New York, 1975.
59. Crank, J. The Mathematics of Diffusion, 2nd ed.; Oxford University Press: Oxford, 1975.
60. Hines, A. L.; Maddox, R. N. Mass Transfer: Fundamentals and Applications; Prentice Hall: New Jersey, 1985.
61. Masaro, L.; Zhu, X. X. Physical Models of Diffusion for Polymer Solutions, Gels and Solids. Prog. Polym. Sci. 1999, 24, 731–775.
62. Mousavi, M.; Desobry, S.; Hardy, J. Mathematical Modelling of Acetic Acid Migration from Packaging into Food via Package Free Space: Part I – Spherical Shape Foods. J. Food
Eng. 1998, 36, 453–472.
63. Guilbert, S.; Gontard, N.; Cuq, B. Technology and Applications of Edible Protective Films. Packag. Technol. Sci. 1995, 8, 339–346.
64. Baldwin, E. A. Edible Coatings for Fresh Fruits and Vegetables: Past, Present, and Future. In Edible Coatings and Films to Improve Food Quality; Krochta, M., Baldwin, E. A.,
Nisperos-Carriedo, M. O., Eds.; Technomic Publishing Company: Lancaster, PA, 1994; p 25.
65. Baker, R. A.; Baldwin, E. A.; Nisperos-Carriedo, M. O. Edible Coatings and Films for Processed Foods. In Edible Coatings and Films to Improve Food Quality; Krochta, M.,
Baldwin, E. A., NisperosCarriedo, M. O., Eds.; Technomic Publishing Company: Lancaster, PA, 1994; p 89.
66. Domnick, J.; Scheibe, A.; Ye, Q. The Simulation of the Electrostatic Spray Painting Process with High-Speed Rotary Bell Atomizers. Part I: Direct Charging. Part. Part. Syst.
Charact. 2005, 22 (2), 141–150.
67. Guilbert, S.; Gontard, N.; Gorris, L. G. M. Prolongation of the Shelf-Life of Perishable Food Products Using Biodegradable Films and Coatings. LWT-Food Sci. Technol. 1996, 29,
10–17.
68. Li, P.; Barth, M. M. Impact of Edible Coatings on Nutritional and Physiological Changes in Lightly Processed Carrots. Postharvest. Biol. Technol. 1998, 14, 51–60.
69. Arvanitoyannis, I.; Gorris, L. G. M. Edible and Biodegradable Polymeric Materials for Food Packaging or Coating in Processing Foods: Quality Optimization and Process
Assessment; CRC Press: Boca Raton, FL, 1999; pp 357–371.
70. Krochta, J. M.; Mulder-Johnston, C-de. Edible and Biodegradable Polymer Films: Challenges and Opportunities. Food Technol. 1997, 51 (2), 61–74.
71. Tharanathan, R. N. Biodegradable Films and Composite Coatings: Past, Present and Future. Trends Food Sci. Technol. 2003, 14 (3), 71–78.
72. Nisperos-Carriedo, M. O. Edible Coatings and Films Based on Polysaccharides. In Edible Coatings and Films to Improve Food Quality; Krochta, J. M., Baldwin, E. A., Nisperos-
Carriedo, M. O., Eds.; Technomic Publishing Co.: Lancaster, PA, 1994; pp 305–335.
73. Krochta, J. M. FAQ about Edible Films and Coatings, 2001. http://www.dairybiz.com/feature.htm (accessed July 5 2002).
74. Kester, J. J.; Fennema, O. R. Edible Films and Coatings. A Review. Food Technol. 1986, 40 (12), 47–59.
75. Baldwin, E. New Coating Formulations for the Conservation of Tropical Fruits, 2001. http://technofruits2001.cirad.fr (accessed Aug 10, 2002).
76. Javanmard, M.; Ahangari, R.; Tavakolipour, H. Effects of Wheat Starch Edible Films on Rancidity and Moisture Uptake of Pistachio Kernels as a New Package. J. Food Process.
Eng. 2011, 34, 1156–1171.
77. Tien, C. L.; Vachon, C.; Mateescu, M. A.; Lacroix, M. Milk Protein Coatings Prevent Oxidative Browning of Apples and Potatoes. J. Food Sci. 2001, 66 (4), 512.
78. Zhang, D.; Quantick, P. C. Antifungal Effects of Chitosan Coating on Fresh Strawberries and Raspberries during Storage. J. Hortic. Sci. Biotechnol. 1998, 73 (6), 763–767.
79. Tamer, C. E.; Copur, O. U. Chitosan: An Edible Coating for Fresh-Cut Fruits and Vegetables. Acta Hortic. 2010, 877, 619–626.
80. Rojas-Argudo, C.; del Rio, M. A.; Pérez-Gago, M. B. Development and Optimization of Locust Bean Gum (LBG)-Based Edible Coatings for Postharvest Storage of ‘Fortune’
Mandarins. Postharvest. Biol. Technol. 2009, 52, 227–234.
www.iran-mavad.com
81. Dilek, M.; Polat, H.; Kezer, F.; Korcan, E. Application of Locust Bean Gum Edible Coating to Extend Shelf Life of Sausages and Garlic-Flavored Sausage. J. Food Process. Preserv.
2011, 35, 410–416.
82. Maqbool, M.; Ali, A.; Alderson, P. G.; Zahid, N.; Siddiqui, Y. Effect of a Novel Edible Composite Coating Based on Gum Arabic and Chitosan on Biochemical and Physiological
Responses of Banana Fruits during Cold Storage. J. Agric. Food Chem. 2011, 59, 5474–5482.
83. Baldwin, E. A.; Baker, R. A. Use of Protein in Edible Coatings for Whole and Minimally Processed Fruit and Vegetables. In Protein-Based Films and Coatings; Gennadios, A., Ed.;
CRC Press: Boca Raton, FL, 2002; pp 501–515.
84. Han, J. H.; Gennadios, A. Edible Films and Coatings: A Review. In Innovations in Food Packaging; Han, J. H., Ed.; Elsevier Academic Press: Amsterdam, 2005;
pp 239–262.
85. Di Pierro, P.; Sorrentino, A.; Mariniello, L.; Giosafatto, C. V. L. Chitosan/Whey Protein Film as Active Coating to Extend Ricotta Cheese Shelf-Life. LWT-Food Sci. Technol.
2011, 44, 2324–2327.
86. Baldwin, E. A.; Nisperos-Carriedo, M. O.; Baker, R. A. Use of Edible Coatings to Preserve Quality of Lightly (and Slightly) Processed Products. Crit. Rev. Food Sci. Nutr.
1995, 35 (6), 509–552.
87. Kester, J. J.; Fennema, O. Tempering Influence on Oxygen and Water Vapor Transmission through a Stearyl Alcohol Film. J. Am. Oil Chem. Soc. 1989, 66, 1154–1157.
88. McHugh, T. H.; Krochta, J. M. Milk-Protein-Based Edible Films and Coatings. Food Technol. 1994, 48, 97–103.
89. Pierro, P. D.; Chico, B.; Villalonga, R.; Mariniello, L.; Damiao, A. E.; Masi, P. Chitosan-Whey Protein Edible Films Produced in the Absence or Presence of Trans-
glutaminase: Analysis of Their Mechanical and Barrier Properties. Biomacromolecules 2006, 7, 744–749.
90. Atarés, L.; Bonilla, J.; Chiralt, A. Characterization of Sodium Caseinate-Based Edible Films Incorporated with Cinnamon or Ginger Essential Oils. J. Food Eng. 2010, 100,
678–687.
91. Sánchez-González, L.; González-Martínez, C.; Chiralt, A.; Cháfer, M. Physical and Antimicrobial Properties of Chitosan – Tea Tree Essential Oil Composite Films. J. Food
Eng. 2010, 98 (4), 443–452.
92. Cha, D. S.; Chinnan, M. S. Biopolymer-Based Antimicrobial Packaging: A Review. Crit. Rev. Food Sci. Nutr. 2004, 44, 223–237.
93. Jamshidian, M.; Arab Tehrany, E.; Cleymand, F.; Leconte, S.; Falher, T.; Desobry, S. Effects of Synthetic Phenolic Antioxidants on Physical, Structural, Mechanical and
Barrier Properties of Poly Lactic Acid Film. Carbohydrate Polym. 2011, 87 (2), 1763–1773.
94. Imran, M.; EL-Fahmy, S.; Revol-Junelles, A. M.; Desobry, S. Cellulose Derivative Based Active Coatings: Effects of Nisin and Plasticizer on Physico-Chemical and
Antimicrobial Properties of Hydroxypropyl Methylcellulose Films. Carbohydrate Polym. 2010, 81, 219–225.
95. Geraldine, R. M.; Ferreira Soares, N. F.; Botrel, D. A.; de Almeida Gonçalves, L. Characterization and Effect of Edible Coatings on Minimally Processed Garlic Quality.
Carbohydr. Polym. 2008, 72, 403–409.
96. Ouattara, B.; Simard, R.; Piette, G.; Begin, A.; Holley, R. A. Inhibition of Surface Spoilage Bacteria in Processed Meats by Application of Antimicrobial Films Prepared with
Chitosan. Int. J. Food Microbiol. 2000, 62, 139–148.
97. Valencia-Chamorro, S. A.; Palou, L.; del Rio, M. A.; Pérez-Gago, M. B. Antimicrobial Edible Films and Coatings for Fresh and Minimally Processed Fruits and Vegetables:
A Review. Crit. Rev. Food Sci. Nutr. 2011, 51, 872–900.
98. No, H. K.; Meyers, S. P.; Prinyawiwatkul, W.; Xu, Z. Applications of Chitosan for Improvement of Quality and Shelf Life of Foods: A Review. J. Food Sci. 2007, 72,
87–100.
99. Coma, V.; Deschamps, A.; Martial-Gros, A. Bioactive Packaging Materials from Edible Chitosan Polymer Antimicrobial Activity Assessment on Dairy-Related Contami-
nants. J. Food Sci. 2003, 68 (9), 2788–2792.
100. Devlieghere, F.; Vermeulen, A.; Debevere, J. Chitosan: Antimicrobial Activity, Interactions with Food Components and Applicability as a Coating on Fruit and Vegetables.
Food Microbiol. 2004, 21, 703–714.
101. Dutta, P.; Tripathi, S.; Mehrotra, G.; Dutta, J. Review: Perspectives for Chitosan Based Antimicrobial Films in Food Applications. Food Chem. 2009, 114, 1173–1182.
102. Kordali, S.; Kotan, R.; Mavi, A.; Cakir, A.; Ala, A.; Yildirim, A. Determination of the Chemical Composition and Antioxidant Activity of the Essential Oil of Artemisia
dracunculus and of the Antifungal and Antibacterial Activities of Turkish Artemisia absinthium, A-dracunculus, Artemisia santonicum, and Artemisia spicigera Essential
Oils. J. Agric. Food Chem. 2005, 53, 9452–9458.
103. Pezo, D.; Salafranca, J.; Nerin, C. Design of a Method for Generation of Gas-Phase Hydroxyl Radicals, and Use of HPLC with Fluorescence Detection to Assess the
Antioxidant Capacity of Natural Essential Oils. Anal. Bioanal. Chem. 2006, 385, 1241–1246.
104. Burt, S. Essential oils: Their Antibacterial Properties and Potential Applications in Foods – A Review. Int. J. Food Microbiol. 2004, 94, 223–253.
105. Yanishlieva, N. V.; Marinova, E. M.; Gordon, M. H.; Raneva, V. G. Antioxidant Activity and Mechanism of Action of Thymol and Carvacrol in Two Lipid Systems. Food
Chem. 1999, 64, 59–66.
106. Kanatt, S. R.; Chander, R.; Sharma, A. Chitosan and Mint Mixture: A New Preservative for Meat and Meat Products. Food Chem. 2008, 107, 845–852.
107. Salmieri, S.; Lacroix, M. Physicochemical Properties of Alginate/Polycaprolactone-Based Films Containing Essential Oils. J. Agric. Food Chem. 2006, 54, 10205–10214.
108. Vargas, M.; Albors, A.; Chiralt, A.; González-Martínez, C. Characterization of Chitosan-Oleic Acid Composite Films. Food Hydrocoll. 2009, 23, 536–547.
109. Lacroix, M. The Use of Essential Oils and Bacteriocins as Natural Antimicrobial and Antioxidant Compounds. Food 2007, 1, 181–192.
110. Delaquis, P. J.; Stanich, K.; Girard, B.; Mazza, G. Antimicrobial Activity of Individual and Mixed Fractions of Dill, Cilantro, Coriander and Eucalyptus Essential Oils. Int. J.
Food Microbiol. 2002, 74, 101–109.
111. Karatzas, A. K.; Bennik, M. H. J.; Smid, E. J.; Kets, E. P. W. Combined Action of S-Carvone and Mild Heat Treatment on Listeria monocytogenes Scott A. J. Appl.
Microbiol. 2000, 89, 296–301.
112. Lee, S. J.; Umano, K.; Shibamoto, T.; Lee, K. G. Identification of Volatile Components in Basil (Ocimum basilicum L.) and Thyme Leaves (Thymus vulgaris L.) and Their
Antioxidant Properties. Food Chem. 2005, 91, 131–137.
113. Sánchez-González, L.; Vargas, M.; González-Martínez, C.; Chiralt, A.; Cháfer, M. Use of Essential Oils in Bioactive Edible Coatings. J. Food Eng. Rev. 2011, 3, 1–16.
114. Moufida, S.; Marzouk, B. Biochemical Characterization of Blood Orange, Sweet Orange, Lemon, Bergamot and Bitter Orange. Phytochemistry 2003, 62, 1283–1289.
115. Fisher, K.; Phillips, C. A. The Effect of Lemon, Orange and Bergamot Essential Oils and Their Components on the Survival of Campylobacter jejuni, Escherichia coli O157,
Listeria monocytogenes, Bacillus cereus and Staphylococcus aureus In Vitro and in Food Systems. J. Appl. Microbiol. 2006, 101, 1232–1240.
116. Goñi, P.; López, P.; Sánchez, C.; Gómez-Lus, R.; Becerril, R.; Nerín, C. Antimicrobial Activity in the Vapour Phase of a Combination of Cinnamon and Clove Essential Oils.
Food Chem. 2009, 116, 982–989.
117. Mantle, D.; Anderton, J. G.; Falkous, G.; Barnes, M.; Jones, P.; Perry, E. K. Comparison of Methods for Determination of Total Antioxidant Status: Application to Analysis
of Medicinal Plant Essential Oils. Comp. Biochem. Physiol. B 1998, 121, 385–391.
118. Singh, G.; Maurya, S.; deLampasona, M. P.; Catalan, C. A. N. A Comparison of Chemical, Antioxidant and Antimicrobial Studies of Cinnamon Leaf and Bark Volatile Oils,
Oleoresins and Their Constituents. Food Chem. Toxicol. 2007, 45, 1650–1661.
119. Msaada, K.; Hosni, K.; Ben Taarit, M.; Chahed, T.; Kchouk, M. E.; Marzouk, B. Changes on Essential Oil Composition of Coriander (Coriandrum sativum L.) Fruits during
Three Stages of Maturity. Food Chem. 2007, 102, 1131–1134.
120. Wangensteen, H.; Samuelsen, A. B.; Malterud, K. E. Antioxidant Activity in Extracts from Coriander. Food Chem. 2004, 88, 293–297.
121. Wenqiang, G.; Shufen, L.; Ruixiang, Y.; Shaokun, T.; Can, Q. Comparison of Essential Oils of Clove Buds Extracted with Supercritical Carbon Dioxide and Other Three
Traditional Extraction Methods. Food Chem. 2007, 101, 1558–1564.
122. Gülçin, I.; Sat, I. G.; Beydemir, S.; Elmastas, M.; Küfrevioglu, Ò;I. Comparison of Antioxidant Activity of Clove (Eugenia caryophylata Thunb) Buds and Lavender (Lavandula
stoechas L.). Food Chem. 2004, 87, 393–400.
www.iran-mavad.com
123. Omidbeygi, M.; Barzegar, M.; Hamidi, Z.; Naghdibadi, H. Antifungal Activity of Thyme, Summer Savory and Clove Essential Oils against Aspergillus Flavus in Liquid
Medium and Tomato Paste. Food Control 2007, 18, 1518–1523.
124. Amakura, Y.; Umino, Y.; Tsuji, S.; Ito, H.; Hatano, T.; Yoshida, T. Constituents and Their Antioxidative Effects in Reply to: Eucalyptus Leaf Extract Used as a Natural Food
Additive. Food Chem. 2002, 77, 47–56.
125. Oyedeji, A. O.; Ekundayo, O.; Olawore, O. N.; Adeniyi, B. A.; Koenig, W. A. Antimicrobial Activity of the Essential Oils of Five Eucalyptus Species Growing in Nigeria.
Fitoterapia 1999, 70 (5), 526–528.
126. Sacchetti, G.; Maietti, S.; Muzzoli, M.; Scaglianti, M.; Manfredini, S.; Radice, M.; Bruni, R. Comparative Evaluation of 11 Essential Oils of Different Origin as Functional
Antioxidants, Antiradicals and Antimicrobials in Foods. Food Chem. 2005, 91, 621–632.
127. Viuda-Martos, M.; Ruiz-Navajas, Y.; Fernández-López, J.; Pérez-Álvarez, J. Antifungal Activity of Lemon (Citrus lemon L.), Mandarin (Citrus reticulata L.), Grapefruit
(Citrus paradisi L.) and Orange (Citrus sinensis L.) Essential Oils. Food Control 2008, 19, 1130–1138.
128. Kulisic, T.; Radonic, A.; Katalinic, V.; Milos, M. Use of Different Methods for Testing Antioxidative Activity of Oregano Essential Oil. Food Chem. 2004, 85 (4),
633–644.
129. Souza, E. L.; Stamford, T. L. M.; Lima, E. O.; Trajano, V. N. Effectiveness of Origanum vulgare L. Essential Oil to Inhibit the Growth of Food Spoiling Yeasts. Food Control
2007, 18 (5), 409–413.
130. Zivanovic, S.; Chi, S.; Draughon, F. Antimicrobial Activity of Chitosan Films Enriched with Essential Oils. J. Food Sci. 2005, 70, 45–51.
131. Gachkar, L.; Yadegari, D.; Rezaei, M. B.; Taghizadeh, M.; Astaneh, S. A.; Rasooli, I. Chemical and Biological Characteristics of Cuminum cyminum and Rosmarinus
officinalis Essential Oils. Food Chem. 2007, 102, 898–904.
132. Moore, M. E.; Han, I. Y.; Acton, J. C.; Ogale, A. A.; Barmore, C. R.; Dawson, P. L. Effects of Antioxidants in Polyethylene Film on Fresh Beef Color. J. Food Sci. 2003,
68 (1), 99–104.
133. Oussalah, M.; Caillet, S.; Salmiéri, S.; Saucier, L.; Lacroix, M. Antimicrobial and Antioxidant Effects of Milk Protein-Based Film Containing Essential Oils for the
Preservation of Whole Beef Muscle. J. Agric. Food Chem. 2004, 52 (18), 5598–5605.
134. Tovar, L.; Salafranca, J.; Sánchez, C.; Nerín, C. Migration Studies to Assess the Safety in Use of a New Antioxidant Active Packaging. J. Agric. Food Chem. 2005,
53 (13), 5270–5275.
135. Wessling, C.; Nielsen, T.; Leufven, A.; Jagerstad, M. Retention of Alpha-Tocopherol in Low-Density Polyethylene (LDPE) and Polypropylene (PP) in Contact with Foodstuffs
and Food-Simulating Liquids. J. Sci. Food Agric. 1999, 79 (12), 1635–1641.
136. Lee, C. H.; An, D. S.; Lee, S. C.; Park, H. J.; Lee, D. S. A Coating for Use as an Antimicrobial and Antioxidative Packaging Material Incorporating Nisin and Alpha-
Tocopherol. J. Food Eng. 2004, 62, 323–329.
137. Ayranci, E.; Tunc, S. A Method for the Measurement of the Oxygen Permeability and the Development of Edible Films to Reduce the Rate of Oxidative Reactions in Fresh
Foods. Food Chem. 2003, 80 (3), 423–431.
138. Zheng, Z. L.; Tan, J. Y. W.; Liu, H. Y.; Zhou, X. H.; Xiang, X.; Wang, K. Y. Evaluation of Oregano Essential Oil (Origanum heracleoticum L.) on Growth, Antioxidant Effect
and Resistance against Aeromonas hydrophila in Channel Catfish (Ictalurus punctatus). Aquaculture 2009, 292 (3–4), 214–218.
139. Rice-Evans, C. A.; Miller, N. J.; Paganga, G. Structure-Antioxidant Activity Relationships of Flavonoids and Phenolic Acids. Free Radic. Biol. Med. 1996, 20, 933–956.
140. Montero, P.; Giménez, B.; Pérez-Mateos, M.; Gómez-Guillén, M. C. Oxidation Stability of Muscle with Quercetin and Rosemary during Thermal and High-Pressure
Gelation. Food Chem. 2005, 93, 17–23.
141. Lee, J. Y.; Park, H. J.; Lee, C. Y.; Choi, W. Y. Extending Shelf-Life of Minimally Processed Apples with Edible Coatings and Antibrowning Agents. LWT-Food Sci. Technol.
2003, 36, 323–329.
142. Baldwin, E. A.; Wood, B. Use of Edible Coating to Preserve Pecans at Room Temperature. HortScience 2006, 41 (1), 188–192.
143. Atares, L.; Perez-Masia, P.; Chiralt, A. The Role of Some Antioxidants in the HPMC Film Properties and Lipid Protection in Coated Toasted Almonds. J. Food Eng. 2011,
104, 649–656.
144. Brandy-Williams, W.; Cuvelier, M. E.; Berset, C. Use of a Free Radical Method to Evaluate Antioxidant Activity. Lebenson Wiss Technol. 1995, 28, 25–30.
145. Hiemenz, P. C.; Rajagopalan, R. Surface Tension and Contact Angle, 3rd ed.; Marcel Dekker: New York, 1997; pp 248–255.
146. Ribeiro, C.; Vicente, A. A.; Teixeira, J. A.; Miranda, C. Optimization of Edible Coating Composition to Retard Strawberry Fruit Senescence. Postharvest Biol. Technol.
2007, 44 (1), 63–70.
147. Rodríguez, M.; Osés, J.; Ziani, K.; Maté, J. I. Combined Effect of Plasticizers and Surfactants on the Physical Properties of Starch Based Edible Films. Food Res. Int.
2006, 39 (8), 840–846.
148. Amarante, C.; Banks, N. H. Ripening Behavior, Postharvest Quality, and Physiological Disorders of Coated Pears (Pyrus communis). N. Z. J. Crop Hortic. Sci. 2001, 30,
49–59.
149. ASTM Standard Test Methods for Water Vapor Transmission of Materials. Standard Designations: E96–95. In Annual Book of ASTM; ASTM: Philadelphia, 1995;
pp 406–413.
150. Min, S.; Krochta, J. M. Ascorbic Acid-Containing Whey Protein Film Coatings for Control of Oxidation. J. Agric. Food Chem. 2007, 55, 2964–2969.
151. Herald, T. J.; Hackmeister, K. A.; Huang, S.; Browers, J. R. Corn Zein Packaging Materials for Cooked Turkey. J. Food Sci. 1996, 61, 415–418.
152. Kleen, D.; Padua, G.; Engeseth, N. Stabilization of Lipids in a Biodegradable Zein-Oleate Film by Incorporation of Antioxidants. Cereal Chem. 2002, 79, 687–694.
153. Chen, S.; Nussinovitch, A. Permeability and Roughness Determinations of Wax-Hydrocolloid Coatings, and Their Limitations in Determining Citrus Fruit Overall Quality.
Food Hydrocoll. 2001, 15, 127–137.
154. Kluge, R. A.; Nachtigal, J. C.; Fachinello, J. C.; Bilhalva, A. B. Fisiologia e manejo pós-colheita de frutas de clima temperado; Livraria e Editora Rural: Campinas, 2002;
p 214.
155. Vargas, M.; Pastor, C.; Chiralt, A.; McClements, J.; González-Martinez, C. Recent Advances in Edible Coatings for Fresh and Minimally Processed Fruits. Crit. Rev. Food
Sci. Nutr. 2008, 48, 496–511.
156. Garcia, M. A.; Pinotti, A.; Martino, M. N.; Zaritzky, N. E. Characterization of Starch and Composite Edible Films and Coatings. In Edible Films and Coatings for Food
Applications; Embuscado, M. E., Huber, K. C., Eds.; Springer Science: France, 2009; pp 169–209.
157. Krochta, J. M. Proteins as Raw Materials for Films and Coatings: Definitions, Current Status, and Opportunities. InProtein-Based Films and Coatings; Protein-Based Films
and Coatings; CRC Press: Boca Raton, FL, 2002; pp 1–41. (Chapter 1).
158. Bourtoom, T.; Chinnan, M. S. Improvement of Water Barrier Property of Rice Starch-Chitosan Composite Film Incorporated with Lipids. Food Sci. Technol. 2009, 15,
149–158.
159. Campos, C. A.; Gerschenson, L. N.; Flores, S. K. Development of Edible Films and Coatings with Antimicrobial Activity. Food Bioprocess Technol. 2011, 4, 849–875.
160. Cagri, A.; Ustunol, Z.; Ryser, E. T. Antimicrobial, Mechanical, and Moisture Barrier Properties of Low pH Whey Protein-Based Edible Films Containing p-Amino Benzoic or
Sorbic Acids. J. Food Sci. 2001, 66 (6), 865–870.
www.iran-mavad.com
4.21 Large-Scale Industrial Coating Applications and Systems
R Tietema, Hauzer Techno Coating BV, Venlo, The Netherlands
4.21.1 Different Equipment Aspects Classified after Their Application 519

