Silicon Nitride - Synthesis, Properties and Application (2012)

CHEMICAL ENGINEERING METHODS AND TECHNOLOGY
SILICON NITRIDE
SYNTHESIS, PROPERTIES
AND APPLICATIONS
No part of this digital document may be reproduced, stored in a retrieval system or transmitted in any form or
by any means. The publisher has taken reasonable care in the preparation of this digital document, but makes no
expressed or implied warranty of any kind and assumes no responsibility for any errors or omissions. No
liability is assumed for incidental or consequential damages in connection with or arising out of information
contained herein. This digital document is sold with the clear understanding that the publisher is not engaged in
rendering legal, medical or any other professional services.
CHEMICAL ENGINEERING METHODS

AND TECHNOLOGY
Additional books in this series can be found on Novas website
under the Series tab.
Additional E-books in this series can be found on Novas website

under the E-book tab.
MATERIALS SCIENCE
AND TECHNOLOGIES
Additional books in this series can be found on Novas website
under the Series tab.
Additional E-books in this series can be found on Novas website

under the E-book tab.
CHEMICAL ENGINEERING METHODS AND TECHNOLOGY
SILICON NITRIDE
SYNTHESIS, PROPERTIES
AND APPLICATIONS
EMILIANO JOSE HIERRA

AND
JESUS ANJEL SALAZAR

EDITORS
Nova Science Publishers, Inc.

New York
Copyright 2012 by Nova Science Publishers, Inc.

All rights reserved. No part of this book may be reproduced, stored in a retrieval system or
transmitted in any form or by any means: electronic, electrostatic, magnetic, tape, mechanical
photocopying, recording or otherwise without the written permission of the Publisher.
For permission to use material from this book please contact us:
Telephone 631-231-7269; Fax 631-231-8175
Web Site: http://www.novapublishers.com
NOTICE TO THE READER
The Publisher has taken reasonable care in the preparation of this book, but makes no expressed or
implied warranty of any kind and assumes no responsibility for any errors or omissions. No
liability is assumed for incidental or consequential damages in connection with or arising out of
information contained in this book. The Publisher shall not be liable for any special,
consequential, or exemplary damages resulting, in whole or in part, from the readers use of, or
reliance upon, this material. Any parts of this book based on government reports are so indicated
and copyright is claimed for those parts to the extent applicable to compilations of such works.
Independent verification should be sought for any data, advice or recommendations contained in
this book. In addition, no responsibility is assumed by the publisher for any injury and/or damage
to persons or property arising from any methods, products, instructions, ideas or otherwise
contained in this publication.
This publication is designed to provide accurate and authoritative information with regard to the
subject matter covered herein. It is sold with the clear understanding that the Publisher is not
engaged in rendering legal or any other professional services. If legal or any other expert
assistance is required, the services of a competent person should be sought. FROM A
DECLARATION OF PARTICIPANTS JOINTLY ADOPTED BY A COMMITTEE OF THE
AMERICAN BAR ASSOCIATION AND A COMMITTEE OF PUBLISHERS.
Additional color graphics may be available in the e-book version of this book.
LIBRARY OF CONGRESS CATALOGING-IN-PUBLICATION DATA
Silicon nitride : synthesis, properties, and applications / editors, Emiliano Jose Hierra and Jesus
Anjel Salazar.
p. cm.
Includes bibliographical references and index.
ISBN: (eBook)
1. Silicon nitride. I. Hierra, Emiliano Jose. II. Salazar, Jesus Anjel.
TP245.N8S46 2011
546'.683--dc23
2011053520
Published by Nova Science Publishers, Inc. New York
CONTENTS
Preface
Chapter 1
Chapter 2
Chapter 3
Chapter 4
vii
High Temperature Oxidation of Silicon Nitride
Based Ceramics: A Review
M. H. Bocanegra-Bernal, A. Aguilar-Elguzabal
and J. Echeberria
Low Temperature Preparation of
Phosphate Bonded Silicon Nitride Ceramics
with High Mechanical Strength
Fei Chen, Qiang Shen and Lianmeng Zhang
Optical and Vibration Properties
of Silicon Rich Nitride
V. A. Volodin
Influence of the Additives Amount upon
SPS-Sintering, Metal Joining, and
High Temperature Oxidation of Si3N4-Ceramics
Leonel Ceja-Crdenas, Jos Lemus-Ruiz,
Sebastin Daz-de la Torre
and Egberto Bedolla-Becerril
33
69
101
Chapter 5
Synthesis and Properties of Silicon Nitride Coatings

Robert B. Heimann
125
Chapter 6
Silicon Nitride as Biomaterial

Cecilia C. Guedes-Silva
149
Index
157
PREFACE
Silicon nitride (Si3N4) ceramic is used in numerous applications because of its superior
properties, such as high-temperature strength, good oxidation resistance and low thermal
expansion coefficient. In this book, the authors present current research in the study of the
synthesis, properties and applications of silicon nitride. Topics include the optical and
vibration properties of silicon rich nitride; high temperature oxidation of silicon nitride based
ceramics; low temperature preparation of phosphate bonded silicon nitride ceramics with high
mechanical strength; the SPS-sintering process and use of ceramics for high-temperature
engineering applications; synthesis and properties of silicon nitride coatings; and the use of
silicon nitride as biomaterial.
Chapter 1 - Silicon nitride-based ceramics are inherently unstable in oxidizing
atmospheres at high temperatures. Generally, oxidation of Si3N4 can be divided into active
oxidation, characterised by mass losses under low pressures, and passive oxidation,
characterised by an increase in net mass by the conversion of Si3N4 to SiO2. The nature and
amount of oxides used as sintering aids (which form the grain boundary glassy phase during
sintering) has a great importance on the oxidation kinetics, related to the outward diffusion of
cations from this glassy phase through the oxide layers. It was found that the oxidation
resistance of silicon nitride ceramics depends mainly of quality and quantity of the sintering
additives. Pressure-assisted sintering of Si3N4 ceramics, such as hot pressed Si3N4 (HPSN) or
hot isostatic pressed Si3N4 (HIPSN), exhibit a higher oxidation resistance than the
pressureless sintered Si3N4 (PSSN), due to their lower content of sintering aids. Other factors,
such the atmosphere (water vapour, CO, Ar) or cyclic oxidation, have minor dependence. In
this review, only available and the most relevant sources were included.
Chapter 2 - Silicon nitride (Si3N4) ceramic is used in numerous applications because of its
superior properties, such as high-temperature strength, good oxidation resistance and low
thermal expansion coefficient. Therefore, it becomes one of the most promising candidates as
structural and functional ceramic materials. Previous studies focused mainly on acquiring
both perfect mechanical properties and complete transformation from alpha to beta phase
Si3N4 porous ceramics. However, better dielectric property of -Si3N4 ceramic whose
dielectric constant () is 5.6 and dielectric loss (tg) is 0.003 at room temperature than Si3N4 (=7.9, tg=0.005) should not be ignored because it can be utilized as an
electromagnetic wave penetrating material. In order to prevent the phase transformation from
-Si3N4 to -Si3N4, a new method of low temperature processing is required for the
preparation of -Si3N4 matrix ceramics. In this case, porous -Si3N4 matrix structure with
viii
Emiliano Jose Hierra and Jesus Anjel Salazar
controlled high porosity (30~50%), micropores and guaranteed strength (>50MPa) is needed
for this application.
It is reported that phosphate is a kind of high-temperature stable and low liner thermalexpansion coefficient binder material, and can be obtained at low temperature. In this chapter,
based on -Si3N4 ceramics matrix, the phosphate bonded silicon nitride porous ceramics,
silicon nitride ceramics using phosphate as sintering additives and phosphate bonded silicon
nitride ceramics applied for broadband radome are introduced.
Chapter 3 - In this chapter optical and vibration properties of hydrogenated silicon-rich
nitride (SRN) films of various stoichiometry (SiNx:H, 0.7<x1.3) were studied. The films
were deposited with the use of plasma enhanced chemical vapor deposition (PECVD)
technique. The properties of as-deposited and annealed films were studied using laser
scanning and spectral ellipsometry, Fourier transform infrared spectroscopy, Raman
spectroscopy, and photoluminescence (PL) techniques. According to the Raman data, the asdeposited film, in which the stoichiometry parameter x is below 1, contains amorphous
silicon nanoclusters. The annealings lead to crystallization of these nanoclusters. No silicon
nanoclusters were observed in the as-deposited films with relatively low concentration of
excessive silicon atoms (the case of SiNx:H, x>1); annealings lead to gathering of excessive
Si atoms in amorphous Si clusters. States of hydrogen in as-deposited and annealed films
were studied. Dramatic changes in the PL spectra of the annealed SRN films were observed.
Femto- and nanosecond laser treatments were applied for crystallization of amorphous silicon
nanoclusters in SRN films. The laser fluences needed for crystallization of the silicon
nanoclusters were found for films of various non-stoichiometric parameters. The effect of
laser assisted formation of amorphous Si nanoclusters in SRN films with relatively low
concentration of additional silicon atoms was observed. The developed approach can be used
for the creation of dielectric films with semiconductor nanoclusters on non-refractory
substrates.
Chapter 4 - Technological developments over the past years have created the need for
developing new materials and processes that can better fit the requirements for advanced
applications. Silicon nitride (Si3N4) is one of the most attractive advanced ceramic materials
for structural applications due to its excellent mechanical properties, either at ambient or
elevated temperatures. Although several techniques for sintering silicon nitride are now
available, most of them still require liquid phase formation to confer the ceramic high-density
level. This chapter is focused on the spark plasma sintering of -Si3N4 powder using different
amount of additives. Both, the - to -phase transformation observed along densification of
Si3N4 specimens and the interfaces they form during the ceramic to metal joining are
discussed as a function of the additives amount. Moreover, the Si3N4-ceramics behavior when
exposed to high-temperature oxidation is presented. This chapter might be interesting for
academic and industrial communities concerned with the sinter processing and use of
ceramics for high-temperature engineering applications.
Chapter 5 - Silicon nitride ceramics show excellent mechanical, thermal and chemical
properties that today are being exploited in diverse, technologically challenging areas. The
superior performance of monolithic silicon nitride in service notwithstanding, problems arise
from the inherently brittle nature of the material as well as the difficulties involved in joining
silicon nitride to metallic parts. To overcome these challenges, several attempts have been
made to develop silicon nitride as protective coatings for otherwise ductile, stress-resilient
metal substrates. Such coatings are designed to combine synergistically the strength and
Preface
ix
toughness of a metal with the protective properties of silicon nitride against mechanical and
thermal attack, and chemical deterioration, as encountered, for example, in highly corrosive
combustion gases of aerospace gas turbines and automotive engines. Unfortunately,
deposition of pure silicon nitride coatings by thermal spray technology is impossible since the
material dissociates and, in turn, sublimates at temperatures above 1800oC, and is also subject
to rapid oxidation at elevated temperature. To combat these technological limitations,
addition of sintering aids such as alumina and yttria, stringent control of the viscosity of the
molten silicon nitride-oxide spray droplets, optimization of the heat transfer from thermal
plasmas to the particles, and high kinetic energy of the deposition process are required. In this
chapter experimental solutions leading to hard, dense, and well-adhering coatings will be
presented that include low-frequency pulsed detonation spraying (LFPDS) and high current
pulsed plasma spraying (HCPPS).
Chapter 6 - Among the ceramic materials investigated for biomedical purposes, those
based on silicon nitride have attracted attention due to high mechanical strength, relatively
high fracture toughness and low density. This review reports the main works about the use of
silicon nitride as biomaterial. Dense and porous bodies as well as particles of silicon nitride
are considered in various papers whose procedure of biological evaluation ranged from
bioactivity tests in Simulated Body Fluid (SBF) to in vitro experiments using osteoblasts likecells and in vivo tests using rabbits as animal model. Although the first study related to the
biocompatibility of silicon nitride published in the 80's has shown poor results, a large
number of later studies indicated that this material might exhibit good biological performance
with great potential for high loaded prosthetic applications.
In: Silicon Nitride

Editors: E. J. Hierra and J. A. Salazar
ISBN 978-1-61942-865-2
2012 Nova Science Publishers, Inc.
Chapter 1
HIGH TEMPERATURE OXIDATION

OF SILICON NITRIDE BASED CERAMICS:
A REVIEW
M. H. Bocanegra-Bernal*,1, A. Aguilar-Elguzabal1
and J. Echeberria2
1
Centro de Investigacin en Materiales Avanzados, CIMAV S.C, Chihuahua, Mexico

2
CEIT and TECNUN (University of Navarra), San Sebastian, Spain
ABSTRACT
Silicon nitride-based ceramics are inherently unstable in oxidizing atmospheres at
high temperatures. Generally, oxidation of Si3N4 can be divided into active oxidation,
characterised by mass losses under low pressures, and passive oxidation, characterised by
an increase in net mass by the conversion of Si3N4 to SiO2. The nature and amount of
oxides used as sintering aids (which form the grain boundary glassy phase during
sintering) has a great importance on the oxidation kinetics, related to the outward
diffusion of cations from this glassy phase through the oxide layers. It was found that the
oxidation resistance of silicon nitride ceramics depends mainly of quality and quantity of
the sintering additives. Pressure-assisted sintering of Si3N4 ceramics, such as hot pressed
Si3N4 (HPSN) or hot isostatic pressed Si3N4 (HIPSN), exhibit a higher oxidation
resistance than the pressureless sintered Si3N4 (PSSN), due to their lower content of
sintering aids. Other factors, such the atmosphere (water vapour, CO, Ar) or cyclic
oxidation, have minor dependence. In this review, only available and the most relevant
sources were included.
Keywords: Silicon nitride, active oxidation, passive oxidation, high-temperature, weight gain
Contact author: MSci Miguel Humberto Bocanegra Bernal, Centro de Investigacin en Materiales Avanzados
CIMAV S.C., Miguel de Cervantes # 120 Complejo Industrial Chihuahua, 31109 Chihuahua, Chihuahua,
Mexico; Tel. +52 614 4394801; Fax. +52 614 439 4823. E-mail: miguel.bocanegra@cimav.edu.mx.
M. H. Bocanegra-Bernal, A. Aguilar-Elguzabal and J. Echeberria
1. INTRODUCTION
Silicon nitride, Si3N4 is a promising material for use as structural components in
advanced heat engines because of its superior oxidation resistance, mechanical strength at
high temperatures, and excellent thermal shock resistance [1, 2]. However, most actual
applications for Si3N4 ceramic materials require exposure to high-temperatures under
oxidizing conditions [3]. Taking into account of their unique combination of properties,
silicon nitride and related materials have probably become the most thoroughly characterised
non-oxide ceramics with wide applications under high-temperature conditions [4] including
ceramic gas turbines, high temperature fuel cells, and ceramic cutting tools [5]. Catalyst
support for some reactions such as Pd/Si3N4 for complete oxidation of methane and Pt/Si3N4
for partial oxidation of methane syngas [6] is another important application. Considering that
the market potential for these applications is very high, the development of engineering
ceramic components could be impeded by the high cost of fine ceramic parts [7]. The
importance of the high-temperature properties is considered as important guidelines in order
to design a plan of structural ceramic members at high temperature [8]. The sintering of
silicon nitride is very difficult and high density bodies by means of solid state sintering are
impossible to obtain due to the low auto-diffusion coefficients of this covalent ceramic [9, 1023]. The addition of some oxides to the pure Si3N4 provides the formation of intergranular
liquid phase which aids the further densification of the silicon nitride during different
sintering routes [24-31].
Oxides such as MgO [32] and Al2O3Y2O3 [33, 34] are generally used to promote
densification by the liquid-phase sintering mechanism [20, 35]. These additives, however,
remain as grain boundary glassy phase, which deteriorate the high-temperature properties of
the ceramics such as creep and high-temperature strength [14, 36, 37] having a major effect
on the oxidation behaviour [8, 9]. The nature and amount of oxides such as Al2O3, Y2O3,
Yb2O3, ZrO2 used as sintering aids at high temperatures (17502000 oC) [13, 24] have a great
importance on the physical properties of the oxidation layer depending of the surface oxygen
content of the powders [15, 38-41]. Studies have shown that oxidation occurs on the surface
and in the bulk of Si3N4 materials [42, 43]. As consequence, mechanical properties degrade
due to the changes in the microstructure. Usually, the oxidation behaviour is described by the
weight gain (W), which is measured and the oxidation constant (Kp) is calculated from the
parabolic increase of the weight gain with time [38]. Generally oxidation of Si3N4 can be
divided into active oxidation and passive oxidation. Active oxidation is characterised by the
loss in mass under low pressures (103 Pa at 1000 oC) while in passive oxidation an increase in
net mass by the conversion of Si3N4 to SiO2 is occurred [3, 44].
Although acceptable oxidation resistance is obtained in the passive regime of pure Si3N4,
there may be major problems resulting from the additives used to promote liquid phase
sintering of Si3N4. Studies on the oxidation of Si3N4 ceramic with additions of Al2O3 and
Y2O3 revealed that the oxidation followed the parabolic rate law but with an oxide layer
formed which could not prevent Si3N4 from being oxidized efficiently. This behaviour
implies that the rate-controlling step was not the outward diffusion of oxygen through the
silicate film, but the outward diffusion of cations in the intergranular glass according to the
reports by Weaver and Lucek [45] and Cubicciotti and Lau [46]. Oxidation kinetic, therefore,
depends on the composition and amount of intergranular phase [7], which can be different
High Temperature Oxidation of Silicon Nitride Based Ceramics
depending on the sintering route used for its processing: pressureless sintering (PS), hot
pressing (HP), hot isostatic pressing (HIP). Other factors, such the atmosphere (water vapour)
or cyclic oxidation, have also minor dependence. Up to now most of studies are concentrated
on the experimental aspects mainly for less than 1000 h indicating that these short duration of
experimentation are not enough for structural members [8]. More quantity and accuracy in
experimental data [3], more rigorous experimental control and more experimental results
under different kinds of conditions are required so as to establish a theoretical system for
oxidation of Si3N4. The more important results on oxidation of Si3N4 materials will be briefly
examinated.
2. OXIDATION
Because of its high-temperature properties, silicon nitride ceramics could eventually
replace various metals in engineering applications [47]. With their low-density, low thermal
expansion coefficient, high hardness, high corrosion-resistance and better mechanical
properties make Si3N4 more appropriate material for such applications compared with metals.
However, few problems like life time, reliability and high cost of fine ceramic parts must be
resolved [1, 47]. In particular, these structural ceramics must retain their high-temperature
stability under several conditions when mechanical stresses and chemical attack are
simultaneously present [1, 48-51]. Nowadays, for various industrial and military applications,
there is an increasing need for structural materials that are able to withstand temperatures
between 1300 and 1500 oC [1, 38].
2.1. Passive Oxidation

The interaction of Si3N4 materials with an oxidizing environment is an important factor in
the behaviour at elevated temperatures according to the reported by Tressler [52], Jacobson
[53] and Rendtel et al. [54]. It is believed that silicon nitride-based alloys are inherently
unstable in oxidizing atmospheres [55]. It is thought [55-60] that oxidation of Si3N4 in air at
temperatures between 1000 and 1400 oC occurs by passive oxidation [61] in which a thin
protective layer of SiO2 (s) is formed on the Si3N4 according to the following reaction:
Si3N4 (s) + 3O2 3SiO2(s) + 2 N2
(1)
with an intermediate thin film of silicon oxynitride (Si2N2O), where the free energy change
[62] associated with the reaction is expressed as:
Go = - 1981 + 0.21T kJmol-1
(2)
where T is the absolute temperature (K). The weight gained on oxidation represents the
weight difference between Si3N4 and 3SiO2. Therefore, complete oxidation of a sample
causes a weight increase of approximately 28.5 % [63]. Literature concerning to silicon
nitride [64], frequently expresses the oxidation by the parabolic equation or general parabolic
rate equation:
W2 + A W = kt + c
(3)
where W is the weight gain per unit area, t is the time, and A, k and c are constants. The first
term in the equation (3) is negligibly small for very thin oxide films. According to Ogbuji
[65], the oxidation of Si3N4 is controlled by the diffusion of oxygen through the SiO2 and
oxynitride layers. Similarly, sintered silicon nitride oxidizes more quickly than silicon nitride
powders or chemical vapour deposition (CVD) deposits. Ziegler et al. [17], Beaume et al.
[66], Clarke and Lange [55], Lewis and Barnard [67], Wang et al. [68], and Meara and
Sjoberg [69] suggested that this is related to the presence of an intergranular phase originating
from the sintering aids used such as Y2O3 and Al2O3 [34, 55].
It is very important to stress that the increase of weight is not continuous since the
formation of the SiO2 superficial layer limits the oxygen flow and thereby reducing the
oxidation rate. During the manufacture of silicon nitride ceramics, Y2O3 is normally used as
sintering additive because it reacts with the SiO2 layer covering the Si3N4 surface during the
sintering, forming Y2Si2O7, which is being the most stable compound in the Y-Si-O system as
reported by Lange et al. [70] and Cinibulk and Thomas [71]. On the other hand, taking into
account that during the passive oxidation of Si3N4 ceramics oxygen is incorporated and which
in turn generates and releases several kinds of gases so the weight of the ceramic samples
does not simply increase, but rather also implies secondary reaction by-products [72].
Important investigations show some databases [73-75] related with the passive oxidation of
Si3N4 based ceramics under different test conditions (temperatures between 1000-1400 oC,
under 1-303 h) and weight gains oscillating between 0.21-2.23 mg.cm-2. When Si3N4 is used
as protective coatings on carbon-carbon hypersonic vehicle airframe components, for
example, it is essential that this material remain in the passive oxidation regime.
2.2. Active Oxidation

The active oxidation in Si3N4 ceramics is mainly characterised by a loss in mass
according to the following reaction:
Si3N4 (s) + 1.5O2 (g) 3SiO (g) + 2N2 (g)
(4)
The above reaction occurs under low oxygen pressures of 103 Pa at a temperature of 1000
C. Therefore, if the oxygen potential in the environment is too low for the formation of the
thin SiO2 surface layer, the active oxidation can result in a catastrophic attack which often
leads to the degradation of mechanical properties [49, 50, 76-81]. In a low oxygen partial
pressure near to atmospheric pressure, however, it is uncertain whether the mass loss results
from decomposition or active oxidation [49]. Kim and Moorehead [76] and Sheehan [82]
suggested that the mass loss behaviour might also be affected by impurities which are used as
sintering aids. Several investigations on active oxidation of silicon nitride based ceramics
have been reported by different authors, for example Tripp and Graham [57], Kim and
o
Moorhead [76], and Sheehan [82]. Likewise, the active oxidation rates [52, 76, 83] as well as
the transition of oxygen partial pressures from active to passive oxidation [76, 82] of silicon
nitride have also been reported in Ar-O2 or N2-O2 atmospheres. It is very important to note
that the active oxidation of Si3N4 limits it use as structural ceramics [78, 84]. Different studies
on the mass loss by decomposition and active oxidation for Si3N4 have been carried out by
Tripp and Graham [57], Kim and Moorhead [76], Vaughn and Maahs [78], Sheehan [82],
Batha and Whitney [85], Goto and Hirai [86], Warburton et al. [87], and Kim [88] and all of
them have reported the decomposition of Si3N4 at low partial pressures.
2.3. Effect of Sintering Additives on the Oxidation of Si3N4

It is well known that both sintering aids and sintering powders have a major effect on the
oxidation and high temperature strength of structural silicon nitride ceramics [9, 34, 38, 47,
66, 72]. As was outlined previously, it is necessary to obtain structural materials that
withstand temperatures above of 1400 oC and silicon nitride ceramics exhibit a remarkable
high level of mechanical properties at elevated temperatures [38].
Klemm et al. [43, 51] undertaken an investigation using additions of SiC and MoSi and
found that an effective mechanism prevented the oxidation in a Si3N4 material and a
Si3N4/MoSi2 composite (10 wt% MoSi2), both fabricated with 8 wt% Y2O3 as sintering aid
and tested at temperature of 1500 oC up to 100 cycles and 10000 h, using four bend bars.
Minimal alterations in the microstructure of the composite leading to a moderate degradation
in strength was obtained under experimental conditions above mentioned, meanwhile, the
microstructure of Si3N4 material was strongly degraded with significant loss in the
mechanical properties. The authors suggested that Si2ON2 was the crystalline oxidation
product in the composite material, which was probably stimulated by an increase in the extent
of formation of crystallization nuclei. Similarly, as a consequence, the grain boundary
composition remained nearly constant and the microstructure of the composite was stabilized.
The weight gain reported by Klemm and Schubert [38] was 3.23 and 3.9 mg.cm-2 for Si3N4
and Si3N4/MoSi2, respectively. On the other hand, Guo et al. [89] reported 3.95 mg.cm-2 and
1.30 mg.cm-2 as weight gain in Si3N4-MoSi2-Y2O3+Al2O3 and Si3N4-MoSi2-Lu2O3
compositions confirming that the use of Lu2O3 as sintering aid leads to and improvement in
oxidation resistance for Si3N4-MoSi2 composites. Meanwhile, Klemm et al. [90] reported
investigations on the oxidation behaviour in various Si3N4 and SiC ceramics after long term
oxidation tests (up to 5000 h) at 1500 oC in air, where the diffusion of oxygen into the
material and its further reaction in the bulk of material were found to be the most critical
factors under experimental conditions specified. It was also observed by Klemm et al. [90] in
Si3N4 ceramics that after a longer exposure at elevated temperatures (> 100 h) the damage
mechanisms are furthermore influenced by the oxidation processes leading to degradation of
the lifetime of the ceramic [42, 52].
Electroconductive ceramics [66] such as silicon nitride-titanium nitride (Si3N4+TiN)
composites may offer new opportunities considering the combination of high performance of
Si3N4-based ceramics with machinability by electrodischarge machining (EDM) [34, 66, 9196]. A phenomenological oxidation model was suggested by the authors in order to describe
the complex oxidation behaviour of such composites at temperature > 1000 oC, where several
oxidation processes act simultaneously and/or successively. The oxidation of titanium nitride
phase involves the formation of a rutile or titanium oxide (TiO2) layer at the outer surface of
the TiN particles [97-100], following the reaction:
TiN (s) + O2 (g) TiO2 (s) + N2 (g)
(5)
It is generally admitted that the TiN oxidation is controlled via oxygen diffusion through
TiO2 [97, 99]. A few percent of high purity Y2O3 to increase high-temperature mechanical
properties and Al2O3 to favour the Si3N4 solubility in the liquid phase at the processing
temperature were added to the initial mixture. The researchers, taking into account their
experimental results as well as bibliographical data, suggested that three oxidation modes,
controlled by distinct diffusion mechanisms, take place simultaneously: i) At an early stage,
the oxidation is controlled by the diffusion of titanium through TiO2 which is oxidized at the
oxide/gas interface, with the formation of a continuous rutile (TiO2) layer at the outer surface,
ii) The oxygen diffusion through rutile and through the cavities of the sublayer, are the
predominant mechanisms. This sublayer reacts with the internal titanium nitride particles and
the Si3N4 phase, iii) for prolonged oxidations times, the silicon nitride phase oxidation leads
to formation of a layer of amorphous silica under the surface TiO2 layer where the dominating
process becomes the diffusion of oxygen through the SiO2-rich phase [66].
The silicon nitride oxidation at temperatures of 1000 and 1200 oC was studied by Pereira
and Ribeiro [9] in -Si3N4 with additions of 7 and 14 vol.% Yttrium disilicate (Y2Si2O7)
produced by thermal treatment at 1400 oC during 32 h with the molar composition SiO2 and
Y2O3 of 2:1. According to experimental studies, the tested samples showed oxidized layers
predominantly of -Si3N4 when compared to -Y2Si2O7 phases. Considering the oxidation
time during exposure of the samples to the processing temperatures up to 256 h, both
compositions (Si3N4 + 7 vol.% Y2Si2O7 and Si3N4 + 14 vol.% Y2Si2O7) presented small
weight changes at 1000 oC suggesting no oxidation, meanwhile , at 1200 oC the weight
change was approximately twice the value obtained at 1000 oC, reaching therefore 0.2 mg.cm2
. On the other hand, after 128 h of exposure the oxidation rate was increased indicating the
possible change in the oxidation mechanism. Similarly, the microcracks observed by the
authors [9] in the oxidized surfaces in both experimented compositions, might be due to the
differential thermal expansion between Si3N4 and SiO2. Similar observations were reported by
Houjou et al. [72] for Si3N4 + 20 wt% SiC + 8 wt% Y2O3 after 2000 h of exposure under
atmosphere environment at 1300 oC. Pereira and Ribeiro [9] concluded from their study that,
both compositions can be used at temperatures of 1000 oC without an adverse effect on
oxidation resistance. Otherwise, applications at 1200 oC must receive a careful analysis due to
the significant increase of oxidation rate under long periods of exposure. Experimental results
at temperatures of 1000 and 1375 oC with similar compositions, have also been reported by
Lange et al. [70] obtaining weight gain of 2.0 and 0.2 mg.cm-2 at exposure times of 20 and
303 h at the aforementioned temperatures.
The effect of sintering aids such as Y2O3 (5-8 wt%) and Al2O3 (3 wt%) on the oxidation
behaviour of Si3N4/SiC (SiC addition was 20 wt%) based ceramics at 1300 oC with exposure
times between 100 and 2000 h, was studied and reported by Houjou et al. [72]. At prolonged
exposure times (2000 h), the authors found many cracks on the oxidized surface caused by
difference of the coefficient thermal expansion between base material (in this case Si3N4/SiC)
and oxidized layer during cooling down procedure. For the composition Si3N4 + 5 wt% Y2O3
+ 3 wt% Al2O3 after 100 h of exposure at 1300 oC, the surface is completely covered with a
glass phase (consisting of Si, O, Al, etc.) with the presence of small crystals recognized by
means of electron probe micro analyzer (EPMA) and X-ray diffraction (XRD), as Y2Si2O7
which was probably generated from the following equations:
2/3 Si3N4 + Y2O3 + 2O2 Y2Si2O7 + 4/3N2 (g)
(6)
2SiC + Y2O3 + 3O2 Y2Si2O7 + CO(CO2) (g)
(7)
Y2O3 + 2 Si02 Y2Si2O7
(8)
Besides the above mentioned, several holes were formed on the surface, generated by gas
escape through the oxidized layer [72] due to the gas generated by the reaction between
Si3N4, SiC and O2. It is important to note that the holes disappear from the surface at high
exposure times. On the other hand, the addition of SiC has very few effects on the oxidation
behaviour. The growth of the thickness in the oxidized layers to several ten of micrometers in
the composition Si3N4 + 5 wt% Y2O3 + 3 wt% Al2O3 seems to be responsible to produce large
crystals of Y2Si2O7 inside the layer accompanied by many pores in the range of 20-40 m
from the oxide/ceramics interface. For compositions with only Y2O3 as an additive, the
oxidized layer was very small according to Houjou et al. [72] suggesting an excellent
oxidation resistance compared to compositions with Y2O3 and Al2O3 as sintering additives,
where the oxidized layer grew very thick to about 30 m after 2000 h. Likewise, the authors
concluded that the rate of increasing the thickness of oxidized layer on Si3N4 + 5 wt% Y2O3 +
3 wt% Al2O3 compositions was controlled by the diffusion of metallic ions (Y, Al) and others
which move through the grain boundary of the base material, in addition to the diffusion in
the oxidized layer. The corrosion behaviour of Si3N4 with constant composition (Y2O3-Al2O3
additives) but with ceramics from different shapes (for example plate, cylinder and ball)
revealed that the corrosion behaviour depends on the samples shape when these ceramics are
subjected to oxidation in 0.5 mol/l H2SO4 at 90 oC [101,102].
The development of a material with superior oxidation resistance is very important to
high-temperature structural applications. Although acceptable oxidation resistance is obtained
in the passive regime of pure Si3N4, some problems may result from the additives used in
order to promote the liquid phase sintering in silicon nitride [47]. An improvement of the
oxidation resistance of the LiYO2-Si3N4 system due to the strong reduction of residual Li by
the annealing treatment at 1650 oC, was reported by Lee and Mager [103] in the mixture of
Si3N4 with additions of 7.5 wt% LiYO2 powder without requiring excessively high
temperature, high pressure without sacrifing the mechanical properties as consequence of
annealing treatment. However, other additives as Y2O3 and CeO2 have been reported to cause
cracking and spalling in monolithic Si3N4 surfaces during oxidation tests [70, 100]. These
problems can be overcome by choosing compositions in the Si3N4Si2N2O-Y2Si2O7
compatibility triangle since Si2N2O and Y2Si2O7 are stable during oxidation process. It is
expected that rare-earth oxide systems such as Si3N4SiO2-La2O3 and Si3N4SiO2-Sm2O3 will
exhibit important advantages in the oxidation of silicon nitride [47] at temperature range of
700-1370 oC. The investigated compositions were Si3N4 + 3.6 SiO2 + 4.3 Y2O3, Si3N4 + 4.0
SiO2 + 4.0 Ce2O3, Si3N4 + 3.6 SiO2 + 6.3 La2O3 and Si3N4 + 3.6 SiO2 + 5.9 Sm2O3. At
Weight Gain, W (mg.cm -2)
temperatures of 700 and 1000 oC insignificant oxidation occurred and it is important to point
out that all four compositions were stables during oxidation at 1000 oC indicating that
stability to Si oxidation for the additives experimented is similar to that of Y2O3 as anticipated
by Lange [100]. The specific weight gains experimented in the above mentioned
compositions ranging 0.028 and 0.056 mg.cm-2, 0.034 and 0.086 mg.cm-2, and 0.395 and
0.752 mg.cm-2 for the temperatures of 700, 1000, and 1370 oC respectively. Interesting
differences in the oxide surfaces were found: i) Y produces an oxide having coarse Y2Si2O7
grain according to the report by Houjou et al. [72]; ii) Sm produces very fine grained
Sm2Si2O7 and iii) Ce and La produce intermediate crystallite size.
The resulting microstructures have been observed by other researchers in the systems
Si3N4-CeO2-SiO2 and Si3N4-Y2O3-SiO2 [104-106]. On the other hand, Y2O3 offered the best
oxidation resistance at 1370 oC. The oxidation rates produced with additions of CeO2, La2O3
and Sm2O3 are within an order of magnitude of that produced by Y2O3 additive. Pomeroy and
Hampshire [107] reported similar results in silicon nitride with additions of 17 mol% of an
equimolar mixture of magnesia and neodymia. However, their experimental results showed a
deviation from the parabolic weight gain per unit area-time relationship. It is thought that this
is due to the devitrification of amorphous grain boundary phases, which in turn results in a
marked deceleration in oxidation rate after long exposure times. Larger volumes of such
phases result in more rapid oxidation with great decreases in high-temperature properties.
Echeberria [108] reported a parabolic nature of the relationship between weight gain per
unit area and time for Si3N4 with high additives contents: Si3N4 + 5 wt% MgO (B) (Baikalox
M15), Si3N4 + 5 wt% MgO (L) (Lycal 95/57S), Si3N4 + 6 wt% Y2O3 + 4 wt% Al2O3 (KY64),
Si3N4 + 6 wt% Y2O3 + 4 wt% MgO (KY604), Si3N4 + 6 wt% CeO2 + 4 wt% Al2O3 (KC64). It
is important to stress that in all temperature range experimented (12501500 oC), the
oxidation kinetic was always parabolic according to the following equation:
Time (h)
(a)
Weight Gain, W2 (mg2.cm-4)
Weight Gain, W (mg.cm -2)
(W.s-1)2 = Kp. t + C
Time (h)
(a)
Time (h)
(b)
eight Gain, W2 (mg2.cm-4)
Figure 1. Plots of specific weight gain (a) and parabolic fitting plots (b) as a function of time for the
Si3N4 + 5 % MgO Baikalox (KM5) composition oxidized at different temperatures. [From ref. 108].
(9)
Time (h)
Weight Gain, W (mg.cm-2)
(a)
Time (h)
(a)
Time (h)
(b)
Si3N4 + 5% MgO Lycal (KM5L) composition oxidized at different temperatures. [From ref. 108].
where W is the weight gain per unit area (s), t is the time, C is a constant, and Kp is the
parabolic rate constant. On the other hand, the observed weight gain was more sluggish when
the exposure time is increased indicating a passivation in the oxidation process. This
behaviour can be seen in Figs. 1 to 5 for all tested compositions. It is important to note in
Figure 3b a slight deviation from parabolic behaviour at the higher temperature (1500 oC) for
exposure times between 48 and 72 h indicating a change in the oxidation kinetic leading to a
passivation as mentioned previously, which is associated with the existent macroporosity in
the oxide layer. The calculated activation energies for the above compositions are
summarized in Figure 6. As can be seen in this figure, two activation energy values of 815
Time (h)
kJ.mol-1 (label 1) and 325 kJ.mol-1 (label 2) for temperature ranges 1250-1350 oC and 1400o
(b)confirm that two different mechanisms would control the oxidation in
1500 C, respectively,
this specific composition (KY64). However, some compositions in the Si3N4-Si2N2O-Y2Si2O7
compatibility triangle presented a logarithmic kinetic at temperatures between 1400-1500 oC
due to the high crystalline character of the oxidized layer.
2.4. Effect of Hot Pressing on the Oxidation of Si3N4

The oxidation of Si3N4 alloys hot-pressed (HP) with different additives has been studied
by several researchers as a function of the phase content of the alloys [5, 34, 55, 82, 89, 96,
109-111]. Although the oxidation mechanism of HP silicon nitride is not well understood, it
is important to point out that the corrosion rates are satisfactory under gas turbine operating
conditions [109]. The oxidation of HP Si3N4 with additions of MgO as sintering aid has been
the subject of considerable study in several years ago [82]. An experiment carried out by
Tripp and Graham [57] at 1400 oC in a Si3N4 sample HP containing MgO was exposed to an
oxygen partial pressure of 10-7 Pa and the oxygen pressure then increased, showed a transition
from active to passive behaviour at about 5 x 10-3 Pa.
We
Time (h)
Time (h)
(a)
-2
Weight Gain, W (mg.cm
)
Gain, W 2 (mg2.cm-4)
Weight
M. H. Bocanegra-Bernal, A. Aguilar-Elguzabal and J.(a)

Echeberria
10
Time (h)
(b)
Weight Gain, W 2 (mg2.cm-4)
2
2
-4
-2
(mg.cmW
Gain, W Gain,
WeightWeight
) (mg .cm )
(h)
Figure 3. Plots of specific weight gain (a) and parabolic fitting plots (b) asTime
a function
of time for the
(a)
Si3N4 + 6 % Y2O3 + 4 % Al2O3 (KY64) composition oxidized at different temperatures.
[From ref. 108].
Time (h)
(h)
Time
(b)(a)
Time (h)
(b)
Weight Gain, W 2 (mg2.cm-4)
Si3N4 + 6 % Y2O3 + 4 % MgO (KY604) composition oxidized at different temperatures. [From ref.
108].
Sheehan [82] studied the passive and active oxidation of HP silicon nitride ceramics
containing 1 wt% and 5 wt% MgO contents under helium atmosphere at temperatures
between 1000 and 1200 oC. The author reported that the oxidation of the specimen containing
5 wt% MgO at 1100 oC produced consistent and significant weight losses where the oxidized
surfaces contained a porous glassy oxide and what appears to be a uniform distribution of
discrete oxide particles. In contrast, the active oxidation that was observed at 1200 oC for both
(h)
compositions is notTime
consistent
with the results reported by Tripp and Graham [57] where
(b)
oxygen partial pressures of 5 x 10-3 Pa or lower were required to produce active oxidation at
1400 oC. However, the reason for such discrepancy is not known [82]. The aforementioned
observations suggest that active oxidation occurred at the porous zones, while oxidation was
still passive or of transitional nature in the regions that consisted of purer Si3N4. On the other
hand, Lange [60] reported that at a constant Si3N4 levels, the oxidation rates commonly
increased with increasing MgO/SiO2 up to molar ratios of about 0.5 to 1.0 decreasing slightly
before increasing again at ratios of 2.0 to 2.5. The composition with 5 wt% MgO was more
11
susceptible to both the passive and active oxidation compared with the 1 wt% composition,
suggesting a probable enhancement of the impurity levels in Si3N4 + 5 wt% MgO. Sheehan
Timeextended
(h)
[82] concluded that the findings obtained by Lange [60] can be
to the active
(a)
oxidation regime in Si3N4 ceramics.
Time (h)
Time (h)
(a)
(b)
Si3N4 + 6 % CeO2 + 4 % Al2O3 (KC64) composition oxidized at different temperatures. [From ref.
108].
T (oC)
log K (K2.m-4.s-1)
Time (h)
(b)
KY64 (2) Q = 325 kJ.mol-1

KY64 (1) Q = 815 kJ.mol-1
KY604
Q = 700 kJ.mol-1
KC64
Q = 325 kJ.mol-1
KM5
Q = 465 kJ.mol-1
1/T.104 (oK -1)
Figure 6. Arrhenius plot to determine activation energy for the Si3N4 + 6 % Y2O3 + 4 % Al2O3 (KY64),
Si3N4 + 6 % Y2O3 + 4 % MgO (KY604), Si3N4 + 6 % CeO2 + 4 % Al2O3 (KC64), and the Si3N4 + 5 %
MgO Baikalox (KM5) compositions oxidized at different temperatures. In the figure: (1) is activation
energy for 1250-1350 oC temperature range and (2) for 1400-1500 oC temperature range. [From ref.
108].
12
Likewise, Cubicciotti and Lue [109], investigated the oxidation of a commercial HP