4.21.1.1 Market Situation of Hard Coatings 520
4.21.1.2 Coatings on Tools 520
4.21.1.3 Tribological (Component) Coatings 521
4.21.1.4 Decorative Coatings 525
4.21.2 Equipment Aspects Based on Applications 527
4.21.2.1 Batch Systems 527
4.21.2.2 Special Systems 527
4.21.2.2.1 Special Batch Systems 527
4.21.2.2.2 Special Systems for Research 528
4.21.2.2.3 In-Line Systems 529
4.21.3 Design Phases 530
4.21.3.1 Substrate Rotation 531
4.21.3.2 Etching Steps 533
4.21.3.3 Sputter Deposition: Sputter Cathode and Target 533
4.21.3.3.1 Target Cooling 534
4.21.3.3.2 Sputter Cathode and Target: Magnetic Field 534
4.21.3.3.3 Sputter Cathode and Target: Unbalanced Magnetron Cathode Coils and Closed Field Configuration 535
4.21.3.3.4 Sputter Cathode and Target: Magnetic Field and Coating Uniformity 536
4.21.3.3.5 Sputter Cathode and Target: Nonreactive and Reactive Sputtering 536
4.21.3.3.6 Cathode Power Supply in Reactive and Nonreactive Sputterings 539
4.21.3.4 Coating Architectures of Carbon-Based Coatings 542
4.21.3.4.1 a-C:H:W Coating 542
4.21.3.4.2 a-C:H Coating Architecture 544
4.21.3.4.3 Ta-C Coatings 545
4.21.3.5 Automation Requirements 546
4.21.4 Maintenance of Batch and In-Line Systems 548
4.21.4.1 Target Lifetime and Cost Considerations 548
4.21.4.2 Dust Formation 550
4.21.4.3 Different Types of Shield Functionalities 550
4.21.4.4 Maintenance per Batch 552
4.21.4.5 Target Erosion of Arc Cathodes 554
4.21.4.6 Target Exchange 556
4.21.5 Project Management Organization 557
Acknowledgment 560
References 560
4.21.1 Different Equipment Aspects Classified after Their Application
The role of large-volume production of wear-resistant coatings and the different aspects related to the processes and systems to
produce these coatings are discussed. Coatings for decorative applications are addressed, as well as coatings for tools, but the main
emphasis is given to tribological coating systems. The differences in cycle time and process temperatures account for principal
differences between these types of systems.
First the market position of hard coatings is discussed. Tool coatings were the first hard coatings introduced. The historical
perspective and general aspects for coatings on cutting tools and forming tools are addressed, followed by a similar discussion of
component coatings. For the component coatings, some application examples are discussed. Initially only wear-protection
applications existed, which were closely related to the introduction of high pressure fuel injection for diesel engines. In the last years,
there is a huge emphasis on friction reduction in view of CO2 emissions reduction.
Decorative coatings are discussed, since here the large-scale production played a decisive role in the introduction of physical
vapor deposition (PVD) technology in this market segment. Today this market has mainly been lost to Asian manufacturers.

www.iran-mavad.com
520 Large-Scale Industrial Coating Applications and Systems
4.21.1.1 Market Situation of Hard Coatings

Today hard coatings are playing an important role in the wear protection for cutting and forming tools and on automotive
components. Functional coatings (for example, switchable coatings, glass coatings, and optical coatings) and decorative coatings
also need wear protection, even though mechanical wear plays here a role in a somewhat different manner. The estimated market
volume worldwide for hard coatings in 2011 was about US$2 billion.
The word ‘tribology’ originates from the Greek word for friction. The definition of tribology is ‘the science that describes the wear
of parts moving along each other.’ Today’s general use is that the description of tribological coatings refers to component coatings,
whereas strictly spoken also tool coatings fall under this definition. Here the expression ‘tribological coatings’ refers to coatings on
components if no other reference is used.
The losses caused by friction and tribological wear are adding up to extremely high figures. Interpolation of data from 1985 (1)
shows that in 2000 economic losses in Germany are adding up to approximately 2–3% of the gross national product. This equates
to between V30 billion and V40 billion per year. These costs for losses are caused by friction losses (Figure 1) by machine stops and
repair and production losses and losses due to an insufficient product quality (Figure 2). In the category of insufficient product
quality are damages included that are not detected directly after production and delivery, but which are occurring before the normal
end of life after longer use of the product.
4.21.1.2 Coatings on Tools

If we simplify the tool coating processes, we might note that they are characterized by a long duration (in the range of a few hours)
and high temperatures (between 450 and 650 C). The long duration is related to the relatively high thickness and the high
temperature is related to the substrate material (often tungsten carbide, WC). The high substrate temperature allows a high
deposition rate and is mostly favorable for the coating properties as well.
For cutting tools, the majority of the market is dominated by arc coatings, whereas a smaller part of this market is covered by
sputter coatings. In the 1970s in East Germany, there were also applications running for TiN and TiAlN with hollow cathode
evaporation (2,3). In the beginning of the 1980s, when PVD was introduced in the western market, TiN dominated the market.
Systems producing TiN coatings were produced and operated by Balzers, using evaporation. Multiarc introduced coatings based
on Russian licensed technology (from KIPT). The market for tool coatings started to grow rapidly, as the end users became aware of
the benefits in tool lifetime (4).
TiAlN coatings were introduced at the end of the 1980s, initially as a sputtering process, later also as an arc process. The benefits
of this coating were discovered by Cemecon, as a highly improved cutting tool performance was realized because Al brackets were
cosputtered during a TiN process and was further deeper investigated by Leyendecker (5), who found the face-cubic centered phase
Ti2AlN (6).
The substitution of Ti for Al in the structure leads to a temperature increase of the oxidation onset and reduces the oxidation rate
of the coating while retaining high hardness because of the binary TiN. From then on (in the early 1990s) also in arc technology the
focus was on the development of TiAlN, ultimately leading to the strong position that these coatings are now taking in the cutting
tool coating market.
Figure 1 Estimated costs of friction losses for Germany in 2000.
www.iran-mavad.com
Large-Scale Industrial Coating Applications and Systems 521
Figure 2 Estimated costs of losses due to wear (repair and maintenance) for Germany in 2000.
Approximately 50% of the tool coating market is taken by job coaters, but particularly the major tool manufacturers now possess
in-house operations. Sometimes end users have set up factory-gate operations with a job coating partner. The reason is that the
development of the coating for a cutting tool cannot be separated from other tool design aspects, such as the shape and dimensions
of the cutting edge, the substrate material properties of the tool, and the properties of the workpiece material to be cut with that
specific tool.
The market for forming tools is mainly developed by job coaters. Since the number of molds to be coated is limited, it is often
not economic for an end user to coat molds in house.
Both cutting and forming tools require special pre- and posttreatments, which many job coaters have specialized in as well. Many
tool job coaters have specialized regrinding centers, where worn tools are reconditioned by removal of the coating, regrinding of the
tool in the desired geometry, coating, and postpolishing.
The market for tool coatings is still taking the biggest share of the hard coating market today.
4.21.1.3 Tribological (Component) Coatings

Work on diamond-like carbon (DLC) coatings started quite early, in the 1970s. Weissmantel et al. (7) analyzed initial DLC coatings
produced by plasma-enhanced chemical vapor deposition (PECVD; also known as plasma-assisted CVD, or PACVD) and showed
their amorphous character. Initial applications were introduced by IBM on hard disks in 1985 (8), but remained until the intro-
duction on diesel fuel injection systems in the mid 1990s the only large-volume application. In the late 1990s DLC-coated razor
blades were the next high-volume application. Currently DLCs are tested or used for many other automotive components and other
applications like micro-electro-mechanical systems (MEMS), knee and hip joints, machine tools, bearings, and dies (9–12).
In the mid 1990s, the first applications for automotive coatings were introduced in the market. In general for most of the
component coating processes, we may note that they are characterized by a long duration (around 8 h) and low temperatures
(between 160 and 200 C). The long duration is related to the relatively low deposition rate needed to keep the substrate material
below the tempering temperature of the substrate material, which is mostly alloyed steel with a low hardening temperature.
This temperature range is also beneficial for the process conditions for most carbon-based DLC coatings (13).
The introduction of automotive coatings was initiated with the development of the high-pressure (HP) fuel injection pump. The
first company that introduced this technology on the market on a large scale was Robert Bosch in Stuttgart/Germany (14,15). The
introduction of this technology marked the beginning of the big breakthrough of diesel engines in passenger cars on the European
market, because diesel cars became very powerful and silent, but also extremely economic. Injection pressures were initially around
1000 bar for the radial-piston distributor pump (1996), but nowadays are greater than 2200 bar for the common rail system
(initially developed by Fiat, but taken to large-scale production by Bosch in 1997), and the unit injector system (1998). Eventually,
the common rail system became the dominating system since the consistently high pressure in the rails allows for multiple
injections and therefore an improved fuel economy. Besides, the common rail system is also a solution with lower cost.
The HP injection technology also contributed in this way significantly to the reduction of CO2 emissions. Due to the extreme
high pressures, tolerances were needed within micrometer range to prevent leakage of injection valve parts. Besides this, it is
www.iran-mavad.com
apparent that high wear stress will occur if an injection needle has small deviations in symmetry, since extreme forces will result
from these high pressures, even if the pressurized areas are relatively small. Consequently, it is not possible to use normal steel parts
anymore since they will simply wear instantaneously. The introduction of coatings solved this problem. In fact, coating technology
was an enabling technology for HP fuel injection.
Initially, the applied coatings were the a-C:H:Me coatings (metal-DLC coatings, as defined in (16)). The coating a-C:H:Ti was
used on smaller scale in the United States, but for the HP fuel injection produced by the European automotive industry, the initially
applied coating was a-C:H:W. In the early 1990s this coating was developed and patented by Dimigen at Philips in Hamburg (later
this research institute became the Fraunhofer Institute for Surface Technology) (17). This patent described the invention of a fric-
tion-reducing coating layer existing from metal and carbon, essentially not being a metal carbide (the carbon content is much higher
than the stoichiometric concentration of the corresponding metal carbide). This amorphous coating showed excellent friction-
reducing behavior and excellent wear reduction against steel counterparts, independent of atmospheric conditions. This was the first
coating used for fuel injection parts. At pressures near 1000 bar (100 MPa), this was a feasible solution, but as soon as the applied
injection pressures increased, the wear resistance of a-C:H:W was inadequate to protect the injection components against excessive
wear. At this point, the introduction of a-C:H coatings (hydrogenated DLC) became necessary. Due to its higher hardness and by
designing the coating with the required ductility, it was possible to overcome the wear problems. From this point on these a-C:H
coatings were the majority of coating types applied for HP fuel injection.
With the introduction of HP injection technology, the load on other engine components increased also, due to higher power
densities and torque. This implied that wear on these components (18) increased as well, which opened new markets for coating
applications. Components under high stress are piston rings, piston pins, tappets/shims, journal bearings, roller and ball bearings,
valves, and gears. Some of these components, in the period before HP injection was introduced, were coated with non-PVD coatings
(often galvanic coatings). As an example, piston rings can be mentioned. Piston rings were previously coated with hard chromium.
It is easy to make thick coatings with galvanic technology, but the coating result is a relatively soft coating with Vickers hardness in
the range of about HV 600 (6 GPa). As long as pressures were not too high, wear was not an issue. This changed, however, as soon as
pressures increased. The higher pressure in the cylinder causes high forces pushing on the running side of the piston ring, which is
consequently pushed to slide against the liner (19). Wear became unacceptably high, which resulted in leakage, especially from the
top ring (the first compression ring). As an initial solution, thick CrN coatings were used. CrN was selected because chemically it is
a material resembling the hard chrome galvanic coating, but it can be made harder and as such more wear resistant. Initially,
designers were convinced that the coating should not be extremely hard since that will promote wear from the cast iron liner as
counterpart in the tribological system. These CrN piston ring coatings were developed to have a high porosity, which helps reduce
wear since it improves lubrication by having oil reservoirs in the pores of the coating. By embedded oil inside the pores of the
coating, the relatively high coefficient of friction of CrN is not an issue anymore and wear is now under control. Wear is, however,
still happening, but by the thickness of the coating (ranging between 15 and 40 mm) the wear is largely compensated to ensure
a long operational life of both piston ring as well as liner (20). Nowadays, other solutions are investigated encompassing harder
coatings that can then also be thinner. Such a case requires, however, a low friction and a smooth surface (to prevent wear of the
relatively soft liner).
Since the losses in the cylinder liner are one of the major contributors to the engine losses, only 10% reduction of the losses here
could already improve the fuel efficiency substantially with 2.5% (21–23). For this reason, a lower friction coating can help improve
CO2 emission values of the engine. Today’s researchers are also looking for other solutions to improve wear. Carbon-based coatings
are consequently in the picture as new-generation coatings to follow the existing CrN coatings.
Another similar example we find is in journal bearings (Figure 3). AlSn coatings are solutions for wear of journal bearings.
Journal bearings are operating against crank shafts in a gear box, where wear particles are always encountered in the oil. Sputtered
AlSn PVD coatings for this application are relatively soft (70–110 HV), but this is about three times harder than casted AlSn, but also
Figure 3 Journal bearing structure: Steel back for rigidity; the intermediate layer is CuPSn (leaded bronze) or CuSnNi and has high fatigue strength.
The top layer has a low friction and high embed ability. Reproduced from Leyendecker, T. Über neuartige Schneidwerkzeugbeschichtungen auf Titan und
Aluminiumbasis. PhD Thesis, RWTH Aachen, 1985.
www.iran-mavad.com
harder than a galvanic coating (Sn6Cu is used as an alternative for lead). The target material is AlSn. They require a barrier layer for
lead diffusion or tin diffusion, which is mainly formed by Ni-containing base layers in the coating with a thickness of 1–2 mm.
Subsequently the soft metallic AlSn coating is deposited. The coatings have thicknesses in a range of 10–20 mm. Hard particles
floating in the oil can cause third particle wear (fatigue wear) when they are crushed between the moving parts of the tribological
system. Due to the thick, and compared with these particles relatively soft, coating, these particles will be embedded in the coating
without causing damage to both the crank shaft and the journal bearing. Sputtered bearings are mainly used as connecting rod
bearings in highly loaded diesel engines with direct fuel injection system.
The coatings on piston rings and journal bearings are nowadays exceptions in the world of component coatings, since most
automotive components needing wear protection nowadays have carbon-based coatings.
The wear protection by these coatings ensures that the increasingly tighter tolerances required stay in compliance with the
component specification during the lifetime of the car. Coatings also help reduce maintenance intervals, as replacement of
components can largely be reduced because of the efficient wear protection by the coatings.
Due to the size of the automotive market, each part that will be coated quite rapidly is part of a high-volume application since for
a certain type of component the number of parts to be coated often rises rapidly into the millions. This is proven by the market size
for coated automotive components. In 2011, the market of automotive components to be coated by PVD/PECVD worldwide was
estimated to add up to 800 million parts (24). An example of a high-volume system for components is shown in Figure 4.
Besides wear protection, a new functional requirement has become extremely important. On 11 December 1997, the conference
of the United Nations Framework Convention on Climate Change in Kyoto was held, and the Kyoto protocol was signed, which
entered into force on 16 February 2005 (25). Agreements were made to reduce the emissions of CO2 substantially to reduce global
heating. The commitments made in Kyoto led the European Commission to introduce measures for the reduction of CO2 emissions
in Europe. An emission trading scheme was introduced in the EU. The contribution of the automotive industry to CO2 reduction is
substantial. In Europe in 1998 voluntary agreements on CO2 reduction were already negotiated between the European Commission
and the European, Japanese, and Korean manufacturer associations (respectively ACEA, JAMA, KAMA), representing approximately
90% of all cars being driven in Europe. These voluntary agreements were initially followed quite well, but in 2004, the rate of
reduction changed since the progress of the technological development was mostly put in increasing the engine power of the cars,
making cars heavier and increasing therefore the fuel consumption, than in following the voluntary agreements (Figure 5). The
misfit between the voluntary agreement and the actual situation was the sign for the European Commission to define regulations
and use penalties for noncomplying manufacturers. The regulations encompassed a level of 120 g CO2 per km in 2012. The
emission is to be measured in the standard New European Driving Cycle (NEDC) test cycle that has been defined where 130 g CO2
per km has to be reached. The remaining 10 g CO2 per km is supposed to be reached by additional measures outside the tested car,
from a higher efficiency of a climate control unit, which is not running during the test cycle, up to mixing traditional fuel with
biofuels (23,26).
Exceeding the required emissions limit is penalized with a fine, called the excess premium. This is calculated according to defined
formulas. Per manufactured car, the excess premium to be paid is calculated as an average of the whole fleet of M1 type cars
(passenger cars for maximum of eight passengers) (27). The premiums to be paid are adding up so fast to such high levels that they
are in fact forcing the manufacturers to take measures for emission reductions in order to stay competitive. For this reason, many car
Figure 4 Large-volume system for components (Hauzer Techno Coating).
www.iran-mavad.com
220
200
180
g CO2 km
160
140 Average EU
ACEA actual
120
ACEA commitment
JAMA
100
KAMA
80
1994 1996 1998 2000 2002 2004 2006 2008 2010
Year
Figure 5 Progress of automotive CO2 emissions reduction after 2003 behind voluntary agreements.
manufacturers are designing all kinds of measures, like improved aerodynamics, reduction of weight, and reduction of energy
losses.
Increasing engine efficiency will lead to downsizing of the engine. As a consequence, the power densities are increasing and the
dimensions and therefore the weight of components will be reduced. The increasing power densities will lead to a higher
mechanical load and therefore an increase in wear.
Additional design improvements on other components will be implemented as well for the reduction of losses. Between possible
measures to reduce losses, carbon-based coatings are playing an important role due to their inherent low coefficient of friction. The
application of lubrication oil with low viscosity is also a method of high importance for the reduction of friction. In that case, wear
protection gets a higher necessity, because the load carrying capacity of the thinner oil film enforces a broader operating range of the
tribosystem in boundary and mixed-mode lubrication regimes (see Figure 6). There is an increased contact because of asperities.
The tribosystem requires increased wear prevention, which can be ensured by the coating.
Besides the reduced viscosity, chemical interactions between coating and lubricant are important for the friction of the tribo-
system. Chemical reactions can take place, thus reducing dangling bonds where the two parts in movement can react and give
a sticking effect. Also the creation of a low friction coefficient can take place by the creation of a low friction transfer film, where the
low coefficient of friction is reached by the catalyst effect of the counterpart material with the coating (28).
Figure 6 Stribeck curve with the different lubrication modes. Source: Fuchs Europe Schmierstoffe GmbH, Mannheim, Germany.
www.iran-mavad.com
With respect to reduced friction, ta-C coatings play an important role. Ta-C coatings are nonhydrogenated carbon-based DLC
films. These films are generally much harder and can also be applied at higher temperatures than hydrogenated DC films (a-C:H
films), but there is a drawback to these films. Since ta-C coatings will be produced by arc technology, they inherently have a high
coating roughness. Besides ionized particles, the arc spot generates also macroparticles. These macroparticles can be small, but are
mostly larger particles that can be several mmeters in size (29,30). Therefore, as asperities they add up to the roughness of the coating.
Asperities cause a mountainous profile of the coating. These roughness peaks of the mountainous profile are not covered by the oil
film in boundary lubrication and mixed lubrication mode. Because the counterpart of the tribosystem is uncoated, the consequence
of a very hard and rough coated part of the tribosystem will be that this part acts as a file and will wear out the counterpart as well as
other contact areas. There is an additional disadvantageous effect. If these hard particles are breaking out, they will act as grinding
slurry/sandpaper and wear out the counterpart as well. Therefore, it is imperative that ta-C coatings either must be produced by
a technology preventing macroparticle generation (which can be filtered arc) or that a solution must be found in polishing the
coating to remove the asperities. Both solutions require additional investments and as such they are increasing the cost. Contrary to
this, a-C:H coatings are very smooth and do not require polishing as a posttreatment.
Kano et al. (31) discovered, however, one very interesting aspect concerning friction in relation to ta-C coatings. Kano investigated
contacts from steel or coated parts against a steel counterpart. He found in the case of standard mineral oil a small difference in
friction coefficient in favor of the ta-C coating compared to a-C:H, but by applying poly alpha olefin fluid (PAO) with glycerol mono-
oleate (GMO) additive a dramatic reduction in friction coefficient was observed. The friction coefficient of the ta-C/steel system was
in this case reduced to 75% compared to a-C:H/steel. The physical effect causing the spectacular friction reduction is thought to be
the formation of an OH-terminated surface, generated by the chemical reaction between alcohol functional group and the friction-
activated ta-C atoms. The very low energy interaction between OH-terminated surfaces causes the superlubricity. The Japanese car
manufacturer Nissan has patented the application of hydrogen-free DLC coating (ta-C) for the camshaft-tappet tribosystem. Nissan
has reported that in practice they reached a reduction of 50% with special oil (32). The full potential of 75% friction reduction is in
practice not achievable because of the requirement for different additives for the functioning of some other engine parts. These
findings for ta-C coatings form an important base for the search to solutions for reduction of friction in the automotive systems.
Besides applications in the automotive industry, many carbon-based coatings are used in components in the machining industry
and for textile machines, forming machines, and others as well.
4.21.1.4 Decorative Coatings