Si3N4 (NC 132, Norton, Co.) in dry oxygen at temperature of 1248 to 1458 oC. The results
obtained showed a parabolic behaviour when the square of the weight gain is plotted as a
function of time, wherein all parabolic lines intersect the ordinate at positive values,
indicating therefore, that the initial part of the curve was slightly more rapid than parabolic.
The authors [109] used for the fitting of their results, the following equation:
W2 = Kpt + C
(10)
in which W is the specimen weight gain, t the time, Kp the parabolic rate constant and C a
constant to account for the initial reaction. These observations are in agreement with earlier
reports of Kofsad [112]. The MgO added to accelerate sintering during hot-pressing as well as
other impurities present in the ceramic body, increased the oxidation rate as was corroborated
by Gugel et al. [113]. Particularly for higher oxidation temperatures (between 1227 to 1377
o
C), the surface presented fissures and pores due to the escape of N2 gas produced at the
Si3N4/SiO2 interface as consequence of the oxidation reaction. These appreciations are also in
good agreement with reported results of Sheehan [82]. An interesting observation is that the
Mg concentration in the unoxidized samples of the silicon nitride increased with the depth in
the sample, but the concentrations of the other impurities such as Na, Fe, and K, did not. In
reason of this, Mg was diffusing out of the material into the corrosion scale and was
intimately involved in the rate-determining step.
The oxidation resistance of silicon nitride alloys HP with magnesia as sintering additive
was studied as a function of the phase content of the alloys by Clarke and Lange [55] at
temperature of 1400 oC. Several works concerning this theme have been reported. For
example, Singhal [59] reported the presence of both cristobalite and steatite phases as well as
Mg, Al, Ca, Fe, Mn, Na and K, in the oxide scale formed by oxidizing a commercial HP
silicon nitride (HS-130, Norton, Co., Worcester, Mass.) between 1100 and 1400 oC. Long
oxidation exposure (approximately 4000 h at 1370 oC) produced a mixed magnesium-calcium
silicate (diopside) which was detected by XRD. Similarly, Cubicciotti et al. [114] reported
similar results for a commercial Si3N4 (NC-132, Norton Co.) also. The majority of previous
works reported that the weight gain vs time curves at all temperatures approximate the
classical parabolic oxidation behaviour, indicating that the rate of oxidation is diffusion
controlled. The system worked by Clarke and Lange [55] was Si3N4-SiO2-MgO taking into
account that its phase equilibria are known and generally agreed up on [60, 115, 116]. The
alloys studied within this system including two series containing 0.833 and 0.755 mol
fractions of Si3N4 and were tested in static air at 1400 oC during 24 and 288 h.
In each of the studied alloys, there was magnesium depletion in the subscale, the extent
of which was again a function of the MgO/SiO2 ratio of the starting composition [55, 82]. It is
noteworthy that the composition at the top of the scale depended on the initial MgO/SiO2
molar ratio, where the crystalline phases were MgSiO3 and Mg2SiO4 when the MgO/SiO2
ratio was > 2 and MgSiO3 was the major phase for MgO/SiO2 < 2. Similarly with other
investigations, the analysis of the weight changes as a function of oxidation time shows that
the oxidation of all samples in two series of alloys obeyed parabolic kinetics up to 288 h
highlighting that the oxidation resistance increases by almost three orders of magnitude as the
MgO/SiO2 molar ratio is changed [55, 82, 114] from 0.1 to 6.0. The secondary phases
therefore control and determine the oxidation of these alloys, where a common finding was
13
that the oxide scale was enriched in both Mg and impurities, as for example Ca and Fe, in
comparison with the unoxidized ceramics. All studies carried out on the Si3N4/MgO alloys
have shown that the Mg and the cation impurities diffuse to the surface during oxidation
process. The initial composition in the Si3N4 alloys HP with magnesia has an important effect
on the oxidation resistance at 1400 oC [55], where the oxidation kinetics will increase as the
Si3N4/MgO ratio increases to 1.6, as observed by Lange [60].
It is well known that the amorphous grain-boundary phase present in HP silicon nitride
has a significant effect on the high-temperature mechanical properties (mainly oxidation
resistance) [34, 110, 117-119] although it is not a substantial phase in terms of the volume
that it occupies. To improve high temperature properties, it is necessary to reduce the amount
of residual amorphous phase and/or to rise the melting/softening point of this phase [111,
117]. On the other hand, increasing the refractoriness of the grain-boundary phase [120-122]
and promoting the crystallization of the amorphous phase [123, 124] it is possible to improve
the oxidation resistance as well as enhance strength retention after oxidation [110]. With these
observations, the Lu2O3-doped HP Si3N4 developed recently has excellent high temperature
strength up to 1500 oC [110, 111, 125, 126]. Guo et al. [110] experimented Lu2O3-containing
HP Si3N4 ceramics with additions of 3.33 and 12.51 wt% Lu2O3 additives in air at
temperatures of 1400 and 1500 oC over 200 h, where the oxidation products consisted of
Lu2Si2O7 and SiO2. In this investigation, the specific weight gain in the 3.33 wt% Lu2O3
composition was always lower than that of the 12.51 wt% one. When is plotted the square of
the weight gain, W2 in equation (10) as a function of oxidation time, t, for the two
compositions, it is observed that the straight lines represent close adherence to parabolic
oxidation kinetics similar to findings of Cubicciotti and Lau [109]. This appreciation suggests
that the diffusion was the rate-controlling mechanism for oxidation. An interesting
observation is that the parabolic rate constants of the two compositions at a temperature of
1400 oC are the lowest in those oxidation rate data reported in RE2Si2O7-Si3N4 materials
(k=0.33 x 10-3 3.3 x 10-3 mg2.cm-4.h-1, RE = Yb, Y, Er, Gd, La, Dy, Sm and Ce) [111, 123]
and RE4Si2O7N2-Si3N4 ceramics (k= 0.15 x 10-3 3.6 x 10-3 mg2.cm-4.h-1, RE= Y, Er, La and
Yb) showing excellent resistance to oxidation [111, 124, 127, 128].
A recent study [129] demonstrated that the phase relationships in the Si3N4-SiO2-Lu2O3
system are similar to those in the Si3N4-SiO2-Yb2O3 system [130]. Guo et al. [110] reported
that also the oxidized surface of 12.51 wt% Lu2O3-Si3N4 consisted mainly of the disilicates
with needlike morphology and with the crystalline Lu2Si2O7 growing out of the silicate phase.
Likewise, cracking of oxide layer clearly appeared for both compositions, behaviour very
similar to the reported by Houjou et al. [72] for Si3N4/SiC + 8 wt% Y2O3. This cracking
behaviour has also been reported in other Si3N4 ceramics oxidized in air at and above 1370 oC
[47, 120, 123, 124] and seem to be due to the different thermal expansion between the bulk
and oxide layer or with volumetric change accompanying transformation of surface
cristobalite from - to -phase on cooling [47, 110, 120, 123, 124]. Moreover, although
cracks were observed in the oxidized layer, these defects did not extended into the bulk due to
that these cracks arrested at the interface between the bulk and the oxide layer. After
prolonged oxidation exposure (200 h), the average reduction in room-temperature flexural
strength than before oxidation was reported as approximately 5% at 1400 oC and 13% at 1500
o
C for 3.33 wt% composition, and it is approximately 7% at 1400 oC and 15% at 1500 oC for
12.51 wt% composition [110].
14
Transmission electron microscopy (TEM) studies carried out by Guo et al. [111] revealed
that the complete crystallization behaviour observed in the Lu2Si2O7-Si3N4 compositions does
not agree with that of glasses segregated at grain boundaries, as predicted by Raj and Lange
[131]. However, Vetrano et al. [132] reported a complete crystallization in silicon nitride
ceramics sintered with additions of Yb2O3 and Cinibulk et al. [133] found a similar behaviour
but with YSiAlON glass additives. Observations made using high resolution transmission
electron microscopy (HRTEM) revealed that the boundaries posses a thin amorphous film
with approximately 0.5-1.0 nm in thickness for the Lu2Si2O7-Si3N4 compositions. Guo et al.
[111] concluded in their investigation that the thinner intergranular amorphous film observed
in Lu2Si2O7-Si3N4 is attributed to forming a more silica-rich intergranular liquid phase during
the sintering process. An important observation in this composition is that the increase in
strength at 1200 oC seems to be due to the relaxation of larger residual internal radial tensile
stresses in the triple-grain junctions induced during the crystallization of the Lu2Si2O7 phase
[111]. Nevertheless, beyond 1200 oC, the strength rapidly decreased and 62% of its strength
was lost at 1600 oC. Tong et al. [134] reported important results about oxidation behaviour of
self-reinforced Si3N4 with additions of Re2O3 where Re = Yb, Lu. For this, they used two
kinds of seeds obtained by heating -Si3N4 powder with Yb2O3 and Lu2O3, and the oxidation
test was developed at 1400 oC during 100 h. The Lu2O3-doped HP Si3N4, as was outlined
above, has excellent high-temperature strength up to 1500 oC because its grain-boundary
phase, Lu4Si2O7N2, has a high melting point and shows extensive crystallization during the
sintering process. The graphs corresponding to the typical specific weight gain curves of
Re2O3-doped Si3N4 were all non-linear fitted results of the raw data. It was found in this
investigation, that the specific weight gain of Lu2O3 compositions was lower than those of the
Yb2O3 case. In other words, Lu-containing Si3N4 materials achieved high oxidation
resistance. The authors revealed that the oxidation behaviour in HP Si3N4 obeyed a parabolic
type of rate law according to the equation (10) [109]. From this it could be concluded that
diffusion was the rate-controlling mechanism for the oxidation [134]. During the oxidation
process at temperature of 1400 oC during long exposure time, the crystallization of the
secondary phase (Re2O3-SiO2 or Re2O3-SiO2-Si3N4 system) probably occurred gradually.
In an excellent work [135], a summary of the mechanical properties and microstructural
stability of silicon nitride blades and nozzles tested in an industrial gas turbine is reported,
taking into account that the major objective in the program, was to improve the fuel efficiency
and reduce NOx and CO emissions in gas turbines by replacement of cooled metallic blades
and nozzles and the combustor liner with uncooled ceramic components. Undoubtedly, the
insertion of advanced ceramic components into the hot sections of existing gas turbine
engines in the United States, mainly in those engines for power-and-steam cogeneration, is of
paramount importance [136, 137]. In order to undertake this investigation, Lin et al. [135]
used commercially available Kyocera SN282 vanes densified via gas pressure sintering (GPS)
and SN281 blades which was densified via gas pressure sintering followed by hot isostatic
pressing (HIP) technology, resulting in a fine-grained matrix structure and substantial
reduction in internal defects. In both materials, Lu2O3 and SiO2 were employed as sintering
additives and the crystalline phases in the two as-received materials were -Si3N4 and
Lu2Si2O7.
The presence of porous Lu2Si2O7 surface layers in SN282 nozzles is probably due to the
recession of silicon nitride grains as the normally protective silica layer and formed because
15
of oxidation reaction, became unstable in the combustion environment [135]. On the other
hand, previous engine testing of SN282 vanes at Roll-Royce Allison also suggested that
Lu2Si2O7 phase was stable in gas turbine environments, and as was reported by Lin et al.
[138], could be a potential candidate for an environment barrier coating with the aim of
protects silicon nitride components in gas turbine environments. Nevertheless, for long term
application within the lifetime of gas turbine engines, the observed silicon nitride material
recession would lead to dimensional instability as well as a possible failure in the
components. Therefore, environmental protective coating systems need to be developed and
engineered so as to protect silicon nitride ceramic parts and ensure long-term microstructure
and chemical stability [135].
2.5. Oxidation of Si3N4 under Different Atmospheres

It is well known that the microstructure of silicon nitride-based ceramics is controlled by
factors such as quality of starting powders, additives amount, manufacturing method,
temperature, sintering and atmosphere of sintering furnace. The selection of a suitable
atmosphere is very important to avoid the decomposition of Si3N4. Likewise, this atmosphere
should not affect the viscosity and composition of liquid phase [139]. With an oxidation
exposure or treatment in impurity and additive cations to the surface from the intergranular
phase, decreasing the volume fraction of that phase and reducing its impurity concentration
[55, 140], the high-temperature strength [141], creep resistance [142], are enhanced. As
mentioned above, there are different alternatives modifying the intergranular phase. One of
them was suggested by McDonough et al. [143], who showed that the heat treating silicon
nitride ceramics in an argon gas atmosphere chemically reduces the crystalline oxide phase
present. A combination of this suggestion with the possibility of combining heat treatments in
both oxidizing and reducing atmospheres [140], at temperature of 1400 oC is reflected in a
decrease of the cation impurities and reduction of the volume of intergranular glass phase.
Some investigations have been made on the oxidation of -Si3N4 in different
atmospheres in the temperature range from 1000 and 1340 oC. Horton [63] studied the
oxidation kinetics of powdered silicon nitride in dry oxygen and dry air at 1 atm pressure and
temperature between 1065 and 1340 oC and obtained oxidation products tridymite and
amorphous silica at 1125 o C and 1067 oC, respectively. From the plot of W2 as a function of
exposure time for several tests, it is observed that the data from a given experiment formed a
straight line for oxidation in two tested atmospheres. Likewise, the linear portion indicates
therefore, the region of parabolic oxidation, where the oxidation rate in dry oxygen was
approximately twice that in dry air at the same temperature. The oxidation in this case, is
represented according to equation (10). Horton [63] concluded that when relatively impure
silicon nitride is submitted to oxidation in dry air or in dry oxygen, a protective oxide layer is
formed which retards further oxidation.
On the other hand, in more recent investigations, Hou et al. [3] worked on the oxidation
of -Si3N4 powder at temperatures ranging 1100 to 1300 oC in air. In their study, they suggest
a new model for predicting the isothermal oxidation process of -Si3N4 powder. It is expected
with the application of this interesting model to get some important message to direct us how
to compare and improve the utilization of -Si3N4 materials. For the oxidation test, the
16
furnace was heated to the oxidation temperature as quickly as possible in the flowing argon
gas. After the thermal equilibrium was reached, the atmosphere was changed to air
atmosphere with the flow-rate of 50 ml. min-1. The major component detected by XRD was
Si3N4 and only SiO2 as crystalline oxide phase. A porous layer formed around Si3N4 powder
as consequence of prolonged oxidation exposure which was probably originated by N2
resulting from a rise in the partial pressure of nitrogen at the interface [144]. Following the
same criteria from several authors, the oxidation curves of -Si3N4 powders at the tested
temperatures, the oxidation behaviour followed a parabolic rate law indicating that the
diffusion of oxygen through the oxide product layer to the oxide/Si3N4 interface is the ratecontrolling step, which is in agreement with the reports by Horton [63], Du et al. [145, 146],
Ogbuji and Bryan [147], Ogbuji [65], Weaver and Lucek [45], Cubicciotti and Lau [109], and
Biswas et al. [75].
With the application of a new model of oxidation of -Si3N4 powder proposed by Hou et
al. [3], it was possible to obtain a good agreement between experimental data and theoretical
model. After using the new model, it was extracted an apparent activation energy value of
325.6 kJ.mol-1 which is close to the value reported by Ogbuji and Bryan [147] and Ogbuji
[65]. For this explicit case, the physical meaning of above equation means that, when the
reaction time reaches a time of 12500 s, the oxidation will be completed. It is reasonable to
say that the rate-controlling step should be localized in the oxygen diffusion range within the
oxide layer, if one keeps in mind that the theoretical calculation results are coincident with
experimental data.
It is essential that the silicon nitride ceramics remain in the passive oxidation regime
when this is used as coatings on carbon-carbon hypersonic vehicle airframe components. In
reason of this, it is necessary that under certain conditions at high-temperature and low
pressure which could be encountered in hypersonic flight, Si3N4 can rapidly oxidize.
Therefore it is essential that the temperature and pressure conditions at which rapid oxidation
occur be well established [78]. The active-to-passive transition in the oxidation of Si3N4 was
studied in a flowing air environment as a function of temperature and total pressure in a
temperature range of 1347 and 1543 oC for partial pressures of oxygen of 2.5 and 123.2 Pa
[78]. This active-to-passive oxidation transition was measured on three sintered silicon nitride
and three hot-pressed silicon nitride with a mass flow rate of 0.56 cm3. s-1. The transition
temperatures ranged from 1366 to 1520 oC for the hot-pressed Si3N4 and 1365 to 1480 oC for
sintered Si3N4. It is noteworthy that the transition points reported here [78] for both sintered
and hot-pressed Si3N4 are significantly different from those previously reported for sintered
silicon nitride [57, 148, 149]. The authors reported that for a given pressure, the transition
temperature moves towards higher temperatures as the flow rate is increased, which is in
accordance with the experimental results of Narushima et al. [150] who suggest that the effect
may be due to the thinning of the boundary layer thickness for CO or SiO transport. With
increasing flow rates, it is expected to shift the transition point to higher temperatures, being a
conservative result for a component on a hypersonic vehicle.
Luthra [151] study new perspectives on why the oxidation rates of silicon nitride are
lower than those of silicon and on the conditions under which CO and N2 gas bubbles can
form from them. Two important observations were highlighted [151]: i) the rate constants for
CVD Si3N4 are much lower than those silicon by factors of 10 to 1000 at approximately 1350
to 1100 oC, respectively, and ii) gas bubbles are observed in some cases, particularly on
17
sintered materials [152-155], while gas bubbles has not been observed with CVD and single
crystal materials [146, 155-158]. Gas bubbles of both CO and N2 can form on silicon nitride
only if the permeabilities or diffusion rates of CO and N2 through silica are substantially
lower than the oxygen. The analysis of the oxidation processes of Si3N4 was carried out by
Luthra [151] considering consequences of various rate-limiting steps. From their investigation
it is extracted that again the observations in the literature do not fit a process that are
exclusively controlled by the oxygen diffusion, or the nitrogen diffusion, or an interface
reaction. On the other hand, the activation energy for oxidation of Si3N4 is of 330 to 485
kJ.mol-1 [146, 156-158] which are much higher than 119 kJ.mol-1 reported for silicon [153,
158, 159].
In order to take advantage of the excellent oxidation resistance of Si3N4 and increase their
sinterability [160], Hanabusa et al. [161] used Si3N4 ultrafine powders as sintering additive
applied on the surface of fine Si3N4 particles instead of common sintering additives used in
conventional sintering process, optimizing CVD conditions to form the ultrafine powder on
the core fine particles with excellent high-temperature properties. Considering the above
mentioned, investigations concerning to oxidation of CVD-Si3N4 have been reported. For
example, Kim [88] reported the effect of temperature and gas flow rates on the active
oxidation rates of CVD-Si3N4 while Narushima et al. [50] investigated on the oxidation of
CVD-Si3N4 in CO-CO2 atmospheres under temperatures between 1550 and 1710 oC. At 1600
o
C a linear mass loss due to the active oxidation of CVD-Si3N4 was observed in a reaction of
PCO 2 /PCO < 1. From plot of mass loss as a function of exposure time (see Figure 1 from ref.
[50]) it is extracted that the slopes of the lines represent the mass-loss rate corresponding to
the activation oxidation rates, which in turn was independent of gas velocity above 9.73 x 10-3
m.s-1, suggesting that the rate-controlling step of active oxidation could be a chemical
reaction at the Si3N4 surface.
After alternative experiments carried out in CO-Ar, CO-N2 and CO-Ar-N2 atmospheres,
with the purpose to study the effect of CO partial pressure (PCO) on the KCO-CO 2N , it is
concluded that the contribution of CO2 gas to the active oxidation of Si3N4 could be
insignificant; therefore, the CO gas was the main oxidant for the active oxidation of CVDSi3N4. On the other hand, the active oxidation reaction may be expressed following the
reaction:
Si3N4 (s) + 3CO (g) 3SiO (g) + 3C (s) + 2N2 (g)
(11)
and the chemical reaction between Si3N4 and CO might be a rate-controlling step. Scanning
electron microscoy (SEM) analysis on the oxidized Si3N4 surfaces revealed that the grain
boundary of CVD-Si3N4 was preferentially oxidized at first, which could be related to the
change of the rate-controlling step from a chemical reaction to a diffusion process with
increasing PCO 2 /PCO. An interesting appreciation obtained from work of Narushima et al.
[50] makes reference to the active oxidation rates for CVD-Si3N4 which was almost in
agreement with those obtained for CVD-SiC in the region of PCO 2 /PCO > 10-1. This behaviour
indicates that the active oxidation rates are independent of substrates (CVD-Si3N4 and CVDSiC), which agrees with the conclusion that the rate-controlling step would be the diffusion of
gas species formed by the reaction expressed as follows:
18

SiO2 (s) + CO (g) SiO (g) + CO2 (g)
(12)
Therefore, the same kinetics was observed in the active oxidation of CVD-SiC in COCO2 atmospheres according to Narushima et al. [162].
In other research work undertaken by Narushima et al. [49] they reported important
findings on the oxidation behaviour of CVD-Si3N4 in N2-O2 and Ar-O2 atmospheres at
b
temperatures of 1550 and 1650 oC. The observed mass-loss rates were constant at fixed P O2
b
during the experiments for several hours. However, at specified P O2 value (approximately 30
Pa for N2-O2 and 111 Pa for the Ar-O2 atmosphere) an abrupt transition from mass loss to
mass gain was clearly observed (active to passive oxidation). The stable solid phase in the
system N2-O2 is Si3N4, meanwhile, the Si is for the Ar-O2 atmosphere. On the other hand, in
the Si3N4- N2-O2, the mass-loss is attributed to equation (4). Thermodynamic calculation may
suggest that silicon nitride decomposes into Si (l) and N2 gas because Si (l) is stable and the
dissociation nitrogen pressure is higher than the ambient nitrogen pressure [49]. It is
important to note that the mass loss of Si3N4 was due to the active oxidation and not to
decomposition in the low oxygen partial pressure range near the atmospheric pressure at 1550
to 1650 oC. In both atmospheres, the active oxidation rates increased linearly with increasing
oxygen partial pressure, and the rate-controlling step was oxygen diffusion through a gaseous
stagnant boundary layer.
Hsu et al. [79] worked on the oxidation of CVD-Si3N4 in high-temperature regime under
environments containing oxygen and chlorine. Earlier experiments developed with
engineering Si3N4 ceramics are confirmed by similar experiments on CVD-Si3N4 (United
Technologies Research Center, East Hartford, CT) without additives that could influence the
oxidation kinetics [163-165]. Considering that CVD-Si3N4 ceramic showed the same
behaviour as the engineering material, it is believed that the resistance to active attack by
chlorine is not due to the sintering aids, but is a characteristic of Si3N4. On the other hand, no
changes in the surface of Si3N4 were observed by means of SEM. For comparison effects, the
surface of the corroded SiC (tested under same conditions than Si3N4) revealed a surface
deeply pitted as it was reported elsewhere [165-167]. According to above mentioned, it is
concluded that for reactions in pure oxygen, Si3N4 is substantially more resistant to attack in
mixed oxygen-chlorine environments than either silicon or SiC. Likewise, the protective
character of the silicon oxynitride layer formed in the interface between Si2O and Si3N4
would explain the increase in oxidation resistance of Si3N4 compared with SiC or Si [145,
168]. The presence of the oxynitride layer could also block the bubble formation during
oxidation of Si3N4 in Ar-20% O2 2% Cl2 at 1300 oC. Moreover, outside of the silicon
oxynitride layer, the higher oxygen potential may not support SiCl4 formation, and, as result,
bubble formation does not occur on Si3N4, and therefore, it is not subject to accelerated
oxidation under these conditions [79].
2.6. Oxidation of Hot Isostatically Pressed Silicon Nitride

As mentioned earlier, for silicon nitride caramics, the oxidation resistance strongly
depends on the quality and quantity of the sintering aids and taking into consideration of this,
19
the design of its microstructure can be tremendously improved [168]. Few studies exist
concerning oxidation resistance of dense, additive-free materials obtained by HIPing [169171]. Hot isostatic pressed (HIPed) Si3N4 based ceramics with low metal oxide additive
(approximately 0.5-1%) and silica content as surface, present as surface film in the starting
powder or introduced on HIPing due to the glass encapsulation [169] show best oxidation
resistance and approach that of single-crystalline or CVD silicon nitride [146, 158, 171, 172175].
An interesting study [69] on oxidation behaviour of HIPed silicon nitride (with and
without sintering aids) has been carried out at 1250 oC in flowing dry air with a flow rate of
10 l.h-1 during 100 h, [176, 177] by using TEM. The investigated materials were HIPed Si3N4,
one with 4 wt% Y2O3 as a sintering aid and one without sintering additive. The authors
reported that both silicon nitride with and without sintering additive, showed a thin
(approximately 1 m) oxide scale indicating a very low, weight gain and high oxidation
resistance [176]. When thin foil was tilted in TEM, the lack of diffraction contrast identified
the oxide layer in each HIPed ceramic to consist primarily of amorphous material. Likewise, a
small volume of crystalline material (identified as -Si3N4) was observed in the scale above
the scale/matrix interface in the sample without sintering aid, meanwhile, for the composition
Si3N4 + 4 wt% Y2O3 the crystalline identified phases were SiO and Y2Si2O7. Only Si and O
were detected in the amorphous and crystalline SiO2 phase with absence of nitrogen,
therefore, no indication was found for the existence of an oxynitride layer in these
materials. Luthra [151, 167] has suggested that, thermodynamically a silicon oxynitride layer
is expected to form in the interface between Si3N4 and SiO2 on the oxidation of pure silicon
nitride, when the process is attributed at diffusion controlled. On the other hand, when the
process is governed by interface-reaction or mixed-interface-reaction and diffusion
controlled, the mentioned oxynitride layer may or may not form. From this study, it is clear
that important differences between silicon nitride with and without sintering aids include the
difference in the degree of crystallinity of the oxide scale and the fact that Si3N4 with
additions of Y2O3 contains a secondary phase. However, the oxidation rates in these tested
materials and scale thickness are very similar [46, 171, 178-181].
It is believed that the oxidation rate of ceramics depends on the allotropic ratio and best
oxidation resistance is achieved with ceramics rich in -phase. Considering this, BackhausRicoult et al. [169] reported in a recent work the results obtained on the oxidation resistance
of silicon nitride without additives with different /- phase content. The starting powder was
a high-purity silicon nitride with 63% -phase HIPed at 1900 oC and 200 MPa of applied
pressure for 2 h. The dense Si3N4 samples were subsequently oxidized at temperatures
between 900 and 1200 oC in air or alternatively under a continuous flow of Ar with residual
oxygen partial pressure on the order of 1 Pa. In all cases, an oxide film formed at a surface of
the ceramic up to approximately 1450 oC with the presence of a large number of gas bubbles
in the silica scale.
Nevertheless, at temperatures above of 1300 oC cristobalite was found as an oxidation
product according to XRD analysis, with grain sizes increased with the -content and outer
oxygen activity. Large columnar cristobalite crystals were formed when the sample was
oxidized for 100 h at 1550 oC [169]. Similar observations have been reported in Si3N4 + 6%
Y2O3 + 4% Al2O3 (named KY64) and Si3N4 + 6% CeO2 + 4% Al2O3 (named KC64) sinterHIPed compositions where some cracking in the oxidized layer was revealed as consequence
20
of differences in the coefficient thermal expansion between different oxidation products (i.e.
SiO2, vitreous silicates and Si3N4) [108]. The generated stresses by volume expansion are
absorbed by the oxidized layer at the oxidation temperatures due to its great plasticity.
However, during cooling cracking is produced. This appreciation is illustrated in Figure 7 for
the first of the above compositions (KY64) oxidized at 1425 oC. It is clearly observed from
this figure the presence of larger Y2Si2O7- crystals, surface cracking and cristobalite crystals
surrounded by glassy phase (yttrium and aluminium-rich silicate) (Figure 7c) [108].
Backhaus-Ricoult et al. [169] concluded that in general, the HIPed samples followed an
oxidation parabolic law after a short initial linear growth period. Similarly, as expected,
oxidation rates in the two environments differed more at low temperature indicating that the
contribution of diffusion in the glass film is larger at low temperatures.
(a)
(b)
Y2Si2O7-
Y2Si2O7-
100 m
100 m
(c)
Cristobalite
Y2Si2O7-
Glass phase
Surface cracks
20 m
Figure 7. SEM micrographs showing the morphologies of oxidized surface for the Si3N4 + 6 % Y2O3 +
4 % Al2O3 (KY64) composition oxidized at 1425 oC. (a) quick cooling, (b) slow cooling, and (c) detail
of cristobalite crystals corresponding to the slowly cooling sample. [From ref. 108].
Oxidation rates in atmospheres with lower oxygen activity, aO
10-6, were
considerably smaller than oxidation in air [169]. In this study silicon oxynitride layer was not
found, which in turn is in agreement with the results obtained by Meara and Sjoberg [69] for
HIPed Si3N4 with and without sintering additives. According to the literature [182-187] the
pure silicon nitride polymorphs are very close in their thermodynamics stability, wherein on
dissolution of small amounts of oxygen, the - phase suffers an important decrease in
stability and decomposes in SiO2 and Si3N4 [183]. From this, the differences showed in
21
oxidation behaviour of -and -Si3N4 phase must then related to the oxygen content dissolved
on oxidation in the allotropic phases.
Table 1. Summary of experimental specific weight change in silicon nitride-based
ceramics
Composition
Si3N4+8 wt% Y2O3+3
wt% Al2O3
Si3N4/MoSi2+8 wt%
Y2O3
Si3N4-MoSi2-Y2O3-Al2O3
Si3N4-MoSi2-Lu2O3
Si3N4+7 vol% Y2Si2O7

Si3N4+3.6 wt% SiO2+4.3
wt% Y2O3
Si3N4+4.0 wt% SiO2+4.0
wt% Ce2O3
Si3N4+3.6 wt% SiO2+6.3
wt% La2O3
Si3N4+3.6 wt% SiO2+5.9
wt% Sm2O3
Si3N4+3.33 wt% Lu2O3
Si3N4+12.51 wt% Lu2O3
Sintering conditions
Oxidation
conditions
Hot pressed at 1840 oC, 1500 oC in air
30 MPa pressure, 60
104
min
Weight gain
(mg.cm-2)
3.2
Hot pressing 1800 oC,

1atm, N2
1atm, N2
Sintered at 1800 oC,
nitrogen atmosphere,
1.8 MPa pressure, 4h
1500 oC in air
128 h
3.95
Sintered at 2140 oC,

nitrogen atmosphere,
2.5 MPa pressure
700, 1000, and

1370 oC in air
200 h
Reference
[38]
3.9
[89]
1.30
1000 and 1200
o
C in air 0-256 h
0.02 at 1000 oC 256 h

0.14 at 1200 oC 256 h
0.02 at 1000 oC 256 h
0.20 at 1200 oC 256 h
0.395 at 1370 oC
[9]
[47]
0.727 at 1370 oC
0.562 at 1370 oC
0.752 at 1370 oC

20 MPa, N2, 1h
1400 and 1500

o
C in air 200 h
1000 oC in Ar2% Cl2
CVD-Si3N4 (Al2O3 and

Y2O3) as sintering
additives
Si3N4-35 vol% TiN+2.63
wt% Y2O3+1.05 wt%
Al2O3

vacuum, 30 MPa
pressure, 10 min
Si3N4+2 wt% Al2O3+2.5

wt% Y2O3
Si3N4+5.3 wt% MgO

1 atm, N2, 30 MPa
pressure, 30 min
Si3N4+6 wt% Y2O3+4

wt% Al2O3
(Different shaped
simples)
GPS at 1800 oC
1.8M Sulphuric
acid at 70 and 90
o
C, 200 h, open
system
1200 oC in air
and water vapour
atmosphere
0.5 mol/l H2SO4

at 90 oC, 0-350 h
0.1128 at 1400 oC
0.1698 at 1500 oC
0.1389 at 1400 oC
0.2236 at 1500 oC
< 0.2
[79]
2.5 at 70 oC
3.6 at 90 oC
[96]
0.2 in air
0.3 in water
atmosphere
1.5 in air
0.6 in water
atmosphere
21 (350 h plate shape)
17 (350 h cylinder
shape)
16 (350 h sphere
shape)
[110]
[34]
[101]
22
It is well known that a major barrier to the utilization of structural ceramics, such as
silicon nitride as components in advanced heat engines is the uncertainty in the long-term
mechanical reliability [184]. With the goal to improve thermal efficiency in power generating,
gas turbines have focused on increasing the turbine inlet temperature by innovative turbine
cooling techniques or alternatively, by the use of high temperature materials [185]. However,
less or no cooling air would be necessary if all metallic components are changed by ceramic
components, but undoubtedly only components which are exposed to high temperature and
low stress are being considered for replacement with ceramics. Nowadays steam cooling is of
considerable interest as an alternative method to air cooling [185]. Unfortunately, in the
practice, the gas turbine environment includes water vapour pressures in excess of 1
atmosphere. Therefore, for a given temperature, oxidation of silicon nitride is increased by i)
the replacement of oxygen by water vapour, and ii) an increase in the pressure of the oxidant.
Researchers found [186] that the presence of water vapour led to the formation of a
gaseous species (Si(OH)4) via reaction with the silica layer where the rate of formation of
such species, and thus the rate at which Si3N4 ceramics are consumed by its continued
oxidation, is expressed by the following expression:
k v1/2 P(H2O)2/(Ptotal)1/2
(13)
where v is the gas velocity, P(H2O) is the pressure of the water vapour, and (Ptotal) is the total
pressure. Rho et al. [185] investigated the oxidation of Si3N4 with additions of Y2O3 and
Al2O3 as sintering additives at temperatures between 1000 and 1350 oC in air with 0, 1.2
and 6.4 v/o H2O up to 500 h exposure. Two commercial Si3N4 ceramics denoted by the
manufacturer as Norton NT 164 were tested under above mentioned conditions. Although the
two batches of NT 164 used were the same, the heat treatment in one of them after machining
was different. With these appreciations, the authors identified the samples as NT 164-95
(1995 vintage) and NT 164-93 (1993 vintage) which were heat treated at 1400 oC during 5 h
after HIP. The oxidation behaviour was characterised by considering the weight change per
unit area as a function of water content (the average water content in air was estimated in 1.2
vol%) in the test environment for NT 164-93 exposed at 1350 oC for 100 h, where it is
observed that the oxidation rate is enhanced as the water content was increased. Volatilization
of reactants and/or products, escape of nitrogen gas in the sample, as well as diffusion
outward of excess oxygen dissolved at the oxide layer and grain boundary during heat
treatments, are the possible causes for the weight loss.
From the graph weight gain as a function of square root of oxidation time, was observed
a significant weight loss for the sample NT 164-95 for the first 100 h which was followed by
weight gain compared with the less weight loss in NT 164-93 sample suggesting that, the
weight loss depends on heat treatment conditions. Chen et al. [177] also observed deviations
from the parabolic kinetic law in Si3N4 ceramics being the composition of their silicon nitride
ceramic similar to that of NT 164. The enhance in oxidation rate in the presence of water
vapour was reported by Maeda et al. [187] which in turn suggested that the pores generated
by crystallization of cristobalite is the reason for this increased oxidation. Equiaxed
cristobalite grains were also detected by means of SEM in NT 164-93 silicon nitride oxidized
at 1350 oC in wet air [185]. A small amount of weight loss in the NT 164-95 sample would be
caused by volatilization of glassy phase in the grain boundary which was contamined by
23
impurities such as Na or K ions that come from the glass encapsulation used in the hot
isostatic pressing. This behaviour has also been reported by Rho et al. [185] in HIPed silicon
nitride with additions of Y2O3 as sintering aid. The trend observed in the oxidation behaviour
for NT 164-93 in the two different test environments were very similar to the previous work
[188] for Si3N4 ceramics with additions of Yb2O3 as a sintering aid. During the oxidation in
wet air, the dissolution of Si3N4 grains is fast but the diffusion of oxygen through the grain
boundary becomes the rate controlling step as the oxidation time increases. Nevertheless, the
dissolution of Si3N4 grains is the rate controlling step in air oxidation according to BackhausRicolut et al. [171] and that is in agreement with the results reported by Rho et al. [185].
Compared to silicon nitride materials, silicon nitride-titanium nitride composites are
machinable by EDM providing thereby the realization of ceramic parts with very complicated
shapes [34, 96]. Therefore, this advantage makes the composite a suitable high temperature
structural material [168]. Several studies showed that the oxidation mechanism is complex
and multilayered oxidation scales form on Si3N4-TiN composites [95, 189-192]. Considering
the possible high-temperature applications of this composite, the oxidation behaviour of TiN
particulate reinforced HIPed [95] silicon nitride composites was investigated in both dry and
humid air in the range 10001400 oC [168] adding a lower amount of sintering aids (0.25
wt% Al2O3 and 0.5 wt% Y2O3) compared to those used in a common hot pressed composite.
The HIPed composite was oxidized for 4, 24, and 100 h at 1000, 1200 and 1400 oC for each
time at above mentioned environments. High-resolution transmission electron microscoy
(HRTEM) studies revealed a thin film of glass phase between Si3N4 grains with a width of
approximately 2 nm. Likewise, after several hours of oxidation exposure at 1000 oC, layered
oxidation scales developed and completely covered all faces of the specimen where the oxide
layer corresponds to TiO2 crystals with uniform surface exhibiting equiaxed grains. However,
under this coverage, a sub-scale containing within the silicon nitride matrix mainly small
rutile crystals and a high density of pores developed at the site of the former TiN particles.
Mazerolles et al. [168] have considered two oxidation fronts as follows: i) the Si3N4
phase one which corresponds to the boundary between the outer layer and the sub-layer, and
ii) the TiN one, deeper and corresponding to the interface between the sub-layer and the
sound material. The oxidation kinetics at 1000 and 1200 oC followed approximately a
parabolic law for the entire investigated time suggesting reaction control by diffusion. These
results show that, during oxidation, important transport processes take place in the
intergranular phase. On the other hand, the temperature of 1000 oC corresponds to typical
glass transition temperatures of the sintered system SiO2-Al2O3-Y2O3 [193]. In consequence
with this, the constituents of the glass and the species dissolved in this glass display an
enhanced mobility. At high temperatures, oxidation of silicon nitride ceramics leads to higher
amount of glass and humidity at a low water vapour pressure has not a big influence. In order
to understand the role of humidity in oxidizing atmosphere, experiments with higher water
pressures (for example air saturated by water at 40 and 60 oC) are being carried out
nowadays. Mazerolles et al. [168] suggested that the oxidation behaviour of such composites
could be optimized and controlled by tailoring the intergranular phase with emphasis in both
chemical composition and quantity. Feldhoff et al. [194] in similar studies, reported that the
oxidation of Si3N4-TiN composites starts around 650 oC with the formation of nanocrystalline
titania particles on TiN particles that are exposed to the outer surface. They observed also that
the high-temperature oxidation resistance is improved if the amount of glass phase is kept
24
low. Table 1 shows a summary of experimental specific weight change during oxidation of
Si3N4 based ceramics.
CONCLUSION
Considering the processing techniques of silicon nitride (Si3N4) based ceramics and
according to contemporary concepts, oxidation resistance is of crucial importance for high
temperature applications generally under oxidizing conditions. The nature of the grain
boundary phases formed during sintering has a great influence on the oxidation resistance
because leads at high temperature, to a decrease in oxidation and corrosion resistance as
consequence of softening effects. The selection of appropriate additives and compositions is
very important in order to obtain a suitable grain-boundary phase with a high-melting point
which in turn crystallizes during sintering process. High oxidation resistance can be achieved
with low amount of silica grain-boundary phase and this is reached by using HIP. From
different studies on oxidation of Si3N4 ceramics, the use of Y2O3 as sintering additive seems
to be favourable because it aids to the formation of a liquid which produces on solidification,
a more refractory intergranular phase than those formed by most oxide additives. However,
the sinter ability of this system is low, and, therefore, requires the use of high temperatures
or/and simultaneous addition of Al2O3 to obtain low oxidation rates compared with the silicon
nitride-magnesia system where the oxidation resistance of silicon nitride alloys hot-pressed
with MgO as sintering aid depends on their initial composition and large changes in the
parabolic rate constant with composition are evidenced. Similar behaviour has been observed
with Si3N4 ceramics when are under acid environments. Electroconductive ceramics based
Si3N4-TiN may offer new and important opportunities because they combine the high
performance of Si3N4-based ceramics with the machinability by electro-discharge machining
(EDM). Further studies on less-expensive processing techniques as well as to improve the
high-temperature oxidation and corrosion resistance by optimizing the grain boundary
characteristics are need.
REFERENCES
[1]
Takatori, T.; Honma, T.; Kamiya, N.; Masaki, M.; Sasaki, S.; Wada, S. ASME paper
91-GT-142, 1991.
[2] Hirosaki, N.; Okada, A.; Mitomo, M. J Mater Sci 1990, 25, 1872-1876.
[3] Hou, X.-M.; Chou, K.C.; Hu, X.-J.; Zhao, H.-L. J Alloy Compd 2008, 459, 123-129.
[4] Nakatani, N.; Ando K.; Houjou, K. J Eur Ceram Soc 2008, 28, 1251-1257.
[5] Zou, B.; Huang, C.; Chen, M. Int J Refract Met Hard Mat 2009, 27, 52-60.
[6] Shang, R.; Sun, W.; Wang, Y.; Jin, G.-Q.; Guo, X.-Y. Catal Commun 2008, 9, 21032106.
[7] Deschaux-Beaume, F.; Frety, N.; Cutard, T.; Colin, C. Ceram Int 2007, 33, 1331-1339.
[8] Gogotsi, Y. G.; Porz, F. Corros Sci 1992, 33, 627-640.
[9] Pereira, S.; Ribeiro, S. J Mater Proc Technol 2004, 147, 336-342.
[10] Yeheskel, O.; Gefen, Y.; Talianker, M. J Mater Sci 1984, 19, 745-752.
25
[11] Fujikawa, T.; Inoue, Y.; Homma, K.; Tatsuno, T.; Okada, H.; Kondo, N. Kobelco
Technology Review 1988, 3, 1-8.
[12] Hoffmann, M.J.; Petzow, G. Pure & Appl Chem 1994, 66, 1807-1814.
[13] Liu, M.; Nemat-Nasser, S. J Mater Res 1999, 14, 4621-4629.
[14] Melndez-Martnez, J.J.; Domnguez-Rodrguez, A. Prog Mater Sci 2004, 49, 19- 107.
[15] Szpvlgyi, J.; Mohai, I.; Gubicza, J. J Mater Chem 2001, 11, 859-863.
[16] Kawaoka, H.; Kusunose, T.; Choa, Y.H.; Sekino, T.; Niihara, K. J Ceram Proc Res
2001, 2, 51-53.
[17] Ziegler, G.; Heinrich, J.; Wtting, G. J Mater Sci 1987, 22, 3041-3086.
[18] Katz, R.N. In Progress in Nitrogen Ceramics; Riley F.L.; Ed.; Nitrogen Ceramics;
Martinus Nijhoff Pub: The Hague, Netherlands, 1983; 1976-1981, pp. 3-20.
[19] Negita, K. J Mater Sci Lett 1985, 4, 755-758.
[20] Hirosaki, N.; Okada, A.; Matoba, K. J Am Ceram Soc 1988, 71, C-144-C-147.
[21] Miyazaki, S.; Tsuji, I.; Mori, M.; Hara, Y.; Furuse, Y.; Hamada, S. Yokohama
International Gas Turbine Congress, October 27th to November 1st, Yokohama, Japan:
1991, I-149.
[22] Tsuji, I.; Aoki, S.; Miyazaki, S.; Hara, Y.; Furuse, Y.; Hamada, S. ASME paper 91-GT311, 1991.
[23] Shimizu, T.; Takama, K.; Enokishima, H.; Mikame, K.; Tsuji, S. SAE Technical
Papers 900656, 1990.
[24] Danforth, S.C.; Symons, W.; Nilsen, K.J.; Riman, R.E. Processing of silicon nitride
powders; Binner J.G.P; Ed.; Advanced Ceramic Processing and Technology; Willian
Andrew Publishing: Noyes, 1990; pp. 39-71.
[25] Santos, C.; Ribeiro, S.; Strecker, K.; Da Silva, L.E.G. Acta Microscopica 2003, 12, 8386.
[26] Bowen, L.J.; Weston, R.J.; Carruthers, T.G.; Brook, R.J. J Mater Sci 1978, 13, 341350.
[27] Messier, D.R.; Riley, F.L.; Brook, R.J. J Mater Sci 1978, 13, 1199-1205.
[28] Sajgalik, P.J.; Rajan, K.; Warbichler, P.; Hoffer, F.; Dusza, J. Key Eng Mat 1999, 159160, 405-410.
[29] Sato, T.; Takeishi, K.; Matsuura, M.; Miyauchi, J. Proceedings of 1983 Tokyo
International Gas Turbine Conress, 83-TOKYO-IGTC-8, 1983, p. 59.
[30] Soma, T.; Matsui, M.; Oda, I. NGK Technical Reports, NGK Technical Reports 1983,
R-83-3, 1-6.
[31] Izhevskiy, V.A.; Genova, L.A.; Bressiani, J.C. Ceramica 1999, 45, 5-23.
[32] Deely, G.G.; Herbert, J.M.; Moore, N.C. Powder Metall 1961, 8, 145-151.
[33] Tsuge, A.; Kudo, H.; Komeya, K. J Am Ceram Soc 1974, 57, 269-270.
[34] Blugan, G.; Wittig, D.; Kuebler, J. Corros Sci 2009, 51, 547-555.
[35] Kingery, W.D. J Appl Phys 1959, 30, 301-306.
[36] Richerson, D.W. Am Ceram Soc Bull 1973, 52, 560-569.
[37] Kossowsky, R.; Miller, D.G.; Diaz, E.S. J Mater Sci 1975, 10, 983-997.
[38] Klemm, H.; Schubert, C. J Am Ceram Soc 2001, 84, 2430-2432.
[39] Hoffmann, M.J.; Becher, P.F.; Petzow, P. Silicon Nitride 93; Key Engineering
Materials; Trans Tech. Publications: Switzerland, 1993, Vol. 89-91.
[40] Szpvlgyi, J.; Riley, F.L.; Mohai, I.; Bertti, I.; Gilbart, E. J Mater Chem 1996, 6,
1175-1186.
26
[41] Symons, W.; Danforth, S.G. Advances in Ceramics; Messing, G.L., Madziyasmi, K.S.;
McCanley, J.W.; Haber, R.A.; Eds.; Ceramic Processing Science; American Ceramic
Society: Westerville, OH, 1987, Vol. 21, pp. 247-256.
[42] Wereszczak, A.A.; Ferber, M.K.; Kirkland, T.P.; Foley, M.R.; Yeckley, R.L. J Am
Ceram Soc 1995, 78, 2129-2140.
[43] Klemm, H.; Herrmann, M.; Schubert, Chr. Ceram Eng Sci Proc 1997, 18, 615-623.
[44] Riley, F.L. J Am Ceram Soc 2000, 83, 245-265.
[45] Weaver, G.Q.; Lucek, J.W. Am Ceram Soc Bull 1978, 57, 1131-1134.
[46] Cubicciotti, D.; Law K.H. J Electrochem Soc 1979, 126, 1723-1728.
[47] Mieskowski, D.M.; Sanders, W.A. J Am Ceram Soc 1985, 68, C-160-C-163.
[48] Gogotsi, Y.G.; Grathwohl, G. J Am Ceram Soc 1993, 76, 3093-3104.
[49] Narushima, T.; Goto, T.; Yokoyama, Y.; Hagiwara, J.; Iguchi, Y.; Hirai.T. J Am Ceram
Soc 1994, 77, 2369-2375.
[50] Narushima, T.; Goto, T.; Hagiwara, J.; Iguchi, Y.; Ira, T. J Am Ceram Soc 1994, 77,
2921-2925.
[51] Klemm, H.; Tangermann, K.; Schubert, C.; Hermel, W. J Am Ceram Soc 1996, 79,
2429-2435.
[52] Tressler, R.E. Corrosion and Corrosive Degradation of Ceramics; Tressler, R.E.;
McNallan, M.; Eds.; Ceramic Transactions; American Ceramic Society: Westerville,
OH, 1990; Vol. 10, pp. 99-124.
[53] Jacobson, N.S. Am Ceram Soc 1993, 76, 3-28.
[54] Rendtel, P.; Rendtel, A.; Hubner, H.; Klemm, H.; Herrmann, M. J Eur Ceram Soc
1999, 19, 217-226.
[55] Clarke, D.R.; Lange, F.F. J Am Ceram Soc 1980, 63, 586-593.
[56] Kiehle, A.J.; Heung, L.K.; Gielisse, P.J.; Rockett, T.J. J Am Ceram Soc 1975, 58, 1720.
[57] Tripp, W.C.; Graham, H.C. J Am Ceram Soc 1976, 59, 399-403.
[58] Singhal, S.C. Nitrogen Ceramics; Riley, F.L.; Ed.; Noordhoff International, Reading,
MS, 1977; pp. 607-626.
[59] Singhal, S.C. J Mater Sci 1976, 11, 500-509.
[60] Lange, F.F. J Am Ceram Soc 1978, 61, 53-56.
[61] Wagner, C. J Appl Phys 1958, 29, 1295-1297.
[62] Torre, J.P.; Mocellin, A. Nitrogen Ceramics; Riley, F.L.; Ed.; Noordhoff International,
Reading, MS, 1977; pp. 63-78.
[63] Horton, R.M. J Am Ceram Soc 1969, 52, 121-124.
[64] Tomashov, N.D. Theory of Corrosion and Protection of Metals; The McMillan Co:
NY, 1995; pp. 51-53.
[65] Ogbuji, L.U.J.T. J Am Ceram Soc 1995, 78, 1279-1284.
[66] Beaume, F.D.; Cutard, T.; Frey, N.; Levaillant, C. J Am Ceram Soc 2002, 85, 18601866.
[67] Lewis, M.W.; Barnard, P. J Mater Sci 1980, 15, 443-448.
[68] Wang, L.; He, C.; Wu, J.G. Oxidation of sintered silicon nitride materials; 3rd
International Symposium on Ceramic Materials and Components for Engines;
American Ceramic Society: Westerville, OH, 1988; pp. 604-611.
[69] Meara, C.O.; Sjoberg, J. J Am Ceram Soc 1997, 80, 1491-1500.
[70] Lange, F.F.; Singhal, S.C.; Kuznicki, R.K. J Am Ceram Soc 1977, 60, 249-252.
27
[71] Cinibulk, M.K.; Thomas, G. J Am Ceram Soc 1992, 75, 2044-2049.