Most of the decorative coating processes are characterized by a short duration (about 1 h) and low temperatures (between 70 and
150 C). The short duration is related to the low coating thickness generally needed. The maximum temperatures are related to the
substrate material (for brass products the maximum is 150 C; for plastics, mostly applied acrylonitrile butadiene styrene (ABS)
plastics, it is 70 C).
PVD coating was applied in large volume by HochVakuum Dresden (HVD) in the 1970s. Applications were watch straps,
jewelry, armatures for lamps, door handles, and so. These systems used hollow cathode evaporation. Ar ions are generated in the
hollow cathode discharge (HCD) source and together with an electron beam bombarding a tile of material to be evaporated by the
electrons, whereas the ions are drawn to the substrates by the negative bias voltage. These processes could also be applied for low-
temperature deposition. An example was TiN coatings on candles.
PVD coating was used in large volume in the 1980s on watches and watchstraps. Timex was one of the leading companies in this
field, using box coaters produced by Vactec. These coaters used arc and sputtering in one process. A gold-colored ZrN or TiN arc
coating with a sputtered gold top finish was used. Because the base layer was very thin, arc technology was useable. As described
earlier, arc technology generates macroparticles, which are in case of metal targets generated by bursts of molten target material.
Similar to carbon arc coating, here the generated macroparticles are also large compared to the coating and result in a high
roughness of the coating. A high roughness will ultimately lead to a dull appearance of the coating, where the coated product should
be shiny. However, by keeping the thickness of the coating low (within a range approximately 0.3 mm) and by adequate control of
the arc spot, the roughness can be limited. In this way, negative effects of droplets on the appearance could be avoided.
The US companies Moen and Delta introduced initially the mass production of decorative coatings on faucets. Moen and Hauzer set
up the first high-volume systems in 1995, whereas Vaportech in Boulder, Colorado, set up similar sized systems for Delta. In all cases, it
was necessary to introduce high-volume systems in order to produce products economically. Hauzer introduced their HTC1500 system
(Figure 7), with a design that was basically an upscale version of the smaller-scaled HTC1000 machine. The system is equipped with
a Polycold water-vapor cryo pump and infrared heaters for fast degassing and pumpdown, which is important to reduce the residual
gas pressure during deposition. This is very important for the prevention of discoloration, especially in case of ABS plastic substrates.
The background of the design steps to be taken is treated later. Comparable chamber sizes were introduced by Vapor Tech-
nologies for Delta relatively soon after Hauzer introduced the first decorative system at Moen. Initially, the faucets received
a galvanic base coating (electroplating multilayer: Cu–Ni–Cr) and a brass-colored top coating (Figure 8). The galvanic coating has
the function of corrosion protection and leveling. Since the substrate material is generally very soft compared to the coating, the
electroplating provides additionally sufficient load carrying support for the coating. Leveling is needed, because the PVD coating will
always follow the starting profile. Before the introduction of PVD, the faucets were coated with a lacquer as a protective top coating
to protect the galvanic top coating against mechanical damage and corrosion caused by perspiration and cleaning agents. Since the
lacquer was relatively soft, scratches damaged the lacquer and the galvanic base coating. This was drastically reducing the protection
www.iran-mavad.com
Figure 7 (a) Machine equipped with four unbalanced magnetrons. The unbalancing coils are located under safety covers for the cathodes and are not
visible. (b) Schematic top view of the system shown in Figure 7(a).
Figure 8 Structure of electroplated layer.
against corrosion, thus negatively affecting the appearance of the faucet. The replacement of the lacquer by a PVD coating was a huge
step forward in quality. The PVD coating does not provide corrosion protection, since a PVD coating in general has pores due to dust
generated during the process in the equipment. Also due to droplets, an arc coating will have pores. Therefore, the galvanic base
coating will still be needed, but by adding the very hard and therefore scratch- and wear-resistant PVD top coating, the corrosion
protection of the galvanic base coating is guaranteed. PVD coating also enabled producing different coatings/colors by changing the
process conditions. The PVD colors are giving the faucets an exclusive shiny metal look.
Later, ABS plastics were applied as base material for faucets. This made the faucets cheaper, and by applying PVD they still kept
a metallic appearance. In the beginning, there were severe problems with coating ABS plastics. ABS plastics contain large quantities
of water vapor, since it has a porous structure with large cavities and relatively small gaps for the water vapor to escape. This resulted
in discoloration of the products, caused by oxidation. Therefore, this substrate material required special measures for pumping and
degassing to prevent discoloration caused by oxidation due to the trapped water vapor escaping during the process. The tendency for
further cost reduction kept driving the market. By producing even larger systems (loading zone dimensions: diameter
1670 mm height 1500 mm), it was possible to reduce the coating cost even further (Figure 9). Besides all big faucet manufac-
turers, job coaters started with decorative PVD coating and a new large market was created for coating door handles. After the start of
this century, the market for high-volume coating of faucets and door handles was mostly taken over by Asian producers.
www.iran-mavad.com
Figure 9 Dr J. Vyskocil (HVM) with loaded tables of a large-volume system for decorative PVD coating of faucets.
Other large-volume applications for decorative coatings appeared a few years later with the large-scale introduction of PVD
coatings on mobile telephones. Besides the color, the coatings needed to protect the substrate against indentation because the
substrate was relatively thin and soft steel without an intermediate electroplating as support for the PVD coating. Coating cost
played a major role here also. The coating should protect the thin substrate against scratches as well as indents, which requires the
coatings being much thicker than the earlier mentioned few tenths of a micrometer in the case of arc-coated faucets. The thicknesses
should be several micrometers. This implied automatically that arc technology would not be feasible anymore since the roughness
would cause a dull appearance. Another related issue is that compared to the state-of-the-art decorative coating (1 h processes) the
deposition time increased enormously and reached approximately 8 h. The deposition rate has to be maximized for cost reduction,
which implies that consequently the temperature of the substrate will also be much higher than in the short-cycle faucet coating
processes. For stainless steel substrates, the temperature can be over 400 C, however.
Initially, the mobile telephone application found a market for these coatings and machines in Europe, but within a few years
Asian producers completely took over this market.
4.21.2 Equipment Aspects Based on Applications

4.21.2.1 Batch Systems
Most of the coating applications for hard coatings, as mentioned previously, are based on batch systems. Batch systems have the
advantage that they are more flexible and redundancy can be relatively easily achieved by expansion of the number of systems in
operation. Batch systems often require biased substrates and the complexity related to this can in the most convenient way be handled by
batch systems. The same is valid if we talk about high-temperature applications, where thermal expansions make the design also much
less critical if you have a regular possibility to access and check critical points. For in-line systems, this is all much more complicated.
Failures have high cost and production capacity impact. On the other hand, if the challenge is met and the in-line system, works failure
free, the cost of the coating can often be much lower. Besides this, there is with a functioning in-line system no need to create buffers in
the production line, which is inherently necessary for batch systems because of the discontinuous flow of products for these systems.
4.21.2.2 Special Systems

Standard systems are for manufacturers the most preferable, since this minimizes the risks for the manufacturer. It is common
practice that the development of new systems incorporates technical and economic risks. This is related to the fact that much
empirical know-how is needed regarding new developments of machines. Risks can be minimized by expanding the lead time and
making good assessments of the technical requirements translated from the existing know-how inside the manufacturer’s orga-
nization as gathered in earlier projects. Typically, costs are high for initial projects and they are dropping faster than linear to
a certain steady state with each produced follow-up system. The cost reduction is a consequence of the progressing level of detail in
engineering, the optimizations becoming apparent from the earlier design and production and the optimization of process,
maintenance, and operation aspects in relation to the specified requirements.
4.21.2.2.1 Special Batch Systems

Figure 10 shows an example of a special batch system for production. Similar system types as this one are used for the production of
several steps in thin coatings for copying machines. The different types range from thermal evaporation, nonreactive sputtering, to
www.iran-mavad.com
Figure 10 Example of a system used for a special application.
PECVD systems. Radio frequency (RF) plasma sources are applied in the PECVD systems, working under low maintenance
conditions where process gases are utilized for cleaning of cathodes and shutters.
Inside a cleanroom, the process operators move the substrate, with a length of approximately 1.50 m, between the different
batch systems on transport carriers that are coupled to a standardized mechanical clamping mechanism, serving as an interface for
the carrier between the separate machines in the cleanroom. The systems have a length of approximately 2.00 m and a diameter of
approximately 1000 mm. For each system, the loading door on the front side is inside the cleanroom, whereas the rest of the
chamber stretches into the service room. The service room is still a cleanroom, but the restrictions coming from the cleanroom
dress code are relieved in the service room. For instance, hand gloves are not needed in the service room and wearing mouth
guards is not required as well, making maintenance work easier. Loading doors inside the cleanroom open pneumatically,
allowing for loading the uncoated substrate by a fast handling design on a rotation thorn with its base mounted on the rear door.
For this reason, all chambers have to be adjusted with tight tolerances in all three dimensions in order to have the same interfacial
adaption for the load carrier versus the rotation thorn in each of the equipment. The rear door of the systems should be opened
for maintenance reasons. Access for cleaning the interior of the system requires the maintenance technician lying inside the
machine. Since this is not the most convenient position to work in, all shutters and gas frames are moved outside together with
the rear door, thus creating more space inside the machine. The rear door itself rolls backward on transport rails when being
brought in maintenance position, in this way also allowing easy access to shutters and gas frames for maintenance checks and
cleaning.
The sputtering cathodes are equipped with a mechanism that allows them to be rotated around their length axis. The flat-
lying target can now be changed easily, even with its length exceeding 1.70 m. Depending on the applied materials, the
weight can be substantial, but handling by one or two persons is in this way no longer an issue. Besides this another
solution to be adapted, relieving the weight in the case of very heavy target materials, is splitting the target in two pieces.
This is possible for flat sputter targets by providing a blunt separation between the two target halves. On the mounted target,
these halves need to be pushed against each other to minimize the slit. Initially these solutions have been introduced for
long vertical cathodes.
Also for the evaporators, where the boats are located in the bottom of the chamber, a roll-in/roll-out mechanism is used. The
boats have to be filled after each batch, and for this reason all boats are located in a drawer and are drawn out of the system at once.
In the meantime another drawer, containing prefilled boats, is loaded to enable the start of a next batch with the shortest possible
interruption.
4.21.2.2.2 Special Systems for Research

Special systems are also applied for research activities. These are systems distinguished mainly by new features. These systems are,
therefore, inherently complicated and mostly nonstandard. Since there is in most cases no dedicated idea on the processes to be
run on these system types, there is also often not a focused process acceptance goal to be reached. Depending on the innovative
character, these systems will in many cases partially be designed during the project, which elongates the lead time of the
equipment. Partially detailed definitions are to be communicated and fixed on an ongoing basis. Therefore, these projects always
include certain economical risks, since in many cases new features are to be implemented. Buying the maximum on features in
the system within the available budget is often the goal. The trade-off will be that downtime for maintenance and even for repair
will be higher, but mostly less critical for the user because he does not have to deliver within short term. Figure 11 shows such
a system.
www.iran-mavad.com
Figure 11 Glass coater produced by Von Ardenne, Dresden, Germany.
4.21.2.2.3 In-Line Systems

The last class of special systems that has to be mentioned concerns dedicated nonbatch type production systems. We mention here
web coaters, which are used to coat thin foils. These machines are applied for instance for coatings on flexible substrates, such as
barrier coatings on packaging materials. The systems can be continuous operating systems or roll-to-roll systems, where in the latter
case the systems are batch systems where a wind-up/wind-off mechanism is used. Sections of these systems can do different process
steps in a sequence, even at different pressures. In such a case, different sections are separated by slit gates, where the flexible
substrate is moving through.
In-line systems are originally provided for coating of thin coatings on flat substrates, where the substrate is moved through with
high transportation speed. Glass coaters are an example of this type of equipment (see Figure 11), being also the most applied type
of in-line systems (33). Relatively thin layers in the range of a few hundred nanometers are commonly produced for influencing the
color, thermal insulation, and reflectivity.
The production speed of an in-line system is determined by the required process time in each process chamber, where one
should think of one or more chambers for each of the process steps of pumping, heating, etching, and deposition. For the depo-
sition chambers, the production speed is determined by the maximum available deposition rate, taking in account that the substrate
will not be overheated. The composition of the coating (which can be a multilayer, built up from several individual layers) and the
required thicknesses of each individual layer determine the number of different cathodes and eventually other plasma sources for
deposition. Process conditions and/or hardware limitations can determine that the maximum available power is or must be limited.
To reach the required deposition thickness, the transport speed must be limited to a maximum vtp as well. If as an example the
number of cathodes for the process steps with the lowest deposition rate will be doubled in the chamber, the transport speed can
possibly be doubled to 2vtp, thus doubling the capacity of the equipment. Of course this must take into account that the deposition
rate/staying time in the other chambers will allow for the doubling of the transport speed as well. Similar considerations are also
valid for web coaters.
Both web coaters and in-line systems are commonly used to produce relatively thin coatings.
There are, however, other possible economical solutions for in-line production as well. If thicker coatings are required, it
makes, with regard to investment cost, from a certain point on, no sense to keep on multiplying the number of cathodes. An
alternative solution can then be found by moving the substrates a defined number of moves forth and back in front of the
cathodes in the concerning process chamber. This is a convenient possibility to make thicker coatings in the range of several
micrometers. Besides, it is also still possible to enlarge the number of cathodes and to increase the throughput. Then for the same
thickness, utilizing the maximum allowed power per cathode, the necessary dwell time for the substrate underneath the cathode
decreases linearly with the increase ratio of the number of cathodes compared to the starting situation. A corresponding increase
of the transport speed is then possible whilst maintaining the same coating properties and the throughput of the machine being
increased.
It is for in-line systems also not necessarily required to apply coatings only on flat substrates, although this is the easiest and most
usual configuration. Instead, also substrate carriers can be used loaded with one or several products. The products on the holders
often need a rotation to expose all substrates uniformly to the cathodes. The combination of this type of rotation with the
translation movement is of course much more complicated than mere translational movement of flat substrates, as would be the
case for glass plates.
Another difference in the case of thick coatings will be the vertical positioning of the cathodes, whereas in glass coaters in earlier
days the cathodes were often in a horizontal position above the substrate. In current equipment designs, the cathodes are preferably
positioned vertically, sometimes with a slight tilting angle. The benefit for such a setup is that dust falling down under the cathode
www.iran-mavad.com
does not contaminate the substrate. This improves the coating quality, since it reduces the defects in the coating caused by dust
particles. The small tilting angle is necessary if large plates (e.g., glass) have to be transported vertically. Such systems allow for also
moving the cathodes forth and back in front of the glass plates, thus reducing the size of the system as another parameter for
optimization the economics.
One important consideration factor for the design of in-line systems is related to the requirement of a bias voltage. In the case of
glass plates, biasing is generally not needed. In exceptional circumstances, the system can be equipped with plasma sources to
bombard the substrate for densification of the coating. Ion bombardment of the substrate might, for example, be needed for the
improvement of the properties of optical coatings. For hard coatings on metal parts, it is, however, in general necessary to have
a biased substrate. Biasing of the substrate is complicated in an in-line system because of the linear movement. It is even more
difficult when multiple rotations are needed for uniformity reasons, related to the line-of-sight process character.
Solutions have to be found also if the bias potential is different between different chambers. It is, for instance, possible to work
with floating potentials, but that requires that the chamber walls of certain sections be isolated and on ungrounded potential. This
makes the equipment design and the safety precautions complicated. Alternatively, potential barriers should be created, taking into
account that there are no electrical connections under any circumstances when the carrier is moved through the different process
chambers.
If an in-line system has to be stopped suddenly, the products that are still left in the system are turned into scrap. It will take time
before the system is emptied. This is to be anticipated in the software design, but it also is to be taken into account that the system
should be emptied in some way if the software is not functioning or usable due to software or hardware failures of the system. After
the system is started up to be brought back into operation, it will be necessary to condition the process chambers.
Often, conditioning of process chambers is required to achieve the correct process atmosphere, which can be very important in
delicate processes.
4.21.3 Design Phases
The trajectory from the initial end user’s wear problem and the initial sample production to equipment delivery is treated in this
section. Specific issues involve the initial sampling setup that must consider the vision for full-scale production. Also, first, cost
aspects must be taken into consideration from the earliest start, being narrowed during the succeeding phases. As an example, the
focus is on equipment for carbon-based coatings.
For the equipment design, the main emphasis here is on the a-C:H process with PECVD. This coating type constitutes the major
part of the coatings produced for components.
The substrates have to rotate, which requires a substrate table and a rotating axis, including vacuum feed through for the rotating
axis. The line of sight of the sputter processes demand for two- or threefold rotation of all substrates, which should equally expose
them to the cathodes.
After a brief discussion of the etching steps, sputtering is addressed. The cooling of the sputter cathode and target is important,
because much heat must be dissipated by the cooling water. Cooling can be designed as direct or indirect cooling, which also brings
advantages in accessibility of the magnet design. The magnetic field and its relation to the sputtering process are explained. Besides
permanent magnets, external electromagnets are applied to influence the magnetic field on the cathode, thus influencing the sputter
behavior as well as the influence on the remote substrate by additional argon ionization in case of a suitable design, called closed
field design configuration. Also the coating uniformity will be influenced by the magnetic field. Compensation of the loss in
uniformity is, however, possible.
Differences between nonreactive and reactive sputtering are explained, where especially the reactive process and the occurrences
on the target and in the vacuum environment are described in the section on the feedback control method. The cathode power
supplies play an important role. This section refers not only to carbon-based processes but also to oxide processes.
The next section describes the coating architectures of carbon-based coatings and begins with tungsten-doped DLC, followed by
amorphous hydrogenated DLC. After this there is a description of nonhydrogenated DLC, which constitutes the hardest coatings,
approaching closely the hardness of diamond.
At the end of this section, a description follows from requirements to the automation of the equipment, followed by
a description of the requirements to the (project) organization to accomplish the transition from first (small-scale) sampling,
through equipment design until the final stage of delivery of a large-scale industrial equipment.
The next section describes the whole trajectory from the initial design of the end user looking for a solution to a wear problem, to
the point where a coating is considered as a solution and ending with the field service and operational coating aspects.
A component designer will initially test his design, which will presumably not yet include a coating. When being confronted with
too much wear in the application, he needs to look for a solution. Initially, he will try to improve the quality by changing his design.
Cost issues will always play a role, which implies that the designer will initially try to minimize cost as far as the quality according to
the system specification. If the solution cannot be found in the traditional mechanical design of the component, an alternative
solution of providing the component with a coating will be one of the considerations. At this point the designer will make initial
contacts between one or more coating providers, where the coating provider can be a job coater, a research institute, or an
equipment manufacturer. At the end, in the case of a successful solution with coating on the component, components have to be
produced in quantities from which the prognosis needs to be taken under early consideration. The final method of loading the PVD
www.iran-mavad.com
equipment has therefore to be considered from the beginning on. The number of parts and the process duration have a strong
influence on the cost of the coating per part. The coating cost is one of the most important arguments for the decision to coat parts in
high-volume applications.
Anticipating the final fixture method for the parts during early sampling tests reduces development risks (i.e., development time
and batches), since it will be easier to make good estimations for the final process parameters. If more parts must be considered for
coating serial products, which is almost always the case if automotive components are concerned, there will be a greater possibility
for price reduction related to large scale. The cost structure is explained in more detail in the Section 4.21.5.
Initially only a small number of test samples will be used since in this phase it is only necessary to show feasibility of the solution
with coating. Best practice is to simulate the conditions as foreseen for serial production. This implies that it is beneficial to produce
the small quantity of initial samples on product holders that simulate the anticipated full load conditions in the closest manner.
Large-scale circumstances are difficult to simulate. Therefore, estimations must be made to translate the sampling results to results
achieved under full load conditions. Modeling is too complicated and there are no adequate mathematical models available.
Available models largely depend on empirical data. It is for this reason helpful to simulate full load conditions by using dummy
loads to fill up the system as well as by configuring a defined environment for the component samples to be evaluated in tribological
tests.
From the factors of influence mentioned here, the deposition rate and the number of parts per batch, respectively, the
throughput speed in the case of an in-line system, are of main importance.
Hydrogenated DLC coatings (a-C:H) represent the majority of the market for automotive coatings due to their hardness. Next
these coatings are used to describe equipment requirements. The a-C:H coatings are characterized by an increase in compressive
stress with the growth of the top coating thickness, which consists of an amorphous hydrogenated DLC (a-C:H) coating.
As a rule of thumb, one can assume that the compressive stress of the DLC coating increases with 1 GPa for each micrometer of
coating thickness. Compressive stress helps in wear protection, since it helps relieve a mechanical load, perpendicular to the
substrate. On the other hand, higher stress in the coating increases chances for the occurrence of adhesion failures (spallation of the
coating). Normal values of thickness of the a-C:H top coating range between 1 and 5 mm. An increased stress should be handled,
however, since for abrasive wear certain thickness values are necessary as well. Commercial brand names for these coatings are
specific per producer.
A graded base coating is necessary to provide a solid base for adhesion of the tribological top coating. In case of a tribological
hydrogenated DLC (a-C:H) coating applied on the component, the traditional adhesion layer will be a Cr sputtered layer followed
by a W-C:H intermediate layer, which is a graded sputter coating as well. Other base layers can be applied as well. Examples are CrN,
CrC, TiC, TiN, or SiC. The top coating will exist of a hydrogenated DLC coating, produced by PECVD, sometimes also referred to as
PACVD.
4.21.3.1 Substrate Rotation

In the case of an open load, the sputter deposition rate will be much higher than in the case of a full loaded system. In a standard
PVD batch system, the cathodes are vertically mounted in flanges in the chamber walls. A general design of a substrate rotation table
is a table where its base is equipped with vertical spindles distributed over the table circumference.
The substrate table (Figure 12(a)) must in general be equipped with threefold rotation (Figures 12(b)–12(e)). The design of the
table is a satellite table, where the sun gear is a gear located on the circumference of the table base. The vertical spindles with product
holders, visible in Figures 12(a)–12(c), are individually driven by planetary gears, needed for the second rotation, which rotate
against the sun gear. Usually the table base contains these gears and is connected to a rotation axis under the table.
It is necessary for the horizontal uniformity of the sputter-coating properties that the components are seeing the active targets
uniformly during the production of sputter- or arc-deposited layers, but also for PECVD the rotation is playing an important role for
the distribution.
The table base acts as a shield protecting the gears from being coated during the process. The table base is under circumstances
on a different potential than the substrate, which implies that there should then be insulators in the design of the shield for the
table base.
Commonly the negative bias voltage will be connected to the substrates through the rotation feed through by software control
(fully or semiautomatic). In some configurations special vacuum suitable plug-in connections can be applied, which may have to be
connected manually. If the electrical contact is made eccentrically, a sliding electrical contact will be needed inside the vacuum,
whereas in case the central axis is also the electrical feedthrough, slide rings will be applied under atmospheric pressure. The latter is
the most applied solution for batch systems.
The concentrical clamping design takes care of the torque transfer and the electrical connection of the substrates to the bias
voltage supply. The axis of the rotation table needs a strong form-closed clamping mechanism to allow the table to be taken out and
brought into the vacuum chamber on a fixed position by a roll-in/roll-out or a lifting mechanism. In the case of the application of
a lifting mechanism, the table will make electrical and mechanical contacts with the rotation axis when going down. In case of a roll-
in/roll-out mechanism, this clamping mechanism is made by a coupling that is lifted to couple the correctly positioned table
mechanically and electrically to the rotation axis inside the vacuum chamber.
Inside the table, the bias current is connected to the spindles from the side of the table base. The electrical current path continues
through brush contacts (silver-brass), connected in parallel with the planetary gears and making contact with the vertical spindles.
www.iran-mavad.com
Figure 12 (a) Loaded substrate table with threefold rotating products. (b) Threefold rotating drills with stationary kicker. (c) View on gears for third
rotation of each product loaded on a product holder. (d) Closeup of (nonrotating) kicker for third rotation of the products (here: hollow pipes) mounted on
twofold rotating holder. (e) Schematic view of the different rotation modes of the products on the substrate table.
www.iran-mavad.com
Silver-brass is selected as brush contact material for its good vacuum and electrical conduction properties. Depending on the desired
plasma properties, the table base should be either on substrate potential or on a floating potential. The table base also functions as
a shield separating the parts of the table, carrying bias potential, from the plasma and preventing undesired parasitic deposition
under the table.
Depending on the substrate holder design and in the case of threefold rotation, small satellite tables are mounted in
several levels on the spindle (Figures 12(a)–12(d)). These small satellite tables are carrying products on their circumference.
In order to get symmetrical deposition, it is required to rotate each product on the small satellite tables. This is obvious, since
one can imagine that substrates will get less coating on the side directed to the spindle than on the side pointing radially away
from the spindle. To distribute the thickness of the coating evenly, there are kickers to rotate each substrate mounted on the
concerning small satellite (Figure 12(d)). On each position for a substrate on the small satellite tables, small gears are
mounted. The teeth of these gears protrude from the circumference of the small satellite table. When the kicker passes the
small gear with its small number of teeth (four to six), each of the small gears touches the kicker one after the other. The gears
rotate by the touch of the kicker one tooth further, thus forcing the substrate in the position above to rotate between 60 and
90 (see Figures 12(b)–12(d)). This movement is the third rotation (see Figures 12(d) and 12(e)). The kickers should be on
substrate potential in order to prevent charge building up in the plasma, since it will otherwise discharge itself in an arc
discharge as soon as the charged kicker arrives at a small and decreasing distance. Such a discharge on the substrate is to be
prevented since it will interrupt the bias voltage and forms a risk for arcing damage (reaching from small star shapes to
so-called crow feet). The practical design is often to mount all kickers on a biased rod at exactly the required distance. This
allows the adjustment of the third rotation kickers if the vacuum chamber is opened. Other more robust, but also more
complicated/expensive designs are using gear-driven third rotation movements. Such a design with gears reduces, however, the
flexibility in loading capacity of the machine.
As discussed, the substrate table mostly exists of a table base configuration driven by a central axis. The axis is in most cases
coupled to a rotating vacuum feedthrough, which is also used as electrical feedthrough for the bias voltage in the generally necessary
case of processes with biased substrates. In many cases, the utilized feedthrough is a magnetic fluid feedthrough with a hollow axis.
Originally, this feedthrough was designed as rotating vacuum feedthrough for space applications. The magnetic fluid forms the
vacuum seal. This central feedthrough is driven outside the vacuum and forms the rotational axis for the first of up to three
rotational movements, applied on standard substrate tables.
In the case of single-fold rotation, all products are loaded on the circumference of the table, at a suitable distance to the cathodes.
In such a loading case, the table load is to be seen as a cylinder.
The loading height has also its importance. In the case of PVD processes, which are processes with sputter cathodes, either planar
or rotatable cylinder cathodes, arc cathodes or even evaporator sources (E-beam or thermal), the process is a line-of-sight process. It
is necessary to load the machine with the highest possible number of parts in order to minimize the coating cost per part, while at
the same time taking into account that the coating on all these parts must be in compliance with the specification. Because of this,
often twofold rotation, and in some cases threefold rotation and seldom single-fold rotation, is applied in the production of
coatings.
4.21.3.2 Etching Steps

Etching steps are done either with assistance of a plasma source (sputter cathode, arc cathode, or gas ion source), giving a line-of-
sight effect for the argon ions (respectively for the metal ions in case of arc etching), or by a glow discharge generated on the
substrate holder by pulsed power, DC-power, or RF-power supplies connected to the substrate. The attraction of the argon ions
(respectively, metal ions) enforces etching of the biased substrates.
Line-of-sight processes are used for base layer steps in PVD/PECVD processes, and these steps are usually sputtered layers and
therefore require a multiple rotation of the substrates. Also for the PECVD steps in these processes, multiple rotations are generally
required. The plasma in these process steps is the source of the reaction product ions. Since in most cases the plasma distribution is
spatially uneven for production of hard coatings, similar effects as in the actual line-of-sight process steps like sputtering and etching
must be taken into account as well.
4.21.3.3 Sputter Deposition: Sputter Cathode and Target

In the sputtering process, the deposited layer distribution over the height of each cathode is flat over a long range, but falling at
the end.
The cause of this thickness distribution relates to the properties of the magnetic field of the sputter cathode. To define for the
reader’s sake the orientation, the following definitions describe the sides of the target. The expression ‘process side’ refers to the side
of the cathode exposed directly to the vacuum inside the process chamber while being assembled in its normal position. The
‘backside’ of the cathode is defined as the side opposite the process side. Therefore, the backside is the side facing the atmospheric
environment of the cathode in the normal assembly position.
A sputter cathode consists of a cathode body carrying permanent magnet materials mounted in a yoke and of a target with
clamping provisions for keeping the target in position. The clamping is also functional for cooling of the target. Magnets and yoke
are generally mounted in the body inside the cooling water (in the case of direct cooled targets) or on the backside of the cathode in
www.iran-mavad.com
normal atmospheric conditions. The yoke maximizes the magnetic field strength on the target surface and also shields the backside
of the cathode against stray fields.
4.21.3.3.1 Target Cooling

The target, which contains the material to sputter, is clamped to the cathode body on the process side. The target is mostly clamped
both in the middle and along the circumference of the target. This is necessary to hold the target in its correct position, to allow
sufficient space for thermal expansion of the target to the sides, as well as to guarantee an efficient target cooling.
In the case of indirect cooling, the cathode body contains channels for the cooling water. The body needs to be made out of
nonmagnetic and thermally good conducting material such as copper. A foil attached between the cooling water channels and the
target has as a function to ensure good thermal conduction between the target and the cooling water. The copper body is a good heat
sink for the cooling water as well as for the foil. The foil needs to be thin in order to ensure good thermal contact between the target
and the copper body and as such to reduce thermal resistance to the cooling water. Additional measures reduce the effects of thermal
bending of the target. Maximum heat is generated on the surface of the target inside the racetrack. Due to thermal resistance of the
target material and the unevenly distributed flow of heat between target surface and cooling water via the foil and the copper body,
the target will be heated much more on the surface inside the racetrack region than in the surface outside the racetrack in the less-
heated target surface sections. The result is that the target will expand unequally over its thickness range, corresponding to the
thermal expansion coefficient of the target material and the temperature differences between the target surface and the backside of
the target. These temperature differences will therefore ultimately result in bending of the target. With a good heat sink, this is an
inevitable effect explaining the reduction of maximum power in relation to the mechanical strength and the maximum applicable
power for this combination of target and cathode.
It is clear that on the other side the thermal resistances transporting the heat flow from target surface to cooling water need to
be minimized. A minimum thermal resistance yields minimum heating of the target. To reduce the thermal resistance, the target
will generally need to be in good contact with the foil. For this reason, the target is clamped in the middle and on the sides of the
target. Additional good practice is to put graphite foil between the foil and backside of the target. The graphite acts as a good
thermal conductor, and higher pressing forces will result in even improved thermal contact between the target and the heat sink,
built by cathode body and cooling water, whereas the graphite can be compressed and distribute in this way the pressing forces
evenly.
4.21.3.3.2 Sputter Cathode and Target: Magnetic Field

The magnetic field of the cathode is needed to trap electrons generated in the plasma. Without this magnetic trap, only diode
sputtering would be possible (34). For sustaining a discharge there is a need to create the so-called avalanche conditions, where the
plasma is capable of generating a sufficient number of charge carriers. In diode sputtering, a discharge under this condition can be
generated only by high voltages on the cathode. The pressure needs to be kept generally low, and the electron density will therefore
also be low. High energy ions are generated in the sheath and will collide on the surface, thus also generating high energy neutrals
sputtered from the target. The generated plasma is of high energy with a low ion density.
The introduction of the magnetic fields increases the argon ionization. Permanent magnets, in some cases combined with
electromagnets, generate the magnetic field strength. This causes a high electron density in the plasma in front of the target. The
maximum of the electron density is coinciding with the maximum of the magnetic fields’ parallel component (where parallel stands
for the direction in parallel with the target). The parallel component creates a drift velocity of electrons in parallel with the target
surface. The normal component increases the electron density, since electrons can escape only with greater difficulty due to the
longer path along the magnetic field lines, making it more difficult for electrons to escape from the electron cloud in front of
the target. The electrons play a role of major importance in generating and maintaining the plasma. It is well known that most of the
energy transfer between particles in plasma takes place in collisions between electrons and heavier neutral, excited, or ionized
atoms. Due to the conversation of momentum and because of the huge difference in weight between the electron and the neutral/
atom, almost all energy is transferred from the electrons to the neutrals/atoms in a nonelastic collision.
The permanent magnetic field the case of a planar cathode consists of an inner row of magnets with a defined strength. The
magnet materials can be magnetic tape material, but usually the magnets will be made of blocks of strong ferromagnetic rare earth
materials, like samarium–cobalt and neodymium–iron–boron. The backside of the cathode carries the yoke. In the magnetic circuit,
the yoke acts as conductor for the magnetic flux. An advantage of the indirect cooling design is that it allows for direct access from
the backside of the cathode. This makes optimizations and changes in magnetic field much easier and practical than in the case of
direct cooled targets.
For the magnetic field, one can distinguish between balanced and unbalanced field design. Both are displayed in Figure 13,
showing the magnetic field and the consequences for the plasma.
The balanced magnetron has the advantage of good uniformity of ion flux, but the degree of ionization in the ion flux is low. The
unbalanced magnetron (35) causes electrons to escape from the magnetic trap of the magnetron and pushes therewith the plasma
forward to the substrate (see Figure 13, right). Also negative bias is initiated on an insulating or floating substrate, thus allowing ion
bombardment on the substrate. Because the plasma needs to be quasineutral, ions will be dragged with the electrons along the
magnetic field lines to the substrate. Stronger unbalancing does also lead to less uniform ion flux distribution, but with the inte-
grating effect of moving substrates this can be flattened out. If the substrate is biased, an ion current increase can be observed due to
the unbalancing effect.
www.iran-mavad.com
Figure 13 Difference in the magnetic field and the plasma distribution for a balanced cathode (left) and an unbalanced cathode (right).
In case of balanced magnetic fields, the strength of the inner magnets is usually twice the strength of the outer magnets, whereas
in case of unbalanced magnetron the magnets in the middle and in the sides are equally strong. Practically, in the case of unbalanced
magnets, the strength of the outer magnets is doubled compared to balanced magnets, thus making these outer magnets as strong as
the center pole magnets of the unbalanced case. Common practice is the additional use of electromagnets to increase the unbal-
ancing effect even further. The strength of magnets can be redoubled by stacking the magnets in double rows on top of each other.
Handling of strong magnets is difficult and must be done carefully to prevent injuries. Commonly, bar-shaped magnets that have to
be lined up are applied, requiring high forces due to the high magnetic strength of these magnets. For test setups, it may be useful to
use duct tape for keeping these magnets in position. In a permanent design, however, the best solution is to use guiding ducts of
nonmagnetic stainless steel, acting as holders for the magnets and enabling the assembly of the magnet configuration in an easy
manner.
For an easy exchange of magnets there are advantages of the indirect cooled cathode design. This design allows for manipulation of
the magnets without the excessive disassembly work needed in the case of direct-cooled cathode. In a direct-cooled cathode the
magnets are mounted inside the cooling water. An additional advantage is the possibility to move the magnets. In this design it is
possible to optimize process parameters and target erosion by allowing movement of the magnets. With a well-designed permanent
magnet configuration and the use of suitable magnetic materials in the cathode body, it is possible to increase the target utilization to
higher values by broadening the erosion track. Practice has shown that an improvement in target utilization from 25 to 40% can be
achieved in case of planar cathodes. If the range of electromagnet current is limited and known upfront, an optimization of the design
of the magnetic field with respect to ionization and erosion for such a range of currents may be possible. The design optimization for
the magnetic design cannot, however, be optimized for the whole range of electromagnet currents. Only for relatively limited vari-
ations in currents of the electromagnets, an appropriate magnetic field design for this range of currents will result in an optimum in
erosion. If other electromagnet current ranges are required for reasons of ionization improvements (see discussion about ionization
increase in the next section), this will inherently limit the target utilization. If the magnets are moveable, it is also clear that this
additional degree of freedom allows optimization of magnetic field and target erosion only within certain limits.
4.21.3.3.3 Sputter Cathode and Target: Unbalanced Magnetron Cathode Coils and Closed Field Configuration
The aspect of the magnetic field of the electromagnets strengthening the outer poles of each cathode has been addressed previously.
The effects on racetrack and target erosion were discussed. Another effect of the electromagnets has great process benefits because it
increases the ionization degree of the plasma. This is the so-called closed field configuration.
The machine setup must be symmetrical, with the cathodes equally distributed around the chamber walls and a substrate table
rotating around an axis in the center of the machine (Figure 7(b)). The closed field configuration requires an even number of
cathodes. In such a configuration, the electromagnets and therefore the outer poles of the cathodes have to be magnetized
sequentially following the circumference of the chamber with alternating field polarity. For a situation with four or more cathodes,
this gives a magnetic field in the chamber that is concentrated around the circumference of the product holder. Since electrons tend
to be trapped in spiraled movements around the magnetic field lines, the ionization probability increases as the pathway of the
electrons is elongated. Since the atmosphere is dominantly consisting of argon, the increase of ionization is mainly attributed to
argon ions. By biasing the substrate, the increased ionization of argon enables manipulation of the structure of the coating on the
substrate. In a situation with only two cathodes, the magnetizing direction is through the table, which results in the highest ion
www.iran-mavad.com
current on the substrate. This is therefore from the ionization point of view the most beneficial configuration. There is, however, also
a large increase in ionization in the most common circumstances, where a PVD system is equipped with four or more cathodes.
Different magnetic field configurations exist as well, like Teer Coatings Ltd. using only permanent magnets supporting each other
in a nonsymmetric setup. This allows a similar increase in ionization as discussed before, but the ionization created around the
substrate increases asymmetrically. The ionization will be more comparable to the situation of the symmetrical two-cathode
configuration compared to the asymmetrical four cathode configuration, as discussed previously.
4.21.3.3.4 Sputter Cathode and Target: Magnetic Field and Coating Uniformity
The magnetic field of the cathode has consequences for the thicknesses distribution of the sputtered adhesion and intermediate
layers. The required thickness should vary in general not more than 15%. An adapted formula to calculate the uniformity is:
ðHmax Hmin Þ
Uniformity ¼ ;
ðHmax þ Hmin Þ
which is corresponding to half of the deviation between the maximum (Hmax) and minimum (Hmin) substrate coating thickness in
ratio to the average substrate coating thickness over the defined uniformity zone (Figure 14). With the majority of cathode designs,
the maximum will correspond to the thickness of the substrates on the loading height in front of the middle of the cathode, whereas
the minimum coating thickness values will be found in close proximity of the end on top and bottom of the cathode. If well
designed, taking the process parameters into account, the loading height where the uniformity will reach 15% is located both
approximately 50 mm under the top end and above the bottom end of the target. The distribution of the coating thickness is to
a large extent related to the magnetic field design of the sputter cathode. This is especially the case with lower process pressures and
under nonreactive circumstances. Optimization of the magnetic field at the ends by changes in the magnetic field (dog bones) can
help elongate the erosion track at the top and bottom of the cathode. This will result in an increase of the uniformity zone height.
Contrary to this in case of unbalanced magnetron (UBM) cathode with external unbalancing coils as the uniformity zone decreases.
As described, the external field strengthens the outer poles of the permanent magnet configuration and generates a closed field
configuration. This results in an increased ionization, which is beneficial for influencing the coating structure. The stronger electron
confinement pushes the plasma vertically forward in the direction of the substrate, causing the ionization effect (Figure 13, right). It
results also in pushing the confined electron cloud in front of the target (‘racetrack’) further to the middle of the cathode with
increasing coil current/magnetic field. The end zones of the target on top and bottom are also pushed to the center of the cathode.
Both the erosion tracks and the focusing effect explain that the distribution over the height decreases.
4.21.3.3.5 Sputter Cathode and Target: Nonreactive and Reactive Sputtering