[72] Houjou, K.; Ando, K.; Chu, M.C.; Liu, S.P.; Sato, S. J Eur Ceram Soc 2005, 25, 559567.
[73] Pan, K. J Am Ceram Soc 1996, 79, 2975-2979.
[74] Echeberria, J.; Castro, F. Mater Sci Technol 1990, 6, 497-503.
[75] Biswas, S.K.; Murkerji, J.; Das, P.K. Key Eng Mater 1994, 89-91, 271-274.
[76] Kim, H.E.; Moorehead, A.J. J Am Ceram Soc 1990, 73, 3007-3014.
[77] Otani, T.; Hirata, M. Thin Solid Films 2003, 442, 44-47.
[78] Vaughn, W.L.; Maahs, H.G. J Am Ceram Soc 1990, 73, 1540-1543.
[79] Hsu, P.P.; Ip, S.Y.; Park, Ch.; McNallan, M.J. J Am Ceram Soc 1993, 76, 1621-1623.
[80] Hinze, J.W.; Graham, H.G. J Electrochem Soc 1976, 123, 1066-1073.
[81] Nickel, K.G. J Eur Ceram Soc 1992, 9, 3-8.
[82] Sheehan, J.E. J Am Ceram Soc 1982, 65, C-111-C-113.
[83] Narushima, T.; Goto, T.; Yokoyama, Y.; Hagiwara, J.; Iguchi, Y.; Ira, T. Solid State
Ionics 1992, 53-56, 265-269.
[84] Singhal, S.C. Ceram Int 1976, 2, 123-130.
[85] Batha, H.D.; Whitney, E.D. J Am Ceram Soc 1973, 56, 365-369.
[86] Goto, T.; Hirai, T. J Mater Sci 1987, 22, 2842-2846.
[87] Warburton, J.B.; Antill, J.E.; Hawes, M.W. J Am Ceram Soc 1978, 61, 67-72.
[88] Kim, H.E. Gaseous corrosion of silicon carbide and nitride in hydrogen. PhD. Thesis,
The Ohio State University, Columbus, OH, 1987.
[89] Guo, Z.; Blugan, G.; Graule, T.; Reece, M.; Kuebler, J. J Eur Ceram Soc 2007, 27,
2153-2161.
[90] Klemm, H.; Herrmann, M.; Schubert, C. J Eng Gas Turb Pow 2000, 122, 13-18.
[91] Suzuki, K.H.; Yoshida, H.; Yamaguki, S. U.S. Pat. No. 4 613 455, 1986.
[92] Murtry, C.H.; Boecker, W.D.G.; Seshadr, S.J.; Zanghi, J.S.; Garnier, J.E. Am Ceram
Soc Bull 1987, 66, 325-329.
[93] Higuchi, M.; Miyake, M.; Kateuchi, H.; Kamjo, E. U.S. Pat. No. 4 659 508, 1987.
[94] Martin, C.; Mathieu, P.; Cales, B. Electrical discharge machinable ceramic composites;
Brook, R.J.; Ed.; Proceedings of Symposium on Ceramic Materials Research,
European Materials Research Society Spring Conference; Society European Materials
Research, Strasbourg, FR, 1988.
[95] Bellosi, A.; Guicciardi, S.; Tampieri, A. J Eur Ceram Soc 1992, 9, 83-93.
[96] Laudisio, G.; Seipel, B.; Ruffini, A.; Nkel, K.G. Corros Sci 2005, 47, 1666-1677.
[97] Demaison, J.; Lefort, P.; Billy, M. Oxid Met 1979, 13, 505-517.
[98] Gogotsi, Y.G.; Porz, F.; Dransfield, G. Oxid Met 1993, 39, 69-91.
[99] Ichimura, H.; Kawana, A. J Mater Res 1993, 8, 1093-1100.
[100] Lange, F.F. Am Ceram Soc Bull 1980, 59, 239-241.
[101] Herrmann, M.; Schilm, J. Ceram Int 2009, 35, 797-802.
[102] Monteverde, F.; Mingazzini, C.; Giorgi, M.; Bellosi, A. Corros Sci 2001, 43, 18511863.
[103] Lee, S.-H.; Mager, R. J Eur Ceram Soc 2009, 29, 3293-3297.
[104] Schlichting, J. Energy and Ceramics; Vincenzini, P.; Ed.; Elsevier: NY, 1980; pp. 390398.
[105] Wu, C.C.; McKinney, K.R.; Rice, R.W.; McDonough, W.J.; Freiman, S.W. J Mater Sci
1981, 16, 3099-3104.
28
[106] Babini, G.N.; Bellosi, A.; Vincenzini, P. J Mater Sci 1984, 19, 1029-1042.
[107] Pomeroy, M.; Hampshire, S. Mat Sci Eng 1989, A109, 389-394.
[108] Echeberria, J. Oxidacion a Elevadas Temperaturas de Materiales Ceramicos Basados
en Nitruro de Silicio: Cinetica y Microestructura, PhD Thesis, Escuela Superior de
Ingenieros Industriales de San Sebastian, Universidad de Navarra, Spain, (1993).
[109] Cubicciotti, D.; Lau, K.H. J Am Ceram Soc 1978, 61, 512-517.
[110] Guo, S.; Hirosaki, N.; Nishimura, T.; Yamamoto, Y.; Mitomo, M. J Am Ceram Soc
2003, 86, 1900-1905.
[111] Guo, S., Hirosaki, N.; Yamamoto, Y.; Nishimura, T.; Kagawa, Y. Mater Sci Eng 2005,
A408, 9-18.
[112] Kofsad, P. High Temperature Oxidation of Metals; Willey & Sons: NY, 1966.
[113] Gugel, E.; Fickel, A.; Kessel, H. Powder Met Int 1974, 6, 136-140.
[114] Cubicciotti, D.; Lau, K.H.; Jones, R.L. J Electrochem Soc 1977, 124, 1955-56.
[115] Wild, S.; Grieveson, P.; Jack, K.H.; Latimer, M.J. Special Ceramics 5; British Ceramic
Research Association: Stoke-on-Trent, England, 1970; pp. 377-384.
[116] Lange, F.F. Nitrogen Ceramics; Riley, F.L.; Ed.; Noordhoff International, Reading,
MS, 1977; pp. 491-509.
[117] Knickerbocker, S.H.; Zangvil, A.; Brown, S.W. J Am Ceram Soc 1985, 68, C-99-C101.
[118] Guo, S.; Hirosaki, N.; Nishimura, T.; Yamamoto, Y.; Mitomo, M. Mater Sci Technol
2003, 19, 544-548.
[119] Cinibulk, M.K.; Thomas, G.; Johnson, S.M. J Am Ceram Soc 1992, 75, 2050-2055.
[120] Park, H.; Kim, H.W.; Kim, H.E. J Am Ceram Soc 1998, 81, 2130-2134.
[121] Kim, H.W.; Koh, Y.H.; Kim, H.E. J Mater Res 2000, 15, 1478-1482.
[122] Herrmann, M.; Klemm, H.; Schubert, Shr.; Hermel, W. Key Eng Mater 1997, 132-136,
1977-1980.
[123] Hong, Z.L.; Yoshida, H.; Ikuhara, Y.; Sakuma, T.; Nishimura, T.; Mitomo, M. J Eur
Ceram Soc 2002, 22, 527-534.
[124] Guo, S.; Hirosaki, N.; Nishimura, T.; Yamamoto, Y.; Mitomo, M. Philos Mag 2002,
A82, 3027-3043.
[125] Guo, S.; Hirosaki, N.; Yamamoto, Y.; Nishimura, T.; Mitomo, M. Scr Mater 2001, 45,
867-874.
[126] Guo, S.; Hirosaki, N.; Yamamoto, Y.; Nishimura, T.; Mitomo, M. J Am Ceram Soc
2002, 85, 1607-1609.
[127] Hirosaki, N.; Yamamoto, Y.; Nishimura, T.; Mitomo, M. Japan Patent 324327, 2000.
[128] Choi, H.J.; Lee, J.G.; Kim, Y.W. J Eur Ceram Soc 1999, 19, 2757-2762.
[129] Hirosaki, N.; Yamamoto, Y.; Nishimura, T.; Mitomo, M.; Takahashi, J.; Yamane, H.;
Shimada, M. J Am Ceram Soc 2002, 85, 2861-2863.
[130] Nishimura, T.; Mitomo, M. J Mater Res 1995, 10, 240-242.
[131] Raj, R.; Lange, F.F. Acta Metall 1981, 29, 1993-2000.
[132] Vetrano, J.S.; Kleebe, H-J.; Hampp, E.; Hoffmann, M.J.; Ruhle, M.; Cannon, R.M. J
Mater Sci 1993, 28, 3529-3538.
[133] Cinibulk, M.K.; Thomas, G.; Johnson, S.M. J Am Ceram Soc 1990, 73, 1606-1612.
[134] Tong, X.; Li, J.; Yang, X.; Lin, H.; Guo, G.; He, M. J Am Ceram Soc 2006, 89, 17301732.
29
[135] Lin, H-T.; Ferber, M.K.; Becher, P.F.; Price Van Roode, M.; Kimmel, J.B.; Jimenez,
O.D. J Am Ceram Soc 2006, 89, 258-265.
[136] Anson, D.; Ramesh, K.S.; DeCorso, M. Applications of ceramics in industrial gas
turbines: Task 1, Technology Background: Task 1 Report, U.S., Department of
Energy Contract No. DE-AC-2-89CE40878, DOE/CE/40878-2, 1991.
[137] Anson, D.; Scheppard, W.J.; Norland, D.L.; DeCorso, M. Application of Ceramics in
Industrial Gas Turbines: Economics Market, Analysis, and R&D Plan: Part 2 ReportTasks 2,,3, and 4, U.S., Department of Energy Contract No. DE-AC-2-89CE40878,
DOE/CE/40878-2, 1991.
[138] Lin, H-T.; Ferber, M.K.; Westphal, W.; Macri, F. Evaluation of mechanical reliability
of silicon nitride vanes after field tests in an industrial gas turbine; Proceedings of
ASME Turbo Expo 2002, June 3-6, Amsterdam, The Netherlands, GT-2002-30629,
2002, pp. 1-8.
[139] Arat, P.; Wber, F. Key Eng Mat 1999, 161-163, 221-224.
[140] Clarke, D.R. J Am Ceram Soc 1983, 66, 92-95.
[141] Clarke, D.R.; Lange, F.F.; Schnittgrund, G.D. J Amer Ceram Soc 1982, 65, C-51-C-52.
[142] Lange, F.F.; Davis, B.I.; Clarke, D.R. J Mater Sci 1980, 15, 601-610.
[143] McDonough, W.J.; Wu, C.Cm.; Morgan, P.E.D. J Am Ceram Soc 1981, 64, C-45-C-47.
[144] Ogbuji, L.U.J.T.; Opila, E.J. J Electrochem Soc 1995, 142, 925-930.
[145] Du, H.H.; Tressler, R.E.; Spear, K.E. J Electrochem Soc 1989, 136, 3210-3215.
[146] Du, H.H.; Tressler, R.E.; Spear, K.E.; Pantano, C.G. J Electrochem Soc 1989, 136,
1527-1535.
[147] Ogbuji, L.; Bryan, S.R. J Am Ceram Soc 1995, 78, 1272-1278.
[148] Singhal, S.C. Proceedings of the Symposium on Properties of High Temperature Alloys
with Emphasis on Environmental Effects; Foroulis, Z.A.; Pettit, F.S.; Eds.; The
Electrochemical Society: Princeton, NJ, 1976; Vol. 77-1.
[149] Du, H. Fundamental Studies of Oxidation of Chemical Vapor Deposited Silicon
Nitride, PhD. Thesis, Pensylvania State University, University Park, PA, December
1988.
[150] Narushima, T.; Goto, T.; Hirai, T. Active to passive transition in the oxidation of CVDSiC; Proceedings of the International Meeting on Advanced Materials, Materials
Research Society: Symposium G: Corrosion Coatings of Advanced Materials;
Pittsburgh, PA, 1989; Vol. 4.
[151] Luthra, K.L. J Am Ceram Soc 1991, 74, 1095-1103.
[152] Schlichting, J. Ber Dtsch Keram Ges 1979, 56, 196-200.
[153] Tripp, W.C.; Hinze, J.W.; Mendiratta, M.G.; Duff, R.H.; Hampton, A.F.; Stround, J.E.;
Rodine, E.T. Final Technical Report No. ARL-75-0130, Wright Patterson AFB, OH,
June 1975.
[154] Mieskowski, D.M.; Mitchell, T.E.; Euer, A.H. J Am Ceram Soc 1984, 67, C-17-C-18.
[155] Costello, J.A.; Tressler, R.E. J Am Ceram Soc 1986, 69, 674-681.
[156] Zheng, Z.; Tressler, R.E.; Spear, K.E. J Electrochem Soc 1990, 137, 854-858.
[157] Euer, A.H.; Ogbuji, L.U.; Mitchell, T.E. J Am Ceram Soc 1980, 63, 354-355.
[158] Choi, D.J.; Fischback, D.B.; Scott, W.D. J Am Ceram Soc 1989, 72, 1118-1123.
[159] Deal, B.E.; Grove, A.S. J Appl Phys 1965, 36, 3770-3778.
[160] Vahlas, C.; Caussat, B.G.; Serp, P.; Angelopoulus, G.N. Mater Sci Eng 2006, R3, 1-72.
[161] Hanabusa, T.; Uemiya, S.; Kojima, T. Chem Eng Sci 1999, 54, 3335-3340.
30
[162] Narushima, T.; Goto, T.; Yokoyama, Y.; Iguchi, Y.; Ira, T. J Am Ceram Soc 1993, 76,
2521-2524.
[163] Park, C.; McNallan, M.J. J Am Ceram Soc 1995, 78, 922-928.
[164] Park, D.S.; McNallan, M.J.; Park, C.; Liang, W.W. J Am Ceram Soc 1990, 73, 13231329.
[165] Ip, Y.S.; McNallan, M.J.; Park, D.S. J Am Ceram Soc 1992, 75, 1942-1948.
[166] Marra, J.E.; Kreidler, E.R.; Jacobson, N.S.; Fox, D.S. J Am Ceram Soc 1988, 71, 10671073.
[167] Luthra, K.L. J Electrochem Soc 1991, 138, 3001-3007.
[168] Mazerolles, L.; Feldhoff, A.; Trichet, M.F.; Backhaus-Ricoult, M. J Eur Ceram Soc
2005, 25, 1743-1748.
[169] Backhaus-Ricoult, M.; Guerin, V.; Huntz, A.M.; Urbanovich, V.S. J Am Ceram Soc
2002, 85, 385-392.
[170] OMeara, C.; Heim, M.; Pompe, R. J Eur Ceram Soc 1995, 15, 319-328.
[171] Backhaus-Ricoult, M.; Gogotsi, Y. J Mater Res 1995, 10, 2306-2321.
[172] Nickel, K. Corrosion of Advanced Ceramics, Measurement and Modelling; NATO ASI
Series, No. E267; Kluwer Academic Publishers: Dordrecht, Netherlands; 1994; pp. 5972.
[173] Butschereit, E.; Nickel, K. Key Eng Mater 1997, 132-136, 1592-1597.
[174] Hirai, T.; Niihara, K.; Goto, T. J Am Ceram Soc 1980, 63, 419-424.
[175] Monteverde, F.; Bellosi, A. Key Eng Mater 1997, 132-136, 1609-1612.
[176] OMeara, C.; Chen, J.C.; Sjoberg, J.; Pejryd, L. Proceedings of the 11th Riso
International Symposium on Metallurgy and Materials Science: Structural CeramicsProcessing, Microstructure and Properties; Bentzen, J,J. et al.; Eds.; Riso National
Laboratory, Roskilde, Denmark, 1990; pp.439-444.
[177] Chen, J.C.; Sjoberg, J.; Lindqvist, D.; OMeara, C.; Pejryd, L. J Eur Ceram Soc 1991,
7, 319-327.
[178] Persson, J.; Kall, P.-O.; Nygren, M. J Eur Ceram Soc 1993, 12, 177-184.
[179] Andrews, P.; Riley, F.L. J Eur Ceram Soc 1989, 5, 245-256.
[180] Knoch, H.; Schlichting, J. Sprechsaal 1981, 114, 99-103.
[181] Babini, G.N.; Bellosi, A.; Vincenzini, P. J Mater Sci 1983, 8, 231-244.
[182] OHare, P.A.; Tomaszkiewicz, I.; Belk, C.M.; Seifert, H.J. J Mater Res 1997, 12, 32033205.
[183] Liang, J.; Topor, L.; Navrotsky, A.; Mitomo, M. J Mater Res 1999, 14, 1959-1968.
[184] Ferber, M.K.; Lin, H.T.; Parthasarathy, V.; Brentnall, W. Degradation of silicon
nitrides in high pressure, moisture rich environments; Proceedings of 1999 ASME
Turbo Expo: Land, Sea & Air, June 7-10, Indianapolis, Indiana, USA, GT1999-0000,
1999.
[185] Rho, H.; Hecht, N.L.; Graves, G.A. J Mater Sci 2000, 35, 3631-3639.
[186] Opila, E.J.; Hann, R.E. J Am Ceram Soc 1997, 80, 197-205.
[187] Maeda, M.; Nakamura, K.; Ohkubo, T. J Mater Sci 1989, 24, 2120-2126.
[188] Rho, H. Effect of water vapor on the oxidation and thermo-mechanical behavior of SiC
and Si3N4 ceramics. PhD Thesis, University of Dayton, Dayton, Ohio, USA, 1997.
[189] Bellosi, A.; Tampieri, A.; Liu, Y.Z. Mater Sci Eng 1990, A127, 115-122.
[190] Sheldon, B.W. J Am Ceram Soc 1996, 79, 2993-2996.
31
[191] Deschaux-Beaume, F.; Cutard, T.; Frety, N.; Levaillant, Ch. J Am Ceram Soc 2002, 85,
1860-1866.
[192] Feldhoff, A.; Mazerolles, L.; Trichet, M.F.; Backhaus-Ricoult, M. High temperature
oxidation of Si3N4-TiN composites in dry and humid air; Proceedings of CIMTEC
2002, 14th-18th July, 2002.
[193] Becher, P.F.; Waters, S.B.; Westmoreland, C.G.; Riester, L. J Am Ceram Soc 2002, 85,
897-902.
[194] Feldhoff, A.; Trichet, M.F.; Mazerolles, L.; Backhaus-Ricoult, M. J Eur Ceram Soc
2005, 25, 1733-1742.
In: Silicon Nitride

ISBN 978-1-61942-865-2
Chapter 2
LOW TEMPERATURE PREPARATION

OF PHOSPHATE BONDED SILICON
NITRIDE CERAMICS WITH HIGH MECHANICAL
STRENGTH
Fei Chen1,2, Qiang Shen2,* and Lianmeng Zhang1,2
1
Key Lab of Advanced Technology for Specially Functional Materials, Ministry

of Education, Wuhan, China
2
State Key Lab of Advanced Technology for Materials Synthesis and Processing,
Wuhan University of Technology, Wuhan, China
ABSTRACT
Silicon nitride (Si3N4) ceramic is used in numerous applications because of its
superior properties, such as high-temperature strength, good oxidation resistance and low
thermal expansion coefficient. Therefore, it becomes one of the most promising
candidates as structural and functional ceramic materials. Previous studies focused
mainly on acquiring both perfect mechanical properties and complete transformation
from alpha to beta phase Si3N4 porous ceramics. However, better dielectric property of Si3N4 ceramic whose dielectric constant () is 5.6 and dielectric loss (tg) is 0.003 at
room temperature than -Si3N4 (=7.9, tg=0.005) should not be ignored because it can
be utilized as an electromagnetic wave penetrating material. In order to prevent the phase
transformation from -Si3N4 to -Si3N4, a new method of low temperature processing is
required for the preparation of -Si3N4 matrix ceramics. In this case, porous -Si3N4
matrix structure with controlled high porosity (30~50%), micropores and guaranteed
strength (>50MPa) is needed for this application.
It is reported that phosphate is a kind of high-temperature stable and low liner
thermal-expansion coefficient binder material, and can be obtained at low temperature. In
this chapter, based on -Si3N4 ceramics matrix, the phosphate bonded silicon nitride
porous ceramics, silicon nitride ceramics using phosphate as sintering additives and
phosphate bonded silicon nitride ceramics applied for broadband radome are introduced.
34
INTRODUCTION
Silicon nitride (Si3N4) ceramic is used in numerous applications because of its superior
properties, such as high-temperature strength, good oxidation resistance, thermochemical
corrosion resistance, thermal shock resistance and low thermal expansion coefficient [1-3].
Therefore, it becomes one of the most promising candidates as structural and functional
ceramic materials. Recently, many studies have been developed to prepare porous Si3N4
ceramics [4-6] which offer exciting opportunities for applications in hot gas filtration, high
temperature separation membranes, and catalyst supports. Previous studies focused mainly on
acquiring both perfect mechanical properties and complete transformation from alpha to beta
phase Si3N4 porous ceramics by using different sintering additives. However, better dielectric
properties [7] of -Si3N4 ceramic whose dielectric constant () is 5.6 and dielectric loss (tg)
is 0.003 at room temperature than -Si3N4 (=7.9, tg=0.005) should not be ignored because it
can be utilized as an electromagnetic wave penetrating material [8]. In order to prevent both
the phase transformation from -Si3N4 to -Si3N4 and the use of sintering additives, a new
method of low temperature processing is required for the preparation of -Si3N4 ceramics. In
this case, pure porous -Si3N4 structure with controlled high porosity (>50%) and guaranteed
strength (>50MPa) is needed for this application.
Nowadays, various processing techniques have been developed to prepare porous silicon
nitride ceramics [9]. A common method of fabricating controlled porosity ceramic materials
is adding organic fugitive substance [10] or using some organic sacrificial template [5] which
is burned out to form pores. However, this method requires an extra heat treatment to remove
the agents, and it is generally difficult to eliminate the impurity of residual carbon. Another
method to adjust the porosity is sintering powder compacts to a certain degree of
densification, the so-called partially sintering process [11]. Sintering temperature and hold
time are usually used to control the porosity in the partially sintering. As examples, porous
Si3N4 ceramics can be prepared by using the partial hot pressing [12], with porosities of
10~20% and strength of 200~300MPa, which may be too dense for our application; or by
controlled sintering of Si3N4 powder with low content of sintering additives [13], with
porosities of 40~50% but strength of less than 50MPa. Porous Si3N4 ceramics can also be
prepared by reaction sintering through nitridation of silicon powders [1], reaction sintering
through carbothermal nitridation of SiO2 [14], freeze-drying [15], combustion synthesis [16],
polymer replica method [17] etc, with porosities of more than 60% but strength of less than
20MPa. It is reported that zirconium phosphate (ZrP2O7) is a kind of high-temperature stable
and low liner thermal-expansion coefficient binder material [18]. In this paper, we motivated
to prepare -Si3N4 matrix porous ceramics of controlled porosity by using ZrP2O7 as a binder,
which utilized a net shape, pressureless sintering technique, and explored the correlation
between the microstructure and mechanical properties.
Low Temperature Preparation of Phosphate Bonded Silicon Nitride Ceramics ...
35
1. PHOSPHATE BONDED SILICON NITRIDE POROUS CERAMICS

1.1. Zirconium Phosphate Bonded Silicon Nitride Porous Ceramics
1.1.1. Synthesis and Sintering Mechanisms
The ceramics were prepared by mixing the ZrO2 and H3PO4 in stoichiometric ratio as a
precursors at room temperature and then adding this mixture to Si3N4 powder with certain
designed content to achieve a final mixture of ZrP2O7 and Si3N4. After cold iso-static pressing
(CIP) at 200MPa, the specimen was firstly heated at 300 oC for 2 hours so as to realize a total
formation of ZrP2O7 bonded Si3N4 and porous structure, and then sintered at 1000 oC and
1100 oC in a furnace protected by nitrogen (N2) atmosphere and holding 2 h.
Phosphoric acid (H3PO4) has three hydrogen atoms bonded to oxygen atoms in its
structure and all three hydrogens are acidic to varying degrees and can be lost from the
molecule as H+ ions. As can be seen from Figure 1, when phosphoric acid (H3PO4) ionized to
H2PO4-, HPO42-, or PO43-, ZrO2 then reacts with them and as a result, Zr(HPO4)2H2O and
Zr(PO4)(H2PO4)2H2O are detected. When two phosphoric acid molecules are condensed into
one molecule, pyrophosphoric acid (H4P2O7) is obtained at about 150 oC as Eq. (1):
Figure 1. X-ray diffraction patterns for reaction products between ZrO2 and H3PO4 heat treated at
various temperatures:
:ZrP2O7, :Zr(HPO4)2.H2O, :Zr(PO4)(H2PO4).2H2O, :c-ZrO2, : t-ZrO2.
2H3PO4H4P2O7+H2O
(1)
Thus, ZrP2O7 is produced at the same time. Pure ZrP2O7 is finally formed at 250 oC and no
phase transformation occurs even at temperatures above 1000 oC, which is the sintering
temperature used in this paper. We write the final reaction between ZrO2 and H3PO4 as
follows:
36

ZrO2+2H3PO4ZrP2O7+3H2O
(2)
The continuous reaction of ZrP2O7 leads to the formation of micropores, and the ZrP2O7
binder bonds well onto the Si3N4 grains.
The sintering process of the ZrP2O7 bonded Si3N4 porous ceramics is described by two
stages: one is from room temperature to 250oC, which is the process mainly caused by the
formation of ZrP2O7 and porosity; the other is after ZrP2O7 totally forms, from 250oC to the
sintering temperature, which is dominated by the sintering of the ZrP2O7 bonded Si3N4.
(1) Stage I: Room Temperature to 250 Oc

After uniformly stirring the Si3N4, ZrO2 and H3PO4 at room temperature and cold isostatic pressing thereafter, Si3N4 and ZrO2 particles are surrounded with H3PO4 due to the
capillary force and pressure of adjacent particles. The reaction between ZrO2 and H3PO4
finishes at 250 oC, forming ZrP2O7. Figure 2 shows the microstructure of a 20 wt.% ZrP2O7
bonded Si3N4 porous ceramic sintered only at 250 oC. It is seen that Si3N4 particles are
surrounded by the ZrP2O7 the binder phase, suggesting that ZrP2O7 can bond well onto the
Si3N4 particles during the reaction. However, the loose structure of the Si3N4 particles
indicates that although the ZrP2O7 phase is totally formed at 250 oC, it does not adhere to the
Si3N4 particles. Only an interparticle necking structure is observed in Figure 2. This is
because no grain growth of ZrP2O7 has occurred at this low temperature and the ZrP2O7 phase
just bonds onto the Si3N4 particles but the ZrP2O7 bonded Si3N4 porous bodies are not
sintered.
Figure 2. Fracture surface of 20 wt.% ZrP2O7 bonded Si3N4 porous ceramic sintered at 250oC.
37
(2) Stage II: 250oC to Sintering Temperature (~ 1000 OC)

At temperatures above 250 oC, the ZrP2O7 bonded Si3N4 bodies sinter and the grains
grow gradually. Figure 3 shows the microstructure of a 20 wt.% ZrP2O7 bonded Si3N4 porous
ceramic sintered at 1000 oC. It is seen that this ZrP2O7 bonded Si3N4 ceramic shows a
continuous porous structure. Si3N4 particles are adhered together and the ZrP2O7 binder phase
performed well in bonding with the Si3N4 grains, indicating a good adhesion between the
grains and, thus, overall good sintering behavior. Generally, a large amount of fine and
uniform micropores with an average pore size less than 0.5 m are observed.
Figure 4 provides an overview of the sintering process for the ZrP2O7 bonded Si3N4
porous ceramic. Firstly, Si3N4, ZrO2 and H3PO4 are mixed uniformly and after the cold isostatic pressing, Si3N4 and ZrO2 particles are surrounded with H3PO4, creating a porous
structure of disconnected particles. Secondly, during sintering stage I, from room temperature
to 250 oC, ZrO2 reacts completely with H3PO4 to form ZrP2O7. During this reaction, due to
the SiO2 amorphous film existing on the surface of the Si3N4 grains, the ZrP2O7 phase bonds
well onto the Si3N4 and an interparticle necking structure is obtained. Thirdly, during
sintering stage II, from 250oC to the sintering temperature (1000 oC in this work), solid state
sintering of the ZrP2O7 is dominant and a continuous open porosity network is obtained.
1.1.2. Phase Compositions and Microstructure
Figure 5 showed the XRD patterns of both the Si3N4 raw material and 20 wt.% ZrP2O7
bonded Si3N4 specimen which was sintered at 1100 oC. The XRD spectra showed that -Si3N4
was the main matrix phase which maintained the character of Si3N4 raw material with just a
little amount of -Si3N4 and no apparent phase transformation from -Si3N4 to -Si3N4 was
detected. The phases of ZrP2O7 and residual ZrO2 were identified from the XRD spectra. Due
to the solid state sintering of ZrP2O7 binder around the Si3N4 grains, the sintering of the whole
composite was finished at 1000 oC or 1100 oC, which was the main sintering mechanism. As
shown in Figure 1 that up to the temperature of 250 oC and even in the sintering temperature
of 1100 oC, no ZrO2 phase existed after the reaction finishing and only the pure ZrP2O7 could
be finally acquired. This residual ZrO2 in Figure 5 can be explained as the lack of H3PO4
which had reacted with SiO2 presenting on the surface of the silicon nitride particles. The
evidence was shown in ref [19]. The reason that the amount of silicon phosphate not shown
by XRD was believed that the majority of silicon phosphate was amorphous [20] and
undetectable by XRD.
Figure 6 showed the SEM microphotograph of the fracture surface of the samples which
were sintered at 1100 oC with 5 wt.%, 15 wt.% and 25 wt.% content of ZrP2O7. As can be
seen clearly that with the increasing content of ZrP2O7, apparent binder effect was well
performed to form the structure of ZrP2O7 bonded Si3N4 porous ceramics. When the content
of ZrP2O7 was low, 5 wt.% as shown in Figure 6(a), porosity was mainly caused by partially
sintered Si3N4 particles and the incompact structure was observed. On the other hand, , when
the content of ZrP2O7 reached at 25 wt.% i.e. in Figure 6(c), ZrP2O7 binder phase performed a
well binding property with Si3N4 grains, indicating a good adhesion between the grains and,
thus, overall good sintering behavior. Generally, a large amount of fine and uniform
micropores with pore size of less than 0.5 m were observed in this kind of porous materials.
38
Figure 3. SEM image of 20wt.% ZrP2O7 bonded Si3N4 porous ceramic sintered at 1000oC.
Figure 4. Schematic of the sintering process for the ZrP2O7 bonded Si3N4 porous ceramic.
39
Figure 5. X-ray diffraction spectra of sintered ZrP2O7 bonded Si3N4 ceramic and Si3N4 raw powders: :
ZrP2O7, : -Si3N4, : -Si3N4, : ZrO2.
Figure 6. SEM of ZrP2O7 bonded Si3N4 porous ceramics. The content of ZrP2O7 were: (a) 5 wt.%; (b)
15 wt.%; (c) 25 wt.%.
Figure 7 shows the HRTEM images of a 20 wt.% ZrP2O7 bonded Si3N4 porous ceramic
sintered at 1000 oC. The images in Figure 7(a) indicate that the bulk material consists
primarily of the -Si3N4 phase, as determined by measuring the interplanar spacing. It can be
seen that virtually no Si3N4 grain growth occurs after sintering. A thin amorphous film
surrounding the Si3N4 particle is observed, which should be the SiO2P2O5 formed by the
reaction between SiO2 and H3PO4 as mentioned by Barta and Manela [21]. Figure 7(b) is the
EDX image corresponding to Figure 7(a), indicating that there are mainly Si, P, and O atoms
and a low concentration of Zr detected on the surface of the sintered Si3N4 grains, providing
further evidence of the existence of the SiO2P2O5 phase. So, it is the SiO2 film surrounding
the Si3N4 grains that makes it easy for H3PO4 to bond onto the Si3N4 grains during the initial
reaction and, as a result, ZrP2O7 can bond well onto the Si3N4 particles after it is formed at
250 oC. Figure 7(c) is a HRTEM image of the 20 wt.% ZrP2O7 bonded Si3N4 porous structure.
40
It is observed that the Si3N4 grains are spherical and are surrounded by the SiO2-P2O5
amorphous film. There is another crystal structure existing between the Si3N4 grains, a higher
magnification of which is shown in Figure 7(d). Figure 7(e) shows the HRTEM image of the
binder phase existing between the Si3N4 grains viewed from the <121> direction. Lattice
parameters of 0.47 nm and 0.37 nm, and the planar angle of 75o, are measured, respectively.
The lattice parameter is quite in agreement with the lattice parameter of ZrP2O7, the planar
spacing for which is 0.47 nm in {111} plane and 0.37 nm in {210} plane. So, it is confirmed
that the binder phase between the Si3N4 grains is ZrP2O7. The presence of crystalline ZrP2O7
indicates that consolidation occurs due to the solid state sintering of ZrP2O7 between Si3N4
grains, which has the potential to improve the mechanical properties of the Si3N4 porous
ceramics.
Figure 7. HRTEM of of 20wt.% ZrP2O7 bonded Si3N4 porous ceramic sintered at 1000oC. (a)
Morphology of the single Si3N4 particle in the 20wt.% ZrP2O7 bonded Si3N4 bulk porous ceramic and
its high resolution lattice image, proving it is -Si3N4. (b) EDX of the single Si3N4 particle in (a),
showing there are Zr, P, and O on the surface of the Si3N4 particle. (c) Morphology of the 20wt.%
ZrP2O7 bonded Si3N4 porous ceramic, including the Si3N4 particles, pores, and ZrP2O7 binder. (d) High
magnification image of the black box in (c), which is believed to be the ZrP2O7 binder phase, a lattice
image is clearly shown. (e) High resolution lattice image of the ZrP2O7 binder phase.
41
Figure 8 presented the porosities of the sintered specimens, as a function to the content of
ZrP2O7 in the starting mixture. In general, with increasing the content of ZrP2O7, a decrease
of the porosity was observed. It could be seen that the porosities ranged from 33 % to 47 %
could be obtained and showed almost linear variation against the content of ZrP2O7 altering
from 5 % to 30 %. Sintering temperature was not the dominating factor although there was a
slight difference between 1000 oC and 1100 oC, which suggested that porosity could be well
controlled by adjusting the amount of ZrP2O7 binder without changing the sintering
conditions. Open porosity measured according to European Standard EN 623-2 (1993) which
was not listed here suggested that the obtained ZrP2O7 bonded Si3N4 porous ceramics
composed of nearly all open pores, which was quite consistent with the fabrication methods
by using H3PO4 as a starting material to acquire ZrP2O7 binder.
1.1.3. Mechanical Properties

Figure 9 showed the bending strength as a function of the content of ZrP2O7. It was clear
that bending strength of the porous ceramics were in the range of 40~114 MPa and had a
close correlation with the porosity. The specimens exhibited relatively high bending strength,
which could reach 114 MPa when the sintering temperature was 1100 oC and the content of
ZrP2O7 binder is 30 wt.%, although it contained 38 % porosity, which enabled us to prepare
-Si3N4 matrix porous ceramics of high strength by adjusting the porosity of the samples.
This high bending strength for the porous material was very likely due to the reaction of
ZrP2O7 binder as could easily be seen in Figure 6(c).
The Youngs modulus is a key parameter in strength and toughness relations. Youngs
moduli of polycrystalline -Si3N4 typically fall in the range of 90-110 GPa depending on the
intergranular phase content, texture, and the relative amount of and phase. In addition,
controlling the level of porosity, the pore size, and pore shape are important in attaining
porous structures with improved mechanical performance. Figure 10 showed the Youngs
modulus as a function of the content of ZrP2O7. Youngs modulus from 20-50 GPa were
obtained and had nearly the same trend with bending strength.
Figure 8. Effect of the content of ZrP2O7 on the porosities of sintered samples.
42
Figure 9. The bending strength of the sintered porous ceramics.
Figure 10. The Youngs modulus of the sintered porous ceramics.
1.2. Phosphate Bonded Silicon Nitride Porous Ceramics with Bimodel Pore
Structure
1.2.1. Synthesis
The experiment was carried out by mixing stoichiometric ratios of the ZrO2 (produced by
H.C. Starck, Germany, with an average particle size of ~ 0.5 m and a purity of 99.9 %) and
H3PO4 (the concentration of H3PO4 of 85 % and the purity of up to 99.9 %) as precursors at
room temperature together with the Si3N4 powder (produced by UBE industries, Co., Ltd,
43
Japan with an average particle size of ~ 0.5 m and /(+)>93 wt.%) at a certain
concentration to achieve a mixture of 25wt.% ZrP2O7 and 75wt.% Si3N4. This mixture was
then mixed with starch powder (with uniform particle size of ~ 10 m) at various
concentrations to achieve a final mixture. After cold iso-static pressing (CIP) at 200 MPa, the
specimen was heated at 250 oC for 2 hours to promote the formation of the ZrP2O7 bonded
Si3N4 with a porous structure, then heated at 550 oC for 2 hours to burn out the rest starch
powder. Samples were then sintered at 1000 oC in a furnace protected by nitrogen (N2)
atmosphere and held for 2 hours.
1.2.2. Phase Compositions and Microstructure

Figure 11 demonstrates the phase compositions of reaction products between ZrO2 and
H3PO4 heat treated at various temperatures. Our previous work on preparation of ZrP2O7
bonded Si3N4 porous ceramics has revealed that a large amount of fine and uniform pores
with pore size of less than 0.5 m are formed during the continuous reaction of ZrP2O7
derived from ZrO2 and H3PO4 at the temperature of ~ 250 oC. The formed ZrP2O7 binder,
which is stable existed of higher than 1000 oC, (confirmed by Figure 11) bonds well on the
Si3N4 grains, which promotes the enhancement of the mechanical strength. It is proved by
therml analysis that the starch powder starts to burn out to release CO2 at ~ 350 oC and
completes at ~ 550 oC, so that the rest of the pores with pore size of as large as the particle
size of used starch (~ 10 m) are expected to form then.
Table 1 lists the measured values of bulk density, open porosity, total porosity and linear
shrinkage for starch processed ZrP2O7 bonded Si3N4 porous ceramics. It is noted that the open
porosity of the obtained samples is almost equal to the total porosity, indicating that nearly all
pores are open pores. It is also obviously observed that the linear shrinkage is surprisingly
low, which is ~ 2 % for each sample. The reason is clear now: the formation of the pores for
starch processed ZrP2O7 bonded Si3N4 porous ceramics includes two processes: (1) The
reaction of ZrP2O7 derived from ZrO2 and H3PO4 completes at ~ 250 oC and pores with pore
size of less than 0.5 m are continuously formed. Due to the high mechanical strength of
ZrP2O7 binder phase, the shrinkage of the Si3N4 ceramics will be restrained. (2) Due to the
burnout of starch completes at ~ 550 oC, the rest of the pores with large pore size are formed,
and the generated CO2 gas passes through the small open pores formed at 250 oC, causing
limited volume change even after the samples are sintered at 1000 oC. Eventually, the total
porosity of starch processed 25 wt. % ZrP2O7 bonded Si3N4 porous ceramics is in the range of
36 ~ 62.3 % when the starch content varies from 0 to 40 vol.%.
Figure 12 illustrates the theoretical porosity and experimental total porosity of the assintered starch processed 25 wt.% ZrP2O7 bonded Si3N4 porous ceramics as a function of
starch content. As the porosity of 25 wt.% ZrP2O7 bonded Si3N4 porous ceramic without
starch is 36.0 %, and the porosity of the matrix is assumed to maintain the same during 550
o
C to 1000 oC, then the theoretical porosity of the starch processed 25 wt.% ZrP2O7 bonded
Si3N4 porous ceramics could be calculated by the following Eq. (3):
theoretical
36% (100% starch ) starch
(3)
where is the volume content of starch. It is indicated in this formula that the starch burns out
totally and the space where all the starch powder occupies will form the pores and no volume
44
change is occurred by neglecting the volume change of ZrP2O7 bonded Si3N4 porous body. It
is obviously seen that the experimental results are in prefect agreement with the theoretical
porosity derived form Eq. (3), indicating nearly all starch powder burns out to form the pores
and nearly no volume change is occurred during the sintering of the porous ceramics, which is
also the reason for limit shrinkage during the burnout of the starch as shown in Table 1. Thus,
it is indicated that the porosity of the as-sintered starch processed 25 wt.% ZrP2O7 bonded
Si3N4 porous ceramics is well controlled and precisely predicted by adjusting the starch
content.
Figure 11. X-ray diffraction patterns for reaction products between ZrO2 and H3PO4 heat treated at
various temperatures.
Table 1. Bulk density, open porosity, total porosity and linear shrinkage of starch
processed 25 wt.% ZrP2O7 bonded Si3N4 porous ceramics
Starch
content
(vol.%)
0
10
16
20
23
30
33.3
40
Bulk density
(g/cm3)
Open porosity
(%)
Total porosity
(%)
Linear shrinkage
(%)
2.03 0.02
1.83 0.02
1.71 0.01
1.63 0.02
1.55 0.01
1.40 0.01
1.35 0.02
1.20 0.02
34.5 0.2
42.0 0.3
46.0 0.1
48.5 0.2
51.2 0.1
55.5 0.3
57.0 0.1
61.5 0.2
36.0 0.3
42.5 0.4
46.1 0.1
48.7 0.3
51.2 0.1
55.8 0.3
57.4 0.2
62.3 0.2
2.0 0.2
2.1 0.3
2.0 0.2
2.0 0.2
1.8 0.4
1.6 0.2
1.5 0.1
1.4 0.2
45
Figure 13 illustrates the typical microstructures of starch processed 25 wt.% ZrP2O7

bonded Si3N4 porous ceramics with starch content of 40 vol.% after sintering for 2 h at 1000
o
C. It is obviously observed from Figure 13(a) that a large amount of sphere pores with pore
size of ~ 10 m are uniformly distributed in the as-sintered sample. Both the shape and the
size of the pores are quite in agreement with the raw starch powder used in this study,
indicating that the formation of such large pores is due to the burnout of the starch powder
completed at 550 oC and the pore structure maintains the character even after sintering at
1000 oC. Figure 13(b) illustrates a higher magnification image (black box area in Figure
13(a)) highlighting the ligament area among large pores formed by starch burnout. It is seen
clearly that the ligament is about ~ 5 m in width and a large amount of fine open pores with
pore size of < 0.5 m are formed in the ligament, which is the typical microstructure of
ZrP2O7 bonded Si3N4 porous ceramics shown in Figure 13(c). It is clearly seen that the
ZrP2O7 binder phase is between Si3N4 particles, which performs like a support due to the
effect of phosphate on high binding strength of porous Si3N4 ceramics, which restrains the
shrinkage of the sintered porous material. On the other hand, the ZrP2O7 bonded Si3N4 porous
structure contains a large amount of small open pores, promoting the increase of the porosity
and supplying a path for the release of the CO2 by burnout of the starch. It is hinted in this
way that not only the porosity but also the pore structure can be freely controlled and
precisely designed according to different application needs.
1.3. Phosphate Bonded Silicon Nitride Porous Ceramics Using a H3PO4 PoreForming Agent
1.3.1. Synthesis and Sintering Mechanism
The ceramics are prepared by mixing the Si3N4 powders and H3PO4 in different content
ratio. After cold iso-static pressing (CIP) at 200MPa, the specimen is firstly heated at 500 oC
in a slow heating ramp to prevent the pore expansion resulting from the volatilization of
H3PO4 and holding 2 hours so as to realize a total formation of Si3N4 porous structure, and
then sintered at 1000~1200 oC in a furnace protected by a nitrogen (N2) atmosphere and
holding 2 hours.
Phosphoric acid (H3PO4) has three hydrogen atoms bonded to oxygen atoms in its
structure and all three hydrogens are acidic to varying degrees and can be lost from the
molecule as H+ ions. When two phosphoric acid molecules are condensed into one molecule,
pyrophosphoric acid (H4P2O7) is obtained at about 150 oC as follows:
2H3PO4H4P2O7+H2O
(4)
Si3N4 can not react with H3PO4, but SiO2 presenting on the surface of the silicon nitride
particles [22] can react with H3PO4. The evidence is shown in Figure 14. SiO2 fine powders is
mixed with H3PO4 in stoichiometric ratio and heat treated at 500 oC for 2 h, only the peaks
that are characteristic of SiP2O7 appear in the patterns, indicating that the totally reaction
between SiO2 and H3PO4 is completed. We use a final reaction between SiO2 and H3PO4 as
follows:
SiO2+2H3PO4SiP2O7+3H2O
(5)
46
Figure 12. Effect of the starch content on the experimental and theoretical porosity of the 1000 oC assintered starch processed 25 wt.% ZrP2O7 bonded Si3N4 porous ceramics.
Figure 13. Fracture surface of starch processed 25 wt.% ZrP2O7 bonded Si3N4 porous ceramics with
starch content of 40 vol.% after sintering for 2 h at 1000 oC. (b) is a higher magnification image (black
box area of (a)) highlighting the area among large pores formed by starch burnout. (c) TEM image of
the pore ligament highlighted in (b).
SiP2O7 is a crystalline network polymer in which every Si atom is joined to six oxygens and
every P to four oxygens, with a connectivity of four and six [23]. The P atoms are joined in
pairs through a shared O atom to form P2O74- units, each of which shares its six remaining
oxygens with the SiO6 octahedra. SiP2O7 has potential applications as a refractory material
due to its high temperature resistance, high thermal conductivity, high electrical resistance,
and low thermal expansion behavior. Thus, it is appropriate to maintain the high temperature
properties of Si3N4 by using SiP2O7 as the binder material.
The mixture of SiO2 and H3PO4 in stoichiometric ratio is also measured by thermal
analysis. Figure 15 presents the results of thermal analysis. The sample firstly starts to lose
weight at about 80 oC and ends at about 300 oC. Correspondingly, there is an endothermal
peak in the same temperature zone in the DTA curve due to the volatilization of H2O in
H3PO4 and the formation of H4P2O7. The sample then continuously loses weight at about 490
o
C and ends at about 600 oC. Correspondingly, there is an exothermal peak in the same
temperature zone in the DTA curve due to the reaction between SiO2 and H3PO4 and
47
formation of SiP2O7. When the temperature is higher than 1000 oC, a losing weight
phenomenon is also seen due to the further volatilization of H3PO4.
The porous structure is formed then by both the volatilization of H3PO4 and continuous
reaction of SiP2O7 which could bond well on to the Si3N4 grains. These two factors are
believed to be the sintering mechanism and Si3N4 ceramics with porous microstructure can be
formed in this way.
Figure 14. X-ray diffraction diagrams of reaction products between SiO2 and H3PO4 at 500 oC.
Figure 15. TG-DTA curve of SiO2 and H3PO4 mixture.
48
1.3.2. Phase Composition and Microstructure