Another aspect concerning deposition rate can be found in the reactive sputter deposition description. If the desired coating consists
of a compound material, there are two possibilities to produce this type of coating: either nonreactive sputtering from a cathode
with the compound material as target material, or reactive sputtering, where the target material is identical to the metallic
compound in the coating composition. The advantage of nonreactive sputtering is an easy process control, whereas the disad-
vantages are the inflexibility for changing the composition and, the main disadvantage, a reduced deposition rate. Especially, the
deposition rate is a factor of high importance from an economics point of view. Therefore, reactive sputtering is almost inevitable
because of its much high deposition rate, even though the process control is much more complex.
l Nonreactive sputtering
In nonreactive sputtering processes, the applied sputtering gas is argon. The coating composition will directly be equal to the
target composition if the atoms of the target are directly bonded to each other. This is the case for real alloys and compounds. If the
target consists of a mixture, a certain running-in phase of the targets needs to be respected. The coating composition will initially be
Figure 14 Coating uniformity.
www.iran-mavad.com
different than the target composition and will depend on the sputter yields of the components in the target material, provided that
the target material will not melt. The composition will change into the desired equilibrium state after the running-in phase of the
target. This steady state will be reached when the sputter depth is approximately equal to the grain size (36). The composition in the
coating will then be identical to the composition of the sputtered compound. One should also note that the sputter yield also
depends on the ion energy, which will not be linearly dependent of the target material. This means that in case of sputtering with
different voltages there will be a different ratio between the sputtered species, since the sputter yields are not linearly dependent of
the voltage.
l Reactive sputtering-target phenomena
Reactive sputtering processes have the advantage to allow the production of different compositions. In the 1950s, this was
initially applied for TaxNy deposition in the semiconductor industry, where by increasing the N2 flow a possibility to produce
different compositions could be created, thus influencing the resistivity of the film (37).
The main reason to apply reactive sputtering in high-volume processes is to accomplish a high deposition rate. The sputtering gas
is argon, as in nonreactive sputtering. High-rate sputtering requires that there will always be metal species sputtered from the target.
This operation mode is called ‘metallic mode,’ but there is also an operation mode called ‘transition mode,’ where sufficient metallic
particles are still available for a high rate. As we see later, transition mode requires sophisticated control methods but offers benefits
from the point of view of coating quality. The third, undesired mode is called ‘poisoned mode.’ In that mode, the target is covered
with reaction product only.
By adding reactive gas species, there will be chemical reactions between the target material species and the reactive gas generating
the coating. This is due to the fact that at the reactive gas forms compounds by reactions with the target material especially outside
the racetrack, since there are fewer argon ions that can clean the target by sputtering than inside the racetrack. Let us assume an
example of reactive sputtering of TiN. In the process, Ti will be sputtered and these Ti atoms will consume a part of the reactive N2
gas flow. Ti will react with gas species (Nþ 2 ions) to be deposited as TixNy. The vacuum pumps of the system will pump another part
of the N2 flow away. The nonreactive sputtering gas (argon) will not be consumed by the TixNy reaction. Generally, the total pressure
will be in the range of 0.2–0.4 Pa, whereas the partial pressure of the reactive gas is generally about 5–10% of the total pressure
(range 10–40 mPa).
In close proximity of the racetrack, the reaction products (TixNy) will be sputtered, together with the metallic components (Ti)
that are sputtered from the center of the racetrack zone. The target is still in metallic mode. With increasing gas flow, the
concentration of reaction products will increase. As a consequence of the open racetrack, the section of the target that is not nitrided
and still consists of metallic Ti will gradually decrease in width with an increasing flow of reactive gas (N2). The total pressure will
hardly be influenced, due to the dominating argon gas atmosphere.
At a certain flow fp, the metallic section will be completely covered with the TixNy reaction product and there will be no
sputtered metallic particle left in the system. In metallic mode, a substantial part of the reactive gas flow is consumed by this reaction
as metal species are still generated on the target. The total pressure will remain almost unchanged. As soon as the target is completely
covered with the reaction product, there will be an abrupt increase in the total pressure, since the consumption of reactive gas by the
metal species is interrupted due to missing metal species for the reaction. In this situation, the target is said to be in the poisoned
state. Compared to the reactive mode, the deposition rate will fall in the poisoned mode. The reason for this is that only reaction
products can be sputtered from the target in a poisoned situation, whereas earlier a substantial portion of the reaction took place in
the plasma or on the substrate.
For characterization of the working points for the reactive process as well as the poisoning mechanism, it is necessary to record
a so-called hysteresis curve. The feedback value for process control is measured against the flow of reactive gas. Several signals can
serve as feedback for process control, such as total pressure, partial pressure of the reactive gas, optical emission control, or voltage of
the target. The hysteresis curve must be recorded with gas flow increments at a slow pace, allowing each working point to be
stabilized after being set. All other conditions important for sputtering have to be kept identical to the process conditions to be
selected in order to find all parameters for the working point of choice. These parameters are magnetic field, location of the gas
supply inside the chamber, argon flow, power on the cathode, target erosion, load density of the machine, and pumping speed,
where eventually existing obstacles in the gas flow will play a not-negligible role.
During the measurement, the total pressure will increase abruptly at a certain flow fp within a second if the target gets poisoned.
The physical effect is that the erosion track gets increasingly narrower, since with increasing flow the reactions on the target surface are
increasing as well and come then to the point fp where the metallic track vanishes and cannot be sputtered free from reaction
products by the argon ions. This situation is called the poisoned mode, whereas the transition situation between metallic and
poisoned modes is called the transition mode. Transition mode is a kind of metastable situation between metallic and poisoned
modes that can be stabilized by a suitable fast-acting control system. Further increase of reactive flow to values above fp in poisoned
mode will lead to a gradual increase of the pressure, starting from the level reached at fp where the target was poisoned. In order to
return from poison mode to metallic mode, the flow must be reduced. In the hysteresis measurement (Figure 15), this is to be done in
slow pace decrements. In the meantime, we have an argon bombardment on the target, which removes reaction products.
If we reach the flow fp where the pressure made a jump due to the target poisoning, we will now continue to stay in a situation
where the target is poisoned. Because there is still an excess of reactive gas that will react with each metal particle appearing under the
place where a sputtered reaction product has been removed, the target will remain poisoned at this flow. This situation does also not
change at lower flows than fp, until at a certain reactive flow value, there is an excess of metal showing up under the poisoned target
www.iran-mavad.com
Figure 15 Hysteresis in reactive process.
surface that will not immediately react with the reactive gas. In this situation, the target returns into metallic mode. The pressure
starts to drop rapidly to the value it had earlier in metallic mode before poisoning. The target will now stay in metallic mode at these
lower pressures, even when the flow will increase to values not exceeding the reactive flow fp where the target went from metallic
into poisoning mode. The process itself will go back into reactive sputter mode.
The width of the hysteresis is indicative of the ease of the controlling operation. The width depends on the target surface, the
target power, and the system’s pumping speed. The getter pump effect of a large target surface, which is running with high power,
represents a relatively high pumping speed compared to lower powers and smaller target surface areas. In addition, hysteresis effects
will be less for systems with high pumping speeds (38), since in such cases the effect of changes in system pumping speed are reduced
because the proportion of getter pumping speed is low compared to the vacuum pumps. Therefore, the total pumping speed is
changing less and the pressure is also changing less when poisoning happens. This allows the system to recover easier from target
poisoning, since poisoning occurs also more gradually. Obviously, with increasing systems vacuum pumping speed compared to the
gettering effect, the poisoning effect reduces. With increasing pumping speed, the hysteresis effect diminishes, therefore, resulting in
a narrower hysteresis loop.
l Reactive sputtering: feedback control
One of the problems is to create in reactive sputtering mode the composition of the coating where it is aimed. In many cases,
the best properties are found in stoichiometric coatings. In almost all reactive sputtering processes, the required properties of the
coating will therefore require to have the process working at the point of transition between metallic and poisoned modes. Due to
the hysteresis effect, the process should run in metallic mode for the sake of the deposition rate at a working point as close as
possible to the border of poisoning at fp. Any disturbance of the process where an excess of reactive gas or shortage of metal from
the target occurs will then lead to a switch of the target into poisoned mode. As described previously, the process then moves into
a completely different regime from where it will not be able to return, unless the flow is heavily diminished. While working at the
edge of poisoning, arcing on the target already could shift the process into poisoned mode, since during arcing there will be no
sputtering from the target. The metallic section of the target will react with the reactive gas and thus change into poisoned mode.
Nevertheless, this also means that the composition of the coating will deviate completely from the required composition, since
there will be an under-stoichiometric coating produced as long as the reactive gas flow is less than fp. Additionally, the pressure
jump will also influence the growth of the coating on the substrate. Therefore, other solutions have to be found. The best solution
is to use one of the following feedback signals with the best signal resolution for applying closed loop control. With a suitable
closed loop control, it is possible to work in transition mode in each point of the feedback signal between metallic mode and
poisoned mode. The fast-acting control system allows for this kind of regulation to work in stable operation in each working
point of the transition zone.
Industrial machines need sophisticated software that can process all available signals and operating with high sampling
frequencies. The sampling frequency is sufficiently high to allow software proportional integral derivative (PID) controllers to do
the job. Other, less preferential possibilities include separate control systems. These solutions are less preferential since they do not
allow an easy optimization of the control system to adapt to changes that are specific for the deposition system.
www.iran-mavad.com
Reactive controls can also be connected to external standalone control systems, like the IRESS control system developed by W.
Sproul (39).
One of the applied feedback systems for pressure control is the spinning rotor gauge. This device uses a spinning magnetic ball
that decelerates during a sampling time due to the viscosity of the gases in the vacuum chamber. An advantage of the spinning rotor
gauge is that it is a very accurate measurement and it provides in the sputtering range a signal that is programmable and therefore
very well useable. The feedback signal for the partial pressure control is here the total pressure. This is a useful control feedback
signal for processes where the total pressure changes noticeably with poisoning of the targets. With a system equipped with
turbomolecular pumps, the pumping speed is very constant in the normal range (2–4 hPa) of sputtering. Nonlinearity, shaped as
a sharp drop in pumping speed with increasing gas flow, appears in the range of pressures around 1 Pa and higher. The partial
pressure of argon as sputter gas depends only on the pumping system, whereas the partial pressure of the reactive gas is determined
by the pumping system and by the getter pump effect effectuated by the metal atoms reacting with the reactive gas. Therefore, the
argon pressure stays unchanged if the target is poisoned, whereas target poisoning will stop the getter pump effect since no more
metal particles are available to sustain the reaction with the reactive gas. The controller recognizes the point where target poisoning
occurs by an increase of the partial pressure of the reactive gas. The controller should be fast enough to reduce the reactive gas flow
when the pressure increases, in order to allow the reaction products to be sputtered away immediately. In this manner, it is possible
to work at the edge between metallic and transition modes, thus generating stoichiometric coatings and allowing for a high
deposition rate since the process will stay a reactive process.
A control method by regular PID controllers does not function in this case because the feedback signal is discrete and measured
over a sampling period. The duration of the sampling period is adjustable, but should be at least a few seconds long to get sufficient
accuracy of the measurement. The pressure signal stays constant during the next sample period and is refreshed after the next
measurement sample has been taken. This control feedback signal can be used only by applying discrete steps of flow to correct the
pressure, since the discrete sample would lead to overregulation in case of analogue controllers. The measurement signal needs be
synchronized with the controller for proper regulation. Other feedback signals can be used as well, like capacitance manometers, but
in general these signals have only a reduced resolution compared to the spinning rotor gauge, and therefore a limited accuracy.
Using total pressure control is a good working method in the case of relatively forgiving reactive processes, where the target is not
changing too abruptly from metallic into poisoned mode. This is the case for most nitride processes, like TiAlN and CrN. These are
well controllable with spinning rotor gauge feedback.
4.21.3.3.6 Cathode Power Supply in Reactive and Nonreactive Sputterings

Much more problematic are processes where heavy arcing occurs on the target when the target poisons. Many oxide processes are
such processes. Reactive oxide processes generally experience strong poisoning effects. Especially alumina or silica processes
experience strong process changes between metallic and poisoned modes. The oxidized target surface has a much higher secondary
electron emission, which results in a substantially lower target voltage. Also heavy arcing occurs on the target surface because the
voltage drops suddenly with 100–250 V whereas the coating is insulating and collects high charges. Applications in the past have
shown that in these processes suitable arcing suppression can help run these processes with fast-acting detection and suppression of
arcs and a minimized interruption (arc pause duration, ranging from 1 ms to less than 20 ms) of the supply of output power to the
cathode (Figure 16(a)). It is very important then that the power supply returns immediately to the working voltage and current after
the interruption in order to prevent overregulation, since overregulation will result in oscillations that may lead to currents
exceeding the maximum allowed power supply current. The consequence is then that the power supply switches off maximum
current and this repeats itself (Figures 16(b) and 17). Special control algorithms, depending on the power supply type, keep the
control circuitry inside the power supply frozen in the time interval between the moment that the arc is detected and the arc pause
duration ends. Especially in power supply types where inductive energy is stored, it is necessary to apply such adaptive control
methods. These types have, however, the advantage to work with constant current, making them inherently well suitable for sputter
applications, since the characteristics of a sputter process are that in the usual operating range a small voltage variation results in
a large current variation. This implies immediately that large current overshoots are the consequence of a small increase in voltage,
which explains that a large current overshoot will be the consequence. Another effect is that the voltage dropped strongly as arcing
occurred, because of a temporary target poisoning due to the formation of a very thin compound layer covering the target surface in
the race track. This can only be overcome by maintaining the same argon bombardment intensity, i.e., the same voltage as before the
arc was switched off. This allows the target to be cleaned and the target recovers from a poisoned state back into metallic mode.
Actually, the target is operating in this way in transition mode. This type of silicon controlled rectifier resp. thyristor (SCR) power
supply was developed by Hauzer in close cooperation with Prof. Dmowski from the Technical University in Warsaw in the period from
1986 to 1988, and they were produced until the end the first half of the 1990s. These power supplies were designed with mains
commutated SCR’s and with fast electronic switches, working with forced commutated SCR’s to interrupt the connection between the
input mains voltage and the DC output. In the DC output, where the inductive energy was stored in large inductances, SCR circuits
(switch S5) trapped the inductive energy of choke L1 in a freewheeling circuit acting as energy storage (Figure 18). At the end of the arc
pause duration, the fast electronic switches go back in the conductive state again. Firing the mains voltage SCR’s provides then the
commutating energy, forcing the current back from the freewheeling circuit into the cathode, being the main output circuit (Figure 18).
After the middle of the 1990s, Hauzer decided not to produce power supplies in house anymore. Also, new developments were
leading to secondary switched mode power supplies equipped with insulated gate bipolar transistors (IGBTs), soon followed by
primary switched mode power supplies. This was developed in Poland at the company AC, later Hüttinger Poland. These power
www.iran-mavad.com
Figure 16 (a) Oscillogram of power supply with arc detection, running in operation mode close to poisoned mode on SiO2. (b) Oscillogram of power
supply with arc detection, running in poisoned mode on SiO2.
supplies had the advantage of operating at higher frequencies, thus allowing smaller filter components. This makes the power
supplies faster, allowing one to switch the arc faster off and at the same time to increase the output energy from zero at a fast rising
rate. An additional advantage is that a fast controlling speed is possible in the case of smaller filters, making it easier to prevent
overshoots by overregulation. Since the power supplies are inherently faster because of the smaller filters and the fast switched mode
operating frequency, the arc pause duration can also be reduced. This is beneficial for the deposition rate and one of the reasons why
current applications are operating with power supplies in DC mode instead of pulsed switching mode.
www.iran-mavad.com
Figure 17 Oscillogram of power supply-output signals at the end of an arc break.
Figure 18 Principal schematic diagram of SCR-based power supply.
Using power supplies in bipolar pulsed mode gives advantages in reactive processes, especially in oxide processes. One
advantage of a pulsed sputter supply is that sputter cycles in a period are relatively short. Since an arc discharge needs some time to
develop, a number of arcs will be suppressed that might otherwise have developed when a DC-sputtered cathode would have been
applied. A bipolar pulsed power supply output reduces arcing and helps actively to prevent target poisoning (40), especially in
processes where insulating coatings are produced (Figure 19). The period of the target voltage has a negative cycle, which is the cycle
where actually sputtering occurs, and a positive cycle to attract electrons to discharge the positive charge collected on the insulating
layer, which formed on the target by the reactive process.
During the positive cycle the electrons collected on the insulating layers cannot be displaced due to the high resistance of the
coating. These electrons must discharge the positive layer accumulated by positive ions during the negative cycle, which requires
correspondingly a sufficient duration of the positive cycle. After renewed change in polarity, the accumulated electron charge will
increase the potential on these places locally to levels above the cathode voltage, which will attract argon ions at higher energy to
these spots. This helps remove the insulating layer by sputtering in the beginning of the negative cycle.
However, in pulsed mode, the duty cycle limits the sputtering yield and therefore reduces the output power.
www.iran-mavad.com
Figure 19 Bipolar effects on sputter cathode, reducing the chances for arcing.
Current optimizations of power supplies in glass coaters have also led to a reduction in overshooting by fast arc detection, which
allows fast detection of an arc on the cathode, but with the fast acting IGBT’s and small filter components, a fast recovery of output
power is possible as well after the end of the arc pause duration. As such the main advantage for productivity (deposition rate) is that
no reduction of sputter yield is happening compared to the case with pulsed power supplies, where a reduced power is inherent to
the pulse duty cycle.
4.21.3.4 Coating Architectures of Carbon-Based Coatings

Sputtering is applied for carbon-based DLC coatings. An overview of the most applied tribological coatings is given in Table 1. As
mentioned earlier, substrates must rotate since sputter and arc processes are line-of-sight processes. Because of the need to maximize
the load, it is often necessary to have the products rotating with either threefold or twofold rotations.
4.21.3.4.1 a-C:H:W Coating

Tungsten-containing amorphous carbon coatings (a-C:H:W) is a coating with relatively high hardness (ranging from 8 GPa to
20 GPa (Young’s modulus ranging from 150 to 200 GPa)). The coating a-C:H:W is not a tungsten carbide, but contains much more
carbon than needed for stoichiometry (patent Dimigen (17)). The consequence is that the coating will be nanostructured, with WC
nanoparticles embedded in an a-C:H matrix. In the next section, a description is given of the process to produce this coating.
The process starts with a pumping and heating step to degas the system and the substrates, thus creating the required vacuum
conditions. Before the coating steps are started, etching is required to create a properly conditioned interface between the adhesion
layer and the substrate. During the etching step, residual oxides and contamination still present on the substrate surfaces after
aqueous cleaning have to be removed.
Etching can be done with metal ions (requiring arc cathodes or high-power impulse magnetron sputtering cathodes that are
expected to be available soon for industrial application) or argon ions (in the past by abnormal glow discharge, in the state of the art
with an argon ion source). Metal ion etching with arc sources has the disadvantage that droplets will create roughness in the base
Table 1 Overview of the carbon-based coatings most often applied to components and their main properties
Carbon-based coatings
Me-DLC (WC-C:H) DLC a-C:H Doped DLC a-C:H-X a-C:H sputtered C ta-C
Method PVD/PACVD PACVD PACVD PVD PVD

Film nr. according VDI 2840 2.6 2.4 2.7 2.4 2.2
Hardness (HV 0.05) 800–2200 1500–3500 1500–2500 2000–4000 3000–8000
Coefficient of friction 0.1–0.2 0.05–0.15 0.05–0.1 0.05–0.1 0.02–0.1
Internal stress (GPa mm–1) 0.1–1.5 1–3 1–3 2–6 1–3
Thickness (mm) 1–10 1–3 1–3 1–3 1–3
Industrial use þþþ þþþ þþþ þþþ þþ
Mass production þ þþþ þþ þþþ þþ
Note: In the row for “Industrial use,” the “þ” signs indicate how suitable the corresponding coating is for use on automotive components at present.
In the row for “Mass production,” the “þ” signs indicate how wide the corresponding coating is already used for coating of automotive components at present.
www.iran-mavad.com
coating, which will be amplified by the subsequent growing sputter coating following the contour of its base. This roughness is the
reason to use argon ion etching instead.
Between the alternatives of using an abnormal glow discharge or a separate argon plasma source, a glow discharge requires the
substrate to be biased with high voltages in the range of 400–700 V. At these voltage levels, argon ions will be implanted in the
substrate and this will cause embrittlement of the substrate. Other issues are that the etching rate is relatively low and there is
a considerable heat load on the substrate. Increasing the etching rate and the etching depth is possible by pressure increase, but then
the power input to the substrate is increasing as well, with the risk of overheating substrates.
The substrate bias can be applied with RF biasing, but generally the bias voltage will be pulsed DC.
The alternative solution for etching is to use an argon ion source. Preferably the plasma source should be a line source. This can
be a Kaufmann source, operated with DC voltage, or a microwave or RF-operated ion source. The higher frequency sources will
create highly ionized plasmas. The use of any plasma source has, however, the possibility to etch with much lower bias voltages.
Generally, it is sufficient to use voltages of approximately 100–300 V to create enough argon ion bombardment for etching the
substrate. Increasing the ion current is possible by increasing the current/power of the plasma sources. Due to the substantially
higher ionization compared to the glow discharge it is possible to reach high plasma densities around the substrate and high bias
currents at pressures in the high 104 mbar range. Adjustment of the etching conditions is now easily possible by increasing the DC
current, which increases the number of argon ions. A higher ionization is favorable for the deeper etching that will be needed when
there are certain substrates with profiled surfaces. This will normally also require longer etching, where overheating of the substrate
must be prevented by the introduction of cool-down periods during these steps. This does not harm the process as long as the
effective etching time will be measured. Effective etching requires that the bias voltage and current are both above a defined level,
ensuring the presence of plasma around the substrate.
The coating must be designed with an adhesion coating as base, where usually Cr is applied but where Ti or Si can be applied as
well. The adhesion layer is relatively thin (a few tenths of a micrometer). The substrate material is relatively soft compared to the
hard PVD/PECVD top coating. In order to achieve good mechanical properties, it is necessary to have a gradual change of hardness
and E-modulus from the steel substrate to the hard top coating. The coating is produced by reactive sputtering, where the reactive gas
will generally be acetylene (C2H2). As a consequence, the sputtered coating will have a nanostructure consisting of WC particles
embedded in an amorphous carbon (a-C:H). In principle, hydrocarbon gases other than acetylene can be used as well, but acetylene
has the advantage that it is a gas with short CH-chains and the lowest H quantity. Additionally, acetylene can be relatively easily
ionized. This enables having the lowest H content, which generally makes the a-C:H matrix harder as long as the optimum hardness
of the matrix is not exceeded (41). Once the optimum flow is exceeded, the coating will be softer. This may be used if a running-in
layer should be applied.
The coating can be tuned to the wear application. It is possible to alter the properties merely by changes in process parameters.
A nice example is visible in Figure 20, showing the differences in structure obtained by changing the rotation speed of the substrate
table. If the rotation speed is high, the structure will be a nanostructured monolayer of WC particles embedded in an a-C:H matrix. If
the rotation speed is substantially reduced, a multilayered coating will be produced, consisting of periods where layers of WC
particles in an a-C:H matrix are alternated by a-C:H layers without embedded WC particles. The monolayered coating will be much
harder than the multilayered coating, which has as a consequence that in a pin-on-disk test the abrasive wear is 5 times less than the
wear of the multilayered coating, whereas the multilayered coating shows twice the ductility in a comparative impact fatigue test.
A test of these coatings on rollers in a roller bearing confirmed an excellent performance of the a-C:H:W coatings. In this test the
dominating wear factor was impact fatigue wear. In comparison to uncoated rollers, the monolayered coatings showed in this test
a lifetime of 350 h compared to 700 h for uncoated rollers, whereas the multilayered coating showed a lifetime of 1125 h. It is to be
noted that the criterion to interrupt the test was wear of the coating or the detection of vibrations on the bearing, meaning that the
Figure 20 Influence of rotation on the structure of a specific W-C:H coating: multilayered coating, created by increase of rotation speed (right) in
comparison with a monolayered structure (left). Mabuchi, Y.; Higuchi, T. Development of Hydrogen-Free DLC Coating and Engine Applications. In Invited
Presentation on World Tribology Congress, Kyoto, 2009.
www.iran-mavad.com
roller system could have survived still much longer since the wear was a gradual wear. Time availability prevented these tests to be
completed accordingly.
It is remarkable that both an a-C:H and a ta-C coating did fail within 1 h. Despite their much higher hardness, the impact fatigue
wear resistance of these coatings was much lower.
4.21.3.4.2 a-C:H Coating Architecture

Amorphous hydrogenated coatings are harder than a-C:H:W coatings. The coating architecture is consisting of a thin sputtered
adhesion layer of Cr, but under circumstances also CrN or CrC as well as Ti- or Si-based adhesion layers can be used. An inter-
mediate layer with similar or higher thickness follows after the adhesion layer, consisting usually of reactively sputtered a-C:W:H in
an argon atmosphere with the reactive gas C2H2 as precursor. Before the sputtered adhesion and intermediate layer are produced,
the substrate is plasma cleaned by argon ions, where the argon ions are either produced in a glow discharge or preferably by an
argon plasma source. On top of the intermediate layer, the process is continued either by sputtering from carbon targets or in
PECVD mode to produce the hard top coating with a thickness varying normally between 1 and 3 mm, The adhesion and inter-
mediate layer are designed usually as gradient layers to adapt the hardness and E-modulus of the substrate gradually to the hardness
and E-modulus of the top coating.
DLC top coatings with thickness exceeding 3 mm can be produced as well, but with growing thickness the stress in the coating will
increase as well (see also (11)).
The hydrogenated a-C:H top coating can be produced from graphite targets in a regular sputter process. This process needs to be
a reactive process, however, in order to give the top coating enough hardness. The hardness values from a sputtered a-C:H is
generally substantially harder than a-C:H produced by PECVD. Sputtering must then, however, be done reactively, since otherwise
the coating will be too soft. For high-volume production, the performance in view of friction can approach ta-C by arc. A big
disadvantage is that the process is slow since that makes the coating expensive, but the friction performance makes that this coating
is used in high-volume production as well.
The widest applied way to produce the hydrogenated amorphous DLC top coating is a PECVD coating, produced with external
microwave or RF-activated plasma sources or by the bias voltage on the substrate creating a glow discharge on the substrate. In the
last case the bias voltage on the substrate acts also as plasma source, where in the PECVD process the plasma sources are cracking the
gas. As mentioned earlier in the discussion about argon ion etching, argon ions would be implemented in the coating in the case of
a glow discharge. They would either resputter coated surfaces, but also they would add to negative effects because of argon ion
implementation at these high voltages, leading to embrittlement of the substrate. In the case of application of a glow discharge, the
substrate is biased either by RF, DC, or DC-pulse power, where DC power is less suitable since the coating is not very conductive and
therefore the process would be disturbed with growing thickness. RF biasing is very well usable if certain configurations can be kept
constant and no need for upscaling exists, since the geometry plays a huge role in the self-bias voltages created during the process.
Another disadvantage of RF biasing is that arc detection is difficult to manage because of the energy stored in the matchbox. These
disadvantages do not exist with pulsed DC biasing. The discharge can be well controlled and arc handling will protect the substrate
from arcing damages. The AC component in the pulsed DC power helps bias voltages maintaining the discharge also when the
coating grows thicker.
In the DLC step, it is very important that the plasma is uniformly distributed. This is important for the thickness uniformity, but
it is also to be noted that local overheating with the consequence that the substrate will be softened when it has been overheated or
the temperature of deposition influences the coating structure of the DLC layer. DLC coatings are consisting of a mixture of graphite
and diamond bonds (sp2, resp. sp3) in an amorphous structure. The higher the ratio of sp3/sp2 bonds, the harder the coating will be.
The coating temperature may not exceed 300 C. Above a range of 250–300 C, graphitization begins to play a role and the coating
loses hardness. Also the H content in the coating plays a role in the final hardness of the coating. It is known that H is needed to
create sp3 bonds (42), but above a certain optimum for the H content the coating softens increasingly. This is also one of the reasons
to use C2H2 as precursor. In principle, all other hydrocarbon gases can be used as precursor gases, but the quantity of hydrogen is the
lowest for acetylene.
For Si-containing DLCs, precursors like HMDSO and TMS are used. These are preferred because they can be vaporized under
vacuum conditions at relatively low temperatures. This makes the precursor supply system easier. Tracing is needed to prevent
condensation in the gas supply lines, but it is not necessary to use complex MFCs to vaporize the precursor.
One of the issues to be solved for the DLC step is to reach a uniform plasma distribution. In a glow discharge, local plasma
densities can occur, where the temperature will rise locally and easily exceeds the graphitization temperature. These plasma densities
are caused by the hollow cathode effect.
The hollow cathode effect can be used to heat up parts or to coat parts internally. A disadvantage is, however, that the control of
temperature distributions is difficult, since these effects are local. It is therefore in most cases an undesired effect, especially in large
volume production of DLC coatings.
The effect occurs if a more or less metallic closed shape exists inside the plasma. The free electrons in the plasma will be
accelerated away from one metal side, since it has a negative potential. Once the electron approaches the other metal side, it will be
bounced back and finally the time that free electrons are staying in this space will be extended. This has as a consequence that the
electron density increases, which leads to the denser plasma.
To prevent this hollow cathode effect, there are a few possibilities to avoid it. For this we should look into Paschen’s law
(Figure 21, (43)), stating that there is a minimum of breakdown voltage as a function of pressure and distance between the
www.iran-mavad.com
Figure 21 Paschen curve: Breakdown voltage as a function of the product of pressure (p) and gap between two plane electrodes (d ). Reproduced from
Leyendecker, T. Über neuartige Schneidwerkzeugbeschichtungen auf Titan und Aluminiumbasis. Ph.D. Thesis, RWTH Aachen, 1985.
electrodes in a homogeneous electrical field. Please note: the latter we do not find in an industrial system. Let us assume that we have
a space surrounded by negatively poled metal conductors at a potential V. Free electrons will be accelerated away from the electrodes
by the electrical forces: F ¼ q$E, where the electrical field is approximately reversely proportional to the distance from the electrodes
to the center between the electrodes, since there the field will be neutral. Let us assume now that the pressure p is constant. An
electron will gain energy by the electrical forces and will lose energy by collisions with neutrals. If the voltage V is high, the electrical
force pushing the electrons to the center is high. If, however, the distance to the center is small, the electron will not gain much
energy that it can transfer to neutrals in order to excite these neutrals. If we increase the distance d, the energy supplied to the electron
will be the same (W ¼ e$V) but the possibility to collide with a neutral will increase because the time of movement is elongated.
Once a neutral is ionized, secondary electrons will increase the chances for collisions with neutrals, and as a consequence the plasma
density will increase if a sufficient number of secondary electrons are added to create a stable discharge. At greater distances, the
acceleration of the electrons for the electrodes will be relatively low. The chances for a collision with a neutral are bigger due to the
longer path, but if the collision takes place with a low electron velocity, there will be insufficient transfer of energy to create the
required emission of secondary electrons for a stable discharge. Let us assume now the influence of pressure with a constant distance
between the electrodes. At a low pressure, the electrons are accelerated but the chances for collisions are low. By an increase in
pressure, this chance is increased as well and the possibility to reach an avalanche effect of secondary electrons for a stable discharge
is growing. If, however, the pressure increases too much, the electrons are already colliding with neutrals before they accelerated
enough to increase the excitation state of the neutrals. By stopping the rotation, the process operator can observe the discharge and
see where the problem is that has to be tackled.
Understanding this, it is now clear which parameters can influence the discharge. Once a local increase of plasma density has
been detected, the pressure can be changed or the local distances between the electrodes can be changed. In this way, the discharge
conditions are moved away from the minimum, similar as described in Paschen’s law. Changing the voltage is also an option. An
increase in voltage will, however, lead to a more intense plasma and lead ultimately to an arc discharge, disturbing the plasma
conditions completely. A decrease in voltage may help, but the consequence will be that at other places in the equipment the voltage
will be insufficient, decreasing the deposition rate there and possibly even extinguishing the discharge. Another process parameter
can also be the use of magnetic fields, for instance as this is created in a closed field configuration by coils around unbalanced
magnetron cathodes. The magnetic field will influence the plasma distribution, increase local plasma densities, and may therefore
help increase the ionization and eventually help prevent nonuniformities.
In a PECVD step, the consequence of a discharge is that the precursor gases are cracked, which leads to an increase in pressure
after the plasma ignites. The discharge should be able to master this and should not be extinguished directly by arcing occurrences.
Besides, it is important to have good vacuum conditions, since degassing of oxides from water vapor is not desired during this
process step. This is also the reason why the system needs heaters. The system will be heated up to the temperatures prevailing
during the deposition steps, thus preventing degassing during the process.
4.21.3.4.3 Ta-C Coatings

The hardest coatings can be made with ta-C (see (44,45)). This is a coating type that has the highest ratio sp3/sp2 and is therefore
among the hardest of today’s coatings. The coating can be produced with regular arc cathodes, with laser-arc, or with filtered arc. Due
to the high hardness, a well-designed adhesion layer is needed here as much as with the other DLC coatings. The deposition rate for
carbon arc is very high, but the roughness is very high as well (when not being filtered and in comparison with sputtering and
www.iran-mavad.com
PECVD production of a-C:H). A big advantage of ta-C is its high potential to produce low-friction coatings, and even superlubricity
is possible in combination with suitable oil additives (46).
Because carbon arc discharges produce high quantities of carbon ions, ta-C processes do not require additional reactive gases.
A problem with standard arc cathodes is that the cathode needs a design with very strong permanent magnets behind the cathode,
which are often moved around. The magnet strength is needed because carbon has a negative resistance temperature coefficient.
Therefore, the cathode spot has the tendency to dig itself in on the spot where it is running, since the resistance decreases here with
increasing temperature. The magnetic field needs to be strong to be able to drag the cathode spot away from where it was. As this is
related to fast erosion, the magnets will generally be moved along the target. Other designs use unmovable magnets, but in all cases
round cathodes are the preferable shape. Rectangular carbon arc cathodes are not used, since controlling the arc spot on those
cathodes is extremely more difficult.
Laser arc is a different design. Here the target is a rotating cylinder, supplied by a pulsed cathode supply with a current
capability of a few thousand amperes. The current pulses are in the range of 1000–2000 A. A laser is used for the cathode
ignition. The laser energy will evaporate locally target material on the spot where the laser is directed. This creates some
pressure increase and local ionization and has the effect that the supply voltage of the power supply will be sufficiently high
to keep the arc running. The arc pulse duration is around 100 ms and is then stopped by the supply, cutting off the energy to
the arc. Once the arc ignites, the high current splits in several repelling cathode spots, moving away from the ignition point in
a starlike shape until the cathode supply extinguishes the spot. The laser scans the target/cylinder mantle over its full height by
means of mirrors. The mirrors are protected against parasitic coating by a foil, which moves forward automatically if the
ignition becomes problematic. Scanning of the laser over the mantle of the cylinder in combination with the rotation of the
cylinder creates uniform erosion. The cylinder does not need any magnetic field and therefore the cylinder can be made very
thick for a long target life time.
Both regular laser arc and standard arc concepts will have a high deposition rate, will create coatings with tunable and very high
hardness levels, but will produce a rough coating that needs posttreatment. By filtering the arc, the disadvantage of droplets can be
avoided. The coating will be very smooth and will not require any further posttreatment. The drawback is loss of deposition rate. By
filtering, the neutral droplets are stopped by blockades (47). Magnetic fields guide the electrons and the ions follow the electron
paths. A disadvantage of filtering is, however, that there is a large loss in deposition rate. Usually, one has to take into account that
only between 1 and 5% of the deposition rate will be possible with curved filters. Current developments with straight arc filters are
promising that the deposition rate will be approximately 40–50%. Since the rate is very important for the cost of the coating, in
most cases the additional cost of posttreatment will make a solution with an unfiltered arc more favorable than a solution with
filtering.
4.21.3.5 Automation Requirements

PVD processes are so complicated that even for the most-experienced experts it is impossible to operate the machine manually while
monitoring all process variables at the same time. During actual running processes it is impossible to monitor the deposition of the
coating layer directly. Some efforts have been made to introduce in situ layer thickness monitors (where the oscillation frequency of
a crystal changes with thickness of the coating and can be evaluated) and in situ ellipsometry (where the optical properties of the
coating are evaluated, thus relating to thickness of the coating, and even in situ optical properties can be tuned for optical coatings).
Mostly in situ measurements are used for calibration of processes and not for direct in situ process control once a process is opti-
mized, since it adds to maintenance requirements. Additionally, it is important that the in situ devices are representative for a whole
batch. This is much easier for flat substrates than for three-dimensional and threefold rotating substrates.
The control system of industrial PVD equipment is normally made up of a programmable logic controller (PLC) or program-
mable controller communicating over a network to a computer running a SCADA (supervisory control and data acquisition)
system. The SCADA system is used for programming of the parameters in recipes, alarm handling, data acquisition in reports, and in
trending system operation as well as displaying the actual state of the equipment parts.
The PLC has built-in communications ports. Usually these are 9-pin RS-232, but also often Ethernet. Other options include
various fieldbuses such as DeviceNet or Profibus. Unlike general-purpose computers, the PLC is designed for multiple inputs and
output arrangements, extended temperature ranges, immunity to electrical noise, and resistance to vibration and impact. Programs
to control machine operation are stored in memory backed up by battery or nonvolatile memory. A PLC is an example of a hard real-
time system since output results must be produced in response to input conditions within a limited time, otherwise unintended
operation will result.
PLCs are well adapted to a range of automation tasks. In the example of PVD/PECVD coating equipment, the software has been
developed in modules, thus keeping the cost of developing and maintaining the automation system low. One of the reasons to
justify this choice was initially the flexibility, which is high. PLCs contain input and output devices compatible with industrial pilot
devices and controls. Less electrical design is required, and the design problem centers on expressing the desired sequence of
operations.
PLC applications are typically highly customized systems, so the cost of a packaged PLC is low compared to the cost of a specific
custom-built controller design. On the other hand, in the case of systems of mass-produced goods, customized control systems are
economic due to the lower cost of the components, which can be optimally chosen instead of a ‘generic’ solution and where the
nonrecurring engineering charges are spread over thousands or millions of units.
www.iran-mavad.com
In the case of PVD equipment, the systems are preferentially split up in functional groups (subsystems). Each group is defined by
its hardware inputs and outputs and its functional requirements. The whole PVD system can then be modularly configured out of
these defined functional groups. The functional groups can then be standardized.
In recent years, ‘Safety’ PLCs have started to become popular, either as standalone models (Pilz PNOZ Multi, Sick, etc.) or as
functionality and safety-rated hardware added to existing controller architectures (Allen Bradley Guardlogix, Siemens F-series,
etc.). These differ from conventional PLC types as being suitable for use in safety-critical applications for which PLCs have
traditionally been supplemented with hard-wired safety relays. For example, a safety PLC might be used to control access to
a robot cell with trapped-key access, or perhaps to manage the shutdown response to an emergency stop on a conveyor
production line. Such PLCs typically have a restricted regular instruction set augmented with safety-specific instructions designed
to interface with emergency stops, light screens, and so forth. The flexibility that such systems offer has resulted in rapid growth of
demand for these controllers.
Most important is that the control system must ensure the reproducibility of the desired coating. It is already known that very
small deviations of signals may indicate process problems resulting in reduced coating quality. Many parameters are more or less
closely interrelated, including parameters that are not measured at all but that do have influences (e.g., composition of residual
gases). It requires very precise checks on the direct measurements to make conclusions about the parameters that cannot be
measured directly.
Also, during actual industrial processes it is impossible to monitor the layer deposition directly. The control system must ensure
the reproducibility of the desired coating.
For this reason, warning bands and alarm bands are used on all relevant measured analogue values and most devices are
monitored by digital signals as well. Trend signals are visualized to allow monitoring of the real-time signal properties in the course
of a running process. Historical trending allows looking back to the history of earlier processes, which can be applied for direct
comparison of present situations with the past. Warnings, alarms, and trend signals are features that are not required for regular
processes, but for failure analysis they are building extremely useful tools.
The equipment manufacturer should also realize that alarms and warnings must be meaningful. It makes no sense if systems are
continuously creating warnings, since that makes the operators much less alert for the really important failures.
In a production environment, the user wants a user-friendly operating interface, which stands for simplicity and reproducibility
of processes. Because of simplicity, every item more than push buttons for START and STOP must have a well-considered purpose.
For reproducibility reasons, the software should provide the equipment user with all required tools to accomplish this
requirement, taking into account that under some circumstances the reproducibility feedback may come only after weeks or even
after months. This can be the case with some tool coaters, for instance, since not everywhere is the performance and lifetime of a tool
checked because the tool is used not always for one and the same operation. In automotive manufacturing, quality assurance is at
a level that includes a much wider field of quality control, which is tightened after each incident that slipped through the system and
this way the quality is improved continuously.
The SCADA software operates with different machine operation modes to operate the system in automatic or manual mode.
Manual mode is meant to enable service technicians access faster to the system operation function. This can be regarded as
a reduction of waiting time, which is required in automatic mode. Also, the service technician can use some settings that would
endanger the process quality during normal processes, but which are not harmful for the equipment. In manual mode, however,
a high degree of protection of the equipment against damage should be guaranteed. Only the protection against abusive settings
cannot be guaranteed, but this is inherent in the fact that the service technician should have been trained. Also, the SCADA software
will include operator access levels. The operator access levels are normally split in levels, relating to a job function. The distribution
of access functions is therefore related to the way how an organization functions and how it has split up work functions. The access
levels are password protected and the passwords given by the responsible manager will be coupled to access levels. The possible
levels have to be organized such that the lowest level of access has the least access functions, whereas each next higher level allows
access to the previous level. The highest level is for the process developer, with the exclusion of one next higher level for the system
administrator.
The SCADA software should include an easy recipe handling system. Loading a recipe and starting it should be available to the
operator, running the system in automatic mode.
Other main functionalities of the SCADA software are the data acquisition functions trending and reporting. Usually trending is
available for checking the running process and making direct comparisons. Trending will always be on a time scale, which should be
expandable or compressible in time as well as in scale. Trending should contain predefined groups of trend signals, whereas the user
should have additional possibilities to define and save new trend groups. Reports should be found back by batch numbers (which
should be generated automatically) or by date and time. For limitation of the amount of data, it is necessary that the user can define
sampling periods for reports.
Modem access through the SCADA computer is a feature that is important for remote software support by the equipment
supplier to the user. This feature is very useful in case of large distances, but also in case of closer proximity between user and
equipment supplier this feature can be useful if the equipment is for instance located in a cleanroom. Such cases can also be very
beneficial for the user of the equipment when their service department is centrally organized.
The alarm handling is located in the SCADA software as well. The alarm condition from the system will be passed from the
equipment hardware to the PLC and from there to the SCADA software. Acknowledgment must be communicated with the PLC,
since real-time handling of the alarm will be needed.
www.iran-mavad.com
The access to available functions should be clearly indicated, but it should also be directly visible which functions cannot be
accessed. In this way, the user gets direct feedback on his actions and he will immediately feel comfortable to operate the system.
4.21.4 Maintenance of Batch and In-Line Systems
Maintenance of batch systems and in-line systems is of great importance for the availability of the systems and has therefore,
especially in case of in-line equipment, a huge influence on the cost of the coating per part. General issues concerning vacuum
conditions, target utilization/target replacement, dust formation, and the consequences of overall maintenance for the coating cost
per part is discussed in the beginning of this section. Subsequently these issues are discussed in more detail.
Shield types are applied to protect system parts for a number of reasons. Electrical shields have electrical shielding functions.
Other shields protect only against parasitic coating on chamber parts, which would otherwise be difficult to clean fast. The design
considerations for the shields are addressed. Dust protection shields protect the substrate table against falling particles. Heat
protection shields, both for heat insulation and for heat transport/cooling, and their requirements are the next topic concerning
shield design.
Maintenance between batches is required. Vacuum cleaning, as well as information about the selection criteria for vacuum
cleaners and checks of insulation/conductivity of electrical shields, is necessary. The importance of shieldings around cathodes is
explained, especially emphasizing how failures in insulation with arc cathodes have to be handled.
The maintenance of the substrate feedthrough is another important issue, since especially at high voltages insulation failures will
start playing a role.
Target replacement is a next topic. Target erosion and the role of controlling the path of the arc spot by magnetic fields are
explained. The replacement of targets is also discussed. Some design requirements to arc and sputter targets are addressed.
Maintenance of in-line systems is much more intensive than maintenance of batch systems. For glass coaters, it is common
practice to design the system such that it will be maintained in two week cycles by complete maintenance crews.
Conditioning of process chambers is often necessary before an in-line system can be restarted for running production.
A practical encountered example (Example of oxide coating) showing the necessity of conditioning included a system where an
oxide coating was used for tuning the electrical resistance in a chamber. The equipment was a two chamber load-lock machine, with
a pump-down/etch chamber and a process chamber. This design was chosen because the oxygen flow accuracy needed to be kept
accurate within a few tenths of an sccm in a range of 20 sccm. The customer’s policy was to switch off all machines on the weekend.
After the weekend there were always quality problems, which was due to coating composition deviations. Tracing this failure was
possible from the registered vacuum measurements and the pumping time and leak rate measurement data. As the relation with
weekend switching off was found, it was clear that the problem was due to the system cooling down and thus related to residual gas,
altering the process atmosphere conditions.
This example and the discussion about loss of capacity with a standing system show that it is necessary to design all system parts
in such a way that the required time for scheduled maintenance should be minimized in order to keep the system running as long as
possible (of course without taking the risk to run into unscheduled interruptions of normal in-specification system operation).
4.21.4.1 Target Lifetime and Cost Considerations

One important possibility to maximize scheduled maintenance is the provision of targets having long life. In this respect it is
nowadays often good practice to replace planar targets with cylindrical targets in rotatable cathodes (Figure 22). The latter can be
used for single-target sputtering (in some cases combined with anode supplies, useful to influence the plasma conditions, inten-
sities, and location), but are also applied in dual magnetron sputtering configurations as well. The biggest advantage of the rotatable
cathodes is their lifetime, due to the high target utilization. In planar sputter targets, target utilization for normal cathodes can reach
25%, whereas special designs including magnet bypasses behind the target are able to increase the target utilization to approxi-
mately 40%. For a cylindrical target, however, the target utilization is generally in a range between 75 and 90% (see Figure 23). The
utilization depends also on engineering solutions for the end blocks; in some cases the end blocks are shielded with target material,
which will be partially sputtered away. Since not all required target materials are available as sputter target for rotatable cathodes,
there are some limitations for application compared to planar targets. This is because the target material must be mounted on
a backing tube (Figure 24). For instance, TiAl is not available as target material for cylindrical cathodes. Often rotatable targets are
produced by plasma spraying, bond sleeved, or extruded. Plasma-sprayed targets contain more oxygen, which is generally unfa-
vorable because of an increased tendency for arcing.
The high target utilization of rotatable targets increases target life. The maintenance interval times of an in-line system will be
increased strongly when using rotatables. The high up-time of the equipment reduces the coating cost of each produced component
substantially. The price per cubic centimeter of a cylindrical target is usually higher than for a planar target, but the improved target
utilization means that from each volume unit of the rotatable target less cost for waste material has to be taken into account. As an
example, based on market prices in 2012: for a rotatable target of 9000 cm3 the cost will be V1.50 cm3, whereas for a planar target
of 2500 cm3 the target cost will be V1.20 cm3. The target utilization of the rotatable is 75% (in practice, figures up to 90% are
mentioned), which is more favorable than for the planar target with a factor of approximately 40% (in practice only achievable with
moveable magnets, with fixed magnet positions the utilization would be 25%). From the rotatable target the effective cost of
www.iran-mavad.com
Figure 22 Dual rotatable magnetron cathode for large-area coating. Source: Von Ardenne Anlagentechnik GmbH, Dresden, Germany.
Figure 23 Smaller poisoning zones for rotatable targets as potential zones for arcing. Courtesy: Soleras Advanced Coatings.
utilized material is V1.125 cm3. The effective cost of the utilized material for the planar target is V0.48 cm3. Recalculating this in
both cases to a target of 9000 cm3 yields: cost for a rotatable target is V13.500, from which 75% (6750 cm3) is effectively used
material. Due to the 40% target utilization of a planar target, the same amount of 6750 cm3 effective sputtered material would
require the purchase of target material for a value of V16.875 (corresponding to 20 250 cm3 planar target material).
So in this case the rotatable target material is 25% more expensive per cubic centimeter, but the better target utilization efficiency
makes that the overall cost for the rotatable target only 60% of the cost for the planar target.
Besides the utilization, rotatables can also run at higher power densities than planar targets. This is because the thermal load is
distributed over the whole surface, whereas with planar targets the thermal load is mainly restricted to the stationary path of the
racetrack. Therefore the deposition rate of the rotatable target will be higher. Rotatables have fewer tendencies for arcing than planar
targets, since only poisoning zones can built close to the end blocks, compared to everywhere around the sides of the racetrack for
planar targets (see Figure 23). This reduction of target poisoning will also permit increasing the power, since arcing will be strongly
reduced. Another reason to permit increased power is that the maximum heating zone is not limited to a small section of the target
surface around the racetrack, but evenly distributed over the whole target surface.
The improved target utilization contributes to an increased equipment up-time. Increasing power and decreasing maintenance
time are increasing equipment capacity. Both effects, as well as the lower relative target cost, will reduce the coating cost per piece for
rotatables compared to planar targets.
www.iran-mavad.com
Figure 24 Cross section (schematic) of a rotatable cathode. Courtesy: Soleras Advanced Coatings.
Another consideration for the cost of ownership is the overall maintenance of the system. The available capacity is one of the
most important cost determining factors. Since maintenance time means down-time, it is loss of capacity. This is in principle equally
important for batch and in-line systems, but with in-line systems the consequences are often greater. The time to run the machine
empty, to acquire access to the points of maintenance, and to get the system operational again after maintenance (e.g., to reach the
required vacuum atmosphere) lead in-line systems in general to much more difficulties and need therefore more time than
maintenance for a batch system. A glass coater as example of a large in-line coater is shown in Figure 9.
Large in-line systems are generally maintained in two week intervals. It is very important that a high-volume production system
be designed in such a manner that maintenance down-time is reduced to the absolute minimum. This applies not only for in-line
systems but also for large batch production systems.
To reduce maintenance in a batch system, a design with two doors and an octagonal shape of the chamber (as introduced by
Hauzer in 1990 (34a)) is very beneficial and later it has been widely spread among different types of PVD batch equipment. The
cathodes are located on rectangular flanges in the doors with an opening angle of approximately 120 , allowing easy access to the
cathodes as well as to the parts inside the system. The open access allows also several maintenance technicians to work on
the system. Though the investment cost of the system increases with double doors, this design reduces system down-time and
therefore pays mostly back in overall coating cost reduction.
4.21.4.2 Dust Formation