Figure 16 shows the XRD patterns of both the Si3N4 raw material and 50 vol.% H3PO4
bonded Si3N4 specimen which is sintered at 1200 oC. The XRD spectra shows that -Si3N4 is
the main matrix phase which maintains the character of Si3N4 raw material with just a little
amount of -Si3N4 and no apparent phase transformation from -Si3N4 to -Si3N4 is detected.
The little phases of SiP2O7 are identified from the XRD spectra, proving that the reaction
between H3PO4 and SiO2 presenting on the surface of the silicon nitride particles is occurred.
The reason that the amount of silicon phosphate not obviously shown by XRD is believed that
the majority of silicon phosphate is amorphous and undetectable by XRD.
Figure 17 shows the SEM microphotograph of the fracture surface of the samples which
are sintered at 1000 oC with 35 vol.%, 50 vol.% content of H3PO4 and at 1200 oC with 10
vol%, 35 vol.%, 50 vol.% content of H3PO4. As can be seen clearly, with increasing the
content of H3PO4, the binder phase takes obviously effects to form the structure of porous
Si3N4 ceramics. When the content of H3PO4 is low, 10 vol.% as shown in Figure 17(c),
porosity is mainly caused by partially sintered Si3N4 particles and the incompact structure is
observed, indicating a low bending strength. On the other hand, when the content of H3PO4
reaches at 35 vol.% i.e. in Figure 17(a)(d), SiP2O7 binder phase performs a well binding
property with Si3N4 grains, indicating a good adhesion between the grains and, thus, overall a
good sintering behavior. As a result, a large amount of fine and uniform micropores with pore
size of less than 0.5 m are observed in this situation. When the content of H3PO4 is 50 vol.%
i.e. in Figure 17(b)(e), large and uniform pores with pore size of more than 1m are observed
and pores are apparently formed by the volatilization of H3PO4.
Figure 16. X-ray diffraction spectra of sintered H3PO4 bonded Si3N4 ceramic and Si3N4 raw powders.
49
Figure 17. SEM of H3PO4-Si3N4 porous ceramics. The sintering temperature and the content of H3PO4
are: (a)1000 oC, 35 vol.%; (b)1000 oC, 50 vol.%; (c)1200 oC, 10 vol.%; (d)1200 oC, 35 vol.%; (e)1200
o
C, 50 vol.%.
Figure 18 presents the porosities of the sintered specimens, as a to the content of H3PO4
in the starting mixture. It can be seen that the porosities ranged from 42 % to 63 % can be
obtained when the content of H3PO4 alters from 0 % to 50 vol.%. Sintering temperature is the
dominating factor although there is a slight difference in the range of 1000~1200 oC, which
suggests that the porosity can be well controlled by adjusting the amount of H3PO4 and the
sintering conditions. In general, with increasing the content of H3PO4, the porosity decreases
firstly and then increases. This phenomenon is quite consistent with the sintering mechanism.
When the content of H3PO4 is lower than 35 vol.%, the reaction sintering process is
dominated with continuous generating of SiP2O7 binder phase and the porosities show slight
decrease as a function of the content of H3PO4, which can reflect the amount of SiP2O7. When
50
the content of H3PO4 is equal or higher than 35 vol.%, the volatilization of an excessive
amount of H3PO4 takes more effects to form the pores and porosities increase observably
against the content of H3PO4. Because of the further volatilization of a larger amount of
H3PO4 at high temperature, the samples sintered at 1200 oC show larger porosity compared
with the samples sintered at 1000 oC. Open porosity measured according to European
Standard EN 623-2 (1993) which is not listed here suggests that the obtained Si3N4 porous
ceramics composes of nearly all open pores, which is quite consistent with the fabrication
methods by using H3PO4 as the starting material to form the pores.
Figure 18. Effect of the content of H3PO4 and the sintering temperature on the porosities of sintered
samples.
1.3.3. Mechanical Properties

Figure 19 shows the bending strength as a function of the content of H3PO4. The
specimens exhibit relatively high bending strength, it is clear that bending strength of the
samples is in the range of 50~120 MPa and has a close correlation with the content of H3PO4.
When the content of H3PO4 is lower than 35 vol.%, the bending strength is lower than 100
MPa and shows a linear variation against the content of H3PO4, as porosities in Figure 15
shows little variation, the resulting mechanical strength is primarily a function of the amount
of SiP2O7. The bending strength increases with increasing the content of H3PO4 because of
the continuous formed SiP2O7 binder taking effects, which is in agreement with Figure 17(c).
On the other hand, when the content of H3PO4 is equal or higher than 35 vol.%, the bending
strength decreases with increasing the H3PO4 contents, namely the porosity seen from Figure
16, porosity dominates the mechanical strength. Contrasted with Figure 16, if the bending
strength is required above 100 MPa, porosities from 43 % to 57 % can realize it by using
different content of H3PO4 and proper sintering temperature. Compared with the results
obtained for the Si3N4-ZrO2-H3PO4 previously concerned, similar bending strength is reached
by using H3PO4 as the pore forming agent but the porosity is higher. This is significant to
guarantee some properties like dielectric properties which are sensitive to the porosity under a
51
certain bending strength. The maximal bending strength is 120 MPa when the sintering
temperature is 1200 oC and the content of H3PO4 is 35 vol.%, although it contains 50 %
porosity. This high bending strength for the porous material is very likely due to the reaction
of SiP2O7 binder as could easily be seen in Figure 17(d).
Figure 19. The bending strength of the sintered porous ceramics as a function of the content of H3PO4.
2. SILICON NITRIDE CERAMICS USING

PHOSPHATE AS SINTERING ADDITIVES
2.1. Synthesis and Sintering Behavior
According to Table 2, the powder precursors in various ratios were ball milled in ethanol
for 10 h in a nylon jar, using sintered Si3N4 grinding media. After the powder mixtures were
dried, they were passed through a 100 mesh sieve to eliminate large agglomerates, then
loaded in a graphite die with the inner diameter of 32 mm and pre-pressed to some extent and
sintered using a SPS system (model-1050, Sumitomo Coal Mining Co. Ltd., Tokyo).
Table 2. The ratio (wt.%) of raw powders and
the theoretical density of every sample
No.
S1
S2
S3
S4
Si3N4
94
92
86
80
MgO
4
4
4
4
AlPO4
2
4
8
16
52
The sintering temperature of 1300~1500 oC of each sample was measured by an optical

pyrometer focused on the surface of the graphite die. An average heating rate of 100 oC /min
was maintained up to the sintering temperature and then the temperature was held constant
for 5 min. A uniaxial pressure of 30 MPa and a nitrogen (N2) atmosphere were applied from
the start to the end of the sintering cycle. The sample was cooled naturally in the nitrogen
(N2) atmosphere after the sintering period finished.
For sake of constraining the conversion of -Si3N4 to -Si3N4, the powder mixtures are
sintered at a low temperature of 1300~1500 oC. The temperature is measured by an infrared
pyrometer at the outer surface of the graphite die. It is believed that a discrepancy
(underestimate) of about 150~200 oC exists between the temperature at the surface and that at
the interior under such circumstances [24]. So, -Si3N4 can be sintered even at 1500 oC by
SPS.
Consolidation behavior is best represented by the shrinkage and shrinking rate
concurrently measured during the sintering. An example of the dependence as a function of
the amount of additive in the powder mixtures is shown in Figure 20. In all these cases the
sintering temperature is 1500 oC. The AlPO4 additive reduces the thermal expansion during
the sintering, the more the AlPO4 is added, the more obvious of this phenomenon is. This can
be due to the low thermal expansion coefficient of AlPO4. Furthermore, the starting sintering
temperature is lower by adding AlPO4, 1100 oC according to Figure 20. With a high AlPO4
level (e.g., S4, 16 wt.%), densification is completed in a very short time and relative low
temperature (1400 oC) after the peak shrinking-rate is reached, which suggests that AlPO4
performs good sintering behavior. While with a low AlPO4 level (e.g., S1, 2 wt.%), sintering
is attributed mainly to the effect of MgO which is not enough to bring the material to high
density at low temperature. In this situation, consolidation of this material is finished at 1500
o
C.
Figure 20. Dependence of sintering behavior on different additive amount at 1500 oC.
53
The relative density of S1-S4 as a function of different sintering temperatures and content
of sintering additives is shown in Figure 21, respectively. It can be seen from Figure 22 that
the relative densities show sharp variation against the sintering temperature from 1300 oC to
1500 oC when the content of the AlPO4 is 2 wt.% (S1) and 4 wt.% (S2), while when the
content of AlPO4 is 8 wt.% (S3) and 16 wt.% (S4), this phenomenon is from 1300 oC to 1400
o
C. For S3 and S4, the relative densities fluctuate little when the sintering temperature is from
1400 oC to 1500 oC. Samples with the relative density from 58 % to 98 % are acquired. We
can conclude that when the content of AlPO4 or the sintering temperature reaches certain
values, this material can be compactly sintered and the relative density changes little. Whats
more significant, nearly the same relative density of the sintered samples can be obtained by
adjusting the content of the sintering additives and sintering temperature which may have
different performances. An example of this is when S1 sintered at 1400 oC, S3 sintered at
1350 oC and S4 sintered at 1300 oC, the relative density of these three samples is about 73 %,
nearly the same, however, the samples may have different properties like mechanical
properties.
Figure 21. Effect of sintering temperature and content of sintering additives on the relative density of
sintered samples.
2.2. Phase Compositions and Microstructure

Figure 22 shows the SEM observation of fractured surface of sintered samples. It can be
seen that the volume of pores decreases remarkably from S1 to S4 when the sintering
temperature is 1400 oC. This result is identical to that from Figure 21. It is worthy to note that
nearly none -Si3N4 is observed and the grain size is about 0.5 m, which indicates there is no
obvious grain growth in the SPS process. Whats more, for sintered sample S4, when the
sintering temperature is 1300, the relative density is 73.5 %, (Figure 22 (c)), while the same
54
relative density can be obtained for sample S1 and sintering temperature of 1400 oC, (Figure
22(a)), whose relative density is 73.4 %. For Figure 22(a), porosity is mainly caused by
partially sintered Si3N4 particles and the incompact structure is observed. On the other hand,
when the content of AlPO4 reaches 16 wt.% in Figure 22(c), AlPO4 binder phase performs a
well binding property with Si3N4 grains, indicating a good adhesion between the grains and,
thus, overall a good sintering behavior and promising mechanical properties.
The identified phases for all compositions of samples S4 sintered from 1300 oC to 1500
o
C are presented in Figure 23, which are -Si3N4, -Si3N4, MgSiO3 and AlPO4. Compared
with the XRD of raw Si3N4 powder, it indicates that the major phase is -Si3N4, and the
sintered samples preserve the characters of the starting powders basically. When increasing
the sintering temperature, a little conversion from -Si3N4 to -Si3N4 is detected.
The product of identified MgSiO3 illustrates one of the sintering mechanisms that MgO
can react with SiO2 presenting on the surface of the silicon nitride particles [25] and the
reaction promotes the mixture powders adherence [26-27] which to some extent assure of
certain mechanical properties of the sintered samples.
Figure 22. The SEM images of sintered samples: (a)S1, 1400 oC; (b)S3, 1400 oC; (c) S4, 1300 oC; (d)
S4, 1400 oC.
55
AlPO4 is chemical inert, thermally stable (melting point=~2000 oC) and highly covalent
which exists stably during the sintering process [28]. D. K. Kim fabricated AlPO4 composites
and found no volatilization of P2O5 from AlPO4 [29]. Due to the solid state sintering of AlPO4
around the Si3N4 grains, the sintering of the whole composite is finished at 1300 oC to 1500
o
C according to different content of AlPO4, which was the other sintering mechanism.
Figure 23. The X-Ray diffraction patterns of (a) raw Si3N4 powder and S4 at different sintering
temperature: (b) 1300 oC; (c) 1400 oC; (d) 1500 oC.
2.3. Mechanical Properties

Figure 24 shows bending strength of the sintered samples as a function of relative
density. The specimens exhibit relatively high bending strength, it is clear that bending
strength of the samples is in the range of 50~420 MPa and has a close correlation with the
relative density. The factors influencing the relative densities which are content of sintering
additives and sintering temperature have been discussed in former part of this paper. From
Figure 24, if the bending strength is required around 200 MPa, relative densities from 75 % to
97% can realize it by using different sintering additives and proper sintering temperature.
This is significant to guarantee some properties like dielectric properties which are sensitive
to the relative densities under a certain bending strength. On the other hand, when the relative
density is confirmed, the bending strength increases with increasing the content of sintering
additives, generally. This can be explained reasonably by Figure 22 (a) (c). It is sintering
behavior that causes different bending strength. The high bending strength for the sintered
samples is very likely due to the solid state sintering of AlPO4 binder as could easily be seen
in Figure 22.
56
3. PHOSPHATE BONDED SILICON NITRIDE

CERAMICS APPLIED FOR BROADBAND RADOME
3.1. Optimal Designs for the Radome Wall Structure
3.1.1. Hypotheses
Before designing for a radome by calculating the transmission efficiency in an effective
way, some hypotheses are made as follows: (1) Plane wave solutions are used in
mathematical descriptions of wave propagation in order to synthesis more complicated
wavefronts. A plane wave is a mathematical but useful idealization because at large distances
from sources and over regions of restricted size, curved wavefronts can be described
approximately by plane wave functions.
(2) The theory of plane wave propagation through a plane dielectric sheet is used for
radome design because a curved radome can be approximated as local plane. Thus, this paper
studies propagation through flat sheets only. The flat sheet is a practically useful and
instructive boundary value problem, which demonstrates quantitatively how wave
propagation depends on the dielectric constant and thickness of the sheet as well as the wave
frequency, polarization, and incidence angle of the wave.
(3) A linearly polarized wave with the polarization either parallel or perpendicular to the
plane of incidence is considered and the calculation methods for the complex valued
transmission efficiency of a homogeneous, isotropic, nonmagnetic and dielectric sheet are
developed.
Figure 24 Bending strength of the sintered samples vs relative density.
57
3.1.2. Calculation Model

Over the past few decades, several simulation methods for broadband electromagnetic
wave penetration property have been developed, including the basic electromagnetic wave
theory, finite elemental analysis, transmission line method, etc. [30-32]. For single layer
structure, it is easy to calculate the wave transmission properties as described by ref. [33]. For
multilayer structure, such as A-sandwich (three layers), C-sandwich (five layers), etc., the
calculation becomes complicated and it is difficult to calculate the transmission efficiency
directly when the layer number is greater than five [34]. In the present study, a microwave
equivalent network method is adopted to simplify the calculation for the graded porous
structure. The calculation model of graded porous radome wall structure is shown in Figure
25. The graded porous radome wall structure is divided into n layers with equal thickness and
the thickness of each layer is d/n (d is the wall thickness). The top and the bottom layers stand
for the outer and inner layer of the radome respectively. As a result, from the top to the
bottom, porosity increases gradedly from 0 (fully dense) to the highest value, while the
dielectric constant decreases from the highest value 1 to the lowest value n. Thus, the
dielectric constant of each layer can be expressed by Eq. (6).
n (1 n )(1 ( x)1/ p )
(6)
where x is the relative thickness, p is the structural coefficient of the graded porous radome
material. The dielectric constant () of each layer as a function of the relative thickness (x)
and structural coefficient (p) is calculated and shown in Figure 26. It is illustrated in Figure
26 that the greater the p is, the higher the relative thickness of layer containing the low
dielectric constant material is, which is severe to the mechanical reliability for the radome. As
a result, a low p value is expected in the design for the broadband radome. And p = 4 is
selected as the initial condition for the following calculation.
Figure 25(b) is the microwave equivalent network map for the model shown in Figure
26(a). Thus, it is easy to calculate the transmission efficiency by using the fundamental matrix
of the transmission line model, which is expressed by Eq. (7).
Zn
Z
Z
cos k1d 1 j 1 sin k1d 1 cos k 2 d 2 j 2 sin k 2 d 2

cos k nd n j Z sin k nd n
Z0
Z0
A B
0
C D Z
Z0
j 0 sin k d cos k d j Z0 sin k d cos k d

1 1
1 1
2 2
2 2
j Z sin k n d n cos k nd n
Z
Z
1
2
where kn
(7)
n sin 2 0 , n n (1 jtg n ) , d is the wall thickness, 0 is the
incident wavelength, and 0 is the incidence angle, Z is the electromagnetic wave impedance.
The calculation of Z for perpendicular polarization is given by Eq. (8).
n sin 2 0
Zn
Z0
n cos 0
While calculation of Z for parallel polarization is given by Eq. (9).
(8)
58
(a)
(b)
Figure 25. The calculation model of graded porous wall structure radome: (a) mathmatical model; (b)
equivalent network model.
Figure 26. The dielectric constant () of each layer as a function of the relative thickness (x) and
structural coefficient (p) for graded porous structure.
Zn
cos 0
Z0
n sin 2 0
59
(9)
From Eq. (8) and (9), it is clearly seen that when
0 0 , The Z of perpendicular polarization
is equal to that of parallel polarization. In order to simplify the calculation, we use
0 0 in
the following calculations and materials power transmission efficiency test.

2
The power transmission efficiency is T , where T is given by Eq. (10).
2
A B C D
(10)
3.3. Numerical Results and Discussion

For the calculation of power transmission efficiency of graded porous structure, the effect
of layer number (n), structural coefficient (p), wall thickness (d), dielectric constant of inner
layer (n) and outer layer (n) should be taken into consideration. Based on the Si3N4 ceramic
which is selected as the radome material in the present paper, the dielectric constant of asprepared dense Si3N4 ceramic is ~ 8, while dielectric constant of the as-prepared porous Si3N4
ceramic with highest porosity is ~ 2. Meanwhile, Si3N4 ceramic radome wall thickness is
evaluated in ref. [35], based on their design results, wall thickness of d = 6 mm is selected.
3.3.1. Effect of Layer Number (N) on the Transmission

Efficiency of the Graded Radome Material
The initial conditions are: p = 4, d = 6 mm, n = 2 and 1 = 8.
The transmission efficiency of the graded porous panel as functions of the frequency and
layer number (n) is shown in Figure 27(a). It is obviously seen that the greater the n is, the
higher the broadband transmission efficiency is, indicating that the continuous gradient
structure has better power transmission efficiency than the multilayer gradient structure.
When n > 5, the power transmission efficiency is higher than 70 % at 1~18 GHz, which is
good for the missile guide control. On the other hand, the transmission efficiency as a
function of the layer number (n) at 10 GHz is shown in Figure 27(b). It is obviously seen that
when n is less than 5, the transmission efficiency increases greatly with the increase of n, but
it fluctuates little when n is greater than 5. Thus, the layer number n = 5 will be more feasible
for the preparation of the gradient porous structure Si3N4 ceramic radome material. However,
in order to precisely calculate the transmission efficiency, we choose n = 10 in the following
calculations.
3.3.2. Effect of Structural Coefficient (P) on the Transmission Efficiency of the Graded
Radome Material
The initial conditions are: n = 10, d = 6 mm, n = 2 and 1 = 8.
60
The transmission efficiency of the graded porous panel as functions of the frequency and
structural coefficient (p) is shown in Figure 28. According to Figure 28, it is seen that the
greater the p is, the higher the broadband transmission efficiency is. It is also observed that p
plays a very important role in the single frequency application. For example, at frequency of
10.5 GHz, p = 0.2 is the best choice with transmission efficiency is close to 100 %, while at
frequency of 12 GHz, p = 0.6 is the best. As a low p value is expected in the design for the
broadband radome, for broadband application at 1~18 GHz, it is seen that the transmission
efficiency shows nearly no difference in the frequency range of 13~18 GHz when p is 4 or
higher. Also when p is 4 or higher, the transmission efficiency at 1~13 GHz is higher than 70
%, which can be practically used. Thus, we choose p = 4 in the following calculation and
preparation of the graded radome material.
Figure 27. Power transmission efficiency of the graded panel material as functions of frequency and
layer number.
structural coefficient.
61
3.3.3. Effect of Wall Thickness (D) on the Transmission Efficiency of the Graded
Radome Material
The initial conditions are: n = 10, p = 4, n = 2 and 1 = 8.
The transmission efficiency of the gradient porous panel as functions of the frequency
and wall thickness (d) is shown in Figure 29. According to Figure 29, it is seen that when d is
less than 10 mm, the transmission efficiency shows almost no difference, and when d is less
than 7 mm, the transmission efficiency is higher than 70 %, which fulfills the requirements
for the broadband radome application and has a similar conclusion with ref. [36]. Specially, if
the radome is used at defined frequency, the thickness can be selected by the applied
frequencies and the mechanical properties. Considering the mechanical requirements for the
broadband radome, a certain thickness is needed. As a result, we choose d = 6 mm in the
preparation of Si3N4 gradient ceramic radome material, after considering both the broadband
transmission property and the mechanical property.
3.3.4. Effect of Dielectric Constant () on the Transmission Efficiency of the Graded
Radome Material
The initial conditions are: n = 10, p = 4, d = 6 mm, n = 2 and 1 = 8.
The transmission efficiency of the graded panel material as functions of the frequency
and dielectric constant of outer layer (1) is shown in Figure 30. It is obviously seen that the
transmission efficiency decreases with the increase of 1. Our previous experiment suggested
that when MgO and AlPO4 are used as the sintering additives, the dielectric constant for the
fully dense sample is ~ 7. In this situation, according to Figure 30, the transmission efficiency
is higher than 75 %, which fulfills the requirements for the broadband radome application.
wall thickness.
62
dielectric constant of outer layer.
On the other hand, the transmission efficiency of the graded panel material as functions
of the frequency and dielectric constant of inner layer (n) is shown in Figure 31. It is
obviously seen that the transmission efficiency decreases with the increase of n. When n is
less than 2.5, the transmission efficiency is higher than 75 %, which fulfills the requirements
for the broadband radome application.
From the above calculations, the as-designed conditions for the preparation of the Si3N4
broadband ceramic radome material are: n = 5, p = 4, d = 6 mm, n < 2.5 and 1 < 8.
Figure 31. Transmission efficiency of the porous graded panel as functions of frequency and n.
63
3.4. Preparation of Porous Gradient

Considering the optimal design results, according to our experimental progress in the past
years, the material used for the preparation of the graded porous Si3N4 ceramic has been
selected. Firstly, Si3N4 ceramics with nearly fully dense have been prepared by using MgO
and AlPO4 as the sintering additive and spark plasma sintering technique, the dielectric
constant of which is ~ 7.0. This material has actual the lowest porosity and the maximum
dielectric constant, and is used as the top layer for the preparation of the graded porous
Si3N4 ceramics. Secondly, several methods have been applied to prepare Si3N4 porous
ceramics with high porosity and low dielectric constant. The sample using 50 vol.% H3PO4 as
the pore-forming agent is tested to have the highest porosity (~ 63 %) and the lowest
dielectric constant (n = 2.2). Thus, in terms of Eq. (1), for 5 layers porous graded ceramics,
the dielectric constant of each layer when p = 4 is shown in Figure 32. As a result, the
dielectric constant for the other three layers can be calculated for the experimental design.
Figure 32. The dielectric constant of each layer for a 5 layers graded porous ceramic design when 1 = 7
and n = 2.2.
The raw materials used in the present study are as follows: commercial fine Si3N4 powder
whose particle shape is sphere, average grain size is ~ 0.5 m and the content of phase is
higher than 93 %; commercial zirconia (ZrO2) powder whose grain size is ~ 0.5 m and the
purity is up to 99.9 %; commercial magnesia (MgO) and alumina phosphate (AlPO4) powder
whose grain size is < 0.2 m and the purity was up to 99.9 % and phosphorus acid (H3PO4)
liquid whose concentration is 85 % and the purity is up to 99.9 %.
Graded porous Si3N4 ceramic with 5 layers is prepared by stacking orderly the 5
individual prepared Si3N4 ceramics, using a phosphate binder by thermally treated at 200 oC.
The preparation method for the dense and porous Si3N4 ceramics have been published and
64
presented in detail. The physical properties and the composition of each layer of porous
graded Si3N4 ceramics are listed in Table 3.
Table 3. Composition and physical properties of
each layer of the graded porous Si3N4 ceramics
tg
1.2
Porosity
(%)
<2
7.0
0.006
1.2
~ 40
3.8
0.006
1.2
~ 46
3.0
0.005
1.2
~ 55
2.5
0.006
1.2
~ 63
2.2
0.002
Composition and preparation method
d (mm)
Si3N4+4 wt.% MgO+16 wt.% AlPO4

(Spark plasma sintering, 1400 oC, 5 min)
Si3N4+20 wt.% ZrP2O7
(Pressureless sintering, 1000 oC, 2 h)
Si3N4+35 vol.% H3PO4
2
3
4
5
Figure 33. SEM image of the obtained Si3N4 ceramic with graded porous structure.
65
3.5. Phase Compositions and Microstructure

SEM image of the obtained Si3N4 ceramic with graded porous structure is illustrated in
Figure 33. From this image, it is observed obviously that the pores are increased gradually
from layer 1 to layer 5, indicating that an expected graded porous structure Si3N4 ceramic is
successfully prepared by the present technique. On the other hand, the pore size can be well
controlled and increased gradually from less than 0.5 m to several m, suggesting that the
obtained graded porous Si3N4 ceramic may have promising mechanical and thermal matching
property. Generally, a large amount of fine and uniform open micropores are observed from
layer 2 to layer 5. Moreover, the Si3N4 grains are fine and maintained almost the same particle
size of the raw Si3N4 powder. For each individual layer, it is isotropic structure, and the
anisotropic effect for each individual layer is considered to be very limited. As a result, the
obtained Si3N4 ceramic with porous graded structure is quite in agreement with the design
model as shown in Figure 25. Thus, it is easy to adjust and improve the experimental
conditions to prepare the graded structure with various layers, porosities (namely dielectric
constant), layer thickness, etc., according to the electromagnetic microwave calculation and
design.
Figure 34. The calculation value and tested value of power transmission efficiency of the graded porous
structure Si3N4 ceramic radome panel at the incident angle of zero.
Both calculation value and tested value of the power transmission efficiency of the
graded porous structure Si3N4 ceramic radome panel at the incident angle of zero is shown in
Figure 34. It is seen that the tested transmission efficiency is higher than 75 % at the
frequency range of 1~18 GHz, indicating that the graded porous structure has an ideal
broadband transmission property. It is feasible to use the graded porous structure for the
broadband radome application. Compared with the calculation results obtained by using the
same physical properties and structural parameters with the experiment, it is clearly seen that
the two are nearly in agreement with each other, suggesting that the electromagnetic
microwave transmission calculation and optimal design are essential and play a significant
66
role before the experiment being carried out. The reason for the differences between the
experiment and simulation is that the microwave equivalent network method used is not
accurate enough to be stand alone tool. Moreover, the effect of temperature and curvature
shape on the transmission efficiency needs further investigation for the graded porous
structure radome application.
ACKNOWLEDGMENTS
This work is supported by the National Natural Science Fundation of P. R. China under
Grant No. 50972111, Research Fund of the Doctoral Program of Higher Education of China
(No. 20110143110006) and self-determined and innovative research funds of SKLWUT.
REFERENCES
F. L. Riley, Silicon Nitride and Related Materials, J. Am. Ceram. Soc., 83 [2] 245-65
(2000).
[2] G. Ziegler, J. Heinrich, and G. Wotting, Relationships between Processing,
Microstructure and Properties of Dense and Reaction-Bonded Silicon Nitride, J.
Mater. Sci., 22 [9] 3041-86 (1987).
[3] A. J. Pyzik and D. R. Beaman, Microstructure and Properties of Self-Reinforced
Silicon Nitride, J. Am. Ceram. Soc., 76 [11] 2737-44 (1993).
[4] J. F. Yang, G. J. Zhang, N. Kondo, T. Ohji and S. Kanzaki, Synthesis of Porous Si 3N4
Ceramics with Rod-Shaped Pore Structure, J. Am. Ceram. Soc., 88 [4] 1030-2 (2005).
[5] A. Diaz, S. Hampshire, J. F. Yang, T. Ohji, and S. Kanzaki, Comparison of
Mechanical Properties of Silicon Nitrides with Controlled Porosities Produced by
Different Fabrication Routes, J. Am. Ceram. Soc., 88 [3] 698-706 (2005).
[6] Inagaki, N. Kondo and T. Ohji, High Performance Porous Silicon Nitrides, J. Eur.
Ceram. Soc., 22 2489-94 (2002).
[7] S. Q. Ding, Y. P. Zeng and D. L. Jiang. Oxidation Bonding of Porous Silicon Nitride
Ceramics with High Strength and Low Dielectric Constant. Mater. Lett., (2006), In
press.
[8] F. H. Simpson and J. Verzemnieks, Controlled Density Silicon Nitride Material,
Proceedings of the 16th symposium on electromagnetic windows, 81-6 (1982).
[9] A. D. Studart, U. T. Gonzenbach, E. Tervoort and L. J. Gauckler, Processing Routes to
Macroporous Ceramics: A Review, J. Am. Ceram. Soc., 89 [6] 1771-89 (2006).
[10] A. Diaz and S. Hampshire, Characterisation of porous silicon nitride materials
produced with starch, J. Eur. Ceram. Soc., 24 413-9 (2004).
[11] D. C. C. Lam, F. F. Lange, and A. G. Evans, Mechanical Properties of Partially Dense
Alumina Produced from Powder Compacts, J. Am. Ceram. Soc., 77 [8] 2113-7 (1994).
[12] J. F. Yang, G. J. Zhang, and T. Ohji, Porosity and Microstructure Control of Porous
Ceramics by Partial Hot Pressing, J. Mater. Res., 16 [7] 1916-8 (2001).
[1]
67
[13] C. Kawai and A. Yamakawa, Effect of Porosity and Microstructure on the Strength of
Si3N4: Designed Microstructure for High Strength, High Thermal Shock Resistance,
and Facile Machining, J. Am. Ceram. Soc., 80 [10] 2705-8 (1997).
[14] S. Y. Shan, J. F. Yang, J. Q. Gao, W. H. Zhang, Z. H. Jin, Porous Silicon Nitride
Ceramics Prepared by ReductionNitridation of Silica, J. Am. Ceram. Soc., 88 [9]
2594-6 (2005).
[15] T. Fukasawa, Z. Y. Deng, and M. Ando, Synthesis of Porous Silicon Nitride with
Unidirectionally Aligned Channels Using Freeze-Drying Process J. Am. Ceram. Soc.,
85 [9] 2151-5 (2002).
[16] D. Y. Chen, B. L. Zhang, H. R. Zhuang, and W. L. Li, Combustion Synthesis of
Network Silicon Nitride Porous Ceramics, Ceram. Inter. 29 [4] 363-4 (2003).
[17] X. P. Pu, X. J. Liu, F. G. Qiu, and L. P. Huang, Novel Method to Optimize the
Structure of Reticulated Porous Ceramics, J. Am. Ceram. Soc., 87 [7] 13924 (2004).
[18] F. Chen, Q. Shen, F. Q. Yan, L. M. Zhang. Preparation of Zirconium Pyrophosphate
Bonded Silicon Nitride Porous Ceramics. Materials Science and Technology, 22 [8]
915-8 (2006).
[19] D. E. Harrison, H. A. Mckinstry and F. A. Hummel, J. Am. Ceram. Soc. 37 (6) (1954)
277.
[20] D. M. Poojary, R. B. Borade, F. L. Campbell, et al, J. Sol. Sta. Chem. 112 (1994) 106.
[21] J. Barta, M. Manela, Proceedings of the 16th symposium on electromagnetic windows.
(1982) 87.
[22] L. U. J. Ogbuji and S. R. Bryan, The SiO2-Si3N4 Interface, Part I: Nature of the
Interphase, J. Am. Ceram. Soc. 78 [5] 1272-8 (1995).
[23] D. K. Agrawal, A. R. Maslowski, J. H. Adair, Evolution of the Formation of Inorganic
Polymers in the CaOSiO2P2O5 System Using Metal Alkoxides, J. Am. Ceram.
Soc., 73 [2] 430-4 (1990).
[24] J. Wan, R. G. Duan and A. K. Mukherjee, Scripta. Mater., 53 (2005) 663.
[25] D. E. Harrison, H. A. Mckinstry and F. A. Hummel, J. Am. Ceram. Soc., 37 (6) (1954)
277.
[26] L. K. L. Falk, J. Mater. Sci., 39 (2004) 6655-6673.
[27] M. Keppeler, H. G. Reichert, J. M. Broadler, et al. J. Euro. Ceram. Soc., 18 (1998) 521526.
[28] D. E. C. Corbridge, Phosphorus: An outline of its chemistry, biochemistry and
technology, Elsevier, New York, (1985) 141.
[29] D. K. Kim, W. M. Kriven, J. Am. Ceram. Soc., 86 (11) (2003) 1962-1964.
[30] U. C. Hasar, O. Simsek and M. Gulnahar, Simple procedure to simultaneously
evaluate the thickness of and resistive losses in transmission lines from uncalibrated
scattering parameter measurements, J. Electromagn. Waves Appl., Vol. 23, No. 8-9,
999-1010, 2009.
[31] A. Kedar and U. K. Revankar, Parametric study of flat sandwich multilayer radome,
Prog. Electromagn. Res., PIER, Vol. 66, 253-265, 2006.
[32] A. Kedar, K. S. Beenamole, and U. K. Revankar, Performance appraisal of active
phased array antenna in presence of a multilayer flat sandwich radome, Prog.
Electromagn. Res., PIER, Vol. 66, 157-171, 2006.
[33] J. A. Kong, Electromagnetic Wave Theory, Wiley-Interscience, published May 2008.
68
[34] P. Fuerholz and A. Murk, Design of a broadband transition using the constant
impedance structure approach, Prog. Electromagn. Res. Lett., PIERL, Vol. 7, 69-78,
2009.
[35] D. J. Kozakoff, Analysis of Radome Enclosed Antennas, Artech House, Norwood,
MA, 1997.
[36] S. Sunil, K. S. Venu, S. M. Vaitheeswaran and U. Raveendranath, A modified
expression for determining the wall thickness of monolithic half-wave radomes,
Microw. Opt. Techn. Lett., Vol. 30, No. 5, 350-352, 2001.
In: Silicon Nitride

ISBN 978-1-61942-865-2
Chapter 3
OPTICAL AND VIBRATION PROPERTIES

OF SILICON RICH NITRIDE
V. A. Volodin*
1
A.V. Rzhanov Institute of Semiconductor Physics,

Siberian Branch of Russian Academy of Sciences, Novosibirsk, Russia
2
Novosibirsk State University, Novosibirsk, Russia
ABSTRACT
In this chapter optical and vibration properties of hydrogenated silicon-rich nitride
(SRN) films of various stoichiometry (SiNx:H, 0.7<x1.3) were studied. The films were
deposited with the use of plasma enhanced chemical vapor deposition (PECVD)
technique. The properties of as-deposited and annealed films were studied using laser
scanning and spectral ellipsometry, Fourier transform infrared spectroscopy, Raman
spectroscopy, and photoluminescence (PL) techniques. According to the Raman data, the
as-deposited film, in which the stoichiometry parameter x is below 1, contains amorphous
silicon nanoclusters. The annealings lead to crystallization of these nanoclusters. No
silicon nanoclusters were observed in the as-deposited films with relatively low
concentration of excessive silicon atoms (the case of SiNx:H, x>1); annealings lead to
gathering of excessive Si atoms in amorphous Si clusters. States of hydrogen in asdeposited and annealed films were studied. Dramatic changes in the PL spectra of the
annealed SRN films were observed. Femto- and nanosecond laser treatments were
applied for crystallization of amorphous silicon nanoclusters in SRN films. The laser
fluences needed for crystallization of the silicon nanoclusters were found for films of
various non-stoichiometric parameters. The effect of laser assisted formation of
amorphous Si nanoclusters in SRN films with relatively low concentration of additional
silicon atoms was observed. The developed approach can be used for the creation of
dielectric films with semiconductor nanoclusters on non-refractory substrates.
A.V. Rzhanov Institute of Semiconductor Physics, Siberian Branch of Russian Academy of Sciences, pr.
Lavrenteva 13, 630090, Novosibirsk, Russia, E-mail address volodin@isp.nsc.ru.
70
V. A. Volodin
INTRODUCTION
Amorphous semiconductor nanoclusters [1-3] and semiconductor nanocrystals (NCs) [4]
embedded in dielectric films attract much interest due to their optical and electrical properties.
Because the sizes of nanoclusters can be smaller than de-Broil wavelength for electrons and
holes, the quantum size effect can modify electron spectra of the heterosystem. The
dependence of energy of optical transition on the sizes of silicon amorphous nanoclusters and
silicon NCs in silicon nitride films was observed earlier [1-3]. For practical use it is important
that the quantum size effect for Si NCs in dielectric matrix becomes brightly apparent at room
temperature. In some experiments a single NC reveals delta-function-like photoluminescence
spectra [5], so, the silicon NCs in dielectric matrix can be called as quantum dots. The NCs
have shown significant promises for a wide range of nanoelectronic and optoelectronic
applications. So, the NCs are interesting for both fundamental and applied points of view. It is
interesting that the quantum-size effect was observed not only for Si NCs but also for
amorphous silicon nanoclusters [1-3]. In metal-dielectric-semiconductor structures based on
silicon-rich nitride (SRN) films an effective electroluminescence was achieved, and red,
green and blue light emitting diodes were demonstrated [2]. SRN films are non-stoichiometric
SiNx films (x<4/3). There are reproducible, not expensive, low temperature technologies for
deposition of SRN films. The additional silicon atoms in some SRN films are randomly
distributed in a SiNx matrix. They do not form nanoclusters (random bonding - RB model).
Some SRN films contain amorphous Si nanoclusters, which depends on growth condition.
This situation is close to random mixture - RM model [6,7]. Because amorphous
semiconductors are meta-stable, semiconductor NCs are more preferable for device
applications. Modern technique of deposition allows producing of SRN films on substrates
with temperature of plasticity as low as 100 oC. The problems of formation of amorphous Si
nanoclusters and its crystallization are topics in current research. The laser induced
crystallization has several advantages comparing to long time high temperature furnace
annealing [8]. This chapter is devoted to studies of amorphous Si nanoclusters in SRN films
and its crystallization. The development of low thermal budget technology of crystallization
of a-Si nanoclusters is actual and is of interest up today, especially for femtosecond laser
crystallization and modification [9, 10].
EXPERIMENTAL
Hydrogenated SRN films (SiNx:H) of different stoichiometry were deposited on the Si
(100) and glass substrates with the use of low-frequency plasma enhanced chemical vapor
deposition (LF PECVD) method. The frequency of discharge was 55 kHz, the films were
deposited from the gas mixture of ammonia (NH3) and monosilane (SiH4). The stoichiometry
x parameter depends on the ratio of ammonia/momosilane flows. At our experimental
conditions, the stoichiometric Si3N4 films are deposited when the NH3/SiH4 flow ratio is
higher than 10. The growth technology is described elsewhere in more details [3]. All SRN
films were deposited at substrate temperature 100 0C or 380 0C, the pressure in reactor during
deposition was 150 Pa, the deposition time was 16 minutes, and the plasma power was 400
W. Film parameters are shown in Table 1.
Optical and Vibration Properties of Silicon Rich Nitride
71
Ellipsometry was applied to detect the thickness and optical constant of the SRN films.
Laser (He-Ne, =632.8 nm) scanning ellipsometer Microscan LEF-801 and spectral
ellipsometer were used. The range of spectral ellipsometer was within 250 to 900 nm, xenon
lamp was used as a light source, and the spectral resolution was 2 nm. Both ellipsometers are
produced by A.V. Rzhanov Institute of Semiconductor Physics, SB RAS.
To study the states of hydrogen and vibration of the Si-N bonds, the Fourier Transformed
Infrared Spectroscopy (FTIR) was applied. The FTIR spectrometer FT-801, produced by
scientific enterprise Simex (Novosibirsk, Russia), was used. The spectral range of the
spectrometer is from 550 to 5500 cm-1, the spectral resolution can be optionally changed from
4 to 0.5 cm-1.
Photoluminescence (PL) spectra were registered using double spectrometer SDL-1
(LOMO, Sankt-Petersburg, Russia). Photomultiplier tube FEU-79 was used as a detector; the
Ar+ laser line (=488 nm) and pulse nitrogen laser line (=337 nm) were used for excitation.
Some PL spectra were registered at temperatures ranging from 80 to 400 K to analyze the
temperature dependence of PL intensity for the samples. The excitation source was the
ultraviolet line of a mercury lamp (=312.6 nm). In this case, the PL measurements were
carried out using a Horiba Jobin Yvon monochromator, the detector being a Si CCD matrix.
In both cases, the spectral resolution was 2 nm. For measuring the temperature dependence of
the PL intensity in the samples, a standard LinKam thermo cell was used, in which the
accuracy of temperature was 0.1 K. Following the measurements, measured data was
normalized by the spectral sensitivity of measuring equipment. On several samples,
measurements of the PL decay time were attempted. The excitation source in those
measurements was a pulse YAG laser operating at the second harmonics, and the detector, an
InGaAs photodiode. The time resolution of the setup was 10 ns. The measured PL lifetimes
proved to be extremely short, less than 10-20 ns.
Table 1. Specifications of the as-deposited SRN films
Sample PECVD technique
Substrate
1
2
3
4
5
6
7
1s-LT
2s-LT
3s-LT
4s-LT
5s-LT
1s
2s
3s
4s
5s
Glass
Glass
Glass
Glass
Glass
Glass
Glass
Si
Si
Si
Si
Si
Si
Si
Si
Si
Si
LF (low-frequency)
LF
LF
LF
LF
LF
LF
LF
LF
LF
LF
LF
LF
LF
LF
LF
LF
T, oC NH3/SiH4
flow ratio
380 10
380 5
380 3
380 2
380 1.5
380 1
380 0.5
100 5
100 2.6
100 1.5
100 0.9
100 0.5
380 5
380 2.6
380 1.5
380 0.9
380 0.5
d, nm
322
280
250
254
259
281
404
240
400
400
450
450
280
430
430
510
510
Stoichio-metry
parameter x
1.33
1.3
1.25
1.1
1.0
0.80.9
0.6
1.3
1.01.1
0.91.0
0.70.9
0.6
1.3
1.01.1
0.91.0
0.70.9
0.6
72
V. A. Volodin
The structural properties of silicon nanoclusters in the as-deposited and annealed SRN
films were studied using Raman spectroscopy technique. All Raman spectra were recorded in
the back-scattering geometry. The 514.5 nm Ar+ laser line was used as a light source. Triple
spectrometer T64000 (Horiba Jobin Yvon) with micro-Raman setup and the liquid nitrogen
cooled CCD matrix detector were used. The spectral resolution was not worse than 1.5 cm-1.
All spectra were measured at room temperature. To avoid heating of the films, the laser beam
was slightly defocused; the spot diameter was equal to about 8 m; the laser power reaching
the sample was within the 2-3 mW range. The spectra were registered in the Z ( XY ) Z
polarization geometry: the incident light was polarized along the 011 crystallographic axis
of the Si (100) substrate. The polarization of scattered light was chosen along the 01 1
crystallographic axis of the Si (100) substrate. This geometry is forbidden for one-phonon
scattering from a mono-crystalline Si substrate with the (100) orientation. Nevertheless,
caused by some polarization leakage, the peak from the mono-crystalline Si substrate was
detected in some spectra, but in the allowed polarization geometry its intensity would be
about 30 times higher. For direct observation of nanocrystals, the High Resolution
Transmission Electron Microscopy (HRTEM) technique was applied. The JEM-4000EX
(JEOL) electron microscope was used. The equipment of centers of collective use
Nanostructures at A.V. Rzhanov Institute of Semiconductor Physics and Novosibirsk State
University was used.
200
-Si Brodsky Phys.Rev.B, v1, p2632

-Si3N4
180
Samle 1
2
3
4
5
6
7
160
120
100
10 , cm
-1
140
80
=0(E-Eg) /E
60
40
20
0
0,5
1,0
1,5
2,0
2,5
3,0
3,5
4,0
4,5
5,0
5,5
6,0
Energy, eV
Figure 1. The spectral dependence of absorption coefficient obtained from reflectance and transmission
spectra of as-deposited a-SiNx:H films. Specifications of the films are in Table 1.
73
Furnace annealing of the samples in Ar ambient for 5 hours at temperature 1130 0 was
performed in Institute of Electron Technology (Warsaw, Poland). Annealing was carried out
in Ar atmosphere, routinely, under atmospheric pressure (105 Pa), as well as in a specially
designed furnace under the conditions of hydrostatic compression at enhanced hydrostatic
pressure, HP, equal to 11 kbar (1.1 GPa).
A Ti-Sapphire laser (FemtoPower Compact Pro, Femtolasers Produktions GmbH) with a
central wavelength of 800 nm and pulse duration of <30 fs was used for laser treatments. The
distribution of energy density in the laser spot has Gaussian form:
2
E r E0 exp 2r 2
r0
(1)
Figure 2. Refraction index (A) and absorption index (B) as-deposited and annealed SRN films (samples
1s-LT, 2s-LT and 4s-LT, Table 1), =632.8 nm.
74
V. A. Volodin
where E0 is the maximum energy and r0 was taken as radius of the laser spot. Scanning
treatments using x-y sample translation by computer-controlled motors were carried out. The
pulse repetition rate was 1 kHz. The diameter of laser spot was about 20 micrometers. The
overlapping of laser spots was varied from 90 to 96%. The laser fluence was changed by
varying the laser pulse energy Epulse: E0
2 E pulse
r02
. For the femtosecond treatments we used
laser fluences from 65 to 250 mJ/cm2.

For nanosecond laser impacts the excimer KrF laser was used. The parameters of KrF
laser were next: wavelength =248 nm, pulse duration 20 ns. The pulse repetition rate was
also 1 kHz. The laser spot in the case of nanosecond excimer laser treatments had rectangular
shape with sizes 20-40 micrometers. For the nanosecond treatments we used laser fluences
from 130 to 170 mJ/cm2.
RESULTS AND DISCUSSION

Optical Properties of as-Deposited and Annealed SiNx:H Films in Visible
Range
It is well-known the optical gap of SiNx films depends on stoichiometry parameter x [6,
7]. The reflectance and transmittance of the light in spectral region from 0.5 eV to 6 eV were
measured for films grown on fused silica substrates. From these data, knowing the thickness
of the films, the absorbance coefficients were obtained. The results are presented in figure 1.
As one can see, for not intensive, probe beam, the absorbance even in the film 7 for photons
with energy 1.5 eV is low compared with absorbance of a-Si (data obtained by Brodsky,
[11]). It should be noted, that for hydrogenated a-Si:H and a-SiNx:H the optic gap should be
blue shifted with growth of hydrogen concentration. According to shift of the optical gap
versus the parameter x, as presented in details in article of Afanasyev-Charkin et al. [12], our
films have parameter x from 0.7 to 1.33.
The optical constant of SiNx films also depends on stoichiometry parameter x [6, 7]. The
refraction and absorption indexes for =632.8 nm of as-deposited and annealed SRN films are
shown, correspondingly, in Figures 2A and 2B. These data were obtained using scanning
laser ellipsometry measurements and determining the thickness and optical constants from the
one-film-on-the-substrate model. The spectral dependencies of the refraction and absorption
indexes of the as-deposited and annealed SRN films are shown, correspondingly, in Figures
3A and 3B. In this case, the multi-angle measurements were carried out for the different
angles of incidence (45, 50, 55, 60, 65 and 700), the sets of experimental ellipsometic
parameters ( and angles) were obtained. Then, a special computer approximation was
carried out. The ellipsometic parameters for the model of one-film on the mono-crystalline Si
substrate were approximated using the experimental ellipsometic data. The optical constants
for mono-crystalline Si were taken from the work of Aspens and Studna [13]. The optical
constants of SRN films were derived using the best fit of calculated and experimental data.
Using the many-angle approach it occurred possible to obtain more reliable optic constants.
However, one can see that, in accordance with the spectral ellipsometry data, the value of
refraction index at =632.8 nm (Figure 3) is slightly larger than that obtained from scanning
75
laser ellipsometry measurements (Figure 2). This little discrepancy can be explained as
follows. Accordingly to the scanning laser ellipsometry data, the thickness of the SRN films
is slightly non-uniform (the thickness shown in Table 1 is averaged). The light spot of
spectral ellipsometer equals to about 10 mm times 4 mm. Within this spot, the film thickness
indicates small gradient and a few nanometer changes (relative changes of about 1-2 %). But
our model, applied for the spectral ellipsometry data interpretation, concerns in fact only the
homogeneous film. So, accordingly to our estimation, a relative error in calculations of the
spectral dependence of optical constants can reach a value up to a few percents. The data
obtained from the scanning laser ellipsometry measurements are based on numerous massive
scans (several thousands of points) and so are more accurate. So, for further estimation of the
SRN film composition from the refraction index at =632.8 nm, we used just these data.
Anyway, basing both on the scanning laser and spectral ellipsometry data, one can see that
furnace annealing leads to an increase of the refraction indexes for all films (Figures 2 and 3).
The absorption indexes for the SRN films grown with the NH3/SiH4 flow ratio equal to 2.6
and 0.9 are also increased as the result of furnace annealing (Figure 3B). Obviously, it is due
to two factors. The first of them is dehydrogenization of the films (evaporation of hydrogen)
during annealing, and subsequent changes in the defect-related electronic density of state. The
second one is redistribution of excessive (additional) silicon atoms, and their accumulation
into clusters. The optical constants of the SiNx films are determined not only by the
stoichiometry x parameter, which depends on the concentration of the excessive (additional,
over-stoichiometric) silicon atoms, but also on the state of these excessive atoms. They can be
randomly distributed in the SiNx matrix, or formed clusters of different sizes in the Si3N4
matrix.
The stoichiometry parameter x was estimated using two approaches, basing on
dependence of the refraction index at =632.8 nm on this parameter and from a shift of the
optic gap versus the x parameter. Because the above mentioned literature data mainly concern
the films grown at relatively high temperatures, and, consequently, containing almost no
hydrogen, we have made our estimates basing on the data obtained for annealed films. As
seen in Figures 2 and 3, the optical constant of the SRN films annealed under atmospheric
and enhanced hydrostatic pressures, are almost identical. The values of stoichiometry x
parameter estimated from the refraction index and from a shift of the optic gap are well
correlated but, nevertheless, they do not coincide exactly. So the x parameters were
determined with some uncertainty. These data are shown in Table 1.
One can easily determine the molar part of silicon (named y) for the all compositions of
silicon/stoichiometric silicon nitride basing on the x value, using the following relation:
(3-2y)SiNx = ySi+(1-y)Si3N4,
so y (4 3x) /(4 2 x)
The parameter y was determined for all SRN films. For the SRN film grown with the
flow ratio NH3/SiH4 below 1, the molar fraction of silicon exceeds 50%. As it has been
already mentioned, in the amorphous sub-stoichiometric silicon nitride (SiNx with x<4/3), the
excessive silicon atoms can be randomly distributed in the SiNx matrix or formed the clusters
of different sizes in the Si3N4 matrix. In the first case, the SRN films consist of the tetrahedral
76
V. A. Volodin
Si-SiN3, Si-Si2N2, Si-Si3N, and Si-Si4 clusters. Probability to find the one specific clusters
depends on the stoichiometry x parameter. This case is described by the so-called random
bonding (RB) model [7]. In the second case the SRN films are composed of the separated
silicon and Si3N4 clusters with the sharp interfaces; the size of clusters depends on the
stoichiometry parameters. This case is described by the so-called random mixture (RM)
model [7]. However, the structure of real SRN films is not exactly corresponding to the RB or
RM models. Accordingly to the experimental data, the structure of real SRN films
corresponds rather to some superposition of the RB and RM models [7]. The structure
depends on the deposition method and conditions, for example, the higher is temperature of
deposition, the real structure is closer to the RM model [6, 7]. So, the real SRN films can
contain the SiNx composed areas and the separate silicon and Si3N4 clusters of different
form and size [7].
So one can see that, while annealing could not change the Si/Si3N4 molar ratio, it leads to
the remarkably changed optical constants of the SRN films with a relatively high
concentration of excessive silicon. Clearly visible peculiarities in the absorption coefficient at
the 400-500 nm range for the annealed SRN films (with the stoichiometry parameter x1.1)
can be related to the fundamental electron transition in Si nanoclusters formed under
annealing. In the case of annealed SiN0.70.9 film, rather sharp peak are observed at
approximately 450 nm; one can assume that it is due to absorption in Si nanocrystals. Specific
broadening of the absorption coefficient of the annealed SiN1.01.1 film is, presumably, due to
absorption in amorphous Si nanoclusters.
Measured PL spectra (ultraviolet line of a mercury lamp =312.6 nm was used for
excitation) were found to depend substantially on the film stoichiometry, showing a red shift
of PL maxima with increasing amount of excess Si in the samples (Figure 4). The latter
observation was indicative of the formation of larger Si nanoclusters in samples with higher
Si supersaturation. With increasing the measurement temperature from 78 320 K the PL
intensity measured at fixed wavelength decreased, while the general shape of the spectrum in
each individual sample remained roughly unchanged (see Figure 5).
Figure 6 shows the curves of I(T) at the PL intensity maximum in each spectrum. The
kink at T=200 K, especially clearly seen in the semi-logarithmic plots of the curves, points to
the existence of two processes contributing to the nonradiative decay of excitations.
Thus, the existence of three thermally activated processes, one radiative and two
nonradiative, can be assumed. With all the three processes proceeding in parallel at
e1 v1 exp( E1 / kT ) , e2 v 2 exp( E 2 / kT ) , and

e3 v3 exp( E3 / kT ) , the PL intensity is defined, apart from the total number N of
characteristic rates
bright
I N
Si
nanoclusters,
by
the
relative
rate
of
the
radiative
process,
e1 (T )
. In view of this, the curves of Imax(T) were fitted with
e1 (T ) e2 (T ) e3 (T )
dependences
77
1
E1 E3 ,
E1 E2
1 f 2 exp(
) f 3 exp(
)
kT
kT
(1)
Figure 3A. Spectral dependences of refraction index as-deposited and annealed SRN films. (samples 1sLT, 2s-LT and 4s-LT, Table 1).
78
V. A. Volodin
Figure 4. PL spectra of all samples measured at T=-190 oC.
Figure 5. PL spectra of sample 3s-LT at different temperatures.
where f2=v2/v1 and f3=v3/v1 are relative frequency factors of the two nonradiative processes.
The varied parameters were f2, f3, E2, and E3. For the energy E1, a value 1.6 meV was adopted
[14]. The values of E2 , E3, f2 , and f3 found in different samples are summarized in Table 2. The
energy E3 being close to the exciton binding energy in Si nanocrystals [15], it can be
hypothesized that the second process proceeds via hops of most mobile carriers, electrons, to
neighboring Si nanoclusters after their thermal pre-activation by energy E3 required for the
79
electrons to overcome the polarization electric field due to the hole left behind. Estimates of
typical electron and hole hopping times between neighboring nanoclusters made with the use
of known values of energy offsets (Ee2 eV for electrons and Eh1.5 eV for holes [16]) show
that, throughout the whole temperature range 80 to 300 K an electron jump into neighboring
Si nanocluster ends, with high probability, in exciton dissociation. The estimated probability
of uncorrelated simultaneous hops of free electron and freehole into neighboring nanocluster
being low, we tentatively identify the second process as tunnel migration of excitonic states
over the system of Si nanoclusters. The factors f2 and f3 in the samples with different x behave
similarly (if the factor f2 grows in value, so does the factor f3), pointing, in line with the
adopted interpretation of the experimental data, to similar influence of film structural
characteristics on the rates of both nonradiative processes. In addition, the frequency factors f2
and f3 have higher values in samples 1s-LT-5s-LT compared to samples 1s-5s. All these
observations comply with the general concept of tunneling phenomena according to which the
migration of tunneling excitations over the system of Si nanoclusters becomes more intense
as the nanoclusters in the system grow in size with simultaneous decrease of the separation
between them.
To summarize the PL data, the obtained data are indicative of the existence of two
processes involved in nonradiative relaxation of excitations optically generated in Si
nanoclusters embedded in SiNx:H. These processes were tentatively identified as nondissociative migration and dissociation of tunneling excitons (Eact=30 and 130 meV,
respectively).
Figure 6. Maximum PL intensity (logarithmic scale) versus temperature in the spectra of samples grown
at T=100 oC (A) and 380 oC (B). To illustrate the fitting quality, curve 3s* is shown, calculated by
formula (1) to fit the experimental data for sample 3s.
80
V. A. Volodin
Vibration Properties of as-Deposited and Annealed SiNx:H Films

The Raman spectra of as-deposited and annealed SRN films are shown in Figure 7.
Raman scattering technique is widely used to identify the structure (amorphous or crystalline)
of Si nanoclusters in different environments. If a film contains notable numbers of the Si-Si
bonds, the Raman-active vibrations of these bonds are manifested in the Raman spectrum. It
is well known, that the Raman spectrum of amorphous silicon contains two broad peaks at
approximately 480 cm-1 and 150 cm-1 appearing as a result of the effective density of the
transversal optical (TO) and transversal acoustical (TA) vibration modes, correspondingly
[17].
Table 2. Parameter values found in examined samples.
Sample
1s-LT
2s-LT
3s-LT
4s-LT
5s-LT
1s
2s
3s
4s
f2
6.8
4.1
6.0
5.53
4.0
3.8
13.2
8.9
17.3
f3
1400
205
771
445
81.6
725
1530
4760
4270
E2 , meV
31
19
27
25
21
29
37
36
40
E3 , meV
150
98
134
122
85
150
167
168
140
Due to that two reasons, the density of transversal-like phonons indicate more
pronounced maxima in Raman spectra compared to the longitudinal-like phonons. First the
number of the transversal-like phonon bands is doubled compared with the number of
longitudinal-like phonon bands. Second the dispersion of the transversal-like phonon bands
is narrower compared with the dispersion of the longitudinal-like phonon bands. Longitudinal
acoustical (LA) and longitudinal optical (LO) phonons manifest in the Raman scattering of
amorphous silicon as very broad background with the weak maximum at 300-400 cm-1. All
these peculiarities are marked in Figure 7B by arrows. The narrow peaks in the range below
120 cm-1 are caused by Raman scattering on the vibration and rotation modes of oxygen and
nitrogen in atmosphere (the spectra were registered under ambient atmosphere). All Raman
spectra contain the peak originating from the narrow bulk Si - 520.6 cm-1 lines. As mentioned
above, the chosen polarization geometry forbids Raman scattering by LO phonons of the
mono-crystalline Si (100) substrate, thus it allows us to reduce the possible Raman signal
around 520.6 cm-1 from the Si substrate. But, for the used stoichiometry, the SRN films are
semi-transparent, and, in spite of the forbidden polarization geometry, the peak from Si
substrate is detected, even though it is reduced for about 30 times compared with that in the
allowed polarization geometry.
One can clearly see the broad peaks at 150 and 480 cm-1in the Raman spectrum of the asdeposited SiN0.70.9 film (Figure 7B). Therefore, this film contains notable number of
amorphous silicon nanoclusters; the concentration and sizes of these nanoclusters are large
enough to be detected by Raman scattering technique. Both high-pressure and atmospheric
pressure anneals lead to dramatic transformation of the Raman spectrum of this film. After
annealing one can hardly see the amorphous peaks but can detect clearly an appearance of
81
the nanocrystalline peak - the narrow line at 518.5 cm-1 indicating the presence of some Si
nanocrystals (figure 7B, dotted and dashed lines). Due to scattering on the optical phonon
modes localized in nanocrystals, the Raman spectrum of nanocrystals is characterized by the
not-symmetric peak between 500-520 cm-1. The position and width of this peak strongly
depend on the size and structure of nanocrystals, in accordance with dispersion of the
localized modes [18-21]. The peak width is mainly determined by size dispersion of the
nanocrystals. The intensity of the nanocrystalline peak depends on content of the
nanocrystalline phase. So, one can determine the volume part of this phase. In our case it may
be assumed that almost all Si clusters are crystallized during annealing. It should be noted,
that the nanocrystalline peak at 518.5 cm-1 cannot be resolved in respect of the monocrystalline peak at 520.6 cm-1 originating from the Si substrate. After differential analysis of
the Raman spectra (by subtracting the signal from substrate), we determined the position of
the nanocrystalline peak at about 517518 cm-1. In accordance with dependence of the peak
position on the nanocrystal average size [18-21], the average size of Si nanocrystals should be
5-7 nm in our case. One can detect the weak peak at 300 cm-1 in some Raman spectra. This
peak appears is due to two-phonon scattering, namely to scattering by the 2 TA phonons near
the edge of Brillouin zone. This peak can be observed as originating from both the Si (100)
substrate and from Si nanocrystals [9]. Supposedly, confinement of the acoustical phonons in
Si nanocrystals leads to enhancement of two-phonon Raman scattering, similarly as reported
for optical phonons [20]. It should be noted, that similar transformation of the Raman spectra
after annealing was observed in the SRN films of similar stoichiometry grown using the Si
target evaporation in nitrogen and hydrogen atmosphere [22].
As seen in Figure 7A, the as-deposited SiN1.01.1 film is not consisting of silicon
nanoclusters, at least in a quantity sufficiently high to be detected by Raman spectroscopy
method. The Raman spectrum contains only the peaks originating from the mono-crystalline
Si substrate and from an air. In the Raman spectra of annealed films one can observe the
amorphous peaks at 150 and 480 cm-1 and the PL background (Figure 7A, dotted and
dashed lines). Therefore, furnace annealing leads to accumulation of excessive silicon in the
form of amorphous clusters. Presumably, concentration of excessive silicon atoms in this film
is not high enough to form nanocrystalline clusters. In the Raman spectra of the annealed
SiN1.3 films (1s-LT sample, table 1) huge PL background exists; its presence does not allow
for proper analyze of the Raman signal. The as-deposited SiN1.3 film (the Raman spectrum is
shown in figure 8) also does not contain silicon nanoclusters, at least in a quantity sufficiently
high to be detected by Raman scattering. But in the Raman spectra of annealed SiN1.3 films
there are some peaks at 150 and 480 cm-1 (against the background of large PL signal), so one
can assume that in this case furnace annealing also leads to accumulation of excessive silicon
in the form of amorphous clusters.
As seen in Figure 7, the Raman spectra of the films with the same stoichiometry but
annealed under atmospheric pressure and under high pressure annealing (HP) are almost
identical. Thereafter, the structure (phase composition and sizes) of silicon naoclusters in such
films is practically identical. It has been supposed, that HP pressure would influence, under
annealing, the formation processes of silicon nanoclusters in accordance with the Le Chatelier
- Brawn principles. It is known, that the density of crystalline silicon is somewhat higher than
the density of amorphous silicon. This leads to an appearance of tensile mechanical stresses in
the polycrystalline films subjected annealing [23]. Therefore, hydrostatic pressure should
stimulate both crystallization of amorphous silicon clusters existing in the as-deposited SRN
82
V. A. Volodin
films and accumulation of silicon in the cluster form. Obviously, in the case of SiN0.70.9 film,
high thermal budget of furnace annealing leads to crystallization of almost all amorphous
silicon, being not dependent on the pressure conditions (HP or 105 Pa). One can assume that,
in the case of SiN1.01.1 film, majority of the excessive silicon atoms is accumulated in the
form of amorphous nanoclusters; this accumulation is, in general, not sensitive to annealing
conditions.
Figure 7. Raman spectra of as-deposited and annealed SRN films: A - SiN1.01.1 (2s-LT), B
SiN0.70.9.(4s-LT).
Intensity, arb. un.
83
Si
substrate
as deposited
atmosperic pressure
11 kBar
Si-Si
local
vibrations
3
2
1
320 340 360 380 400 420 440 460 480 500 520 540
-1
Raman shift, cm
Figure 8. Raman spectra of as-deposited and annealed 1s-LT films (PL background is reduced).
The electron microscopy data confirm the data of Raman scattering (see Figure 9). In
upper case of figure 9 one can see dark field image of Si-nanocrystals, where the light spots
are corresponded to Si-nanocrystals. The nanocrystals have round shape and the sizes below
10 nm. In the electron microscopy image with high resolution (at the bottom) the
nanocrystalline areas are seen. Interplanar distances in these areas correspond to the
crystalline silicon phase. Thus, from the direct electron microscopy observation, it follows
that annealed the SiN0.70.9 film contains silicon nanocrystals with the sizes of 6-8 nm.
The FTIR spectra of the as-deposited and annealed SRN films are shown in Figure 10.
The spectra were registered with 4 cm-1 resolutions. The main peaks are related to absorption
caused by stretching vibrations of the Si-N, Si-H, and N-H bonds. The as-deposited films
contain hydrogen in the both Si-H and N-H forms. Furnace annealing leads to complete
dehydrogenization of the SRN films with the stoichiometry x1.1 parameter; the absorption
peaks related to stretching vibrations of the Si-H and N-H bonds disappear. Surprisingly, the
SRN film with the x parameter close to that of stoichiometric silicon nitride (SiN1.3) still
contains hydrogen after annealing with high thermal budget, mainly in the form of the Si-H
bonds. In this film, the peak, corresponding to absorption on stretching vibrations of the Si-H
bonds (~2150 cm-1) remains nearly the same after annealing under HP. This peak even has
grown a little after annealing under 105 Pa. In the FTIR spectra of the SRN film (SiN1.3) the
weak peak corresponding to absorption on stretching vibrations of the N-H bonds still
remains detectable after annealing. It has been detected earlier [22] that the closer is the
stoichiometry parameter of the SRN films to that of stoichiometric Si3N4, the higher
temperature is needed for effusion of hydrogen (de-hydrogenization) from these films.
However, short-time annealing at temperature 850 0C [22], 950 0C [24], and annealing at
800 0C for 2 hours [25] lead to total de-hydrogenization of the SRN films, with the
stoichiometry x parameter below 1. One can assume, that for the case of nearly stoichiometric
84
V. A. Volodin
silicon nitride film, furnace annealing leads to so called hardening of silicon nitride inhibiting
hydrogen out-diffusion.
Figure 9. Si-nanocrystals into SiN0.70.9 (4s-LT) - film annealed under high hydrostatic
pressure: a, b -dark field TEM-image (a) and corresponded electron microdifraction (b);
c-HRTEM image.
The transformations evidenced by the absorption-related peaks on stretching vibrations of
the Si-N bonds occur due to the Si and Si3N4 phases separation during annealing. The
separation leads to an appearance of high-frequency shoulder of the Si-N peak, and to blueshift of the peak position. Similar transformations were observed earlier for the SRN films
with different stoichiometry [12, 22, 25]. Herewith, some peculiarities have appeared in the
FTIR spectra of the SRN films with the stoichiometry x1 parameter, which can be related to
absorbance by stretching vibrations of the Si-N bonds in crystalline silicon nitride.
This spectral range was studied in more details using 2 cm-1optional spectral resolution;
the results are shown in Figure 11. Some peculiarities in the Si-N peak are marked by arrows
and numbers. Specification of typical FTIR absorption peaks is shown in Table 3.
Peculiarities marked 1, 2 and 3 can be related to absorbance by stretching vibrations of the SiN bonds in the -modification of crystalline silicon nitride [26, 27]. The peak 4 is possibly
related to absorbance by stretching vibrations of the Si-O bonds, because the surface area of
the films can be oxidized partially. The peak 5 can be related to absorbance by bending
vibrations of the N-H bonds [28]. It has been confirmed by disappearance of this peak after
annealing.

Table 3. Specification of FTIR absorption peaks.
1
2
3
4
5
Frequency, cm-1
684
858
895
1100
1175
Type of vibration, reference

(F), [26]
1(A), [26]
3(F2), [27]
Stretching SiOSi, [28]
Bending N-H, [28]
Figure 10. FTIR spectra of as-deposited and annealed SRN films.
85
86
V. A. Volodin
Figure 11. FTIR spectra of as-deposited and annealed SiN1.01.1 films: frequency range for stretching
vibration of Si-N bonds.
The optical microscopy image of the SiN1.3 film annealed under HP is shown in Figure
12. A net of cracks is observed; these cracks appeared due to mechanical stresses in the film.
The mechanical stresses resulted in de-hydrogenization of the films and in their consequent
hardening. Another possible reason is crystallization of silicon nitride and concomitant
increase of the film density leading to tensile mechanical stresses. Note that, in accordance
with the ellipsometric data, anneals lead to some shrinkage of the films, the film
thicknesses being reduced by several percents. In the SiN1.3 film annealed at atmospheric
pressure the net of cracks has appeared, as it was the case also in SiN1.01.1.
As seen in Figure 13, anneals lead to dramatic changes in the PL spectra of some SRN
films. As it has been mentioned above, the PL signal for annealed SRN films with the x
parameter close to that of stoichiometric silicon nitride (SiN1.3) becomes of higher intensity
high. However, its background does not allow studying the Raman signal. For the SRN films
with relatively high concentration of excessive silicon (SiN0.70.9), the dependence of PL
signal on annealing is reversible. The as-deposited film contains amorphous silicon
nanoclusters and reveals the strong PL signal. Assuming that this signal with the maximum at
1.65 eV is caused by radiative recombination in amorphous silicon nanoclusters, then,
according to the dependence shown earlier [1, 2], the average size of nanoclusters should be
4-5 nm. As it has been mentioned above, in accordance with the Raman and HRTEM data,
the annealed (both under HP and 105 Pa) SiN0.70.9 films contain silicon nanocrystals of 6-7
nm average sizes. So annealing leads to growth and crystallization of silicon clusters. The PL
signal in such annealed films was reduced for about 10 times, and its maximum indicates
red-shift down to 1.5 eV. Some authors [4] have observed the quantum-size effect in PL
of silicon nanocrystals embedded in the silicon nitride matrix. From the dependence, reported
in [4], one can conclude that the average size of silicon nanocrystals is 6-7 nm in our case,
what coincides perfectly with the estimation based on an analysis of the Raman and HRTEM
87
data. The reduction of the PL signal after annealing can be due to several reasons. It is known
that, for indirect band gap semiconductors, the rate of radiative recombination is reduced with
increasing the size of nanocrystals [29, 30]. Besides that, and according to the FTIR data,
annealing leads to de-hydrogenization of these films. It is known, that hydrogen can passivate
the dangling bonds, thereby reducing the rate of irradiative recombination. So, dehydrogenization can lead to the increased rate of irradiative recombination and, consequently,
to the reduced PL signal.
Figure 12. Optical microscopy image of SiN1.3 film annealed under high hydrostatic pressure.
It has been above mentioned, that the as-deposited SiN1.01.1 film does not contain silicon
nanoclusters in an amount sufficient to be detected by Raman scattering technique. The PL
signal in the as-deposited film in the green-yellow spectral range is presumably caused by the
presence of defects, including defects originating due to excessive silicon. The anneals (both
under HP and under atmospheric pressure) lead to an appearance of the PL signal in the
yellow-red spectral range. Accordingly to the Raman data, the annealed films contain
88
V. A. Volodin
amorphous silicon nanoclusters. In accordance with the reported quantum-size dependence

for amorphous silicon nanoclusters [1-3], the average size of amorphous silicon nanoclusters
in this case should be about 5 nm. As one can see, short-wave shoulders exist in the PL
spectra of the annealed SiN1.01.1 films, so the amorphous silicon nanoclusters with lower
sizes seem to be also present in such films. But one can hardly see any influence of pressure
on the PL spectra of the SRN films with relatively high concentration of excessive silicon.
Figure 13. PL spectra of as-deposited and annealed SRN films.
89
The obvious effect of pressure on optical properties is seen only in the PL spectra of the
annealed SRN films with the x parameter close to that in stoichiometric silicon nitride
(SiN1.3). In the case of annealing under HP, beside the PL signal of high intensity in the
green-yellow spectral range, one can also observe very strong PL in the yellow-red spectral
range. As it has been mentioned above, PL in the green-yellow spectral range is presumably
caused by the presence of defects, including defects originating due to excessive silicon. PL
in the yellow-red spectral range seems to be related to radiative recombination in amorphous
silicon nanoclusters. Accordingly to the FTIR data, the annealed films contain hydrogen in
the Si-H bonds. The reason of such behavior of hydrogen can be influence of
dehydrogenization on residual mechanical stress in silicon nitride films. For example,
Gorokhov and co-workers have proposed a mechanism for explaining the development of
residual stress in silicon nitride films [31]. In electronegativity scale nitrogen, silicon and
hydrogen have values 3.07, 2.25 and 2.10 accordingly. When Si-bonded hydrogen is
evaporated and Si-N bond appears instead of Si-H bond, the tensile stresses grow. When Nbonded hydrogen is evaporated and Si-N bond appears instead of N-H bond, the tensile
stresses relax. So, in the first case, elastic energy grows, and free energy grows accordingly.
From thermodynamics principles, it is more preferable for hydrogen to be in Si-H bond
instead of N-H bond.
Hydrogen can passivate defects and dangling bonds, reducing irradiative recombination
and, consequently, enhancing intensity of the PL signal. Because of the high PL background
it was not possible to observe and analyze the Raman signal from the Si-Si bonds in small
amorphous silicon nanoclusters. But, accordingly to the quantum-size dependence reported
for amorphous silicon nanoclusters [1-3], in this case the films should contain amorphous
silicon nanoclusters with 2-4 nm sizes. Therefore one can assume that, accordingly to the Le
Chatelier - Brawn principles, high hydrostatic pressure applied at annealing results in
enhanced accumulation of excessive silicon in the SRN films with the formation of relatively
small amorphous nanoclusters.
Figure 14. Temperature dependence of PL intensities for as deposited and annealed 1s-LT sample.
90
V. A. Volodin
Figure 15. The examples of optical microscope images for nanosecond (figure A - left) and
femtosecond laser (figure B right) scanning treatments. The size of side of rectangle (left) is 40
micrometers, the step between lines is 50 micrometers. The spot in the center is Ar laser spot (the image
was photographed in T64000 spectrometer microscope).
It should be noted that PL signal from the film annealed under HP was very stable with
temperature. Temperature dependence of PL was studied in range from 83 to 373 K. The
intensity of the PL signal at 373 K was only 20% lower than the intensity of the PL signal at
83 K in this case. In the case of film annealed at atmospheric pressure, the changes of PL
intensity for such temperature range was about two times (see figure 14). As for as-deposited
film, the changes of PL intensity for 83 K - 293 K diapason was about eight times. Therefore,
the observed temperature stability of PL signal in the case of HP annealing can be useful for
optoelectronic applications.
So, furnace annealing of the SRN films has modified their structural and optical
properties. Anneals change the refractive and absorbance indexes in the spectral diapason and
change the PL spectra. When the as-deposited films contain amorphous silicon nanoclusters,
the annealing (according to the Raman data analysis) lead to crystallization of these clusters.
The formation of amorphous silicon nanoclusters at annealing was observed in the SRN films
with relatively low concentration of excessive silicon atoms (SiNx:H, x>1), accordingly to the
Raman data. Any effect of pressure on the structure and optical properties of the SRN films
with relatively high concentration of excessive silicon (x<1.1) was not observed. However,
the dramatic difference in the PL spectra of the SRN films with the x parameter close to that
of stoichiometric silicon nitride (SiN1.3), annealed under atmospheric pressure as well as
under high pressure was observed. One can assume that high hydrostatic pressure applied at
annealing leads to accumulation of excessive silicon into the form of amorphous nanoclusters
with the 2-4 nm sizes. The developed approach can be applied for a creation of dielectric
films containing semiconductor nanocrystals or amorphous semiconductor nanoclusters.
Pulse Laser Annealings of SiNx:H Films

The question of homogeneity of laser treatments is very important for practical uses. The
optical microscope images for nanosecond and femtosecond laser scanning treatments are
shown in figure 15. In the case of excimer nanosecond laser treatments spot overlapping was
91
practically absent. One can see the non- homogeneity of laser fluence the rectangles of laser
spots is not homogeneous (figure 15A). Because of non- homogeneity the laser fluence is
reduced to upper right corner. So, a special optics should be used in this case for
homogeneous treatments.
In the case of femtosecond laser treatments the overlapping of laser spots during scanning
was about 96%. So, one can see quite homogeneous lines (figure 15B). But to treat whole
area the step between lines should be smaller. Anyway, the treatments with quite big
overlapping are more homogeneous even in the case of not- homogeneous distribution of
energy density in the laser spot (Gaussian shape).
NANOSECOND LASER TREATMENTS
Intensity, counts/sec
Some results of structural changes of SRN films after nanosecond laser impacts are
shown in figures 16-19. In the case of SRN films with relatively big concentration of excess
silicon (SiNx, when x lower than 1.2-1.25) initial films contain amorphous silicon
nanoclusters. As it was mentioned above, the Raman spectrum of a-Si contains two broad
peaks at approximately 480 cm-1 and 150 cm-1 appearing as result of effective density of
transversal optical (TO) and transversal acoustical (TA) vibration modes correspondingly.
Due to scattering on optical phonon modes localized in the nanocrystals, the Raman spectrum
of nanocrystals is characterized by narrow peak at position between 500-520 cm-1. The
position and the width of the peak strongly depend on size and structure of the nanocrystals
according to dispersion of the localized modes [18]. The peak width is mainly determined by
size dispersion of the nanocrystals. The intensity of the nanocrystal peak depends on the
amount of nanocrystal phase. So, one can determine the volume part of nanocrystal phase:
Sample 6 initial
2
130 mJ/cm
2
150 mJ/cm
2
170 mJ/cm
2
170 mJ/cm Anti-Stokes
-600
-400
-200
200
400
600
Raman shift, cm-1

Figure 16. Raman spectra (Stokes and Anti-Stokes) of as-deposited SRN film (sample 6, Table 1) on
glass substrates and treated by nanosecond pulse XeCl laser with various fluences.
92
V. A. Volodin
Sample 3 initial
170 mJ/cm2
100
200
300
400
500
600
Raman shift, cm-1

Figure 17. Raman spectra of as-deposited SRN film (sample 3, Table 1) on glass substrates and treated
by nanosecond pulse XeCl laser with various fluences.
As one can see in figure 17, when the SRN film contain relatively low concentration of
additional silicon (x is about 1.25) the laser treatments lead to formation of new amorphous
silicon nanoclusters. There is no nanocrystalline peaks in treated samples, but one can
clearly see the growth of amorphous peak.
Sample 4s-LT initial

2
130 mJ/cm
2
150 mJ/cm
2
170 mJ/cm
100
200
300
400
500
600
Raman shift, cm-1

Figure 18. Raman spectra of as-deposited SRN film (sample 4s-LT, Table 1) and treated by nanosecond
pulse XeCl laser with various fluences.
93
So, as one can see from figure 16, the beginning of crystallization for XeCl laser
nanosecond treatments took place for energy density 130 mJ/cm2 and higher only for SRN
films that contain high concentration of excess silicon. The nanosecond laser treatments lead
to growth of amorphous nanoclusters in the case of SRN films with stoichiometric parameter
x between 1.2 and 1.25. The results of structural modification of the SRN films deposited on
Si substrate is similar to that results for SRN films deposited on glass substrate (see figures
18, 19). Because laser radiation (wavelength 248 nm, photon energy is about 5 eV) is mainly
absorbed in films. Only some difference in absorbance and in film and substrate thermoconductivity can take place in this case. For example, SRN films grown at low temperatures
(figure 18) have more nanocrystalline part after nanosecond laser treatments than SRN grown
at higher temperatures (figure 19). Because the films are semi-transparent, one can see in
some spectra the narrow peak at 520 cm-1 from Si substrate.
FEMTOSECOND PULSE LASER TREATMENTS
In the case of femtosecond laser pulse crystallization of amorphous Si films on glass

substrates the threshold for full crystallization of the films with thickness about 100 nm was
found to be of about 65 mJ/cm2 [32]. Because absorbtion in the SRN films is lower, the laser
treatments with fluences up to 250 mJ/cm2 were applied. Some Raman scattering data for
initial and treated films are shown in figures 20-23.
Sample 4 initial
130 mJ/cm2
150 mJ/cm2
170 mJ/cm2
100
200
300
400
500
600
Raman shift, cm-1

Figure 19. Raman spectra of as-deposited SRN film (sample 4, Table 1) and treated by nanosecond
pulse XeCl laser with various fluences.
94
V. A. Volodin
Sample 6 initial
2
110 mJ/cm
2
120 mJ/cm
2
130 mJ/cm
2
150 mJ/cm
100
200
300
400
500
600
Raman shift, cm-1
Figure 20. Raman spectra of as-deposited SRN film (sample 6, Table 1) on glass substrates and treated
by femtosecond laser pulse treatments with various fluences.
Sample 3 initial
2
200 mJ/cm
2
250 mJ/cm Anti-Stokes
2
250 mJ/cm
-600
-400
-200
200
400
600
Raman shift, cm-1

Figure 21. Raman spectra (Stokes and Anti-Stokes) of as-deposited SRN film (sample 3, Table 1) on
glass substrates and treated by femtosecond laser pulse treatments with various fluences.
General trend in laser treatments of this SRN films is the more additional Si atoms
contains the film the less laser fluences were needed to crystallize or to modify silicon
nanoclusters. It is obvious, because the absorbance coefficient grow with growing of silicon
95
amount in the SRN films. Even the femtosecond laser treatments with fluences up to 250
mJ/cm2 was not enough to makes visible structural changes in the sample number 1 (Table 1).
In sample number 2 the laser treatments stimulate only creation of small amorphous
nanoclusters. In sample number 3 the initial films contains amorphous Si nanoclusters, and as
one can see in figure 21, the laser treatments not only stimulate the growth of amorphous
silicon nanocluster (amorphous peak is grow), but also crystallized some nanoclusters
(nanocrystalline peak is appeared). For SRN films with stoichiometric parameter x lesser
that 1.0 it was possible to crystallize silicon nanoclusters using laser treatments with fluences
120 mJ/cm2 and higher (figure 20). In the case of the SRN films on silicon substrates in
Raman spectra of some semi-transparent films one can see narrow peak at 520 cm-1 from Si
substrate (figures 22, 23). It is also remarkable, that femtosecond laser treatments lead not
only crystallization of existed amorphous silicon nanoclusters, but to growth or formation of
new amorphous nanoclusters in SRN films (one can see growth of amorphous peak in figures
22, 23). The data about threshold for crystallization of SRN films using femtosecond laser
annealing are presented in figure 24.
Lets discuss some effects of laser heating during Raman spectra registration. One can
see that the femtosecond laser impact leads to dramatic changes in structure of the sample 3
(figure 21). One can see both growth of amorphous peaks (150 and 480 cm-1) and appearing
of nanocrystalline peak with maximum at 513 cm-1. As it was mentioned, from position of
this peak one can estimate the average size of Si NCs. But for our conditions of
measurements it is important to take into account possible heating of the film and shift of
Raman peak due to anharmonic effects. It is well known that Raman scattering intensities in
I S and anti-Stokes I A spectral regions depend on temperature:

Sample 2S-LT initial
2
150 mJ/cm
Stokes
100
200
300
400
500
600
Raman shift, cm-1

Figure 22. Raman spectra of as-deposited SRN film (sample 2s-LT, Table 1) and treated by
femtosecond laser pulse treatments with various fluences.
96
V. A. Volodin
Sample 2S initial
2
120 mJ/cm
2
150 mJ/cm
2
200 mJ/cm
100
200
300
400
500
600
Raman shift, cm-1

Figure 23. Raman spectra of as-deposited SRN film (sample 2s, Table 1) and treated by femtosecond
laser pulse treatments with various fluences.
Figure 24. Crystallization threshold for amorphous nanoclusters in SiNx films for femtosecond near-IR
laser treatments.
97
Figure 25. HRTEM image of SiN0.6 (5s-LT) - film after femtosecond laser treatments.
(0 ) 4 kT
I A / IS
e
(0 ) 4
where
is Plank constant,
is frequency of initial photon,
(2)
is frequency of phonon
is Boltzmann constant, T is absolute temperature in Kelvin scale. So, from formula 2,

one can determine temperature as:
T ( K ) 1.36 (cm 1 ) / ln(1.23 I S / I A )
(3)
The Raman intensity in the same measurement condition was measured for sample 3 and
according to formula 3 the temperature was determined as 425 K. In bulk Si heating at 425 K
(according to our experimental data) leads to 2.5 cm-1 red shift compared with Raman shift
for bulk Si at room temperature (300 K). One can assume, that anharmonic effects for Si NCs
have the same effects, so taking into account the shift due to heating, the shift due to phonon
confinement effects is about 3.5 cm-1. Using model of folding of vibration modes in NCs, one
can determine the average size of Si NCs in film 3 (laser treated) as 4 nm. The HREM data
have confirmed Raman data about Si nanocrystals sizes, some HREM data are presented in
figure 25.
98
V. A. Volodin
To investigate, what part of energy in pulse was absorbed by the films, the next
experiments were carried out. The reflectance and transmittance of the light in spectral region
were measured. As it was mentioned above, from this data the absorbance coefficients were
determined. The results are presented in figure 1. As one can see, for not intensive, probe
beam, the absorbance even in the film 7 for photons with energy 1.5 eV (femtosecond laser)
is low. So, non-linear effects two or three photon interaction should play role in absorbance
of very intensive femtosecond pulse.
In the case of ultra-power femtosecond pulse impact the non-linear effects play an
important role in light absorption. The optical gap of silicon nitride is about 5 eV. Average
energy of photons in Ti-sapphire laser is 1.5 eV. The photon energy is not enough for
effective absorbance even in a-Si:H films with optical gap 1.5-1.9 eV [32]. Supposedly,
multi-photon processes can take place. Mechanism of phase transition can also be not purely
thermal: absorbance heating melting cooling crystallization. The pulse duration in our
case is much lower than time of electron-phonon interaction in semiconductors (about 1-2
picoseconds). So, during pulse, hot electron-hole plasma does not excite vibration modes in
silicon and does not relax. According to some theoretical calculations, when concentration of
hot electrons approaches to 81021 cm-3, silicon becomes unstable [33]. This metastable
stage can relax to more stable crystalline phase without melting, but with generation of latent
heat of crystallization. So, this process is similar to explosive crystallization. In the case of
laser assistant gathering of additional silicon into cluster, one can assume some laser assistant
diffusion mechanisms. But, because the pulse duration is less than period of atomic vibration,
post-laser pulse diffusion due to heating of the films is more preferable. For example, the time
of cooling of a-Si films down to melting temperature after nanosecond laser treatments is
about one hundred nanosecond [34]. This can be enough for diffusion of additional silicon
and formation of amorphous nanoclusters, as one can see in experiment.
SUMMARY OF THE CHAPTER