PVD and PECVD processes generate dust. The cause for dust formation is primarily related to depositions on shields and carriers. For
the coating quality it is necessary to prevent dust falling onto the substrates, since dust will disturb the coating growth and lead to
undesired defects and pores (48,49), adding up to roughness of the coating and reducing the performance of the coating. Pinholes,
for example, degrade the corrosion protection performance strongly (50).
Initially the topic of dust in general had wide attention in semiconductor applications, where the tiniest defects have large
performance consequences, but nowadays there is also much broader attention to dust issues in the field of hard coatings.
Dust formed inside the plasma is also playing a role. Langmuir discovered and studied dust in plasmas in 1924 (51). Problems of
dust formation in plasmas were brought under broad attention with the discovery of the spokes in the rings of Saturn, discovered in
2005 by the Voyager II. Earlier problems in semiconductor production were already related to dusty plasmas. Much attention in
recent research is directed to dusty plasmas (52). Dust particles can be created as nanometer- or micrometer-sized particles, sus-
pended in the plasma. They can be charged in the plasma and behave as plasma (for dimensions up to 10 nm), following the
electromagnetic laws. In the plasma, dust particles can form clusters, resulting in ‘grain plasmas.’ Grains can change from weakly
bonded (gaseous) to crystalline structures. Since electrons are transferring charge, the dust particles will be charged negatively in the
plasma.
4.21.4.3 Different Types of Shield Functionalities

Shields are applied on the walls and cathodes to protect parts from being covered with undesired coating. Other functions of the
shields include creating an electrical shielding function with a defined or a floating potential. Thermal shielding is applied for
www.iran-mavad.com
protection of radiation heat, where the shield must keep the heat losses to the wall low. This is the main function in tool machines.
In component applications, water-cooled shields can be applied to cool the substrate and as such to enable an increased deposition
rate. Finally, dust shields are sometimes applied for protection against dust falling directly on the substrates.
All shields must be easily replaceable, and since often several shields have to be maintained, they should be clearly identifiable.
This allows a fast shielding change and increases the availability of the system. It is very important that shields are installed only on
their intended locations, since mixing of shield positions may create gaps and/or lacking expansion possibilities, leading to bending
and even short circuits.
Shields must be designed in such a manner that they can be easily cleaned by bead blasting, without influencing its designed
shape.
l Shields with electrical shielding function
If shields are designed to have an electrical shielding function, they will be applied to influence the electron paths and thus create
different quasistationary plasma conditions. This influences the plasma distribution inside the process chamber. Shields having
these functions have to be checked regularly. Special attention is needed in case of floating shields since they tend to lose their
floating potential once the process chamber is getting contaminated with parasitic coatings (especially in the case of conductive
coatings). If shields are designed to have a defined potential, the designer can measure and monitor the shield potential during the
process using the control software.
In the perspective of shielding parts from undesired parasitic coatings, the function of the shields is either to prevent short
circuits, which can create plasma on undesired locations and will damage system parts, or to prevent parasitic coatings from
appearing in locations where these parasitic coatings cannot be easily removed.
For the design of the shield, the engineer must consider that the shields will be heated to the process temperature and/or by
guidance of electrical current. They must therefore be designed in such a way that they are allowed to expand. This prevents strong
bending, which would influence the shielding function and could eventually create short circuits.
l Shields for protection against parasitic coatings
In a batch system as well as in an in-line system, the shields have to be exchanged at a certain moment when the coating layers
grown on the shields are getting so thick that they start to flake because of stresses inside the coating. Additionally, the pump-down
time of the system (plays a role for batch systems) elongates due to the accumulation of water vapor during the exposure to ambient
conditions. This is especially important for batch systems producing carbon coatings, since carbon strongly absorbs water vapor and
since here degassing plays an important role because of the exposure to atmospheric conditions after each batch. As shown in the
oxide example, conditioning of the process chamber must eventually also be considered for in-line systems to reach stable process
environments.
Besides the wall shields, also substrate tables will be covered with parasitic coatings. Similar to the wall shields, the coatings will
start to flake if they get too thick because of repetitive overcoating. Depending on the coating type, flaking will start sooner or later.
Here the rule of thumb is that the harder the coating, the more internal stress in the coating, and the thicker the coating, the more the
stress adds up as well. The coated parts of the rotation tables as well as the shields have to be bead blasted.
As mentioned previously, the engineer must consider that the shields will be heated and can reach process temperature. The
shields must therefore be designed in such a way that they allow for expansion to prevent strong bending. In this way the shielding
function will not be influenced.
For cost reasons it is desirable to clean the shields by bead blasting, since it will reduce the maintenance cost. Shieldings need to
be manufactured with sufficient precision at lowest cost, since these are wear parts. Laser cutting is one of the preferential production
methods for this reason. Other requirements for the shields are that they need to be of a material with a proper strength (i.e.,
sufficient thickness) and vacuum compatibility. Additionally, the shields should in most cases be electrically conductive, have
a good thermal conductivity for a proper temperature distribution in case of heat shielding or heat conduction, and not be made out
of magnetic materials. Stainless steel is therefore one of the most suitable materials to be used for shieldings. With proper strength,
shields can be used several times and allow bead blasting to remove the coatings.
Bead blasting tends also to deform the shields, which also requires therefore a thicker sheet material to be used for the shieldings.
Sometimes chemical cleaning is used as well, but this is limited to a few coating types.
In the case of bead blasting, it is recommended that the blaster uses defined blasting material. One example includes a blaster
using alumina grit as blasting material, contaminated with lead. Since the shields are grounded, they will be heated by electron
current as well as process heat. In this case, it was difficult to find the cause of process problems existing at a certain moment, but
much later after the initial occurrence of the problems component analysis showed the problem as being related to lead
contamination. Due to its low vaporization temperature, the vapor pressure of lead was high during the process pressure and
therefore lead could contaminate the coating atmosphere. Further investigation showed then that the root cause was to be found
with the blaster. Future problems were avoided by the blaster using from there on always clean grit for blasting the shields.
l Dust protection shields for the substrate
Another type of shield is a dust protection shield that is sometimes used for dust directly falling from the bearings and other
rotating and moving parts on the substrate. If applied, these types of shields are part of the spindle on the substrate table. They
protect flat surfaces on the substrates to be coated from collecting falling dust. One of the primary sources of dust are the bearings on
www.iran-mavad.com
the top of the substrate table, where the spindles are rotating (see the Section 4.21.3.1). These collect parasitic coating material that
tends to fall due to the rotating spindles. Of course the collected dust is to be removed in regular maintenance before each batch.
l Heat protection shields
Heat protection shields are used for reduction of heat radiation to cooling surfaces in the case of high-temperature systems. Heat
shields are also applied in other cases for protection of components against overheating by exposure to radiation of hot spots or
plasma radiation.
To reduce radiant cooling, heat shields are brought inside the system before the chamber wall.
Calculation of the heat transfer by radiant cooling q (heat flow per time unit) is done by the Stefan–Boltzmann law:
q ¼ kT 4 A;
where k is the Boltzmann constant, A is the surface being heated up by radiation, and T is the temperature of the radiant body.
Because the relationship is proportional to the fourth power of the temperature, it can be shown that for a heat shield the
temperature will be approximately 84% of the temperature of the substrate as source of radiation. By multiplying the number of
shields, this temperature increases further again. Losses through convection have to be avoided. Heat shields help in this way keep
the substrate temperature high, allowing for increasing the deposition temperature strongly, which is especially beneficial in batch
systems for cutting tool coating application. Insufficient heat shielding would in this case require additional heater power, which is
not beneficial for the chamber design, since it consumes space and might in that case even reduce the required number of cathodes
in the equipment design. Tools are in most cases consisting of cemented carbide and they need to be resistant to high temperatures
anyway because of the application. On the other hand, a higher deposition temperature often results in many cases for nitride
coatings in an improved coating quality.
On the other hand, component coatings are in most cases limited by the maximum allowed substrate temperature. Here it is
often necessary to have the highest possible cooling capacity, allowing the deposition rate to be increased, thus decreasing at the
same time the coating cost. Certain systems are even equipped with shields containing active substrate cooling, allowing the
deposition rate to be increased.
In other cases an active substrate cooling can even be used with temperature-controlled cooling media, allowing for additional
cooling and keeping the temperature low enough for the specific process requirements. In such a design, fast coupling of water hoses
must be available to allow a rapid exchange of the substrate carrier without water leakage.
4.21.4.4 Maintenance per Batch

After each batch, it is recommended to clean the system from parasitic coatings with a vacuum cleaner. All exposed shields should be
cleaned. This reduces defects caused by falling dust.
Normally for a high production system the product carrier will be exchanged completely after each batch by loading another
carrier with fresh parts. The table parts covered with parasitic coating will then be cleaned outside the equipment before loading the
next batch.
The vacuum cleaner to be used needs high-capacity cleaning power, but it should also be noted that the design of the vacuum
cleaner must be dust explosion proof; fine metal dust is highly explosive. In Figure 25 this is shown: a wall shield covered with
parasitic coating is cleaned by steel brushing which generates large sparks. To prevent dust collection in large quantities around the
equipment, it is also necessary to have an easy cleanable environment (53). In practice, it is advisable to have an equipment room
with a floor sealed with a painted coating.
In a production environment, the whole table will be loaded and unloaded completely outside the PVD equipment to minimize
the effective production time losses. Sometimes if the products are relatively light, only the holders are exchanged inside the
equipment. This requires, however, also that the process temperatures are low enough to handle the coated parts without risk of
burning. Managing such risks is often easier to do with a roll-in/roll-out table design, where the risk of product falling from the table
is very low. Concerning the effective production time losses, one should note that especially in high-volume production, loading the
substrate carrier is very labor-intensive work. Therefore, in some applications, where continuously similar products are loaded, this
work is automated and uses industrial robots.
The electrical potential for the shields has to be checked carefully as well. Floating shields should really be floating and not be
connected to either ground or bias potential. This can be checked with a standard ohmmeter.
A different issue is the check of insulators as well as floating shields around the cathodes. These shields may lose their required
electrical potential either by improper assembly or by parasitic conductive coatings, making undesired electrical contact with parts
on a deviating potential.
The floating wall and table shields can be checked with an ohmmeter, but this is different for the floating shields that are used
around the cathodes and which play a role in the plasma confinement of the cathode. The insulators get coated during the batch and
should be checked by insulation measurement devices (Megger, which is an abbreviation for mega-ohm resistance measurement).
These Meggers are to be used with 500 or 1000 V measurement voltage, contrary to the measurement voltage of a few volts in
a normal ohmmeter. The difference is that creep resistance failures are measured, which will never be indicated by regular
ohmmeters. Also, flakes will be measured by the high voltage, which would not be noticed in case of using a low measurement
voltage for the insulation resistance.
www.iran-mavad.com
Figure 25 Metal dust is combustible, as shown in this picture where wall shields are cleaned by steel brushing.
The role of floating shields gets increasingly important with increasing currents and power. In the case of arc cathodes, the arc can
run on the backside of the cathode and will lead to severe damage within seconds. One of the possibilities to prevent this is to apply
software monitoring of the cathode voltage. Normal operating voltages will be in the range between 20 and 25 V. Short circuits
should be monitored (voltages below 10–15 V). If the arc runs off the target, the voltage increases to higher values, ranging between
25 and 35 V. If this situation occurs, it is necessary to switch the cathode off for damage prevention. If the failure disappeared, the arc
cathode is reignited a few times. The number of repeating ignitions should be limited, however, and ultimately be followed by
a complete switch-off of the cathode. In a sensitive reactive process, this may be very disruptive and the process must eventually be
interrupted. If such a highly reactive processes uses flow control mode, the targets will be poisoned as soon as one of the cathodes
falling out disturbs the process equilibrium.
Another example can be found in processes using round cathodes. Depending on the distribution requirements of the cathodes
in the case of round cathodes, it may also be necessary to stop the complete process instead of only the failure cathode. Due to the
point source character, the overall thickness of the coating will be out of the specification limits for the thickness distribution.
Similar considerations are valid for sputter cathodes. Since the currents are generally lower, there is less harm for equipment
damage in case of failures. The voltage drops dramatically in case of a failure and the power supply is equipped with arc handling
capability, in contrary to arc sources where the inherent function of the cathode power supply is to maintain the arc discharge. In
case of failure, the voltage deviates little from the normal running voltage, making the detection more difficult than in the case of
a sputter cathode.
The insulation in the substrate feedthrough is designed to insulate the bias voltage from ground potential. Here the creep
resistance plays an important role. During the process this feedthrough is exposed to high voltages during metal ion-etching process
steps. Voltages are in these step set at 1000–1200 V and, additionally, transients are occurring that can reach values of 2.5–3 kV. The
feedthrough is to be designed in such a way that it should be protected against dust and parasitic depositions. The feedthrough
insulators are for this reason designed with creep way elongation paths. By measuring the insulation resistance with a Megger it is
necessary to have a resistance of >10 MU, preferably even N.
Practical experiences show that lower insulation values lead to substrate arcing problems during the plasma processes. These
arcing effects can be detrimental to the coating adhesion when this happens during the etching steps, whereas it may influence the
www.iran-mavad.com
coating growth if it happens during deposition steps. During etching process steps, the duration of arcing should normally be
compensated by longer process duration, since only the effective ion-etching time should be counted in the equipment software.
This reduces the detrimental effects, but it is clear that the process is running under less stable conditions and therefore this is
a situation that needs to be optimized by maintenance of the insulators.
4.21.4.5 Target Erosion of Arc Cathodes

In the case of an arc target, the target is eroded by the cathode spots moving over the target. The cathode spot can be steered by
magnetic fields. These fields will be generated by permanent magnets or electromagnetic coils. Also, solutions are practiced where
permanent magnets are combined with electromagnetic coils. This gives some security for the arc confinement: if there would be
a failure of the electromagnetic coils, the permanent magnets would still be controlling the arc.
In Figure 26 is a rectangular target shown with the center of the erosion track for a particular case. To clarify the explanations,
a three-dimensional Cartesian coordinate system has been defined with its origin on the target surface inside the erosion track,
indicated by a dashed line.
The steering of the arc spot can be done either by permanent magnets (not drawn) or by electromagnetic coils behind the target
(Figure 27), assuming that only one out of the assumed three coils is active at a certain moment. In this example, the middle coil is
activated.
Close to the position where the activated coil is located, the magnetic field will have only a horizontal component and no vertical
component. The horizontal component is responsible for the movement of the arc spot on the erosion track in the y-direction.
The strength of the normal magnetic field component Bn influences the width of the erosion track (Figure 28). The horizontal
component of the magnetic field Bh will be maximum or close to its maximum at the points where Bn will be zero (Figure 27).
Figure 26 View on target surface with erosion track (dashed oval). The assumed origin of the coordination system has been indicated.
Figure 27 Influence of coils on the erosion track position for arc cathodes.
www.iran-mavad.com
Figure 28 Qualitative diagram of the magnetic field (Bn, displayed on the vertical axis) around the main erosion track position (x ¼ 0), showing the
influence on the width of the erosion track, caused by the normal component. The horizontal axis displays the position x on the target, perpendicular to the
erosion track. Units on the normal and vertical axis are arbitrary.
On the negative x coordinates, the direction of Bh on the racetrack in the middle of the target (between y ¼ 0 and the corner
points) will be pointing to the center of the target, assuming the field direction as indicated by the magnetic field lines in Figure 27.
Near the corner points the direction of Bh bends from the positive x-axis direction gradually into the negative x-direction on the
racetrack on the opposite side of the target.
As mentioned, the normal component of the magnetic field pushes the discharge track to the center of the cathode and prevents
the cathode spot from running off the target. The possibility exists that the arc, once it breaks out, can run off the target. To prevent
this, it is state of the art to confine the cathode spot additionally on the target by using boron nitride or floating shields around the
target. Other cathode types use also special machined shapes of the target to confine the arc spot.
The speed and the movement direction of the cathode spot are determined mainly by the horizontal component of the magnetic
field. The cathode spot will experience a Lorentz force:
F ¼ I$l B:
The Lorentz force drives the movement of the cathode spot in the direction perpendicular to the Bh component, thus enforcing
the cathode spot to travel over the points where Bn ¼ 0. Looking into the effect of Lorentz forces, it can be shown that the
electrons escaping to the wall will be forced into a helix shaped track circling around the magnetic field lines. The arc will
therefore run with lowest possible energy on the positions of the target where Bn ¼ 0. The normal component confines the
cathode spot to this point.
The electron current will never leave the target perpendicularly. Other phenomena like local pressure also play a role and lead
among others to the retrograde movement of the arc (54). This means that the Lorentz force equation will not lead to a force in the
y-direction, but also forces will be experienced in the x-direction. The consequence is that a tendency exists for the arc spot to make
a zigzag movement round the position where the normal component equals zero.
In Figure 28 the dependence of the normal magnetic field Bn in x-direction near the center of the erosion track (x ¼ 0) is shown
qualitatively.
The normal component of the magnetic field Bn confines electrons, as mentioned earlier. At a higher value of Bn more electrons
will only be able to reach the anode (normally this is the chamber wall) after following a path that is more or less perpendicular to
the target, as defined by the magnetic field lines. At a lower value of Bn the number of electrons that will escape in parallel to the
target will be substantially higher. The decreased electron confinement influences the movement of the cathode spot. As a result the
cathode spot deviates wider from the basic point x ¼ 0 at weaker normal components, whereas the spot will be confined to a narrow
track close to position x ¼ 0 if Bn is high. This behavior has been explained in Figure 28.
Suppose that for a weak field the breakout of the spot to the side is limited to the point where B ¼ þ/2 (no units). For the low
normal component the freedom of the arc spot to move around the center of the erosion track is then between position 4 and þ4
(blue dashed line). For a strong field, however, the strength þ/2 is already reached at the x-positions 2 and þ2 (red dashed line).
This shows that the erosion track will be narrowed strongly if the normal component increases.
This erosion behavior shows also one of the difficulties encountered when the steering of the arc is made stronger by magnetic
fields in order to reduce the number of droplets in the coating. Work in this field was done in the 1980s by Ramalingam (55) and
Daalder (56). Ramalingam worked on arc coatings and focused on effects of the magnetic field on the molten pool, created by the
arc spot. Daalder worked on the reduction of arc damage on the contacts in high voltage switching gear. Strong magnetic fields are
www.iran-mavad.com
used here to extinguish the switching arc quickly, and Daalder stated that above a certain speed of the arc spot (100 m s1) the phase
change of contact material into plasma was direct sublimation without a liquid phase of molten material. In that case the number of
droplets reduced strongly, since the droplets are created by molten material, and at the same time the erosion of the contact is
limited substantially. The consequence of a high arc spot speed is that the horizontal component Bh needs to be strengthened. This
required the use of permanent magnets (neodymium–iron–boron, samarium-cobalt, etc.) because of their high strength and small
dimensions. For a certain magnetic field, this implies that the normal component Bn will be linearly increased as well. As mentioned
previously, the erosion track will be very sharp. This will rapidly lead to a deep groove in the target. The consequence is that the arc
will prefer to stay running inside the groove, since the magnetic field increases strongly because the distance between the magnets
and the arc spot is much smaller then, even further confining the arc spot to the groove.
In developments of arc steering, the strong reduction of target lifetime was the main reason not to apply this in PVD technology.
The use of permanent magnets is required, as all design solutions had to include magnets behind the target. The developments in
the 1980s included very complex movement patterns for the stacks of permanent magnets to improve the target lifetime. Finally,
this technology became unprofitable because of target erosion issues and was abandoned by the companies working on this.
Other field configurations require three-dimensional fields. These type of solutions were only recently reinvestigated, and some
solutions have been found (Kobe, Hauzer), but these coatings never will be completely droplet free.
After the initial abandoning of the steered arc because of the target utilization issues, other solutions have been tried out by
having several (usually two or three) coils installed per cathode (inner, middle, and outer coil, according to their position relative to
the target).
If one coil only is switched on, this location will be near the position where this coil is located (see Figure 27). One way to utilize
the coils is to magnetize them all in the same direction, where the direction is determined by the polarity of the current flowing
through the coil and operate them one by one. The arc spot will be running on the position determined by the circumference of the
coil, and the erosion can be influenced by alteration of the coil current (determines the magnetic field) and the time when a coil is
switched on (duration of the arc running in a defined track). Alternative solutions include switching several coils on at the same time
and reversing polarities of individual coil fields. Note that parallel switching of coils with one common power supply can be risky,
since it is unknown how the currents are distributed over the parallel circuitry. Therefore it is preferable that each coil should have its
own power supply.
By optimizing the magnetic fields per coil and the duration of the time that coils are switched on, one can substantially increase
the efficiency in consumption of target material. Up to 70% utilization of the target has been realized by optimizing the magnetic
field for certain cathode types. Optimization of the target erosion can be a determining factor for cost efficiency in processes where
thick coatings are to be produced and target costs are high.
4.21.4.6 Target Exchange