Furnace annealing of the SRN films has modified their structural and optical properties.
Anneals change the refractive and absorbance indexes in the spectral diapason and change the
PL spectra. When the as-deposited films contain amorphous silicon nanoclusters, the
annealing (according to the Raman data analysis and HRTEM data) lead to crystallization of
these clusters. The formation of amorphous silicon nanoclusters at annealing was observed in
the SRN films with relatively low concentration of excessive silicon atoms (SiNx:H, x>1),
accordingly to the Raman data. Any effect of pressure on the structure and optical properties
of the SRN films with relatively high concentration of excessive silicon (x<1.1) was not
observed. However, the dramatic difference in the PL spectra of the SRN films with the x
parameter close to that of stoichiometric silicon nitride (SiN1.3), annealed under atmospheric
pressure as well as under high pressure was observed. One can assume that high hydrostatic
pressure applied at annealing leads to accumulation of excessive silicon into the form of
amorphous nanoclusters with the 2-4 nm sizes. The developed approach can be applied for a
creation of dielectric films containing semiconductor nanocrystals or amorphous
semiconductor nanoclusters.
99
The possibility of the nanosecond and femtosecond laser pulses to crystallize amorphous
Si nanoclusters in SRN films on glass and silicon substrates was shown. Non-linear effects in
light absorbance may play some role in femtosecond treatments. It should be noted that
femtosecond laser treatments were applied for the first time for crystallization of amorphous
silicon nanoclusters in SRN films. The laser fluences for crystallization were found for the
films of various non-stoichiometries. The effect of laser assistant formation of amorphous Si
nanoclusters in SRN films with relatively low concentration of additional silicon atoms was
observed. The developed approach can be used for the creation of dielectric films with
semiconductor nanoclusters on non-refractory substrates.
ACKNOWLEDGEMENTS
The author is grateful to Dr. Aleksandr Popov for some experimental samples. DAAD
has supported visit of V.A.V. to Laser Zentrum Hannover for laser treatments. Author is
grateful to Professor Boris Chichkov (Laser Zentrum Hannover) for help in laser treatments.
REFERENCES
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
Nae-Man Park, Chel-Jong Choi, Tae-Yeon Seong, Seong-Ju Park. Phys. Rev. Lett., 86
(2001), p. 1355.
Nae-Man Park, Tae-Soo Kim, Seong-Ju Park. Appl. Phys. Lett., 78 (2001), p. 2575.
T. T. Korchagina, D. V. Marin, V. A. Volodin, A. A. Popov, and M. Vergnat.
Semiconductors, 43 (2009), p. 1514.
Tae-Youb Kim, Nae-Man Park, Kyung-Hyun Kim, Gun Yong Sung, Young-Woo Ok,
Tae-Yeon Seong, Chel-Jong Choi. Appl. Phys. Lett., 85 (2004), p. 5355.
I. Sychugov, R.Juhasz, J.Valenta, and J.Linnros, Phys. Rev. Lett., 94 (2005), p. 087405
(1-4).
Silicon Nitride in Electronics, edited by A.V. Rzhanov, (Elsevier, Amsterdam, 1988; in
Russian - Nauka, Siberian Branch, Novosibirsk, 1982).
V. A. Gritsenko. Phys. Usp., 51 (2008), p. 699.
V.A.Volodin, M.D.Efremov, V.A.Gritsenko, S.A.Kochubei. Appl. Phys. Lett., 73
(1998), p. 1212.
V.A. Volodin, T.T. Korchagina, J. Koch, B.N. Chichkov. Physica E, 42 (2010), p.
1820.
T. T. Korchagina, V. A. Volodin, and B. N. Chichkov. Semiconductors, 44 (2010) p.
1611.
M. H. Brodsky, R. S. Title, K. Weiser, and G. D. Pettit. Phys. Rev. B, 1 (1970) p. 2632.
I.V. Afanasyev-Charkin, L.G. Jacobsohn, R.D. Averitt, and M. Nastasi. J. Vac. Sci.
Technol. A, 22 (2004) p. 2342.
D.E. Aspens, A.A. Studna, Phys. Rev. B, 27 (1983) p. 985.
A.Yu. Kobitski. Phys. Rev. B, 63, (2001) p. 115423.
T. Takagahara, K. Takeda. Phys. Rev. B, 46, (1992) p. 15578.
100
V. A. Volodin
[16] Electronic Structure of Amorphous Dielectrics in MOS Structures, V.A. Gritsenko. (in
Russian - Nauka, Siberian Branch, Novosibirsk, 1993).
[17] M.H. Brodsky, M. Cardona, and J.J. Cuomo. Phys. Rev. B, 16, (1977) p. 3556.
[18] V. Paillard, P. Puech. J. Appl. Phys., 86, (1999) p. 1921.
[19] S. Hernndez, A. Martnez, P. Pellegrino, Y. Lebour, B. Garrido, E. Jordana, and J.M.
Fedeli. J. Appl. Phys., 104, (2008) p. 044304 (1-5).
[20] L.V. Mercaldo, E.M. Esposito, P.D. Veneri, G. Fameli, S. Mirabella, and G. Nicotra.
Appl. Phys. Lett., 97, (2010) p. 153112 (1-3).
[21] I.V. Antonova, M.B. Gulyaev, Z.S. Yanovitskaya, V.A. Volodin, D.V. Marin, M.D.
Efremov, Y. Goldstein and J. Jedrzejewski. Semiconductors, 40, (2006) p. 1198.
[22] M. Molinari, H. Rinnert, M. Vergnat. J. Appl. Phys., 101, (2007) p. 123532 (1-10).
[23] M.D. Efremov, V.V. Bolotov, V.A. Volodin, L.I. Fedina, E.A. Lipatnikov. J. Phys.:
Condens. Matter, 8 (1996) p. 273.
[24] S.P. Singh, P. Srivastava, S. Ghosh, S.F. Khan and G.V. Prakash. J. Phys.: Condens.
Matter, 21, (2009) p. 095010 1-6.
[25] G. Scardera, T. Puzzer, G. Conibeer, and M.A. Green. J. Appl. Phys., 104, (2008) p.
104310 (1-7).
[26] Y.N. Volgin, Y.I. Ukhanov. Optics and Spectroscopy, 38, (1975) p. 412.
[27] N. Wada, S.A. Solin, J. Wong, S. Prochazka. Journal of Non-Crystalline Solids, 43,
(1981) p. 7.
[28] D.V. Tsu, G. Lucovsky, and M. J. Mantini. Phys. Rev. B, 33, (1986) p. 7069.
[29] Y.M. Niquet, G. Allan, C. Delerue, and M. Lannoo. Appl. Phys. Lett., 77, (2000) p.
1182.
[30] M. Lannoo, C. Delerue, G. Allan. Journal of Luminescence, 70, (1996) p. 170.
[31] A.G. Noskov, E.B. Gorokhov, G.A. Sokolova, E.M. Trukhanov, S.I. Stenin. Thin Solid
Films, 162, (1988) p. 129.
[32] V.A.Volodin, M.D.Efremov, G.A.Kachurin, A.G.Cherkov, M. Deutschmann, N.
Baersch. JETP Letters, 86, (2007) p. 119.
[33] J.Bok. Phys. Letters A, 84, (1981) p. 448.
[34] T.Sameshima, M.Hara, N.Sano, and S.Usui. Japanese Journal of Applied Physics, 29,
(1990) p. L1363.
In: Silicon Nitride

ISBN 978-1-61942-865-2
Chapter 4
INFLUENCE OF THE ADDITIVES AMOUNT UPON

SPS-SINTERING, METAL JOINING,
AND HIGH TEMPERATURE OXIDATION
OF SI3N4-CERAMICS
Leonel Ceja-Crdenas1,2, Jos Lemus-Ruiz1,
Sebastin Daz-de la Torre2 and Egberto Bedolla-Becerril1
1
Instituto de Investigaciones Metalrgicas,

Universidad Michoacana de San Nicols de Hidalgo,
Morelia, Mich., Mxico
2
Instituto Politcnico Nacional, Centro de Investigacin e Innovacin Tecnolgica,
Azcapotzalco, D.F., Mxico
ABSTRACT
Technological developments over the past years have created the need for
developing new materials and processes that can better fit the requirements for advanced
applications. Silicon nitride (Si3N4) is one of the most attractive advanced ceramic
materials for structural applications due to its excellent mechanical properties, either at
ambient or elevated temperatures. Although several techniques for sintering silicon
nitride are now available, most of them still require liquid phase formation to confer the
ceramic high-density level. This chapter is focused on the spark plasma sintering of Si3N4 powder using different amount of additives. Both, the - to -phase transformation
observed along densification of Si3N4 specimens and the interfaces they form during the
ceramic to metal joining are discussed as a function of the additives amount. Moreover,
the Si3N4-ceramics behavior when exposed to high-temperature oxidation is presented.
This chapter might be interesting for academic and industrial communities concerned
with the sinter processing and use of ceramics for high-temperature engineering
applications.
Keywords: Si3N4-ceramics, spark plasma sintering, additives, metal joining, high temperature
oxidation
102
L. Ceja-Crdenas, J. Lemus-Ruiz, S. Daz-de la Torre et al.
1. INTRODUCTION
The silicon nitride ceramics were first produced in the mid-nineteenth century due to its
lightweight and the possibility to be used at temperatures larger than metallic materials. This
situation, in fact attracted a considerable attention of the ceramics community. Thus,
researchers have been focused on these ceramics over the years by conducting systematic
experimental work, trying to further develop the physical and mechanical properties of these
ceramics; such as strengthening, toughening, hardening, and also aiming chemical and
thermal stability. Silicon nitride ceramics are produced for a great number of structural
applications including: turbochargers, engine components, turbine gas, cutting tools,
mechanical sealing, bearings, and rolling elements. In spite of this, cost effective Si3N4ceramic components are today sought as to trigger local markets, whereby innovative
manufacturing processes are needed. Further research work focused on the synthesis of highdensity silicon nitride composites; either pure or coupled with certain alloys, whose fine
microstructure can govern their resulting properties is still an important issue for the materials
scientists [1-3].
1.1. Sintering Mechanism of Si3N4-Based Ceramics Having Additives

Apparently, there are no experimental reports that specify full densification of silicon
nitride powder without the usage of chemical additives. These chemicals are known as
sintering aids, which promote formation of a liquid phase sintering. Although a number of
sintering aids that can be used to densify silicon nitride are known today, it is important to
realize that its concentration affects the liquid phase formation and thus the ultimate
properties of the ceramic body. The existence of this liquid can in turn leads to determine the
grain growth kinetics, the morphology of silicon nitride grains, as well as the type of grainboundary structure. Therefore, knowing the sintering mechanism of Si3N4-based ceramics is
essential. A schematic representation of silicon nitride ceramics sintering having additives is
shown in Figure 1. This Figure represents two possible scenarios, from left to right to end up
with different microstructures. Those mechanisms are briefly described next.
a) Starting powder. Inorganic compounds are commonly used as additives to promote
densification of silicon nitride ceramics at elevated temperatures; such as alumina,
magnesia, yttria, nitrides and/or different rare earth oxides. In the first stage, -Si3N4
powders together with the desired amount of additives are mixed using the
equipment and procedure for ceramic materials.
b) Oxidation stage. Once the powders were mixed and compacted under pressure. At
first stage of the sintering process, the oxide layer on surface of the silicon nitride
crystallites is formed due at the heat treatment, which is known as silica (SiO2). Also
in this stage of sintering process, the rearrangement of the particles play an important
role and compressed body can attain a fairly high relative density. Its very important
to point out that usually nitrogen atmosphere is used during the sintering process of
Si3N4-based ceramics.
Influence of the Additives Amount Upon SPS-Sintering
103
c) Liquid phase formation. At a later stage, the powder mixtures are heated at high
temperatures then the additives oxides react with SiO2 to promote the eutectic liquid
formation. At the same time the liquid phase formed wets the particles which
eliminate the bridges between them and aids densification. Also, during this stage
small -Si3N4 particles dissolves in the liquid and precipitated as -Si3N4 phase by
solution-precipitation process.
d) Si3N4-ceramics. Finally, the excessive liquid phase used during sintering process can
be perceived as the remaining liquid which is kinetically frozen and it might then
remain as amorphous grain boundaries upon cooling and is called conventionally as
intergranular glassy film. The resulting microstructure of these ceramics after
sintering is similar to that of whisker-reinforced composites, exhibiting rod-like Si3N4 grains surrounded by a amorphous phase which act as strengthening/
interlocking elements [4-6].
STARTING POWDER
OXIDATION STAGE
Heating
Additives
Additives
-Si3N4
-Si3N4 + SiO2
LIQUID PHASE FORMATION
Si3N4-CERAMICS
-Si3N4
Amorphous phase
Cooling
-Si3N4
Liquid
Additives+-Si3N4+ SiO2
Figure 1. Sintering mechanism of Si3N4-based ceramics having additives.
1.2. Grains Growth of -Si3N4

As it was mentioned before, the formation of the resulting microstructure in silicon
nitride ceramics is dictated by the densification mechanism which in turn leads to an
interlocking microstructures consisting of elongated grains surrounded by a amorphous phase
which is considered responsible of mechanical properties in the ceramic material. Ideally, a
104
fine microstructure with high proportion of elongated -grains is necessary for obtained the
superior mechanical properties, namely the so called high aspect ratio of -grains (lengthwidth relationship). Different factors have contributed at -grain growth such as the liquid
phase viscosity formed during the sintering process [7,8] and/or the incorporated of -Si3N4
seeds within the starting powder before the sintering treatment [9,10]. However, there will be
some other elements that must be considered before sintering treatment to better -grain
growth but also to make possible - to -phase transformation, which can bring as a
consequent densification improvement. These elements, including:
1. The composition of the starting powder. The impurities amount of the starting
powder due not only to its influence on liquid phase viscosity but also high -Si3N4
content to favour the phase transformation;
2. Oxygen amount. Low oxygen content is necessary but sufficient to assist the liquid
phase through silicate formation;
3. Details of the processing parameters. Maximum temperature applied, holding time,
pressure and atmosphere used.
The nature of such -grain growth in Si3N4 ceramics greatly determines the final strength
of bulk materials. Therefore, it is important to make not only the appropriate selection of
additives but also the amount (mixture) to be used. Although, the cause of the -grain growth
is not yet known, it is generally accepted that growth of elongated grains is predominantly
diffusion-controlled but the factors that govern these grow is still a matter of discussion. The
Figure 2 is a fracture surface micrograph of Si3N4 ceramic etched in hydrofluoric acid, where
evidence has shown the typical rod-like shape of hexagonal -Si3N4 crystals [11].
10 m
Figure 2. Micrograph of hexagonal -Si3N4 crystals sintered at 1500C for SPS (92wt%Si3N45Y2O3
3Al2O3) [11].
105
1.3. Amorphous Phase

It is generally accepted that the mechanical properties of silicon nitride ceramics are
closely linked to the - to -Si3N4 phase transformation. However, the amorphous phase of
grain boundary plays an important role on material properties such as creep resistance,
thermal conductivity and oxidation resistance which in turn leads to deterioration of the
ceramic materials at elevated temperature. The amorphous phase is usually present as a thin
glassy film (about 1-2 nm) characterized for a nearly constant thickness which is present at
almost all grain boundaries, independent of the grain orientation. This phenomenon is
considered an equilibrium stage in sintering treatment [12-13]. The amorphous phase could
be crystallized by a post-sintering heat treatment which in turn leads to several heat treatment
processes and different additive compositions in order to crystallize the amorphous phase at
the same time that can bring as a consequent properties improvement as evidenced by
Wasanapiarnpong, T. et al., [14] they found a significant improve of the thermal conductivity
and Vickers hardness of Si3N4 ceramics with elimination of the grain-boundary glassy phase
by post-sintering heat treatment due to mass loss (evaporation of SiO2, Mg, and N2) which
promote decreasing in amount of amorphous phase, however the bending strength and
fracture toughness of these Si3N4 ceramics decreased due to microstructural changes. They
argued that there is an extensive elongated grain growth in all specimens which is derived
from heat treatment.
At this point it is not clear as to what the common factors are amongst amorphous phase
forming in Si3N4 ceramics, nor are the key differences well understood. Therefore, at this
stage complete understanding of this glassy phase still eludes the academic community.
1.4. Si3N4-Ceramics Manufacture Process

Initially, the silicon nitride ceramics were formed by nitriding silicon powder compacts
(reaction sintering). Subsequently, interest began to grow for potential use of Si3N4-ceramics
as engineering materials which in turn leads to increase the density in the ceramics then hot
pressed silicon nitride (HPSN) technique was used. Nowadays the interest in silicon nitride
based engineering materials and the search for improved powder manufacture and the cost of
production has led to the development of a great number of techniques for sintering,
including:
Reaction-Bonded Silicon Nitrite (RBSN) the densification is helped by a chemical
reaction between the solid powder and gas phase to obtain the ceramic. Namely, silicon
powder is compacted into a desired shape then heated in nitrogen atmosphere at temperature
about 1400C during the sintering treatment, the nitrogen gas have permeated the silicon
porous and reacts to form silicon nitride which grows within the preexisting pores of the
powder compact. According to the following chemical reactions:
3Si(s) + 2N2(g) Si3N4(s)
(1)
3Si(g) + 2N2(g) Si3N4(s)
(2)
Si(s) + SiO2 2SiO(g)
(3)
106
To make a loss of silicon during the nitriding treatment (reaction 3) can be to promote a
slight shrinkage <0.1% accompanied by an increase in density on specimen. This feature is
the biggest benefit of RBSN technique, due to allow technical personal to do complex shapes
without the need of subsequent expensive manufacture operations. The reaction bonding
process is suitable for mass production because this technique is relatively cheaper than other
production processes. Nevertheless, the end product contains porosity (from 15 to 25%
porosity) which in turn leads to the strengths are consistently lower than with other sintering
methods (usually about 200 MPa). A consequence of this porosity its application is limited
particularly, include static structural components.
Hot Pressed Silicon Nitride (HPSN) involves the application of uniform pressure at high
temperature for -Si3N4 powder sintering. Typically, RE2O3 are used as additives, the
pressure required (~15-50 MPa), the temperature range is from 1600 to 1800C and the time
involved is a few hours (~1-4 h). Although, the equipment and tooling required are more
expensive than traditional processes, attractive properties can often be achieved such as high
strength (until 800 MPa) and excellent density (<1% porosity). However, the process is
restricted to fairly simple shapes as cutting tools and wear components. In this process, rare
earths (RE2O3) often oxides are used as additives (e.g. magnesia, alumina, yttria, etc.) in
combination with the native oxide present in Si3N4 powders to -Si3N4 phase formation and
normally an intergranular glassy film. According to the following:
-Si3N4 + SiO2 + RE2O3 -Si3N4 + RE-Si-O-N (IGF)
(4)
Sintered Reaction-Bonded Silicon Nitride (SRBSN) is an improvement on RBSN

technique aimed at reducing the final porosity of the end product which determines the final
properties of bulk materials. This is achieved by adding sintering additives to the starting
powder which in turn leads to the material can be sintered during the reaction-bonding stage
through the liquid phase formation which enhances densification of Si3N4-ceramics (<5%
porosity) and the strength (~700 MPa). As with RBSN technique, the starting raw materials
for SRBSN are silicon powders, which in turn leads to the green powder compact can be
manufacture to final size before nitriding treatment which can bring a significant influence on
cost of Si3N4-based products at industrial scale. Applications of SRBSN include cutting tools.
Sintered Silicon Nitride (SSN) is also called pressureless sintering of silicon nitride.
Usually requires some sintering aids and the temperature range is from 1700 to 1850C.
However, the silicon nitride decomposition can be a problem; therefore a nitrogen overpressure atmosphere is necessary which minimize loss of material by evaporation. SSN
generally offer high density (<1% porosity) which in turn leads to the best mechanical
properties available such as strength (about 750 MPa). The end products are the most widely
used in industrial applications such as industrial wear, molten metal handling, gas and
chemical process industries.
Hot isostatic pressing of silicon nitride (HIP-SN) is also called gas pressure sintering
involves an inert gas (nitrogen) which is used to transmit over-pressure at elevated
temperature in the silicon nitride green body. This process required two-step: In the first step
a nitrogen pressure low is applied (1-10 atm) which in turn leads to the Si3N4-ceramics
formation with closed porosity and the second step a nitrogen pressure of (10-200 atm) is
applied. HIP-SN permits to remove residual porosity due to high pressures are applied
uniformly in all direction via an inert gas which has shown to be effective in sintering
107
process. A main advantage of HIP-SN technique as compared to conventional sintering

process (SSN) is that the additive amount used can be reduced to a minimal level.
Microwave sintering of silicon nitride (MS-SN) the heat is generated at the surface of the
material by radiation and then conducted to the centre, microwave heating is applied
volumetrically due to interactions within the bulk of the material, therefore many insulation
ceramic materials such as Al2O3, Y2O3, SiO2, MgO and glasses can be absorbed microwaves
at above a certain critical temperature. Although, silicon nitride is a poor microwave absorber
a low temperatures in microwave sintering the heating is associated with heating of the oxide
additives. Microwave sintering can be more economical than conventional sintering technique
because heating rates are microwave radiation which in turn leads to heating 50 times greater
than conventional process.
Also, the rapid heating rate causes ceramic processed by microwave radiation to have
finer and more uniform grain size than conventionally processed ceramics. All these are
factors to consider in the future studies on Si3N4 sintering [15-17].
Spark Plasma Sintering (SPS) sometimes referred to by other names, such as pulsed
electric current sintering (PECS), field assisted sintering techniques (FAST), and plasma
assisted sintering (PAS), has emerged as a non-conventional powder consolidation method,
which has shown to be effective in sintering a number of novel materials difficult or
impossible to sinter by ordinary techniques.
One of the major differences found between other sintering process and SPS is the
heating rate. Powder densification takes place in SPS through the usage of a pulsed direct
current which passes thought the electrically conducting die and sample holder, so that the
Joule-heat effect takes place almost instantaneously promoting fast sintering of powders. It is
reported that Si3N4 ceramic can be sintered at low temperatures (~1500C) and short holding
time (< 20 min) for SPS technique [18-21].
Table 1. Relative density comparative results of silicon nitride ceramics sintered with
Table 1.
different techniques and conditions
Relative density comparative results of silicon nitride ceramics sintered with different techniques and conditions.
Powder (wt%) Additives (wt%)
Sintering conditions
-Si3N4
Technique
92
92
91
90
92
92
96
92
90
Al2O3
3.0
4.0
6.67
3.0
3.0
1.5
3.0
3.58
Y2O3
3.0
5.0
5.0
3.33
5.0
5.0
2.5
5.0
5.0
MgO
5.0
1.42
PLS
PLS
HP
HP
HIP
HIP
SPS
SPS
MICROWAVE
T( C)
1700
1850
1750
1750
1750
1850
1500
1500
1450
Relative
t(min)
P(MPa)
density (%)
60
120
120
60
120
120
20
3
60
100 a
97.0
99.9
99.9
98.2
99.9
99.9
90.3
99.6
98.0
200a
50
20
160
140
38
38
500 b
References
[22]
[14]
[10]
[2]
[23]
[23]
[11]
[11]
[24]
Code: PLS=pressure-less sintering, HIP=hot isostatic pressing, HP=hot pressing, SPS=spark plasma sintering.
a
Cold-pressing treatment.
b Cold-isostatically
pressed
108
For comparison purposes, Table 1 summarizes the relative density of Si3N4 specimens
developed by using different techniques and experimental conditions, including SPS. It is
worth saying that depending on the kind of application sought for those specimens the most
convenient sintering technique is to be chosen.
1.5. Densification of Si3N4-Ceramics

In the absence of additives during sintering stage in silicon nitride ceramics can bring
difficult to reach full density of the ceramic due to different factors such as:
The strong covalent bonding between silicon and nitrogen atoms;

The onset of Si3N4 decomposition in place of melting at high temperature;
The solid-state diffusion is very slow.
On the other hand, usually the size and shape of the particles determine their mobility in
the presence of liquid phase during sintering process which has shown to be effective in
densification of the materials. However, several factors play a important part in the stage of
the densification such as applied pressure, wettability, penetration of the liquid between the
solid particles, and the partial solubility of the solid phase in the liquid but also the condition
necessary for optimum densification is the presence of the right amount of the liquid phase
[25-27].
Table 2. Code and chemical composition of sintered specimens (wt%)
Table 2. Code and chemical composition of sintered specimens (wt%).
Code
-Si3N4
M1
100
0.0
0.0
0.0
M2
96
2.5
1.5
4.0
M3
92
5.0
3.0
8.0
M4
88
7.5
4.5
12.0
Y2O3
Al 2O3
Additives
Figure 3 Research shows the relative density of Si3N4-Y2O3-Al2O3 ceramics treated by

SPS at 1500oC, plotted versus the holding time upon sintering, using different amount of
additives (see, Table 2). In all studied cases, the relative density further developed by
increasing the amount of additives. It is confirmed that if no additives were used into the Si3N4 powder matrix, case M1, there will be neither more densification nor phase
transformation (Figure 3 and 4a). To be noted is the fact that allowing holding time along the
sintering course, and once the silicon nitride matrix is at 1500oC denser bodies are obtained,
provided a given amount of additives are supplied into the matrix. Thus, 3 and 6 min soaking
time is needed for M4 and M2 specimens to densify, respectively.
Analyzing Figure 3 it is evident that the more amount of additives the larger densification
level developed on those specimens. Evidently, the usage of more amounts of additives might
increase the sintering cost of Si3N4-based products at industrial scale, apart from spoiling
109
mechanical properties due to an excess of glassy phase. Setting few amount of additives and
expanding the soaking time, case M2, however, does not solve the economic problem since
more energy is then required.
Relative density, d/%
100
90
80
Total Additives:
70
M1
60
M2
M4
50
0
Time, t/min
Figure 3. Relative density of Si3N4 ceramics SPS-sintered at 1500C plotted as a function of soaking
time (Table 2).
1.6. Si3N4 Phase Transformation

From literature it is known that the silicon nitride ceramics have two major
crystallographic modifications: and , both hexagonal where the unit cell of the -Si3N4 is
twice as length as the -Si3N4. However, now known that silicon nitride to undergo a third
crystalline form: (cubic spinel), which can be obtained under high temperature and high
pressure conditions. Jiang, J. Z. et al., [28] reported for cubic phase an average hardness 35.31
GPa, slightly larger than SiO2 stishovite, therefore they referred to cubic phase as the third
hardest material after diamond and cubic boron nitride. Also, they found that cubic spinel is
stable up 1400C in air therefore on the basis of their observations and the reported data, they
suggest a phase transformation sequence of to to -Si3N4 phases.
On the other hand, silicon nitride based ceramics are mostly prepared with -rich Si3N4
powders, also is assumed that the initial temperature of the -transformation is about 1400C
(depends on the nature of the material being sintered and additives composition) which in turn
leads to morphology changes on microstructure namely from -equiaxed grain to an
idiomorphic -rod like grain. A significant degree of the phase transformation and the
subsequent -grain growth is excellent for mechanical properties of Si3N4-ceramics (eg.
fracture toughness, thermal conductivity) [29, 30]. Although, sintering Si3N4-ceramics is a
complex process, during which densification, phase transformation and grain growth take
place simultaneously. Therefore, it is not easy task to clarify the mechanism determining the
rate of grain growth however evidence has shown significantly microstructural changes when
the different amount of additives are used (see, Figure 4). As can be seen from Figure 4 for
110
specimens sintered at 1500C for SPS, the fractured surface of Si3N4 samples having different
amount of additives (see, table 2) showed changes on development grain growth with
increased additives amount. Figure 4a corresponds to the M1 specimen after sintered reveals
that if no additives were used into the -Si3N4 powder matrix, there will be not rheological
difference between particles [11]. Figure 4b case M2 specimen, at the center of this picture
not only a small -Si3N4 grain is observed but also some -Si3N4 crystals on its way to adopt
hexagonal forms, both of them are surround by equiaxed grains. However, the majority of
grains were equiaxial and its microstructure almost the same as that of the starting powder,
indicating little phase transformation. Figure 4c correspond to the case M3 specimens, it is
possible to observe a significantly grain growth of -Si3N4 rod. Therefore, it is evident that
the amounts of additives have a significant effect on grain growth of -Si3N4 [31].
1 m
1 m
(a) M1-specimen, without additives
(b) M2-specimen, having 4wt% total additives
1 m
(c) M3-specimen, having 8wt% total additives

Figure 4. Si3N4-ceramics sintered for SPS-technique. (a) Fracture surface of M1-specimen (100
wt%Si3N4) [11]; (b) Fracture surface of M2-specimen (96 wt%Si3N4-2.5Y2O3-1.5Al2O3) [31] densified
at 1500C for 20 min. Irregular grains seen in (a b) are -Si3N4 grains. (c) Fracture surface of M3specimens (92 wt%Si3N4-5Y2O3-3Al2O3) densified at 1500C for 3 min. Hexagonal rod seen in (c) is Si3N4 grain.
On the other hand, the X-ray diffraction analysis reveals the phase transformations in
Si3N4 ceramics. In this research, the evidence has shown that -phase transformation could
achieve once the amount of additives is sufficient. Figure 5 shows X-ray diffraction patterns
111
of the M2- and M4-specimens sintered at 1500C for 5 min. While in M2-specimens, having
4 wt% total additives the X-ray diffraction pattern indicates that the - to -transformation is
not yet achieved on the precursor powder unlike M4-specimens having 12 wt% total additives
where the X-ray diffraction analysis showed fully crystal phase. These results suggest that
the M2-series specimens, cannot undergo complete - to -phase transformation,
independently of sintering conditions. Possibility, due to insufficient eutectic liquid amount
formed.
-Si3N4
Intensity, I/a.u.
-Si3N4
M4-specimens
M2-specimens
10
20
30
40
Diffraction angle, 2/ (Cu K)

Figure 5. XRD-patterns of M2- and M4-specimens sintered at 1500C for 5 min. - and -Si3N4 phase
seen in M2-specimens, In case, M4-specimens the -Si3N4 phase is observed.
1.7. Fracture Behavior of Si3N4-Ceramics

It is known that the high fracture toughness observed in silicon nitride based ceramics is
attributed to the needle-like grain morphologies that are evidenced in the microstructure
which depend not only on the development of large grains within the microstructure, but also
on the interfacial bonding between grains and amorphous grain boundaries. The presence of
large grains can increase the fracture toughness providing that interfacial unjoining occurs,
which implies that the intergranular glassy film must have a weak plane which debonds
leading to crack propagation. (intergranular fracture). If the interfacial bonding is too strong
then the crack propagation becomes more transgranular and the effect of large grains is
reduced. This bonding between grains and amorphous grain boundaries is controlled by the
chemistry of the glassy phase. In the Si3N4 system is given that the strength of the bulk glass
is similar to that the ceramic [32-34].
Evidently, the crack traveled around the -Si3N4 grains, forming the so called crack
deflection which is clear evidence of so that more energy is absorbed during fracture of the Si3N4 columnar grains. This indicate that the toughening contribution from crack bridging
strongly increases with an increase in larger elongated -Si3N4 then the mechanical properties
112
of sintered silicon nitride such as fracture strength and fracture toughness are closely linked to
the -Si3N4 content, due to the contribution to mechanisms such as crack bridging and crack
deflection. Figure 6 is a fracture surface micrograph of Si3N4 ceramic etched in hydrofluoric
acid, where evidence has shown the bridging effect due to an interlocking network formation
of elongated grains was occurred.
1 m
Figure 6. Fractured surface of Si3N4 sample sintered at 1500C for SPS.
2. JOINING OF SILICON NITRIDE CERAMICS

Although, the several sintering techniques of silicon nitride can be used for easier and
better densification of the ceramic material, however when the task is obtaining fully dense
microstructure bulk samples, these has received additional attention in various fields of the
materials engineering due to a limitation imposed by the sintering process which in turn leads
to obtain Si3N4-ceramics of simple shape and small size. Therefore, the need for advance in
engineering applications of silicon nitride ceramics require them to be fabricated as large and
complex shaped components or more often to be functionally or structurally interfaced with
otherwise metallic structures. Nevertheless, the design and engineering of Si3N4-ceramic
components require special attention due to its nature predominantly ionic-covalent which in
turn leads to have microstructures with high hardness and brittle. Therefore the conventional
machining using cut tools coated with diamond required time consuming and unsuitable for
complex shapes. This suggests that Si3N4-ceramic parts can be manufactured by electrical
discharge machining however this technique is very expensive and require a low electrical
resistivity in the ceramic (10-2-10-3) cm namely an electroconductive secondary phase [35,
36].
All these factors have generated a continued interest from the academic and industrial
communities in the use of joining technologies as an alternative to increase the potential of
these ceramics but also to produce complex configurations where properties of a ceramic
need to be combined with those of a metal, due to that both are essential to the functioning of
the device. In the last decades, various methods of ceramic joining have been developed to
113
make Si3N4/Si3N4 and Si3N4/metal joints. Different ways to achieve the interactions between
ceramic and metal can be used and the techniques divided into three major groups:
1. Mechanical attachment, technique is used to form dissimilar joints and often quite
simple in design and commonly employ bolts or screw threads. Although, this
joining technique is used in mass productions process at industrial scale when the
process to involve ceramic materials, then the design required expensive machining
and its application is limited to the components used at low temperature due to high
thermal stress which act as starting point in crack formations.
2. Solid state diffusion bonding, also called direct bonding, is a solid phase process
achieved by diffusion of atoms that take place at the interface. In the absence of a
liquid phase, the joint has to be made at high temperatures, prolonged times and
using pressure to produce intimate contact and to provide sufficient thermal energy to
permit diffusion and chemical reaction between the ceramic and metallic surfaces.
Generally, diffusion bonding technique is used for joints that must withstand high
temperature services.
3. Liquid phase joining, also called indirect bonding,is a technique that involve the
liquid formation in the interaction between materials, but also is considered the most
common technique to achieve high integrity joints using a wide range of intermediate
bonding materials such as organic adhesives, glasses, welding, soldering, and
brazing. Sometimes is necessary to use pre-treatment on the ceramic surface to
improve the wet [37-39].
2.1. Si3N4-Ceramic to Metal Interface

The choices about materials and joint designs usually have to be made based on the
nature of the components and the proposed service performance of the bonded assembly.
However, as mentioned previously the role that different joining techniques used in Si3N4ceramics can play during the assembly not only is an important factor to consider but also the
numerous parameters that at this stage of the joining process could be damaging to strength
of the assembly such as, bonding time, bonding temperature, atmosphere, etcetera [40].
Nevertheless, others physics factors have influenced on strength in ceramic to metal joining
which is necessary to know.
In the first place is very important to know that the silicon nitride ceramic at high
temperature to promote the decomposition in silicon and nitrogen, according to the reaction
(5), therefore is important to have knowledge not only what kind of a reaction takes place at
the interface but also the possible formations of the new species which decisively to have
influence on strength joint.
Si3N4 3Si + 2N2
(5)
Subsequently, the contact between Si3N4 to metal shows the possibility that Si and N gas
react with the metal (Me) promoting one of the following three reactions:
Me + Si3N4 Me silicide + Me nitride
(6)
114

Me + Si3N4 Me silicide + N2(g)
(7)
Me + Si3N4 Me silicide + Si
(8)
However, silicide formations may or may not be observed depending on the joining
parameters as joining temperature, holding time, atmosphere, etc., it is due at affect of
diffusing Si into the interface which promotes the nature of resulting silicide. In case of the
nitrogen can be react with the metal to form a metal nitride (reaction 6), another option is that
nitrogen produced by reaction can be released from the system as nitrogen gas, or that
diffuses into the metal when it has a high solubility for nitrogen (reaction 7). On the other
hand, if the Si atoms not exceed the solid solution limit then Si atoms could disseminate into
the reaction interface (reaction 8) [41-42]. During the joining process the nitrogen
volatilization and silicon diffusion from Si3N4-ceramic to metal is a result of the reaction at
elevated temperatures to form metal silicide followed in many systems by the reaction of the
metal with the free nitrogen to form binary metal nitrides on the ceramic side of the join.
However, in the few cases where ternary phases occur, these are formed by nitridation of
primary metal silicides. These nitride phases act as diffusion barriers for further silicon
diffusion, resulting more complex structures for ceramic to metal junctions such as:
Si3N4/metal nitride/metal silicide/metal [43].
Another important factor that reduces joiningstrength and promotes residual stresses
creation in ceramic/metal interfaces is the different physic-chemical properties between the
materials, due not only to the coefficient of thermal expansion of the base materials but also
the plastic flow of the metallic members which in turn leads the microcrack formations on the
interface. Generally, this event could be induced by the thermal stress during cooling process.
Although, these residual stresses reduce the strength of the bonded material and in some case
lead to joint failure during or after the joining process, therefore minimizing residual stresses
significantly contributes to increasing the mechanical strength which in turn leads to have
reliability of the joint components. The concentration and distribution of residual stresses in
ceramic/metal interfaces depends of the different factors, such as: thermal cycle of the joining
process, joint geometry, thickness and composition of reaction layers. It is known that a large
number of variables that to control the level of residual stresses have been identified, however
the most of these cannot be modified because of fixed joining practices and workpiece
geometries. Therefore, the use of intermediate or buffer metallic inserts of appropriate
elastic/thermal properties and thickness that decrease the critical stresses in the ceramic has
been proposed such as; copper and aluminum, whereas some refractory metals and
expansion-controlled alloys (Kovar and Invar) can be used to promote a gradual transition
from the ceramic to the metal, thus reducing coefficient of thermal expansion mismatches
[44, 45].
As it is mentioned previously, numerous factors have attracted considerable attention
from the academic and industrial communities due to decrease the strength on ceramic to
metal joining and these effects can be observed for Si3N4/Ti joint in the Figure 7. This figure
shows the microstructure formed at the cross section of the Si3N4/Ti interfaces, which consist
of one layer with three intermixed phases, EPMA performed on this sample indicated that
these phases were Ti5S3, TiSi and the zone (B), which is a mixture of TiN containing Y, Al,
O, and Si. This is associated with the segregation of the sintering aids, Y2O3 and Al2O3 during
the decomposition of the ceramic to form alumino-silicate. Also, several cracks can be
115
observed in the reaction layer, mainly in the Ti5Si3 phase. The high affinity of Ti for N and Si
resulted in immediate formation of TiN and a silicide, initially Ti5Si3. The TiSi phase was the
result of the rapid transformation of Ti5Si3, which indicates considerable decomposition of
Si3N4, and a large supply of Si.
Ti
TiSi
Ti5 Si3
Si3N4
10 m
Figure 7. Cross-section of the Si3N4/Ti interface for sample hot-pressed at 1400C for 120 min in
vacuum.
2.2. Conventional Brazing

In ceramic to metal joining various techniques have been used in many laboratory
studies, and some of the processes have progressed to production of components, however,
the main problem is chemical incompatibility as well as thermal mismatch which are crucial
factors for the development of the joint and its applications. At the beginning is possible to
use the different ceramic to metal joining techniques such as: mechanical attachment and
diffusion bonding however, the range of possibilities is more restricted for these joining
process due not only to service temperatures (mechanical attachment) but also elevated cost
of the processing (diffusion bonding), therefore brazing is the most favored technique.
Brazing is one of the joining processes associated with liquid filler formation between the
base materials. This process has been defined as joining technique that take place above
450C which is based on melting and solidification of an interlayer filler metal or filler alloy
which flows by capillary forces between the materials at join and whose melting temperature
is lower than the solidus temperature of the base materials. Compared with other joining
technologies, brazing present some advantages that can be summarized as follow:
Low-cost at industrial scale due to the equipment requires relatively little capital;
Complex geometric and large can be jointed
116
Hybrid materials and with different thicknesses can be bonded in a single step
operation.
However, the main disadvantage found in the brazing process is the melting point of
fillers or filler alloys which in turn leads to limit the service temperature of assembly [46-47].
2.3. Reactive Brazing

During the brazing process for ceramics to metals joining, it is essential that the molten
filler metal wets the ceramic material. To promote wetting, the surface of the ceramic must be
previously metalized with active element as Ti, Hf, Ni, Nb, Ta, and Cr etcetera. DoNacimento, R. M. et. al, [48] used the mechanical metallization with Ti as an alternative route
to deposit active metallic film on Si3N4-ceramics surface prior to brazing process. They
argued that addition of small amount of Ti into convectional alloys favored the development
of the join. They concluded that the Ti films mechanically deposited onto Si3N4 surfaces
reacted with the fillers alloys, improving their activity and wetting, where there is no evidence
of thermal cracking. As a different alternative the filler metal compositions should contain an
active element which in turn leads to develop of the new commercial filler alloys for high
temperature service. Paulasto, M. et. al, [49] used two kind of commercial filler alloys: AgIn1-Ti1 and Ag-Cu26.5-Ti3 (in wt%) for self-Si3N4 joining. They found a higher growth of
reaction layer provided Ti-activity changed from 2 to 5 m. They argued that the activities of
both Ti and Si decrease strongly with increasing Cu-content in the liquid which in turn leads
to have lower equilibrium activities of Ti and Si in Cu-rich brazes than in Ag-rich, apparently
the reaction products between Cu-rich and Ag-rich filler alloys are different. Nevertheless, the
active element can destabilize the ionic-covalent bond of the ceramic and form an
intermediate reaction layer which in turn leads to achieve chemical bonding between ceramic
and metal. The content of active element in filler must be reasonable; otherwise, the
brittleness of joint is increased due to the excessive reaction which is associated with
liberation of Si-atoms to form brittle silicides [50].
Till now, several factors that could weaken the interface of join have been studied.
However, evidence has shown that when differents amount of additives are used in Si3N4ceramics, this have effect on rate of descomposition in Si3N4 during ceramic to metal joining
process. In Si3N4/Cu-Zn/Nb bonding at 1000C for 40 min (see, Figure 8 and Table 2) while
with 4wt% of additives used in Si3N4-ceramics is possible to observe Si-based components
near of the bonding line between reaction interface and Nb (see, Figure 8a). But when used
the same joining conditions with Si3N4-ceramics having 8wt% of additives no detect silicide
formations on interface (see, Figure 8b).
It is evident that in our case Si-based components formation should be avoided, if the
amount of additives used is 8wt%, therefore the amounts of additives have significant effect
on the silicon nitride decomposition which in turn leads to silicide formations on interface in
Si3N4-ceramic to metal joining.
117
Nb
Si-Nb-Cu
Cu-rich
Nb-rich
Si3N4
(a) M2-specimen, having 4wt% total additives.
Nb
Cu-rich
Nb-rich
Si3N4
(b) M3-specimen, having 8wt% total additives.
Figure 8. Cross-section of the Si3N4/Cu-Zn/Nb interface for sample brazed at 1000C for 40 min in
argon.
2.4. Partial Transient Liquid Phase Bonding (PTLPB)

The active metal brazing is one of the more flexible technique and cost effective in
ceramic to metal joining. However, when the joining temperature is elevated (>1000C), the
brittle compounds formation into the interface has been observed which in turn leads to
changes in microsructure and thickness of the interface which may to lead joining strength to
decrease. This factor has led to explore non-conventional methods of joining that would
exploit a higher diffusivity transport path at lower temperature such as partial transient liquid
phase bonding (PTLPB). The partial transient liquid phase bonding technique is a promising
118
joining technology due to that advantage as much brazing as diffusion bonding are used. This
technique would exploit the liquid film formation, and rapid diffusion of the metal with
higher-melting-point within the liquid to facilitate joint formation. Namely, during ceramic to
metal joining, the metals at edges melt directly or form liquid through eutectic reaction then
the liquid phase accelerates joining due to that the movement of atoms in liquid is quicker
which in turn leads to wets of the ceramics and eliminates the holes at interface which may
reduce the joining time and pressure. These factors could be an indicative of a strong joint
formed by diffusion of atoms and isothermal solidification. Although, the complete
disappearance of the liquid formed would not be required, only rearrangement of the liquid
phase so that the joint region no longer contained a continuous layer of the liquid former. By
dispersing the liquid former, melting would have a minimal effect on the high-temperature
properties of the joint. This alternative approach opened up a much wider range of alloy
systems as interlayer candidates for the future, ones in which the liquid former has minimal
solubility in the host material [51].
3. HIGH TEMPERATURE OXIDATION OF SI3N4 CERAMICS

Silicon nitride (Si3N4) has generated considerable interest as a potential material for hightemperature engineering applications. However, the high-temperature properties such as
oxidation resistance have been not understood. Although, the excess of oxide additives has
been widely used to make densification easy for sintering process, nevertheless the formation
of intergranular glassy film upon cooling to believe has a profound influence on the hightemperature oxidation behavior of Si3N4-ceramics [52, 53]. Commonly, the product oxide
film in Si3N4-ceramics is partially crystalline, and may contain cristoballite (SiO2), together
with crystalline silicates formed by reaction of secondary metal oxides phase with the silicon
dioxide as primary oxidation product.
Namely, initially the oxide layer is controlled by oxygen diffusion in Si3N4-ceramics,
according to the following chemical reaction:
Si3N4 + 3O2 3SiO2 + 2N2(g)
(9)
Therefore, when both SiO2 is obtained and oxidation time prolonged, it is possible that
crystalline phase (SiO2) can be reorganized as silicate, according to:
RE2O3 + 2SiO2 RE2Si2O7
(10)
When Y2O3 and Al2O3 are used as additives in the sintering process, the liquid is formed
by the reaction of the additives with SiO2 present on the surface of Si3N4 grains, which
commonly forms an amorphous phase upon cooling (containing: Si, Al, O, N, and Y).
However, after heat post-treatment in Si3N4-ceramics, the results by electron probe
macro-analysis (EPMA) and X-ray diffraction analysis (XRD) have shown that the oxide
layer is constituted by SiO2 and Y2Si2O7 [54-56]. The evidence has shown rod-like shape
crystal of Y2Si2O7 on oxidized surface of Si3N4 (see, Figure 9).
119
100 m
Figure 9. Oxidized surface of M4-specimen (88wt%Si3N4-7.5Y2O3-4.5Al2O3) at 1200C for 24 d in air.
3.1. Cristallization of the Amorphous Phase

Some authors have explained the oxidation of Si3N4-ceramics in terms of oxygen
potential and as an inward diffusion of oxygen, with a concurrent outward diffusion of cations
[57-60], therefore the amorphous phase has influence on oxide layer formation in the surface
which in turn leads to changes chemistry inner ceramic. The evidence from scanning electron
microscopy (SEM) has shown that the crystallization of the amorphous phase in Si3N4ceramics after heat post-treatment is possible (see, Figure 10).
Oxide layer
2 m
Figure 10. Cross-section of Si3N4-ceramics oxidized at 1200C for 1 d.
Although, the complete crystallization of the amorphous phase in Si3N4-ceramics cannot

be attained and residual glass remains at internal interfaces has been observed [61]. It is
considered that the partial crystallization of the amorphous phase involves a mechanical
degradation in the ceramic due to development of tensile stresses at the grain boundaries
caused by volume reduction and change of density as a consequence of crystallization of the
120
intergranular phase. If this is true, easy crack propagation in the grain boundaries of the Si3N4
and therefore fracture toughness could be altered. The evidence in Si3N4 specimens fractured
after oxidation process has shown microstructural changes into ceramic. Figure 11a
corresponds to the as-sintered sample. It can be observed a representative view of those
sintered bodies disclosing the typical rod-like shape of hexagonal -Si3N4 crystals. Figure 11b
shows same sample after oxidation process at 1200C for 24 d. Elongated Si3N4 grains with
random distribution was covered by a different kind of morphology like a spider network.
2 m
2 m
(a)
(b)
Figure 11. Fractured surface of M4-specimens (88 wt%Si3N4-7.5Y2O3-4.5Al2O3). (a) as sintering, (b)
after oxidation process at 1200C for 24 d.
CONCLUSION
All these facts indicate that the formation of a liquid phase is necessary for obtained
Si3N4-composites. However, the excessive additives amount used during sintering process
could be kinetically frozen then to promote the vitreous phase formation which could not be
beneficiated on properties of the Si3N4-ceramics. If additives amount used is no-sufficient
then the phase transformation and the relative density could not be achieve. To have
121
equilibrium between the additives amount and the sintering conditions are factors to consider
in the future studies.
REFERENCES
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27]
Vikulin, V. V. Adv. Sci. Tech. 2006, 45, 1751-1758.

Jiang, Y.; Wu, L.; Wang, P.; Huang, Z. Mater. Res. Bull. 2009,44, 21-24.
Hampshire, S. Mater. Sci. Forum 2009, 606, 27-71.
Amato, I.; Martorana, D.; Silengo, B. Mater. Sci. Eng. 1977. 28, 215-220.
Tien, T. Y.; In Phase Diagrams in Advanced Ceramics; Alper, A. M. Academic Press
Limited: London U. K., 1995; pp127-156.
Xue-Jian, L.; Zhi-Yong, H.; Qi-Ming, G.; Xing-Wei, S.; Li-Ping, H. J. Eur. Ceram.
Soc. 2005, 25, 3353-3359.
Vuckovic, A.; Matovic, B.; Boskovic, S. Mater. Sci. Forum 2005, 494, 429-434.
Bressiani, J. C.; Izhevskyi, V.; Bressiani, A. A. H. Mater. Res. 1999, 2, 165-172.
Chien-Cheng, L. Ceram. Int. 2003, 29, 841-846.
Belmonte, M.; De Pablos, A.; Osendi, M. I.; Miranzo, P. Mater. Sci. Eng.A. 2008, 475,
185-189.
Ceja-Crdenas, L.; Lemus-Ruiz, J.; Jaramillo-Vigueras, D.; D. de la Torre, S. J. Alloy
Comp. 2010, 501, 345-351.
Shibata, M.; Kinomura, N.; Koizumi, M. Ceramurgia Int. 1980, 6, 146-147.
Vetrano, J. S.; Kleebe, H. J.; Hampp, M.; Hoffmann, M. J. J. Mater. Sci. Lett. 1992,
11, 1249-1252.
Wasanapiarnpong, T.; Wada, S.; Imai, M.; Yano, T. J. Eur. Ceram. Soc. 2006, 26,
3467-3475.
Benseyei, E.; Refractory Metals and Hard Mater.s. 1992, 11, 121-126.
Chawla, K. K.; Ceramic Matrix Composites; Chapman and Hall: Cambridge U. K.,
1993; pp28-31.
Lee, W. E.; Rainforth, W, M.; Ceramic Microstructure, Property Control by
Processing; Chapman and Hall: Sheffield U. K., 1994; pp396-410.
Shen, Z; Nygren, M. J. Mater. Chem. 2001, 11, 204-207.
Chen, F; Shen, Q; Yan, F; Zhang, L. Mater. Chem. Phis. 2008, 107, 67-71.
Lu, X; Ning, X; Xu, W; Zhoy, H; Chen, K. Key Eng. Mater. 2005, 280-283, 1259-1262.
Julin, W; Ren-Guan, D; Amiya, K. M. Scr. Mater. 2005, 53 663-667.
Ling, G.; Yang, H. Mater. Chem. Phis. 2005, 90, 31-34.
Bhandhubanyong, P.; Akhadejdamrong, T. J. Mater. Process. Technol. 1997, 63, 277
280.
Jones, M. I.; Valecillos, M. C.; Hirao, K.; Yamahuchi, Y. J. Eur. Ceram. Soc. 2002, 22,
2981-2988.
Coutinho, A. C. S.; Bressiani, J. C.; Bressiani, A. H. A. Mater. Sci. Forum 2003,
416.418 567-572.
Suganuma, M.; Kitagawa, Y. J. Am. Ceram. Soc. 2003, 86 (3), 387-394.
Dae-Ho, C.; Byung-kyu, M.; Rock-Joo, S.; Seung-Ho, K.; Niihara, K. Mater. Sci.
Forum 2005, 486.487 181-184.
122
[28] Jiang, J. Z.; Kragh, F.; Frost, D. J.; Stahl, K.; Lindelov, H. J. Phys: Condens. Matter.
2001, 13, L515-L520.
[29] Wei, P.; Chen, L.; Okubo A.; Hirai, T.; Mater. Lett. 2001, 49, 239-243.
[30] Kramer, M.; Hoffmann, M. J.; Petzow, G. Acta Metall. Mater. 1993 41(10), 2939-2947.
[31] Ceja-Crdenas, L.; Lemus-Ruiz, J.; D. de la Torre, S.; Bedolla-Becerril, E. Ceram.
Trans. 2012, (under review).
[32] Melndez-Martnez, J. J.; Domnguez-Rodrguez, A. Progress Mater. Sci. 2004, 49, 19107.
[33] Xu, X.; Huang, L.; Liu, X.; Fu, X.; Ceram. Int. 2002, 28, 279-281.
[34] Bjorklund, H.; Falk, L. K. L.; Rundgren K., Wasn, J. J. Eur. Ceram. Soc.1997, 17,
1285-1299.
[35] Klein, R.; Desmaison-Brut, M.; Ginet, P.; Bellosi, A.; Desmaison, J. J. Eur. Ceram.
Soc. 2005, 25, 1757-1763.
[36] Lemus-Ruiz, J.; Ceja-Crdenas, L.; Verduzco, J. A.; Flores, O. J. Mater. Sci. 2008, 43,
6296-6300.
[37] Yong, Z.; Di, F.; Zhi-yong, H.; Xi-chun, C. J. Iron and Steel Res. Int. 2006, 13(2), 0105.
[38] Fernie, A. J.; Sturgeon, A. J. Metals and Mater. 1992, 212-217.
[39] Martinelli, A. E.; Hadian, A. M.; Drew, R. A. L. J. Canadian Ceram. Soc. 1997, 66(4),
276-284.
[40] Lemus-Ruiz, J.; Ceja-Crdenas, L.; Bedolla-Becerril, E.; Lpez-Morelos, V. H.
Production, Characterization, and Mechanical Evaluation of Dissimilar Metal/Ceramic
joints; Cuppoletti, J. InTech: Rijeka, CR., 2011; pp 205-224.
[41] Polanco, R.; De Pablos, A.; Miranzo, P.; Osendi, M. I.; Applied Surface Sci. 2004, 238,
506-512.
[42] Lemus, J. Drew, R. A. L. British Ceramic Transaction. 2000, 99(5), 200-205.
[43] Schuster, J. C.; Weitzer, F.; Bauer, J.; Nowotny, H. Mater. Sci. Eng.A. 1998, 105/106,
201-206.
[44] Do Nacimento, R. M.; Martinelli, A. E.; Buschinelli, A. J. A. Ceramica. 2003, 49, 178198.
[45] Lemus, J.; Drew, R. A. L. Mater. Sci. Eng.A. 2003, 352, 169-178.
[46] Nicholas, M. G. Joining Processes, Introduction to brazing and diffusion bonding,
Springer-Verlag: New York U.S., 1998; pp. 22-24.
[47] Schwartz; M. M. Ceramic Joining; ASM International: Ohio, U.S., 1990; pp 99-103.
[48] Do Nacimento, R. M.; Martinelli, A. E.; Buschinelli, A. J. A.; Reisgen, U.; Remmel, J.;
J. Mater. Sci. 2005, 40(17) 4549-4556.
[49] Paulasto, M.; Kivilahti, J. K. Scripta Metall. et Mater.1995, 32(8) 1209-1214.
[50] Peteves, D. S. Ceram. Int. 1996, 22, 527-533.
[51] Sugar, J. D.; McKown, J. T.; Akashi, T.; Hong, S. M.; Nakashima, K.; Glaeser, A. M. J.
Eur. Ceram. Soc. 2006, 26, 363-372.
[52] Houjou, K.; Ando, K.; Chu, M. C.; Liu, S. P.; Sato S. J. Eur. Ceram. Soc. 2005, 25,
559-567.
[53] Vetrano, J. S.; Kleebe, H. J.; Hampp, E.; Hoffmann, M. J.; Cannon, R. M. J. Mater. Sci.
Lett. 1992, 11, 1249-1252.
[54] Houjou, K.; Ando, K.; Liu, S. P.; Sato S. J. Eur. Ceram. Soc. 2004, 24, 2329-2338.
[55] Gubicza, J.; Wber, F. Mater. Sci. Eng.A. 1999, 263, 101-104.

[56]
[57]
[58]
[59]
[60]
[61]
123
Nakatani, M.; Ando, K.; Houjou, K. J. Eur. Ceram. Soc. 2008, 28 1251-1257.
Andrews, P.; Riley, F. L.; J. Eur. Ceram. Soc. 1989, 5 245-256.
Williams, R. O; Tennery, V. J. J. Mater. Sci. 1979, 14, 1567-1571.
Cubicciotti, D.; Lau, K. H.; J. Am. Ceram. Soc. 1978, 61 (11-12), 512-517.
Clarke, D. R.; Lange, F. F. J. Am. Ceram. Soc. 1980, 63 (9-10), 586-593.
Subramaniam, A.; Koch, C. T.; Cannon, R. M.; Ruhle, M. Mater. Sci. Eng.A. 2006,
422, 3-18.
In: Silicon Nitride

ISBN 978-1-61942-865-2
Chapter 5
SYNTHESIS AND PROPERTIES OF

SILICON NITRIDE COATINGS
Robert B. Heimann
Professor emeritus of Applied Mineralogy and Materials Science
ABSTRACT
Silicon nitride ceramics show excellent mechanical, thermal and chemical properties
that today are being exploited in diverse, technologically challenging areas. The superior
performance of monolithic silicon nitride in service notwithstanding, problems arise from
the inherently brittle nature of the material as well as the difficulties involved in joining
silicon nitride to metallic parts. To overcome these challenges, several attempts have
been made to develop silicon nitride as protective coatings for otherwise ductile, stressresilient metal substrates. Such coatings are designed to combine synergistically the
strength and toughness of a metal with the protective properties of silicon nitride against
mechanical and thermal attack, and chemical deterioration, as encountered, for example,
in highly corrosive combustion gases of aerospace gas turbines and automotive engines.
Unfortunately, deposition of pure silicon nitride coatings by thermal spray technology is
impossible since the material dissociates and, in turn, sublimates at temperatures above
1800oC, and is also subject to rapid oxidation at elevated temperature. To combat these
technological limitations, addition of sintering aids such as alumina and yttria, stringent
control of the viscosity of the molten silicon nitride-oxide spray droplets, optimization of
the heat transfer from thermal plasmas to the particles, and high kinetic energy of the
deposition process are required. In this chapter experimental solutions leading to hard,
dense, and well-adhering coatings will be presented that include low-frequency pulsed
detonation spraying (LFPDS) and high current pulsed plasma spraying (HCPPS).
Email: robert.heimann@ocean-gate.de.
126
Robert B. Heimann
INTRODUCTION
Silicon nitride belongs to the very few primary products of stellar nucleosythesis in
intermediate-mass stars, i.e. stars heavier than eight solar masses. To form the constituting
elements, three 4He nuclei (-particles) produced by H burning combine in a triple- process
to form first 12C that evolves by two routes into still heavier elements. First, it can react with a
proton and by neutron capture to 14N. Secondly, it generates 16O by -capture. Two 16O nuclei
then fuse to form an excited 32S* nucleus that disintegrates into 28Si and an -particle. If the
nuclear fuel is exhausted such massive stars end their lives in gigantic explosions as type-II
supernovae and expel their charge of newly synthesized elements into the interstellar medium
(ISM) [1]. Consequently, presolar grains of Si3N4 crystallize as rare constituents in addition to
more abundant SiC, diamond, graphite, and several oxides (spinel, corundum, hibonite, rutile)
and silicates [2], and can be found as the minerals nierite (-Si3N4) and sinoite (Si2N2O) in
ordinary and enstatite chondritic meteorites.
On earth, silicon nitride was first synthesized by Weiss and Engelhardt [3] through direct
synthesis from the elements by reacting gaseous nitrogen with fine-grained silicon powder at
1400oC, and subsequently studied by Funk [4] and Whler [5]. Turkdogan et al. [6]
discovered that there exist two modifications and with hexagonal lattices. For a long time
this only incompletely known compound was just a scientific curiosity until in the early
1950s silicon nitride was utilized as a bonding phase for blast furnace refractory bricks [7].
Silicon nitride monolithic ceramics were researched and eventually technically applied in the
1960s based on their ability to maintain strength to very high temperatures. Indeed, up to
1400oC silicon nitride ceramics show excellent mechanical (low density, high bending
strength, high elastic modulus and fracture toughness, high abrasion wear and solid particle
erosion resistance), thermal (low coefficient of thermal expansion, high thermoshock
resistance), and chemical properties (stability against most acids and bases, many corrosive
gases, and liquid metals) [7-9].
Today technological applications of silicon nitride and silicon nitride-aluminum oxide
ceramic alloys (SiAlONs) include roller and ball bearings, coil and disc spring materials,
cutting tools for high speed machining of hard steels and Ni-based superalloys, heat
exchangers and heat pumps, inert gas welding and brazing fixtures and pins, stationary blades
and burner nozzles of gas turbines, dies for wire drawing, moulds for pressure casting of light
metals, ladles and tundishes for pouring molten metals, and sealing rings for horizontal
casting of steel tubes. Particular promising areas of application exist in parts of gasoline and
diesel engines including exhaust valves, valve spring retainers, bucket tappets, rocker arm
pads, pistons of internal reciprocating combustion engines, and, in particular turbocharger
rotors [10]. Turbochargers utilize exhaust waste heat to provide extra push whereas pistons
made from silicon nitride allow higher cylinder operating temperatures and thus to run the
engine thermodynamically and energetically more efficiently. However, major barriers to the
incorporation of these parts into power trains and engines of standard passenger cars relate to
the cost of processing, and the severe technological problem of mass-producing complex
ceramic components with a very high degree of reproducibility and long term reliability. The
problem of joining the ceramic to the metallic parts is particularly challenging. Hence
monolithic silicon nitride parts pose problems both during manufacturing and in service. For
example, experimental turbocharger rotors and pistons of internal combustion engines are still
127
affected by failure in service based on their inherently brittle nature. If such parts could be
fashioned from tough high-temperature resistant metals that provide superior mechanical
stability, and in turn could be protected against the attack of corrosive combustion gases by a
mechanically, thermally, and chemically compatible silicon nitride coating than longer
service life as well as increased combustion temperature and hence environmentally
beneficial engine performance would be expected. Consequently, in the past effort has been
expended to design such silicon nitride-based coatings for application under tough
environmental conditions [11-15]. Some of the research strategies involved will be discussed
below.
In contrast to future thick crystalline silicon nitride-based coatings thin amorphous films
produced by chemical vapour deposition (CVD) or plasma-enhanced chemical vapour
deposition (PECVD) are already ubiquitously used in semiconductor processing. Such films
serve as (i) masking layers for growth of electrically insulating silicon oxide layers on
semiconductor silicon wafers, (ii) chemical barriers to prevent cross-diffusion of doping
atoms such as boron, phosphorous, arsenic or gallium, (iii) for protection of doped wafers
from sodium ion contamination and mechanical damage, as (iv) dielectric material to
manufacture thin film capacitors, gate insulators for IGFETs, and components for gate
dielectrics in metal-insulator semiconductor (MIS) memory devices, as (v) UV- and blue-light
transparent passivation layers for complementary metal oxide-semiconductor (CMOS)
imaging sensors, as (vi) humidity barrier for organic light emitting diodes (OLEDs) as well as
(vii) for antireflection coatings for photovoltaic cells [7,9].
Despite an extraordinary amount of research performed on silicon nitride and silicon
nitride-aluminum oxide ceramic alloys (SiAlONs) over the last 50 years, much still remains
to be done. A plethora of research has been recorded in several monographs (for example [1619]) and handbooks (for example [20-23]).
In this chapter, the structure, synthesis and processing ways, and basic properties of
silicon nitride will first be reviewed to set the stage for a detailed description of research and
development strategies to deposit silicon nitride-based coatings by several high-kinetic
coating techniques. Such coatings are being designed for protecting engineered structures
from environmental damage by mechanical stress and abrasion, thermal destruction, and
chemical attack.
ADVANCED STRUCTURAL CERAMIC: A COMPARISON

The performance-application diagram shown in Figure 1 compares some mechanical and
thermal properties of the most widely utilized structural ceramic materials zirconia, silicon
carbide, silicon nitride and alumina [9]. It shows that zirconia ceramics excel in high
temperature applications where high density, high bending strength, high thermal expansion
and low thermal conductivity are required. Silicon carbide ceramics possess high thermal
conductivity and high microhardness but are deficient in strength and thermal shock
resistance. Hence the area of application of such ceramics is in heating elements for furnaces
operating up to 1600oC in ambient atmosphere. Silicon nitride on the other hand shows high
thermal shock resistance, sufficient strength and hardness, and intermediate values of density
and thermal conductivity suggesting applications in automotive engines and heat exchangers
128
Robert B. Heimann
as well as ceramic cutting tools for high-speed machining of hard metals. Lastly, alumina is
hard and shows a rather high thermal expansion but performs less superior in terms of
flexural strength and thermal shock resistance.
Figure 1. Performance-application diagram of advanced structural ceramics: zirconia, alumina, silicon

nitride, and silicon carbide [9].
Figure 2. Structures of hexagonal - and -silicon nitride. Black circles: silicon atoms, white circles:
nitrogen atoms.
129
STRUCTURE OF SILICON NITRIDE

Silicon nitride crystallizes in two low pressure hexagonal (- and -Si3N4) and a high
pressure cubic -modification with spinel structure (>15 GPa, >2000 K; [24]). The -phase
occurs in nature as a rare mineral (nierite) in ordinary and enstatite chondritic meteorites [25].
The crystal structures were determined by Hardie and Jack [26], and Turkdogan et al. [6].
The space group of low temperature -Si3N4 is P31c (154) with a = 774.8(1) pm and c =
561.7(1) pm. Structure refinement of -Si3N4 using synchrotron radiation powder diffraction
was more recently accomplished by Toraya [27], yielding a = 775.193(3) pm, c =
561.949(4) pm, Z = 4.
The space group of high temperature -Si3N4 is P63/m (173) with a = 760.8(1) pm and c
= 291.1(1) pm. The symmetry of high pressure -Si3N4 is
(227) with a = 773.8(1) pm.
The cubic -Si3N4 unit cell has a spinel-type structure in which two silicon atoms each
coordinate six nitrogen atoms octahedrally, and one silicon atom coordinates four nitrogen
atoms tetrahedrally.
As shown in Figure 2 the crystalline lattices of both - and -Si3N4 consist of slightly
distorted corner-sharing SiN4 tetrahedra that form hexagonal rings arranged in layers with
ABCDABCD stacking order (- Si3N4) or ABAB stacking order (-Si3N4). In -Si3N4
the bond lengths between Si and N1 are 173.0 and 173.9 pm, respectively, and between Si
and N2 174.5 pm. In -Si3N4 the bond lengths vary between 156.6 pm (N1-Si1) and 189.6 pm
(N1-Si2). While the AB layer is identical in both - and - phases, the CD layer in the phase is related to AB by a c-glide plane. Hence the double layer in -Si3N4 can be thought of
as a superposition of a -Si3N4 layer and its counterpart inverted by 180o. Consequently there
are twice as many atoms per unit cell in -Si3N4 (Z=4) than in -Si3N4 (Z=2). The Si3N4
tetrahedra in -Si3N4 are interconnected to form large open channels that extend parallel to
the c-axis of the unit cell. As a result of the c-glide plane the -structure contains two isolated
interstitial sites per unit cell instead of channels. The open channel structure of the -phase
supports easy diffusion of ions through the lattice thus greatly promoting sintering at high
temperatures.
SYNTHESIS AND PROCESSING OF SILICON NITRIDE

Powder Syntheses
Synthesis of silicon nitride powders can be achieved by reaction of various silicon
carriers such as elemental silicon, silicon dioxide, and gaseous silicon tetrachloride and
monosilane with nitrogen or ammonia. Routes to silicon nitride powders include the
following technologies [22]:
Direct nitridation of silicon

Carbothermal nitridation of silica including rice hulls
Diimide synthesis
Vapor phase synthesis
Plasma-chemical synthesis
130
Robert B. Heimann
Laser-induced reactions
Pyrolysis of organo-silicon compounds
The widely utilized diimide process involves reaction of silicon tetrachloride or organyl
aminosilane with ammonia and subsequent thermal decomposition of the silicon diimide
according to
3 SiCl4 + 18 NH3 3 Si(NH)2 + 12 NH4Cl
(1a)
3 Si(NH)2 Si3N4 + 2NH3
(1b)
Alternatively, high temperature ammonolysis of SiCl4 in the gaseous phase between 800
and 1100oC leads to largely amorphous silicon nitride that on subsequent annealing at 1500oC
crystallizes to form -Si3N4 [28].
Processing of Silicon Nitride Powders

Figure 3 shows a flow chart of the production routes of three silicon nitride
specifications: (i) porous reaction-bonded silicon nitride (RBSN), (ii) dense sintered silicon
nitride (SSN) and (iii) sintered reaction-bonded silicon nitride (SRBSN).
Processing of RBSN
Direct nitridation of fine-grained silicon powder compacted by cold isostatic pressing at
200 MPa leads to a mixture of - and -Si3N4 by prolonged heating in either nitrogen or
ammonia between 1000 and 1450oC.
Figure 3. Flow chart of silicon nitride processing.
131
The nitridation mechanism proceeds mainly through vapor phase reactions involving
silicon monoxide generated by reaction of silicon with contaminant oxygen occluded in the
powder according to
2 Si(s) + O2(g) 2 SiO(g)
(2a)
3 SiO(g) + 2 N2(g) Si3N4 + 3/2 O2 (g)
(2b)
The volume increase of about 22% will be completely accommodated by the interparticle
void space of the compacted silicon powder. Hence there is no shrinking during nitriding so
that the original dimensions of the green component will be retained. This is the underlying
cause for using RBSN for complex shaped components. Since expensive diamond grinding
and machining can be avoided, parts manufactured from RBSN are economically highly
competitive. Although the porosity of RBSN is between 20 and 30% its mechanical strength
is remarkable: flexural strengths in the range of 200 to 300 MPa can readily be attained. An
additional benefit is that this strength can be retained to about 1400oC so that designing with
RBSN leads to generally high Weibull moduli. Safe design of silicon nitride ceramics
subjected to tensile or flexural loads requires knowledge of the necessary engineering safety
factor. To properly estimate the safety factor, information on the statistical probability of
failure is necessary, using Weibull statistics [29]. The Weibull equation is expressed by
ln 1/ 1 Pf = ln V m ln f m ln 0
(3)
where Pf is the probability of failure of a part subject to the stress, V is the volume of the
sample under tensile stress, is the stress of zero probability of failure, usually taken as zero,
0 is a normalizing parameter of no physical significance, and m is a constant known as the
Weibull modulus. Equation (3) may be linearized by plotting ln ln (1/(1 Pf)) vs. ln f, thus
allowing the probability of failure at any given stress to be estimated by extrapolating a
straight line the slope of which is the Weibull modulus, m.
Processing of SSN
Silicon nitride powder can be economically pressureless sintered to near full density by
adding sintering aids such as magnesia, alumina and yttria, or a combination thereof. Critical
parameters are nano-sized powders in the range of 100 to 200 nm, the ability of the added
oxides to form a film of liquid oxynitride phase surrounding the silicon nitride grains, an
atmosphere of high pressure nitrogen to counteract dissociation, as well as sintering
temperatures between 1700 and 1850oC [7] (Figure 3). In the presence of alumina and yttria,
a significant portion of Al and O enter the Si3N4 lattice forming members of the SiAlON
family as - and -phases. On cooling the intergranular liquid solidifies as glassy or
crystalline (yttrium aluminum garnet, YAG) phase. In particular, the quinary Si-Al-N-O-Y
glass is generally the continuous phase that controls the high temperature properties of the
silicon nitride ceramics such as strength, fracture toughness, as well as creep and oxidation
behavior. The degree of in-situ reinforcement of the structure of pressureless sintered silicon
132
Robert B. Heimann
nitride by the glass-forming intergranular phase depends on the ionic radius of the REE
cation. There exists a (negative) linear dependence of the flexural strength on the radius of the
glass-forming REE ion such as = -18.2r + 2238 MPa at 25oC and = -20.9r + 2261 MPa
at 1200oC, when r is measured in picometers.
While a minimum amount of glass of low viscosity is required to achieve high volume
densities, large amounts of glass degrade the important high-temperature properties of the
material as the quinary glass soften beyond 1350 oC. Hence, an optimum must be sought
between high density that results in good low-temperature strength, and additives leading to
sufficient high-temperature strength. Ideally, the additive after sintering would enter into solid
solution with silicon nitride, but thus far only limited success has been achieved in finding
such alloys, for example for the families of - and -SiAlONs.
Table 1 shows some mechanical and thermophysical properties of reaction-bonded
silicon nitride (RBSN) and pressureless sintered silicon nitride (SSN).
Table 1. Selected properties of RBSN and SSN [3]
Property
Density (Mgm-3) (% of theoretical density)
Flexural strength (4-point, 25oC) (MPa)
Fracture toughness (25oC) (MPam)
Fracture energy (Jm-2)
Modulus of elasticity (25oC) (GPa)
Thermal conductivity (25oC) (W/mK)
Thermal shock resistance R (K)
Thermal shock fracture toughness R (Wm-1)
Coefficient of thermal expansion (10-6 K-1)
RBSN
70-88
150-350
1.5-3
4-10
120-220
4-30
220-580
500-10,000
3.2
SSN
95-100
500-1000
5-8
~60
300-330
15-50
300-780
7,000-32,000
3.2
Processing of Fully Dense Ceramic Bodies

To achieve fully dense silicon nitride monolithic bodies, hot pressing or hot isostatic
pressing of silicon nitride powder with added metal oxides is being employed. At
temperatures above 1550oC and in an inert atmosphere these additives form with contaminant
silicon dioxide films around individual silicon nitride grains a liquid silicon oxynitride phase
in which the silicon nitride readily dissolves. While these hot pressed (HP) and hot
isostatically pressed (HIP) techniques are very involved and restricted to simple geometrical
shapes, pressureless sintering is now most frequently utilized.
SILICON NITRIDE-BASED COATINGS

Deposition of silicon nitride coatings by conventional thermal spraying technologies has
proven highly challenging since the material dissociates above 1800 C and oxidizes in air at
elevated temperature even though attempts have been made to reduce oxidation by applying
protective coatings [30]. Experiments performed to deposit mechanically and chemically
stable silicon nitride coatings by conventional thermal spraying using metallic [31-33] or
133
silicate glass binders [34-36], in-flight nitridation [37] or reactive spray processing [38,39]
was met with limited success since the coatings contained only small amounts of silicon
nitride but large amounts of unwanted embrittling silicide phases instead. More successful
was work to deposit Si3N4-based coatings starting from SiAlON powders [40] or using cladtype powder consolidation in conjunction with NiAl or NiCr bond coats [41].
Work performed by the author's research group between 1996 and 2004 confirmed that
the combination of optimized heat transfer into the powder particles to control the critical
viscosity of the oxide binder phase and high particle velocities generated by detonation
spraying (DS), Top Gun spraying or other high kinetic spraying technologies were the most
decisive factors to deposit thick, dense, homogeneous silicon nitride coatings with desirable
mechanical properties. Special powder preparation [42] and application of extremely high
powder particle velocities up to 3 km/s, generated by electromagnetically accelerated plasmas
[43] yielded coatings that adhere well even to polished stainless steel surfaces (see below).
In this contribution deposition of silicon nitride-based coatings by high kinetic energy
spraying technologies will be described and salient properties of the resulting coatings
highlighted.
POWDER PREPARATION AND PROPERTIES

Several types of silicon nitride-based powder (Table 2) were synthesized as a precursor
material to deposit dense, well-adhering coatings on mild steel and Inconel 600 substrates by
low-frequency pulsed detonation spraying (LFPDS) as well as on stainless steel substrates by
high-current pulsed plasma spraying (HCPPS). Figure 4 shows a close-up of the highly
porous powder particles produced by mixing commercially obtained -Si3N4 powder (Silzot
7038, SKW Trostberg, Germany; [48]) with common sintering additives such as alumina and
yttria. For some powder compositions AlN was added to promote formation of -SiAlON,
and SiO2 or MgO to control the viscosity of the binder matrix (Figure 5). The powders were
agglomerated by spray drying using an organic binder, and subsequently sintered at 1450 C
in nitrogen. The sintered powders were subjected to a mild milling process, and fractionized
by dry sieving into three size classes, large (63+45 m), medium (45+32 m) and fine
(32+22 m). The compositions of the five types of powder are shown in Table 2.
Table 2. Composition of silicon nitride-based feedstock powders (mass%)
Powder #
1
2
3
4
5
-Si3N4
68
60
61
64
64
Al2O3
16
16
16
16
20
Y2O3
16
16
16
16
16
AlN
MgO
4
4
4
SiO2
4
The sintering of mixtures of -Si3N4 and binder oxides initiates a series of reactions
characterized by partial phase transformation of -Si3N4 to high temperature -Si3N4 as well
as formation of -SiAlONs and crystalline Y3Al5O12 (YAG) (see eqs. 4 and 5). YAG forms
134
Robert B. Heimann
with silicon nitride a eutectic at about 1450 C that during the spray process promotes
generation of a liquid phase that solidifies as a quinary glass. The fact that residual -Si3N4
exists in the as-sintered powder instead of the thermodynamically stable -Si3N4 indicates
that the process and thus the phase composition are far from equilibrium. The coexistence of
these phases can be depicted in a hexanary phase diagram, the so-called Jnecke prism
(Figure 6) that describes the compositional complexity of the SiAlONs as ceramic alloys of
silicon nitride and aluminum oxide.
The reactions between silicon nitride, and alumina and yttria can be schematically
expressed by a two-step process. The first step involves alloying of silicon nitride with the
binder oxide additives resulting in formation of -SiAlON and a quinary glass according to
Si3N4+Al2O3+Y2O3+AlN+SiO2 Si6-zAlzOzN8-z + glass I
(4)
The second step describes the consecutive reactions of -SiAlON (for example, z = 1)
with glass I to form a second -SiAlON with a different z-value, for example z = 2,
crystalline YAG and a glass II with reduced Al and O contents according to
Si5AlON7 + glass I Si5-zAl1+zO1+zN7-z + Y3Al5O12+ glass II
(5)
Figure 4. Porous, raspberry-shaped spray powder (powder #1) used to deposit Si3N4-based coatings.
Figure 5. Agglomerated composite 61-Si3N4/16Al2O3/16Y2O3/4AlN/3MgO spray powder #3 showing

surface porosity owing to impeded sinterability.
135
Figure 6. Jnecke prism (hexanary system Si3N4-Al2O3-SiO2-AlN-YN-Y2O3) to visualize major

crystalline phases (empty circles) and minor accessory phases (full circles). On the light shaded plane
Si2N2O-Y4O6-Al4O6 several crystalline solid solutions are located in which oxygen is partially replaced
by nitrogen such as N-apatite, N-wollastonite, and N-melilite (not shown). The darker shaded plane
Si3N4-Y3Al5O12-Al4O6 depicts the compositional range of the -SiAlONs formed during sintering.
In this way a stepwise increase of the degree of replacement of respectively, Si and N by

respectively, Al and O is achieved as shown in Figs. 7 and 16. The -SiAlONs with z = 1, 2
or 3 in spray powders #1 and 2 (Figure 16A, B) transform by reaction with the quinary glass
to -SiAlONs with z = 3 and 4 in the coatings (Figure 16C, D).
LOW-FREQUENCY PULSED DETONATION SPRAYING (LFPDS)

Principle of Technique
This deposition technique involves cyclic detonation of a mixture of high-enthalpy
combustible gasses such as propan, acetylene, propylene and others with oxygen in a watercooled tube (Figure 8).
The tube will be filled with the combustible gas and oxygen through a well-controlled
metering system, and the spray powder is axially injected with a stream of nitrogen from the
back of the tube. After igniting the gas mixture with an electric spark delivered by a spark
plug the advancing flame front converts to a detonation wave with supersonic velocity, owing
to the geometry of the tube.
The powder particles will be accelerated to a velocity in excess of 1000 m/s, comparable
to that of a high velocity oxyfuel (HVOF) spraying system, and melted at the high
combustion temperature that in the case of acetylene/oxygen can reach 3900oC.
136
Robert B. Heimann
Figure 7. Relative amounts of silicon nitride, YAG and -SiAlONs in powders and coating (see Figure
16).
Figure 8. Schematic of a detonation spraying device.
After detonation the tube will be flushed with a stream of pressurized air or nitrogen.
With the equipment used in the present study (Perun P, Tampere University of Technology,
Tampere, Finland) this cycle is repeated about 7 times per second.
The following parameters applied: length of the tube 660 mm, width 20 mm, gas flow
rate 10 m3/h, power 100-300 kW.
COATING PROPERTIES
Composition and Porosity
Figure 9 shows an x-ray diffraction pattern of a silicon nitride coating deposited by lowfrequency pulsed detonation spraying of powder #1 with a grain size distribution 45+32 m.
The starting -Si3N4 has been partially transformed to thermodynamically stable -Si3N4.
The elevated background between 22 and 38 2 (centroid: 30) points to the presence of
an amorphous phase attributed to a quinary glass SiAlYNO in which the crystalline
YAG detected in the precursor powder (see Figure 16A) has completely dissolved.
137
The coating cross-sections (Figure 10) show some porosity that was determined by
automated image analysis to be in the range of < 10 vol%. Application of fine powder
particles (32+22 m) yielded coatings with porosities < 2 vol%.
Wear Behavior
Coating produced from powders #2, 3, and 4 with different grain size ranges on Inconel
600 superalloy substrates with an intermediate NiCrAlY bond coat were subjected to a wear
test according to ASTM G65-04 [44]. Table 3 shows the critical wear path lengths (wear path
to destruction of coating; load 130 N) of selected coatings. Wear path lengths of silicon
nitride coatings deposited by HVOF (Top Gun) on low carbon steel substrates without
bond coat are given for comparison (loads: 130 and 45 N). The microhardness of the coatings
measured by indentation with a Vickers diamond loaded to 0.05 N is rather low and varied
between 40 and 60 MPa owing to the presence of a quinary glass in which the hard silicon
nitride particles are suspended.
Thermoshock Resistance
Qualitative values of the thermoshock resistance of silicon nitride coatings with
comparable thickness (samples 1, 3 and 5 of Table 3) were obtained by heating the samples to
800C and subsequent quenching either to about 150C in air (gradient: 8 K/s) or to room
temperature in water (gradient: 700 K/s). The coatings of samples 1 and 3 withstood without
visible cracking or delamination > 10 cycles in air and > 6 cycles in water whereas coatings
of powder composition 4 failed the test. Also, coatings without a NiCrAlY bond coat caused
immediate spalling due to the large stress gradient at the interface.
Figure 9. Phase composition of LFPD-sprayed silicon nitride coating (powder #1).
138
Robert B. Heimann
Figure 10. Silicon nitride coatings deposited on mild steel substrates using powder #1with different
grain size ranges.
Figure 11. Results of salt-fog test to determine the corrosion resistance of mild steel (St 37-2) coated
with silicon nitride. For details see text.
139
Table 3. Critical wear path length in km of silicon nitride coatings deposited by LFPDS
onto Inconel 600 substrates with a NiCrAlY bond coat (samples 15) and by HVOF
(Top Gun) spraying on low carbon steel (St 37-2; S235JR+AR) (samples 67)
Sample
#
Substrate
Powder
type
Grain
size (m)
Thickness
(m)
1
2
3
4
5
6
7
IN 600
IN 600
IN 600
IN 600
IN 600
St 37-2
St 37-2
2
2
3
4
4
3
3
-32+22
-45+32
-32+22
-32+22
-45+32
-45+32
-45+32
77
200
75
150
75
150
150
Wear
path
length
[km]
0.1
0.1
1.0
3.0
1.0
0.5
4.3
Load
(N)
Method
130
130
130
130
130
130
45
LFPDS
LFPDS
LFPDS
LFPDS
LFPDS
HVOF
HVOF
Corrosion Resistance
A salt-fog test according to ASTM B117-73 [45] was performed by treating two
representative silicon nitride coatings, deposited onto low carbon steel (St 37-2) substrates, at
35C with a solution of 50 g/l NaOH for 23 h and subsequent storage in a moist chamber at
room temperature for 35 h. The backside and edges of the steel coupon were protected by
epoxy resin. Low carbon steel was used to easily visualize rust-colored corrosion products
formed underneath the protective coatings. Coatings produced from powder 2 (45+32 m;
200 m thickness) showed only marginal corrosion of the steel substrate (Figure 11 A,B;
corrosion area: 2%) whereas coatings produced from powder 5 (32+22 m; 60 m
thickness) revealed substantial corrosion (Figure 11 C; corrosion area: 55%) accompanied by
cohesive cracking and delamination (Figure 11 D).
HIGH-CURRENT PULSED PLASMA SPRAYING (HCPPS)

Principle of Technique
This novel deposition technique and the design and working mode of related equipment
have been described in detail elsewhere [46, 43].
The sequence of events is shown schematically in Figure 12. The apparatus consists of a
coaxial accelerating channel evacuated to p < 10-2 Pa that comprises a pair of parallel
electrodes connected to a high current power source by a switch, and a powder container
pressurized by argon gas.
By activating the fast-opening valve of the powder container the fine powder (maximum
grain size < 10 m) is injected by the working gas through a nozzle into the evacuated coaxial
accelerating channel of the apparatus (Figure 12A) and subsequently subjected to a very high
current pulse (pulse length: 300 s) of about 100 kA. The supersonic argon plasma created
with a peak pressure in excess of 1 MPa is being accelerated by a shock wave that heats and
140
Robert B. Heimann
propels the suspended fine powder particles in the wake of the plasma shock wave front
towards the substrate surface to be coated (Figure 12B,C). This technology is akin to the
working of a railgun in which a single projectile is replaced by a multitude of fine powder
particles (Figure 13).
Figure 12. Working principle of high-current pulsed plasma spraying (HCPPS). A. Fine powder
particles will be injected by pressurized argon gas into the evacuated plasma channel. B. A high current
pulse creates argon plasma that travels with high velocity and thus accelerates the powder particles with
supersonic velocity towards a substrate to be coated. C. Close-up of the traveling plasma. Behind the
shock front uniform gas density, viscosity, and velocity prevail.
141
Figure 13. Principle of a projectile launcher (railgun) using electromagnetically accelerated plasma.
The powder particles accelerated by the plasma shock wave are heated adiabatically.
Since the PV adiabat (Hugoniot equation of state) of the porous powder particles always lies
above of that of dense material both the thermal energy, Ut and the thermal pressure
component, Pt increase as a result of the collapse of the pore spaces during compression
(Figure 14). Moreover, the propagating shock wave collides many times with crystalline
components within the porous grains thus creating strong local pressures and, in turn,
additional thermodynamic heating and also crushing of the particles [47].
The overall thermodynamic pattern of propagation of the shock wave through the Si3N4
powder is highly non-uniform and creates micropockets of high pressure and temperature
(hot spots [48]) in the voids between the grains adjacent to regions of (almost) ambient
pressure and temperatures.
Figure 14. P-V shock adiabats (Hugoniot EOS) for solid (PH(V0) and porous (PH(V00) materials. Pc cold
pressure, Pt thermal pressure, Uc cold (elastic) internal energy, Ut internal energy due to thermal
pressure, Ut excess internal energy of porous material, Pt excess thermal pressure due to the porous
nature of the shocked material [47].
142
Robert B. Heimann
Figure 15. Observed (full symbols) and calculated (empty symbols) particle velocities as functions of
the product of particle density and particle diameter for B4C, Al2O3, Cu (circles) [49] and Si3N4
(hexagons) [43].
Figure 16. Phase compositions of powders #1 and #2 (A,B) and coatings deposited by HCPPS using
these powders as feedstock (C,D).
143
Figure 17. Sintered agglomerated powder prepared from 57% -Si3N4 + 16% Al2O3 + 23 % didym + 4
% AlN. Note the formation of elongated -Si3N4 crystals [50].
As shown in Figure 15 for several ceramic coating materials the particle velocities
attained depend parabolically on the product of the specific density and the particle diameter
of the injected powders.
COATING COMPOSITIONS
The phase compositions of typical spray powders as well as the coatings produced
therewith are shown in the x-ray diffraction patterns of Figure 16.
The as-sintered powder consists of -Si3N4, noticeable amounts of -Si3N4, -SiAlON
with z = 1-3, and yttrium aluminum garnet (YAG). The -Si3N4 forms typical elongated
needle-like crystals in a matrix of -Si3N4 + amorphous phase. The development of -Si3N4 is
particularly accelerated when the viscosity of the binder oxide phase is low as observed
during sintering experiments in which the Y2O3 was replaced by a neodymium/praseodymium
mixed oxide (didym) (Figure 17).
The amount of YAG differs strongly in the two powder types #1 and 2. In the coating, the
YAG has almost completely dissolved into the quinary glass the existence of which is
indicated by the elevated background of the XRD pattern.
COATING PROPERTIES
Silicon nitride coatings obtained from powders # 1 and 2 (Table 2) are almost completely
dense (Figure 18) and adhere well to polished stainless steel surfaces. The average adhesive
strength of the coatings measured according to ISO 14916: 1999 designation [51] exceeds 75
MPa, the Vickers indentation hardness was determined to be 4612 MPa (powder #1) and
5512 MPa (powder #2), respectively. Figure 18 shows a cross-section of a typical coating
144
Robert B. Heimann
obtained as the result of spraying powder #1 onto a mirror-polished stainless steel (SUS304;
DIN X5 CrNi 18-10) substrate of dimensions 15151 mm3. The interface between the
polished steel substrate and the silicon nitride coating is extremely smooth and homogeneous
attesting to the excellent adhesion of the coating by a process akin to explosive cladding [52]
caused by the high-velocity impact of solid or semi-molten particles accelerated by a plasma
shock wave.
Experiments to determine the friction coefficient of coatings by a pin-on-disk test [53]
yielded surprisingly large values of 1.0 in water and about 1.4 in air [43] that suggest
applications as tribological surfaces for automotive brake pads, grip-and-release surface for
printing rolls or high friction coatings to damp vibrations of precision micro-positioning
equipment.
Figure 18. Cross-section of a silicon nitride-based coating of composition shown in Figure 16C
deposited by HCPPS onto the surface of a mirror-polished stainless steel (SUS304; DIN X5 CrNi 1810) substrate [15].
In conclusion, while at a first glance the HCPPS technology appears to be the method of
choice to deposit almost theoretically dense silicon nitride coatings with high adhesive
strength, reasonable hardness and wear resistance, and high friction coefficient the process is
still far from any economic reality as the deposition technology is of batch type. This involves
time consuming replacement of both the powder-containing vessel and the gas feeding
connector after each single shot. Since many shots are required to build up coatings
sufficiently thick for industrial applications, much scale-up and optimization work is required
to render this promising technology economically viable.
145
REFERENCES
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
Lauretta, D.S., Elements, 2011, 7, 11-16.