The targets on the cathodes must be exchanged when they are at their end of life. The erosion of the target is determined by the
design of the cathode (see the section on sputter cathode and target: magnetic field/UBM coil influence). For arc targets, target usage
and target lifetime are generally expressed in kiloampere hours (kAh), whereas for sputter targets the useable lifetime of the target is
expressed in kilowatt hours (kWh).
First sputter targets are discussed, followed by a description of arc target exchange requirements. Depending on the sputter
cathode size and the target weight, targets are made from one piece or from several pieces. In the example of WC cathodes, the
targets are designed as an assembly of tiles. The reason for this design was the reduction of target cost and lead time for delivery, but
also handling of the smaller and therefore relatively light tiles is easier. Alternative solutions but at much higher cost would have
been tungsten targets out of one or two sections. Tungsten is a very heavy material and such targets cannot be changed by a single
person. For WC targets the cobalt content as binder material is important. Some Co will be needed to make the targets mechanically
stable enough, but for higher Co contents the magnetic properties of Co will influence the behavior of the cathode, as it will weaken
the magnetic field of the cathode.
For cooling purposes, sputter targets need tight clamping in the middle of the target as well outside in order to run the target at its
maximum power density. For a good contact in the middle, screws through central holes guarantee proper cooling. Bending of the
target should also be taken care of, since the racetrack will reach much higher temperature than the fixed cooling water temperature
on the backside of the target.
In case of targets built up from tiles, central clamping must be provided with sufficient clamping force. The brackets clamping the
target on the outer side and in the middle must be replaced by fresh ones on a regular basis and need to be sand blasted since they
collect parasitic coatings. Insulators should be checked for parasitic coatings, since conductivity of these parasitic coatings influences
the designed creep paths. Shapes of shields around the cathodes, especially those shields having a shielding function for insulators,
should be checked for dimensional deviations from their intended design, which can be due to expansion or stress from deposited
parasitic coatings. Bending through thermal stress can again create undesired short circuits.
For arc targets, the cooling does not require central mounting screws since both the efficiency and the actual supplied power to
the target is much lower than in case of comparable sputter targets. For rectangular arc targets it is, however, recommended to have
a boron nitride insulator strip in the middle of the target (Figure 29(a)), which prevents the arc jumping from left to right before
completion of its full regular path over the length of the target on one side, before turning to the other side and proceeding there on
its path in the opposite direction over the target length (Figure 29(b)).
www.iran-mavad.com
Figure 29 Central boron nitride strips in an arc cathode (a) and arc spot movement (b) on a rectangular arc target. Source: Hauzer Techno Coating BV,
Venlo, The Netherlands.
The boron nitride insulators surrounding the target on the outer side and in the middle must be replaced by fresh ones and need
to be brushed until they are clean again since they will be covered by parasitic coatings. If cleaning is not done here, the arc can cross
the middle strip and jump prematurely from left to right. If that happens, the arc spot will evaporate the parasitic coating on the
boron nitride, which will mostly be undesired as well.
Insulators and shields should be checked for coatings. This applies especially for those ones that are influencing creep paths.
Critical maintenance for arc cathodes includes that the shapes of shields around the cathodes should be checked for not being
bended due to expansion or due to stress from deposited parasitic coatings. This applies especially for those shields that have
a shielding function to protect insulators underneath from being contaminated with conductive parasitic coatings.
4.21.5 Project Management Organization
For project management, there are a few phases to distinguish. The first phase is the quotation phase, followed by the operational
phase and ending with the customer service phase.
In the quotation phase the sales department is leading. Initially after first contacts between customer and equipment supplier,
samples are produced for the customer’s evaluation of the coating performance. Preferably, samples are produced on industrial-
scale systems, since it makes upscaling to full loaded systems much easier. First, rough cost calculations are discussed in order to
have the feeling that further cooperation will end with equipment purchase. If the customer is satisfied with the performance, more
detailed technical specifications are set up and a quotation is prepared, encompassing these specifications and cost calculations.
The calculations as shown in Figure 30 connect the process productivity figures for a certain machine type with the investments
and operational system costs.
Figure 30 shows that the capacity is calculated in number of batches per year. This also takes the up-time of the system into
account. The up-time is the time that the system is available for running batches with products. The process duration determines the
number of batches per day.
The operational cost to run the machine consists of several cost groups, which are all recalculated to cost per batch. The
depreciation cost and the interest cost include the investment in equipment, including cleaning line and quality assurance
equipment. The depreciation costs are determined over a certain number of years.
Floor costs include the cost for the building.
Constant costs per batch are costs for maintenance, targets, and utilities. Maintenance costs are assumed to be related to
a percentage of the machine investment and can be used for spares and consumables. The costs for utilities are cost for power, water,
and gas consumption.
Fixturing cost includes the depreciation cost for product holders, where a separate depreciation time is taken into account.
Finally the cost for labor must be calculated. The calculation result will be the cost per batch. After inclusion of a calculation of the
number of parts per batch, the coating cost per part will be the result of the calculation.
The cost calculation and the final specification of equipment and process are quoted. This quotation goes in the final negoti-
ations, leading to a purchase order and order acceptance by the equipment supplier.
www.iran-mavad.com
CALCULATION HAUZER TECHNO COATING SYSTEM Machine XXXX RFQ YYY

CUSTOMER AAA PROCESS TRIBO DATE 1-1-2012
TARGET COST Number of targets 4
CAPACITY Labour days 250 days Target cost material A 500 EURO
Process time 8 h Target cost material B 6500 EURO
Uptime 90 % Number of batches material A 10
Number of shifts per day 2 Number of batches material B 200
Number of batches per day 2.70 UTILITIES Power consumption per batch 200 kWh
Number of batches per year 675 Power costprice 0.1 EURO per kWhr
RUNNING COST Investment PVD 1500000 EURO Cooling water/batch 50 m3
Investment cleaning 250000 EURO Cooling water cost price 0.5 EURO per m3
Investment infrastructure 50000 EURO Gas 1 consumption/batch 10 ltr Ar
Depreciation schedule 5 yr Gas 1 cost 10 EURO per m3
Interest rate 6 % Gas 2 consumption/batch 1500 ltr N2
Investment fixturing 100000 EURO Gas 2 cost 7.5 EURO per m3
Depreciation schedule fixturing 3 yr Gas 3 consumption/batch 700 ltr CxHy
Operators/shift 1 Gas 3 cost 8 EURO per m3
Persons for handling/shift 2 BATCH LOAD Useful diameter 900 mm
Operator cost/hour 35 EURO per hour Useful height 1500 mm
Handling person cost/hour 25 EURO per hour Number of spindels 18
2
Floor area 100 m Spindle diameter 120 mm
2
Floor cost 50 EURO per m per year PRODUCTS Parts/batch 13500
Spare parts (% of investment) 2 % Production capacity per year 9112500 parts year
Depreciation 533 EURO/batch
Interest rate 160 EURO/batch
Floor cost 7 EURO/batch
CONSTANT COST/BATCH Maintenance 52 EURO/batch TOTAL COST PER BATCH 1600 Euro
Target cost per batch 233 EURO Naked cost price/piece 0.1185 Euro
Power cost/batch 20 EURO
Water cost/batch 25 EURO
Gas cost/batch 17 EURO
VARIABLE COST/BATCH Fixturing 49 EURO
Labour 504 EURO
Figure 30 Cost calculation sheet for calculation of coating coat per part.
Directly after order acceptance the quotation phase is concluded and the operational phase starts with a kick-off of the project.
The operational phase ends with the site acceptance at customer’s site. This is indicated in Figure 31 showing the project structure
(including subsystems) and the project handling activities during the manufacturing of the equipment in the operational phase.
During the operational phase the project leader is the main responsible individual for the project. A project team is set up, where
representatives of mechanical, electrical, software and process engineering, purchasing, electrical and mechanical production, and
sales are included. The main responsibility for the project leader is to lead the project team, to communicate technical issues to the
customer, and to monitor the project budget for cost as well as planning. Each of the project members needs to get priority for the
Figure 31 Schematic overview of the project structure and the operational project handling.
www.iran-mavad.com
Figure 32 Engineering activities and required interrelations between different engineering groups.
project and will be hired to the project from his department. Each member of the project team also has the responsibility to report
technical issues to his department manager, who has an advisory background task. Project leaders should normally report directly to
the technical director, who is in charge of the operational activities.
The operational phase starts with an engineering phase, where at first the contract is translated into specifications for each of the
modular subsystems. The engineering phase can be split into a conceptual engineering phase, so that long-lead items can be
ordered, followed by a detailed engineering phase. During this phase, drawings and part lists are produced and ordering of all
required parts can be finished. The conceptual engineering phase can be different for different engineering groups. The flow of
information between the different engineering groups is shown in Figure 32. The basics start from the process engineers, passing
their requirements for the equipment and the software directly to the mechanical engineering and to the software group. The
communication with the software group concerns mainly the communication about the operator interface and parameter
requirements for the SCADA software as well as the definitions of process steps with their algorithms as to be programmed in the
PLC software.
The communication with the mechanical engineering concerns the chamber design, flanges, gas control system, cathodes, and
substrate carriers.
There are direct relations with the electrical engineering group (mainly power supply related). As shown by the arrows in
Figure 32, there are also many feedbacks included in opposite directions, where design choices have consequences for how the
process is to be operated.
When the hardware parts for the system are delivered, the system will be assembled. Cables and hoses are connected between the
different subsystems. The assembled system will be tested mechanically (including vacuum), followed by an electrical test of all
connections. The next step is the system test, which mainly consists of test of the software and adjusting the correct settings of each
part of each subsystem according to its specified values. The system should be tested here on its boundary performance. This step is
followed by the process test, where the parameters for all processes to be supplied will be tuned by the responsible process engineer.
From here on the lead is with the process engineer. The customer will come after this period for the factory acceptance test (FAT),
which takes normally 1–2 weeks and which is mostly combined with training of customer’s personnel. During training, all engi-
neering and production team members are involved, each bringing in his specific knowhow. The customer will normally also have
a group of mechanical electrical/software and process specialists attending the FAT. If the FAT is satisfactory, the customer will sign
the acceptance document and the system will be released for shipment. Finally, the system will be installed at customer’s site and the
FAT will be repeated by the process engineer. During the FAT, the process engineer will give also training on site to the customer’s
representatives. The completion of site acceptance test (SAT) is an acceptance document, which concluded the operational phase of
the project. Sometimes some open issues require that rework needs to be done, which will stay then under responsibility of the
project leader.
After the SAT, the customer service phase starts. During this phase, the customer service department is leading and responsible for
handling all issues with the customer. This can include also modifications, which are often done to keep the equipment technically
up to date. The technical lifetime is often much less than the time of wear of the equipment. The lifetime of a system can add up to
over 20 years, whereas technical developments are following each other in a much faster pace (maximum 5 year periods).
www.iran-mavad.com
Acknowledgment
I want to thank Hauzer Techno Coating for giving me the opportunity to write this chapter. Special thanks for Thomas Krug, Dave
Doerwald, and Werner Fleischer for their feedback, helping to bring forward the quality of the chapter, as well as to all parties
involved externally in providing material.
References
1. Friedrich, C. Dissertation, Tribologische Problemlösungen mit PVD Hartstoffschichten zum Verschleissschutz. Shaker Verlag, 1998, ISBN: 3-8265-3543-X.
2. Fleischer, W.; Bollinger, H.; Göring, E.; Petzold, K. Herstellung von durchkontaktierten Leiterplatten nach dem VGS-Verfahren. Fernmeldetechnik 1974, 14 (6), 194ff.
3. Fleischer, W.; et al. In 6th Int. Cryogenics Materials Conf.-Proc., Kyoto, 1974.
4. Vergason, G.; et al. Early Years of Commercial Cathodic Arc Vapor Deposition: Hot Deals during the Cold War. In Presentation at the 11th Int. Symp Sputtering Plasma
Processes, Kyoto.
5. Leyendecker, T. Über neuartige Schneidwerkzeugbeschichtungen auf Titan und Aluminiumbasis. PhD Thesis, RWTH Aachen, 1985.
6. Knotek, O.; Bohmer, M.; Leyendecker, T. On Structure and Properties of Sputtered Ti and Al Based Hard Compound Films. J. Vac. Sci. Technol. A 1986, 4, 2695.
7. Weissmantel, C.; Bewilogua, K.; Schurer, C.; Breuer, K.; Zscheile, H. Preparation of Hard Coatings by Ion Beam Methods. Thin Solid Films 1979, 61 (2), L1.
8. Arnoldussen, T. C.; Rossi, E. M. Materials for Magnetic Recording. Annu. Rev. Mater. Sci. 1985, 15, 379–409.
9. Erdemir, A.; Donnet, C. In Modern Tribology Handbook; Bhushan, B., Ed.; CRC Press: Boca Raton, FL, 2000; pp 871–908.
10. Grill, A. Tribology of Diamondlike Carbon and Related Materials: An Updated Review. Surf. Coat. Technol. 1997, 94–95, 507–513.
11. Lettington, A. H. Applications of Diamond-Like Carbon Thin Films. Carbon 1998, 36 (5), 555–560.
12. Hauert, R. A Review of Modified DLC Coatings for Biological Applications. Diamond Relat. Mater. 2003, 12, 583–589.
13. Erdemir, A.; Fenske, G. R. Tribological Performance of Diamond and Diamondlike Carbon Films at Elevated Temperatures. Tribol. Trans. 1996, 39 (4), 787–794.
14. Christoph, P. O. Treutler: Industrial Use of Plasma-Deposited Coatings for Components of Automotive Fuel Injection Systems. Surf. Coat. Technol. November 21, 2005,
200 (5–6), 1969–1975.
15. Kuhlgatz, D. Bosch Automotive, A Product History. J. Bosch Hist. 2011. www.history.bosch.com. Suppl. 2 (Robert Bosch GmbH, Historical Communications (C/CCH)).
16. VDI 2840: Carbon Films – Basic Knowledge, Coating Types and Properties, ICS 25.220.99, November 2005.
17. Dimigen, H.; Hubsch, H. Carbon-Containing Sliding Layer. Patent US4525417, June 25, 1985.
18. On-line Version of GOETZE® Piston Ring Handbook, 2008, http://www.federalmogul.com/korihandbook/en/section_33.htm.
19. Röhrle, M. D. Pistons for Internal Combustion Engines – Fundamentals of Piston Technology; MAHLE GmbH. Verlag Moderne Industrie: Landsberg/Lech, Germany, 1995.
20. Araujo, J. A.; Marques, G. A. Engineered PVD Coatings for Piston Rings Application. In SVC Proc., 2010, ISSN: 0737-5921.
21. Andersson, B. S. Company Perspectives in Vehicle Tribology – Volvo. In 17th Leeds-Lyon Symp Tribology – Vehicle Tribology, Tribological Series; Elsevier, 1991; pp
503–506.
22. Mishra, P. C.; Rahnejat, H.; King, P. D. Tribology of the Ring-Bore Conjunction Subject to a Mixed Regime of Lubrication. Proc. IMechE Part C: J. Mech. Eng. Sci. 2009,
223, 987
23. Schnagl, J. Reibungsreduzierung im Antriebsstrang von Fahrzeugen durch DLC-Schichten. In Presented on EFDS Workshop “Amorphe Kohlenstoffschichten –
tribologische Anwendungen und industrielle Herstellungsverfahren,” December 6, 2011.
24. Source: Hauzer Market Estimation, 2011.
25. http://www.kyotoprotocol.com – Kyoto Protocol.
26. EU Press Release IP/07/155. Commission Plans Legislative Framework to Ensure the EU Meets Its Target for Cutting CO2 Emissions from Cars, February 7, 2007.
27. European Commission. Setting Emission Performance Standards for New Light Commercial Vehicles as Part of the Community’s Integrated Approach to Reduce CO 2
Emissions from Light-Duty Vehicles; SEC, October 28, 2009, 1454, pp 21–31.
28. Podgornik, B.; Vizintin, J. Tribological Reactions between Oil Additives and DLC Coatings for Automotive Applications. Surf. Coat. Technol. 2005, 200, 1982–1989.
29. Anders, S.; Anders, A.; Yu, K. M.; Yao, X. Y.; Brown, I. G. On the Macroparticle Flux from Vacuum Arc Cathode Spots. IEEE Trans. Plasma Sci. 1993, 21, 440–446.
30. Monteiro, O.; Anders, A. Vacuum-Arc-Generated Macroparticles in the Nanometer Range. IEEE Trans. Plasma Sci. 1999, 27, 1030–1033.
31. Kano, M. DLC Coating Technology Applied to Sliding Parts of Automotive Engine. New Diamond Front. Carbon Technol. 2006, 16 (4), 201–210.
32. Mabuchi, Y.; Higuchi, T. Development of Hydrogen-Free DLC Coating and Engine Applications. In Invited Presentation on World Tribology Congress, Kyoto, 2009.
33. Pilkington. Surface Coatings. http://www.pilkington.com
34a Münz, W.-D.; Schulze, D.; Hauzer, F. J. M. A New Method for Hard Coatings: ABS™ (Arc Bond Sputtering). Surf. Coat. Technol. 6 February 1992, 50 (2),
169–178.
34. Grove, W. R. On the Electrochemical Polarity of Gases. Phil. Trans. R. Soc. Lond. 1852, 142, 87–101.
35. Sproul, W. D. Ion-Assisted Deposition in Unbalanced-Magnetron Sputtering Systems. Mater. Sci. Eng. A 1993, 163, 187–192.
36. Cohen, Y.; Riess, I. Study of the Time Dependence of Differentially Sputtered Currents from Multiphase Powder Targets. J. Vac. Sci. Technol. A. 1994, 12, 718.
37. Westwood, W. D. Reactive Sputtering, Physics of Thin Films; Academic Press: New York, 1989; Vol. 14, p 1.
38. Berg, S., et al. Modeling of Reactive Sputtering of Compound Materials. J. Vac. Sci. Technol. A 1987, 5 (2), 202–207.
39. Sproul, W. D. Sylvia, B. E. Multi-Level Control for Reactive Sputtering. In Proc. 45th Annual SVC Technical Conf., 2002, p 11.
40. Sellers, J. Asymmetric Bipolar Pulsed DC; ENI, 12/1995. Technical Note.
41. Strondl, C. Nanocomposite W-C: H Diamond-Like Carbon Coatings, 2007, ISBN: 978-90-367-3253-6.
42. Liu, Y.; Erdemir, A.; Meletis, E. I. A Study of the Wear Mechanism of Diamond-Like Carbon Films. Surf. Coat. Technol. 1996, 82, 48–56.
43. Paschen, F. Ueber die zum Funkenübergang in Luft, Wasserstoff und Kohlensäure bei verschiedenen Drucken erforderliche Potentialdifferenz. Annalen der Physik 1889,
273 (5), 69–96.
44. Scheibe, H.-J.; Schultrich, B.; Ziegele, H.; Siemroth, P. Deposition of Superhard Amorphous Carbon Films by Pulsed Arc Sources. IEEE Trans. Plasma Sci. 1997, 25 (4),
685–688.
45. Haubold, L.; Becker, M.; Packer, D.; Schuelke, T. ta-C:X – Depositing Doped Superhard Amorphous Carbon Films Using the Laserarco PVD Platform. In Society of Vacuum
Coaters, Proc.: Albuquerque, NM, 2008.
46. Kano, M.; Yoshida, K. Ultra Low Friction of DLC Coating with Lubricant. J. Phys. Conf. Ser. 2010, 258.
47. Hakovirta, M.; Walter, K. C.; Wood, B. P.; Nastasi, M. Graphite Macroparticle Filtering Efficiency of Three Different Magnetic Filter Designs Used in the Filtered Cathodic
Vacuum Arc Deposition of Tetrahedral Amorphous Carbon Films. J. Vac. Sci. Technol. A 1999, 17 (5), 3077.

48. Cekada, M.; Panjan, P.; Kek-Mer, D.; Panjan, M.; Kapun, G. SEM Study of Defects in PVD Hard Coatings. Vacuum October 29, 2007, 82 (2), 252–256.

49. Panjan, P.; Cekada,
M.; Panjan, M.; Kek-Merl, D.; Zupanic, F.; Curkovic, L.; Paskvale, S. Surface Density of Growth Defects in Different PVD Hard Coatings Prepared by
Sputtering. Vacuum January 27, 2012, 86 (6), 794–798.
www.iran-mavad.com
50. Korhonen, A. S. Corrosion of Thin Hard PVD Coatings. Vacuum October–November 1994, 45 (10–11), 1031–1034.
51. Merlino, R. L. Experimental Investigations of Dusty Plasmas. AIP Conf. Proc. 2005, 799, 3–11.
52. Anderson, H. M.; Radovanov, S. B. Dusty Plasma Studies in the Gaseous Electronics Conference Reference Cell. J. Res. Natl. Inst. Stand. Technol. July–August 1995,
100 (4), 449.
53. NFPA 654. Standard for the Prevention of Fire and Dust Explosions from the Manufacturing, Processing, and Handling of Combustible Particulate Solids, http://www.
osha.gov.
54. Drouet, M. G. The Physics of the Retrograde Motion of the Electric Arc. IEEE Trans. Plasma Sci. 1985, 13, 235–241.
55. Ramalingam, S.; Qi, C. B.; Kim, K. Controlled Vacuum Arc Material Deposition, Method and Apparatus. US Patent 4,673,477, 1987.
56. Daalder, J. E. Cathode Erosion of Metal Vapor Arcs in Vacuum. PhD Thesis, Eindhoven, The Netherlands, 1978.
www.iran-mavad.com

CMP v4 PDF

Hochgeladen von

Dokumentinformationen

Originaltitel

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

CMP v4 PDF

Hochgeladen von

Copyright:

Verfügbare Formate

‫هب انم خدا‬

Gilmar Ferreira Batalha

FILMS AND COATINGS: TECHNOLOGY AND RECENT DEVELOPMENT

AMSTERDAM • BOSTON • HEIDELBERG • LONDON • NEW YORK • OXFORD

British Library Cataloguing in Publication Data

Library of Congress Cataloging-in-Publication Data

For information on all Elsevier publications

Printed and bound in Italy

Project Manager: Gemma Tomalin

Joseph A. McGeough is an honorary professorial fellow in the School of Engineering at the

Syed H. Masood is professor of Mechanical and Manufacturing Engineering at Swinburne

Mustaﬁzur Rahman is currently a professor at the Department of Mechanical Engineering at

4.01 Introduction to Films and Coatings: Technology and Recent Development 1

4.15 Advanced Carbon-Based Coatings 389

Comprehensive Materials Processing, Volume 4 http://dx.doi.org/10.1016/B978-0-08-096532-1.00400-3 1

Comprehensive Materials Processing, Volume 4 http://dx.doi.org/10.1016/B978-0-08-096532-1.00409-X 3

4.02.2 Thermal Evaporation

jvap ¼ aav ðpeq pi Þ$ð2pmkTÞ1=2 [1]

4.02.2.2 Evaporation Geometry

jvap ¼ aav ðpeq pi Þ$ð2pmkTÞ1=2 cosðQÞcosðfÞ=pR2 [2]

jvap ¼ aav ðpeq pi Þ$ð2pmkTÞ1=2 =4pr 2 [3]

4.02.2.3 Evaporation Sources

4.02.2.4 Web Coating

4.02.2.5 Corrosion Coating

4.02.2.6 Solar Cells

Figure 7 Array of aluminum evaporation. With permission of Leybold Optics GmbH.

4.02.2.7 Polymeric Film Deposition

4.02.2.8 Spotless Vacuum Arc Process

4.02.3 Cathodic-Arc Deposition Process

Figure 10 Schematic illustration of a spotless arc-activated deposition.

4.02.3.2 Unique Features of Cathodic-Arc Deposition

where Ec is the cohesive energy of the material.

4.02.3.3 Ignition and the Cathode Spot

High impedance Cooling down

Hot zone with

4.02.3.3.1 Rapid Sequence of Microexplosions

4.02.3.4 Arc Plasma

4.02.3.5 DC Vacuum Arc

Figure 18 Basic setup of a DC cathodic-arc plasma source.

4.02.3.6 Pulsed Vacuum Arc

Figure 22 A simple pulsed cathodic-arc device.

4.02.3.7.2 Control of Particle Evolution

4.02.3.7.3 Particle Filtering

4.02.3.7.3.1 Physical principle of ﬁlters

4.02.3.7.3.2 Different ﬁlter designs

4.02.3.8 Typical Coatings by Cathodic-Arc Deposition

Deposition rate (nm s1) Transition temperature ( C)

4.02.3.8.2.4 Other ternary nitrides

4.02.3.8.3.5 CrOx and TaOx

4.02.3.9 Ultrathin Films and Nanolaminates

4.02.3.9.1 Multilayers, Superlattices, and Nanolaminates

4.02.3.9.2 MAX Phase

4.02.3.10 Applications of Cathodic-Arc Deposited Films

Table 3 Commercial uses of various typical cathodic-arc coatings

Coating category Application Typical coating material

Hard coatings Cutting tools TiN, TiAlN, TiAlYN, multilayers, ta-C

4.02.3.10.1.2 Nitride coatings

4.02.3.10.1.4 The ta-C coatings

Type of DLC a-C:H a-CH or a-C ta-C or a-C

Friction range < 0.02 0.1 – 0.2 > 0.5

Nature of interaction Van der Waals Hydrogen or

Energy (eV/bond) 0.08 0.2 0.4 – 0.8

Environment Inert of UHV Humid UHV

Deposition rate (nm s1) Transition temperature ( C)