Nguyen, A.N., Messenger, S., Elements, 2011, 7, 17-22.
Weiss, L., Engelhardt, T., Z. anorg. Allg. Chem., 1910, 65, 38-104.
Funk, H., Z. anorg. Allg. Chem., 1924, 133, 67.
Whler, L., Z. Elektrochem. Angew. Phys. Chem., 1926, 32(9), 420-423.
Turkdogan, E.T., Bills, P.M., Tippett, V.A., J. Appl. Chem., 1958, 8(5), 296-302.
Riley, F.L., Application of silicon nitride ceramics. In: Advanced Ceramic Materials;
Mostaghaci, H.; Ed.; Key Eng. Mater.; Trans Tech Publications Ltd: Uetikon-Zuerich,
Switzerland, 1996; 122/124, pp 479-88.
Michalowski, L., Neue Keramische Materialien, Deutscher Verlag fr
Grundstoffindustrie: Leipzig, Stuttgart, Germany, 1994.
Heimann, R.B., Classic and Advanced Ceramics. From Fundamentals to Applications;
Wiley-VCH: Weinheim, Germany, 2010; pp 457-468.
Mrgenthaler, K.D., Bhl, H. Ceramics for Engines. In: Tailoring of Mechanical
Properties of Ceramics, Hoffmann, M.J., Petzow, G.; Eds.; NATO ASI Series E.
Applied Sciences, Kluwer: Dordrecht, The Netherlands, 1994; pp. 429-441.
Heimann, R.B., Thiele, S., Wielage, B., Zschunke, M., Herrmann, M., Berger, L.M.
Entwicklung von Schichten auf Siliziumnitrid-Basis als HT-erosionsbestndiger
Werkstoff fr thermische Anlagen. In: Freiberger Forschungshefte (Maschinenbau),
A848; TU Bergakademie Freiberg: Freiberg, Germany, 1997; pp 166-182.
Heimann, R.B., Thiele, S., Berger, L.-M., Herrmann, M., Nebelung, M., Wielage, B.,
Schnick, T.M., Vuoristo, P. Thermally sprayed silicon nitride-based coatings on steel
for application in severe operation environments: preliminary results. In:
Microstructural Science: Analysis of In-Service Failures and Advances in
Microstructural Characterization, 26; Abramovici, E., Northwood, D.O., Shehata,
M.T., Wylie, J.; Eds.; ASM International: Materials Park, OH, 1998; pp 389-94.
Heimann, R.B., Thiele, S., Berger, L.-M., Herrmann, M., Vuoristo, P. Detonation
spraying of silicon nitride for high temperature erosion-resistant ceramic coatings. In:
Proc. Coatings for Aerospace and Automotive Industries (CAAI-6); Kleiman, J.I.; Ed.;
Toronto, Ontario, Canada, 1999; pp 14-15.
Heimann, R.B. Plasma Spray Coating. Principles and Applications. 2nd ed., WileyVCH: Weinheim, Germany, 2008; pp 267-270.
Heimann, R.B. Surf. Coat. Technol., 2010,205, 943-948.
Belyi, V.I., Vasilyeva, L.L., Ginovker, A.S., Gritsenko, V.A., Repinsky, S.M., Sinitsa,
S.P., Smirnova, T.P., Edelman, F.L. Silicon Nitride in Electronics, Materials Science
Monographs, Elsevier: Amsterdam, The Netherlands, 1988, vol. 34.
Hoffmann, M.J., Becher, P.F., Petzow, G. Silicon Nitride 93. Key Eng. Mater.; Trans
Tech Publications Ltd: Uetikon-Zuerich, Switzerland, 1994; 89/9.
Sangster, R.C. Formation of Silicon Nitride: From the 19th to the 21st Century. In:
Materials Science Foundations (Monograph Series), Trans. Tech. Publications Ltd,
Uetikon-Zuerich, Switzerland, 2005; 22/24, pp 948-952.
Bocanegra-Bernal, M.H., Matovic, B., Mater. Sci. Eng. A, 2009, 527(6), 1314-1338.
146
Robert B. Heimann
[20] Gmelin Handbook of Inorganic and Organometallic Chemistry. Silicon Supplement

Volume B5e, Non-Electronic Applications of Silicon Nitride; Kurtz, W., Schrder, F.;
Eds.; Springer: Berlin, Germany, 1994; pp 429-441.
[21] Gmelin Handbook of Inorganic and Organometallic Chemistry. Silicon Supplement
Volume B5c, Silicon Nitride in Microelectronics and Solar Cells; Pebler, A., Schrder,
F.: Eds.; Springer: Berlin, Germany, 1991.
[22] Petzow, G., Herrmann, M. Silicon nitride ceramics. In: High Performance Non-Oxide
Ceramics II. Structure and Bonding. Jansen, M., Mingos, D.M.P., Haubner, R.; Eds.:
Springer: Berlin, Germany, 2002, 102, pp 47-168.
[23] Cannon, W.R., Gugel, E., Leimer, G., Woetting, G., Heimann, R.B. Ceramics,
Advanced Structural Products, In: Ullmanns Encyclopedia of Industrial Chemistry;
Wiley-VCH: Weinheim, Germany, 2011, chapter 4.4.
[24] Leitch, S., Moewes, A., Ouyang, L., Ching, W.Y., Sekine, T., J. Phys. Condens.
Matter, 2004, 16, 6469-6476.
[25] Lee, M.R., Russell, S.S., Arden, J.W., Pillinger, C.T., Meteoritics, 1995, 30, 387-398.
[26] Hardie, D., Jack, K.H., Nature, 1957, 180, 332-333.
[27] Toraya, H., J. Appl. Cryst., 2000, 33, 95-102.
[28] Ptzold, C. Herstellung, Charakterisierung und neue Einsatzgebiete von keramischen
Nitrid- und Oxinitridpulvern (Production, characterization and novel applications of
ceramic nitride and oxynitride powders), Unpublished Ph.D. thesis, Technische
Universitt Bergakademie Freiberg: Freiberg, Germany, 2001,161 pp.
[29] Weibull, W., Proc. Ing. Vetensk. Akad. (Stockholm), 1939, 151, 5-45.
[30] Moldovan, M., Weyant, C.M., Johnson, D.L., Faber, K.T., J. Therm. Spray Technol.,
2004, 13(1), 51-56.
[31] Lugscheider, E., Limbach, R., DVS Berichte, 130, Deutscher Verlag fr
Schweisstechnik, Dsseldorf, 1990, p. 224, (in German).
[32] Lugscheider, E., Limbach, R., Liden, A., Lodin, J. In: Proc. Conf. High Temp. Mater.
Powder Eng., Liege, Belgium. Bachelet, F.; Ed.; Kluwer: Dordrecht, The Netherlands,
1990, p. 877.
[33] Cheney, R.F., Plasma Spray Powder, US Patent 4, 508, 788 (April 2, 1985).
[34] Kucuk, A., Lima, R.S., Berndt, C.C., J. Mater. Eng. Perform.; 2000, 9, 603-608.
[35] Liu, J., Zhang, T., Bao, Y., Gawne, D.T. In: Advances Applied Ceramics, Edirisinghe,
M.J.; Ed.; Maney Publishing: UK, 2010, 109, pp 71-78.
[36] Gawne, D.T., Bao, Y., Zhang, T. In: Proc. Intern. Thermal Spray Conf., Hofe, D.,
Lugscheider,E.; Eds.; DVS Berichte, Deutscher Verlag fr Schweisstechnik, Essen,
Germany, 2002, p. 225.
[37] Eckardt, T., Mallner, W., Stver, D. In: Thermal Spray Industrial Applications,
Berndt, C.C., Sampath, S.; Eds.; ASM International: Materials Park, OH, 1994, p. 515.
[38] Eckardt, T., Mallner, W., Stver, D. In: Proc. Thermische Spritzkonferenz: TS96, E.
Lugscheider, E.; Eds.; DVS Berichte, 175, Deutscher Verlag fr Schweisstechnik:
Dsseldorf, 1996, pp 309312 (in German).
[39] Yamada, M. , Fukumoto, M., Yasui, T., J. Vacuum Soc. Japan, 2003, 46 (7) 563-566;
Mater. Trans. 2004, 45 (12) 3304-3328.
[40] Sodeoka, S., Ueno, K., Hagiwara, Y., Kose, S., J. Therm. Spray Technol. 1992, 1, 153159.
147
[41] Tomota, T. Miyamoto, N., Koyama, H., Patent JP 63 169371, A. Int. Cl. 4: C23C 4/10,
Filing date 29.12.1986 (in Japanese).
[42] Berger, L.-M., Herrmann, M., Nebelung, M., Heimann, R.B., Wielage, B., U.S. Patent
6,110,853 (August 23, 2000).
[43] Usuba, S., Heimann, R.B., J. Therm. Spray Technol. 2006, 15 (3), 356-363.
[44] ASTM G65-04, Standard Test Method for Measuring Abrasion Using the Dry
Sand/Rubber Wheel Apparatus, American Society for Testing and Materials, 2004.
[45] ASTM B117-73, Standard Method of Salt Spray (Fog) Testing, American Society for
Testing and Materials, 1973.
[46] Kitamura, J., Usuba, S.. Kakudate, Y., Yokoi, H., Yamamoto, K., Tanaka, A.,
Fujiwara, S., J. Therm. Spray Technol. 2003, 12 (1) 70-76.
[47] Heimann, R.B., Kleiman, J. In: Crystals. Growth, Properties and Applications,
Freyhardt, H.C.; Ed.; Vol. 11, Springer: Berlin, 1988, pp 1-73.
[48] Kleiman, J., Heimann, R.B., Hawken, D., Salansky, N.M., J. Appl. Phys., 1984, 56(5),
1440-1454.
[49] Usuba, S., Kitamura, J., Kakudate, Y. In: S. Kuroda (Ed.), Proc.1st Tsukuba Intern.
(JapaneseGerman) Coating Symp., Recent Trends in Coating Technology to
Challenge Harsh Environments and to Open New Applications, Kuroda, S.; Ed.; May
17, National Institute for Materials Science, Japan, 2004, p. 20.
[50] Heimann, R.B., Thiele, S., Herrmann, M., Berger, L.M., Nebelung, M., Wielage, B.,
Zschunke, M., Schnick, T., Final report Investigation on thermal spraying of powders
based on silicon nitride, DFG projects He 923/4, He 2457/1, Wi 688/15, October 1999
(in German).
[51] ISO 14916, Determination of Tensile Adhesive Strength, 1999.
[52] Carpenter, S.H., Wittman, R.H., Ann. Rev. Mater. Sci., 1975, 5, 177-199.
[53] ASTM G99-05, Standard Test Method for Wear Testing with a Pin-on-Disk Apparatus,
American Society for Testing and Materials, 2005.
In: Silicon Nitride

ISBN 978-1-61942-865-2
Chapter 6
SILICON NITRIDE AS BIOMATERIAL

Instituto de Pesquisas Energticas e Nucleares IPEN/CNEN, Brazil
ABSTRACT
Among the ceramic materials investigated for biomedical purposes, those based on
silicon nitride have attracted attention due to high mechanical strength, relatively high
fracture toughness and low density. This review reports the main works about the use of
silicon nitride as biomaterial. Dense and porous bodies as well as particles of silicon
nitride are considered in various papers whose procedure of biological evaluation ranged
from bioactivity tests in Simulated Body Fluid (SBF) to in vitro experiments using
osteoblasts like-cells and in vivo tests using rabbits as animal model. Although the first
study related to the biocompatibility of silicon nitride published in the 80's has shown
poor results, a large number of later studies indicated that this material might exhibit
good biological performance with great potential for high loaded prosthetic applications.
INTRODUCTION
Ceramics formed predominantly by covalent bonds are very appreciated owing to the
high hardness, refractoriness and high fracture toughness. An example of these ceramics is
boron nitride, which is extremely hard and useful as abrasive when is in cubic form; silicon
carbide, largely used as abrasive, refractory, wear resistance structural ceramic and as
semiconductor at high temperatures; and silicon nitride which is used as machine
components, mainly when high wear resistance in aggressive conditions is required [1].
However, the covalent character and high vapor pressure at high temperatures difficult
the densification of pure silicon nitride by solid state diffusion mechanisms, what requires the
use of aids to promote liquid phase sintering. Many different techniques might be used to
sinter Si3N4, such as reactive sintering, hot pressing, hot isostatic pressing, gas pressure
sintering and conventional sintering. By technological view point, the last three techniques
are the most interesting, but the first three can produce components with better properties than
conventional sintering by minimizing the amount of additives [2].
150
Initially, silicon nitride was used in metallic baths and subsequently in coatings of
thermocouples and crucibles for melting metals. However, interest in this material proceeded
to grow primarily for use in gas turbine components, which offer better efficiency in burning
fuel, lower emissions of gases and low levels of particulates, compared with the conventional
spark-ignition engines [3]. Moreover, features such as low density, low inertia and high wear
resistance allowed the use of silicon nitride in conventional internal combustion and adiabatic
diesel engines [4].
Nowadays, silicon nitride has been investigated to be used in clinical applications mainly
in situations where bone replacement is required. It is because those ceramic that are
generally used in orthopedic and dental prostheses, i.e. alumina and zirconia, have presented
some limitations due to brittleness or aging process with consequent phase transformation
into the body [5].
Many different components and prosthetic devices can be improved in function of silicon
nitride properties, either in maxillofacial surgery, plates and screws or even in applications
proved for metallic implants, but that cause interaction with surrounding tissues. Besides,
these ceramics have potential for use as intervertebral spacers and spinal surgery, in view of
their shear strength and implant/bone fusion [6].
SILICON NITRIDE STRUCTURE

Silicon nitride presents two crystalline forms: and . Both structures can be described
as a stacking of Si-N layers in ABCDABCD sequence for -Si3N4 or ABAB for Si3N4. The AB layer is the same both in and -Si3N4, but the CD layer in -Si3N4 is related
to AB layer by c slip plane [7,8].
The idealized structure of -Si3N4 has a trigonal symmetry with space group P31c and
with each unit cell containing four units of Si3N4, c/a = 2/3 and density = 3.168 g/cm3. The Si3N4 structure has a hexagonal symmetry with space group P63/m and each unit cell contains
two units of Si3N4 with c/a = 1/3 and density = 3.192 g/cm3. The and -Si3N4 structures are
geometrically related for a = a = 2c [7,8].
The -Si3N4 phase is formed in the temperature range between 1500 and 1800oC and is
more favorable energetically than phase, formed at lesser temperatures than 1500oC.
Therefore, the transition is irreversible [9].
Differences among unit cell dimensions and density related to and phases led
researchers to investigate the phases compositions with details. Grieveson et al. [10]
proposed that the differences are associated with oxygen contents in -Si3N4 structure which
would be an oxynitride with composition between Si11.4N15O0.3 and Si11.5N15O0.5 and with
oxygen content between 0.90 and 1.48 % wt. According to this model, electrical neutrality is
maintained by replacement of some nitrogen atoms by oxygen, or by the creation of silicon
vacancies [8].
Other studies showed that, although the oxygen can actually dissolve in the -Si3N4
lattice, most appear as a superficial oxide. However, Galasso et al. [11] and Niihara et al. [12]
observed that there is a very close relationship between the measured oxygen content and the
structural parameters of the crystal, since the c axis dimension and unit cell volume increases
151
with decreasing oxygen content, while the size of the b axis is almost independent of the
oxygen content.
Later, it was concluded that even if oxygen is not a necessary requirement for the stability
of the -Si3N4 structure, this phase is able to accommodate large amounts of this element
depending on its availability in the processing environment. Density and dimensions
variations of the unit cell in -Si3N4 are assigned to oxygen amount in its structure, which is
associated with different routes of powder preparation [8].
OBTAINING SILICON NITRIDE COMPONENTES

Due to the high degree of covalent bonds of its constituent elements and its high vapor
pressure at high temperatures, it is very difficult to sinter Si3N4 by state solid diffusion. As
consequence, many different additives are usually used in varied amounts depending on
fabrication technique, which includes: Reaction Bonded Silicon Nitride (RBSN), Hot-Pressed
Silicon Nitride (HPSN), Hot Isostatic Pressed Silicon Nitride (HIPSN), Gas Pressure
Sintering Silicon Nitride (GPSSN) and Normal Sintering Silicon Nitride (SSN) [2,4].
The sintering aids react with the silica on -Si3N4 surface, forming a liquid phase which
promotes the sintering in three stages [7]. The first stage is characterized by the -Si3N4
particles rearrangement, while the second one involves the material dissolution from phase,
followed by silicon and nitrogen diffusion and subsequent phase precipitation. In the third
stage, called coalescence, the grains grew and the residual closed porosity is eliminated.
The final microstructure is usually characterized by elongated grains of -Si3N4 with a
secondary intergranular phase in their grain boundary [4].
The kind and amount of additives have a major influence on the final material properties
because, after cooling, the liquid phase remains at the grain boundaries as an amorphous or
crystalline phase. In addition, the characteristic of the liquid phase in the sintering
temperature, such as quantity and viscosity, influence the densification of the material as well
as the morphology of -Si3N4 grains [13].
Reaction Bonded Silicon Nitride (RBSN) materials did not have good mechanical
properties because they contain 15-20 vol. % of porosity. In order to improve the mechanical
properties of RBSN ceramics, a technique of post sintering after nitridation has been
developed [14].
Different compositions of Hot-Pressed Silicon Nitride (HPSN) can be produced with a
minimum of densification additives what promotes tougher silicon nitride ceramics. In the
Hot Pressing, a mixture of powders formed by silicon nitride and additives is introduce into a
graphite matrix and pressed at temperatures between 1700 and 1800oC. However, this
technique presents limitations in regard to the part geometry will be produced, and generate
products with anisotropic properties due to preferred orientation of grains, perpendicular to
the pressing direction [14].
Hot Isostatic Pressing and Gas Pressure Sintering are very similar techniques considering
processing, but the sintering mechanisms are completely different. In GPS, the greatest factor
to densification is the high temperature with the pressure varying from 0.2 to 10 MPa, applied
only to supply the silicon nitride decomposition. Whereas in HIP, the high pressure is the big
152
factor that contributes for densification. In this case, pre-sintering or encapsulation is

required, but the sintering aids might be eliminated [15,16].
The Normal Sintering is the most economical method to obtain silicon nitride in complex
shapes. The sintering temperature varies in the range between 1700 and 1800oC, but higher
additive contents should be used. Moreover, the dissociation of silicon nitride becomes a
problem at temperatures greater than 1700oC what difficult the densification due to weight
loss. The use of so-called "powder beds", in which the specimens to be sintered is surrounded
by a mixture of powders of same composition, has been shown effective in reducing the
decomposition of Si3N4 [4].
TRIBOLOGICAL BEHAVIOR
The use of bioceramics in some applications such as artificial joints requires this material
presents propitious tribological behavior with low friction coefficient and wear rate. It is
because one of the biggest problems associated with this kind of device is the large volume of
wear debris that could be generated. Hence, the materials combinations whose surfaces come
in contact influence the amount of debris and, consequently, the prosthesis performance. A
low wear is desirable because particulate wear is associated with inflammation, perprosthetic
bone loss and premature implant loosening [17].
The high wear resistance and low friction coefficient become silicon nitride a promising
material for tribological applications. As an example, Saito et al. [18] demonstrated that this
material can reach a friction coefficient of 0.01 when distilled water is used as lubricant. In
these cases, the experiments were performance in Si3N4/Si3N4 pairs by means of a pin-onplate device.
Also, investigations of the friction coefficient and wear rates of Si3N4/Si3N4 pairs
conducted by Zhou et al. [19], led to conclude that this material can be so efficient to replace
joints components. The experiments performed in a pin-on-disc testing machine, using water
as lubricant, showed that the tribological properties of polished surfaces of silicon nitride
might be superior to alumina. Hence, even when the sliding rate is not enough to result in
tribochemical polishing of joints, the contact surfaces can be polished before the implantation
into the human body.
Thereafter, Wong et al. [20] studied the surface roughness effect on friction of three
ceramics pairs: Si3N4/Si3N4, SiC/SiC and Al2O3/Al2O3. The lowest friction coefficient ( =
0.002) was obtained by silicon nitride, followed by silicon carbide ( = 0.006) and alumina (
= 0.007), respectively. And in 1999, Kusaka et al. [21] performed tests in alumina, silicon
carbide and silicon nitride, using bovine serum solution as lubricant. In steady state,
Si3N4/Si3N4 pairs presented higher value of friction coefficient ( = 0.2) and higher wear rate
than Al2O3/Al2O3 and SiC/SiC pairs. The unsatisfying results obtained by Si3N4/Si3N4 were
justified by the adhesion of the hydrated silicon oxide tribofilm on silicon nitride surface
during the test.
Bal et al. [17] studied Si3N4/Si3N4 and Si3N4/CoCr pairs with the THA bearing of Si3N4
with a mean Ra surface roughness value of less than 0.05 m. The results demonstrated that
Si3N4/Si3N4 and Si3N4/CoCr bearings produce wear volumes of 0.2 mm3 and 0.18 mm3,
respectively, at 1 million cycles. Extrapolating these data to 10 million cycles, they found
153
wear volumes of 0.65 and 0.47 mm3, respectively. As the clinical wear volume of CoCrpolyethylene THA is 62 mm3 and the experimental wear volumes of CoCr/CoCr and
Al2O3/Al2O3 are 6.5 mm3 and between 0.35 and 0.6 mm3, measured in other hip simulators,
respectively, the authors concluded the Si3N4/Si3N4 and Si3N4/CoCr pairs produced wear rates
comparable to Al2O3/Al2O3.
BIOCOMPATIBILITY OF SILICON NITRIDE

The first study about silicon nitride biocompatibility accomplished in 1980 was not
encouraging. In fact, Griss et al. [22] performed researches with silicon nitride containing iron
and magnesium as sintering aids and concluded that this material was unsatisfactory for
biomaterial applications. However, in 1989, Howllet et al. [23] showed that porous silicon
nitride obtained by reaction bonded favored bone growth in vivo, although cell propagation in
vitro has not been noted.
After Howllets work, there were a large number of studies demonstrating the interest in
evaluating the biological behavior of silicon nitride and its use in clinical applications. In
1999, Wang et al. [24] carried out a research about silicon nitride coating on titanium dental
implants in order to avoid adherence problems between the titanium and porcelain. In the
same year, Kue at al. [25] performed an in vitro study in polished and non-polished discs
obtained by RBSN and SRBSN (sintered after Reaction Bonded Silicon Nitride), using MG63 cell line. The non-polished discs had lower cell yield than the cells propagated on the
polystyrene control, but they promoted a reduction in osteocalcin production. Considering the
polished discs, the cells presented proliferative capacity similar to that control, producing
higher levels of osteocalcin.
Biocompatibility of silicon nitride particles as well as polished RBSN and SRBSN discs
was also evaluated in 2000 by Sohrabi et al. [26]. While cells propagation on all materials
surfaces has been observed, variation in regard to release of 1L-1 and TNF- cytokines were
also noted. The silicon nitride particles did not activated the 1L-1 cytokine, but high content
promoted release of TNF- cytokine. The cells propagated in silicon nitride discs surfaces
obtained by SRBSN had similar behavior to the control, considering the release of 1L-1 and
TNF- cytokines. Silicon nitride discs obtained by reaction bonded (RBSN) resulted in an
increase of 1L-1 and TNF- cytokines compared with the control.
Si3N4/bioglass composites were also investigated. In this case, wettability experiments
were performed using distilled and deionized water, diiodomethane and simulated body fluid
(SBF) in order to determine the contact angle by sessile drop technique [27]. The results
showed the composites are hydrophilic with good wettability behavior, being the contact
angle lower than those of the conventional ceramics, such as HA, TCP and bioglass.
Latter, Guedes e Silva et al. [28] demonstrated that silicon nitride containing ytterbium,
yttrium and aluminum oxides as sintering aids is a noncytotoxic material. The authors carried
out in vitro tests of cytotoxicity according to ISO 10993-part 5 [29]. Still in 2004, Sorrel et al.
[30] reported results of a clinical study of 15 years using intervertebral spacers of silicon
nitride. For startup of the studies in the late 80s, thirty patients with spinal degeneration
underwent anterior spinal intercoporeatly using RBSN. A control population of 10 patients
underwent the same procedure using autogenous bone. One year after surgery, 14 patients
154
reported significant reduction of pain, seven patients reported moderated reduction, and four
patients reported no change in pain. Five years after surgery, there was implant displacement
in the examined cases, indicating reaction with the implant in two cases and implant loss in
one.
In 2006, Si3N4 prototype minifixation system was manufactured and implanted for
osteosynthesis of artificial frontal bone defects in minipigs. Si3N4 miniplates and screws
showed satisfying intraoperative workability. No implant loss, displacement or fracture was
observed. Unlike, there was complete bone healing in all animals and bone/implant direct
contact was found [31].
Monolithic materials and a Si3N4/TiN eletroconductive composite were studied by
Massocchi and Bellosi [32]. The authors concluded that the eletroconductive composite is
very advantageous because it can be machined by electrical discharge providing to obtain
complex shapes, such as screws and miniplates. Also, the cytotoxicity analysis showed that
the materials are nontoxic. In a second work, Mazzochi et al. [33] evaluated those materials in
regard to (i) wetting behavior in aqueous media, including physiological solutions; (ii) the
chemical stability in water and physiological solutions and (iii) the wear resistance under
similar conditions to the hip joint prosthesis. The results demonstrated that the contact angle
of silicon nitride materials decreases with the phases amount in the grain boundary.
A sponge-cancellous-structured silicon nitride, characterized by a porous and
nonresorbable structure was developed by Anderson and Olsen [34]. They quantified the
extent and rate of bone ingrowth into structure. For that, cylindrical implants were inserted in
the medial femoral condyle of sheep. After 3 and 6 months, the implants were retrieved and
prepared to bone growth analysis by SEM.
Bone growth around silicon nitride implants were analyzed after 8 weeks of implantation
into rabbits tibia. Tissue tracers were administrated in the animals and, after euthanasia,
histological analysis was performed. In this work, the authors observed that the osteogenesis
process happened during the entire period defined by tracers and no adverse reaction occurred
around the implant [35].
SUMMARY
Although silicon nitride ceramics have been initially used in high temperature
components, researches have encouraged their use in biomedical applications. This review
shows that the propitious mechanical properties associated with the biocompatibility and
promising tribological features (low friction coefficient and high wear resistance) become
theses ceramics excellent candidates to replace joints components. Also, they are promising in
many other cases of bone replacement and as aids in regenerative process.
Additionally, studies suggest that the biological behavior of silicon nitride is function of
its forms (particles, porous or dense), processing techniques (RBSN, SRBSN, HIPS, HPSN,
SSN) and composition (kind and amount of sintering aids).
155
REFERENCES
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
Chiang, Y.M.; Birnie III, D.; Kingery, W.D. Physical Ceramics. John Wiley: New
York, 1997.
Ziegler, G.; Heinrich, J.; Wtting, G. J Mater Sci. 1987, 22, 3041-3085.
Swain, M.V. Structural and properties of ceramics. New York, N.Y: VCH, 1994
(Materials Science and Technology. 11).
Hampshire, S. Engineering properties of nitrides. In: Schneider, s.j. Ceramics and
Glasses. [S.I.]: ASM International, 1991. P. 813-829 (Engineered Materials Handbook,
4).
Kue, R.; Sohrabi, A.; Nagle, D.; Frondoza, C.; Hungerford, D. Biomaterials. 1999, 20,
1195-1201.
Mazzocchi, M.; Bellosi, A. J Mater Sci Mater Med. 2008, 19, 2881-2887.
Ziegler, G.; Heinrich, J.; Wtting, G. J Mater Sci. 1987, 22, 3041-3085.
Wang, C.M.; Pan, X.; Rhle, M. J Mater Sci. 1996, 31, 5281-5298.
Liang, J.; Topor, L.; Navrotsky, A.; Mitomo, M. J Mater Res. 1999, 14, 1959-1968.
Grieveson, P.; Jack, K.H.; Wilde, S. Special Ceramics. Manchester: George Faulkner
Press, 1968.
Galasso, F.; Kuntz, L.; Croft, W.J. J Am Ceram Soc. 1972, 55, 431-435.
Niihara, K.; Hirai, T. J. Mater Sci. 1977, 12, 1233-1242.
Sanders, W.A.; Mieskowski, D.M. Am Ceram Soc Bull. 1985, 64, 304-309.
Wtting, G.; Ziegler, G. Ceram Int. 1984, 10, 18-22.
Tajima, Y. Development of high performance silicon nitride ceramics and their
application. In: Mat. Res. Soc. Symp. Proc., v. 287, p. 189-196, 1993.
Pullum, O.J.; Lewis, M.H. J. Eur Ceram Soc. 1996, 16, 1271-1275.
Bal, B.S.; Khandkar, A.; Lakshminarayanan, R.; Clarke, I.; Hoffman, A.A.; Rahman,
M.N. The Journal of Arthroplasty, 2009, 24, 110-116.
Saito, T.; Imada, Y.; Honda, F. Wear. 1997, 205, 153-159.
Zhou, Y.S.; Ohashi, M.; Tomita, N.; Ikeuchi, K.; Takashima, K. Mater Sci Eng C.
1997, 5, 125-129.
Wong, H.C.; Umehara, N.; Kato, K. Wear. 1998, 218, 237-243.
Kusaka, J.; Takashima, K.; Yamane, D.; Iekeuchi, K. Wear. 1999, 225-229, 734-742.
Griss, P.; Wener, E.; Heimke, G. Alumina ceramic, bioglass, and silicon nitride: a
comparative biocompatibility study. In: Mechanical Properties of Biomaterials. Edited
by Hasting, G.W. and Williams, D.F., pp. 217-226, 1980.
Howlett, C.R.; Mccartney, E.; Ching, W. Clin Orthop. 1989, 244, 293304.
Wang, R.R.; Welsch, E.G.; Monteiro, O. J Biomed Mater Res. 1999, 46, 262-270.
Kue, R.; Sohrabi, A.; Nagle, D.; Frondoza, C.; Hungerford, D. Biomaterials. 1999, 20,
1195-1201.
Sohrabi, A.; Holland, C.; Kue, R., Nagle, D.; Hungerford, D.S.; Frondoza, C.G. J
Biomed Mater Res. 2000, 50, 43-49.
Amaral, M.; Lopes, M.A.; Santos, J.D.; Silva, R.F. Biomaterials. 2002, 23, 4123-4129.
Guedes e Silva, Higa, O.Z.; Bressiani, J.C. Mater Sci Eng C. 2004, 24, 643-646.
International Standard: Biological Evaluation of Medical Devices. Part 5. Tests for
Citotoxicity: In vitro Methods. ISO 1993; 10993-5.
156
[30] Sorrell, C.; Hardcastle, P.H.; Druitt, R.K.; Howlett, C.R.; McCartney, E.R. Results of
15-years clinical study of reaction bondend silicon nitride intervertebral spacers. IN:
7th WORD BIOMATERIAL CONGRESS, May 17-21, 2004, Sydney. Proceedings...
Sydney: Australian Society for Biomaterials, 1004, p. 1872.
[31] Newmann, A.; Unkel, C.; Werry, C.; Herborn, C.U.; Maier, H.R.; Rago, C.; Jahnke,
K. Prototype of a silicon nitride osteofixation system for the midface. OtolarygologyHead and Neck Surgery, v. 134, p. 923-930, 2006.
[32] Mazzocchi, M.; Bellosi, A. J Mater Sci Mater Med. 2008, 19, 28812887.
[33] Mazzocchi, M.; Gardini, D.; Traverso, P.L.; Faga, M.G.; Bellosi, A. J Mater Sci Mater
Med. 2008, 19, 28892901.
[34] Anderson, M.C.; Olsen, R. J Biomed Mat Res A. 2009, 92, 1598-1605.
[35] Guedes e Silva; Konig, Jr.; Carbonari, M.J.; Yoshimoto, M.; Allegrini, Jr., S.;
Bressiani, J.C. J Biomed Mat Res A. 2008, 84A, 337-343.
INDEX
A
acid, 21, 24, 35, 45, 63, 104, 112
acidic, 35, 45
activation energy, 9, 11, 16, 17
additives, vii, viii, 1, 2, 7, 8, 9, 13, 14, 15, 17, 18, 19,
20, 21, 22, 24, 33, 34, 53, 55, 61, 101, 102, 103,
104, 106, 107, 108, 109, 110, 116, 117, 118, 120,
132, 133, 134, 149, 151
adhesion, 37, 48, 54, 144, 152
adhesive strength, 143, 144
adhesives, 113
aerospace, ix, 125
aging process, 150
alters, 49
aluminium, 20
aluminum oxide, 126, 127, 134, 153
ammonia, 70, 129, 130
annealing, 7, 70, 73, 75, 76, 80, 81, 83, 84, 86, 89,
90, 95, 98, 130
argon, 15, 16, 117, 139, 140
arsenic, 127
ASI, 30, 145
atmosphere, vii, 1, 3, 6, 10, 15, 16, 18, 21, 22, 23,
35, 43, 45, 52, 73, 80, 81, 102, 104, 105, 106,
113, 114, 127, 131, 132
atmospheric pressure, 4, 18, 73, 80, 81, 86, 87, 90,
98
atoms, viii, 35, 39, 45, 46, 69, 70, 75, 81, 82, 90, 94,
98, 99, 108, 113, 114, 116, 118, 127, 128, 129,
150
attachment, 113, 115
B
band gap, 87
barriers, 114, 126, 127
base, 6, 7, 102, 103, 108, 114, 115, 133, 134
baths, 150
Belgium, 146
bending, 41, 42, 48, 50, 51, 55, 84, 105, 126, 127
binding energy, 78
biochemistry, 67
biocompatibility, ix, 149, 153, 154, 155
Biocompatibility, 153
biological behavior, 153, 154
biomedical applications, 154
body fluid, 153
Boltzmann constant, 97
bonding, 37, 70, 76, 106, 108, 111, 113, 115, 116,
117, 122, 126
bonds, 36, 37, 43, 71, 80, 83, 84, 86, 87, 89, 149,
151
bone, 150, 152, 153, 154
bone growth, 153, 154
boundary value problem, 56
Brazil, 149
brittle nature, viii, 125, 127
brittleness, 116, 150
Broadband, 56
bulk materials, 104, 106
burn, 43
burnout, 43, 44, 45, 46
by-products, 4
C
calcium, 12
candidates, vii, 33, 34, 118, 154
capillary, 36, 115
carbon, 4, 16, 34, 137, 139
casting, 126
catalyst, 34
cation, 13, 15, 132
cell line, 153
ceramic, vii, viii, ix, 2, 3, 4, 5, 12, 14, 15, 18, 19, 22,
23, 27, 33, 34, 36, 37, 38, 39, 40, 43, 48, 59, 61,
62, 63, 64, 65, 101, 102, 103, 104, 105, 106, 107,
158
Index
108, 110, 111, 112, 113, 114, 115, 116, 117, 118,
119, 126, 127, 134, 143, 145, 146, 149, 150, 155
ceramic materials, vii, viii, ix, 2, 33, 34, 101, 102,
105, 107, 113, 127, 149
Ceramics, v, 1, 25, 26, 27, 28, 29, 30, 33, 35, 42, 45,
51, 56, 66, 67, 101, 102, 105, 108, 111, 112, 118,
121, 145, 146, 149, 155
challenges, viii, 125
chemical, viii, 3, 4, 15, 17, 23, 55, 69, 70, 102, 105,
106, 108, 113, 114, 115, 116, 118, 125, 126, 127,
129, 154
chemical properties, viii, 114, 125, 126
chemical reactions, 105
chemical stability, 15, 154
chemical vapor deposition, viii, 69, 70
chemical vapour deposition, 4, 127
chemicals, 102
China, 33, 66
chlorine, 18
cladding, 144
classes, 133
clinical application, 150, 153
clusters, viii, 69, 75, 81, 86, 90, 98
CO2, 7, 17, 18, 43, 45
coatings, vii, viii, 4, 16, 125, 127, 132, 133, 134,
135, 137, 138, 139, 142, 143, 144, 145, 150
cogeneration, 14
combustion, ix, 15, 34, 125, 126, 135, 150
combustion environment, 15
commercial, 12, 22, 63, 116
communities, viii, 101, 112, 114
community, 102, 105
compatibility, 7, 9
complexity, 134
composites, 5, 23, 27, 31, 55, 102, 103, 120, 153
composition, 2, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15,
19, 20, 22, 23, 24, 64, 75, 81, 104, 108, 109, 114,
134, 137, 144, 150, 152, 154
compounds, 102, 117, 130
compression, 73, 141
computer, 74
conductivity, 46, 93, 105, 109, 127, 132
confinement, 81, 97
CONGRESS, 25
connectivity, 46
consolidation, 40, 52, 107, 133
constituents, 23, 126
contaminant, 131, 132
contamination, 127
cooling, 6, 13, 20, 22, 98, 103, 114, 118, 131, 151
cooling process, 114
copper, 114
correlation, 34, 41, 50, 55
corrosion, 3, 7, 9, 12, 24, 27, 34, 138, 139

cost, 2, 3, 102, 105, 106, 108, 115, 117, 126
covalent bond, 108, 116, 149, 151
covalent bonding, 108
covering, 4
cracks, 6, 13, 86, 114
creep, 2, 15, 105, 131
crystal structure, 40, 129
crystalline, 5, 9, 12, 13, 14, 15, 16, 19, 40, 46, 72,
74, 80, 81, 83, 84, 98, 109, 118, 127, 129, 131,
133, 134, 135, 136, 141, 150, 151
crystallinity, 19
crystallites, 102
crystallization, viii, 5, 13, 14, 22, 69, 70, 81, 86, 90,
93, 95, 98, 99, 119
crystals, 7, 19, 20, 23, 104, 110, 120, 143
CT, 18
cubic boron nitride, 109
CVD, 4, 16, 17, 18, 19, 21, 29, 127
cycles, 5, 137, 152
CYTOKINES, 153
cytotoxicity, 153, 154
D
data analysis, 90, 98
decay, 71, 76
decomposition, 4, 15, 18, 106, 108, 113, 114, 116,
130, 151, 152
defects, 13, 14, 87, 89, 154
degradation, 4, 5, 119
degree of crystallinity, 19
Denmark, 30
dental implants, 153
Department of Energy, 29
deposition, viii, ix, 4, 69, 70, 76, 125, 127, 133, 135,
139, 144
deposits, 4
depth, 12
destruction, 127, 137
detectable, 83
detonation, ix, 125, 133, 135, 136
deviation, 8, 9
dielectric constant, vii, 33, 34, 56, 57, 58, 59, 61, 62,
63, 65
dielectrics, 127
diesel engines, 126, 150
diffraction, 7, 19, 35, 39, 44, 47, 48, 55, 110, 118,
129, 136, 143
diffusion, vii, 1, 2, 4, 5, 6, 7, 12, 13, 14, 16, 17, 18,
19, 20, 22, 23, 84, 98, 104, 108, 113, 114, 115,
118, 119, 122, 127, 129, 149, 151
diffusion mechanisms, 6, 98, 149
Index
diffusion process, 17
diffusion rates, 17
diffusivity, 117
diodes, 70, 127
direct observation, 72
discs, 153
dispersion, 80, 81, 91
displacement, 154
dissociation, 18, 79, 131, 152
distilled water, 152
distribution, 10, 73, 91, 114, 120, 136
doping, 127
drawing, 126
drying, 34, 133
DTA curve, 46, 47
E
economic problem, 109
effusion, 83
electric current, 107
electric field, 79
electrical properties, 70
electrical resistance, 46
electrodes, 139
electroluminescence, 70
electromagnetic, vii, 33, 34, 57, 65, 66, 67
electron, 7, 14, 17, 23, 70, 72, 76, 79, 83, 84, 98,
118, 119
electron microscopy, 14, 83, 119
electrons, 70, 78, 98
e-mail, 1
encapsulation, 19, 23, 152
energy, ix, 3, 9, 11, 16, 17, 70, 73, 74, 78, 89, 91, 93,
98, 109, 111, 113, 125, 132, 133, 141
energy density, 73, 91, 93
engineering, vii, viii, 2, 3, 18, 101, 105, 112, 118,
131
England, 28
environment, 3, 4, 6, 15, 16, 22, 151
environmental conditions, 127
environments, 15, 18, 20, 23, 24, 30, 80, 145
equilibrium, 16, 105, 116, 121, 134
equipment, 71, 72, 102, 106, 115, 136, 139, 144
erosion, 126, 145
ethanol, 51
euthanasia, 154
evaporation, 75, 81, 105, 106
evidence, 37, 39, 45, 104, 109, 110, 111, 116, 118,
119, 120
excitation, 71, 76
exciton, 78
experimental condition, 5, 65, 70, 108
159
experimental design, 63
exposure, 2, 5, 6, 7, 8, 9, 12, 13, 14, 15, 16, 17, 22,
23
F
fabrication, 41, 50, 151
Fabrication, 66
families, 132
feedstock, 133, 142
field tests, 29
fillers, 116
film formation, 118
film thickness, 75, 86
films, viii, 4, 69, 70, 71, 72, 73, 74, 75, 76, 77, 80,
81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 93, 94,
96, 98, 99, 116, 127, 132
filtration, 34
Finland, 136
flame, 135
flight, 16, 133
fluid, 153
force, 36
formation, viii, 2, 4, 5, 6, 18, 22, 23, 24, 35, 36, 43,
45, 46, 69, 70, 76, 81, 89, 90, 92, 95, 98, 99, 101,
102, 103, 104, 106, 112, 113, 115, 116, 117, 118,
119, 120, 133, 134, 143
formula, 43, 79, 97
fracture toughness, ix, 105, 109, 111, 112, 120, 126,
131, 132, 149
free energy, 3, 89
friction, 144, 152, 154
FTIR, 71, 83, 84, 85, 86, 87, 89
fuel cell, 2
fuel efficiency, 14
funds, 66
fusion, 150
G
gallium, 127
geometry, 72, 80, 114, 135, 151
Germany, 42, 133, 145, 146
glass transition, 23
glass transition temperature, 23
glasses, 14, 107, 113
good oxidation resistance, vii, 33, 34
GPS, 14, 21, 151
grain boundaries, 14, 103, 105, 111, 119, 151
grain size, 19, 53, 63, 107, 136, 137, 138, 139
graph, 22
graphite, 51, 52, 126, 151
160
Index
growth, 7, 20, 36, 39, 53, 70, 74, 86, 92, 93, 95, 102,
104, 105, 109, 116, 127, 153, 154
guidelines, 2
H
hardness, 3, 105, 109, 112, 127, 143, 144, 149
healing, 154
heat pumps, 126
heat transfer, ix, 125, 133
heating rate, 52, 107
helium, 10
hexagonal lattice, 126
high current pulsed plasma spraying (HCPPS), ix,
125
High Resolution Transmission Electron Microscopy,
72
high strength, 41, 106
hip joint, 154
homogeneity, 90
host, 118
hot pressing, 3, 34, 132, 149
hot spots, 141
House, 68
HRTEM, 14, 23, 39, 40, 72, 84, 86, 97, 98
human, 152
human body, 152
humidity, 23, 127
hydrofluoric acid, 104, 112
hydrogen, viii, 27, 35, 45, 69, 71, 74, 75, 81, 83, 87,
89
hydrogen atoms, 35, 45
I
ideal, 65
idealization, 56
image, 38, 39, 40, 45, 46, 64, 65, 83, 84, 86, 87, 90,
97, 137
image analysis, 137
images, 39, 54, 90
implants, 150, 153, 154
impurities, 4, 12, 13, 15, 23, 104
in vitro, ix, 149, 153
in vivo, ix, 149, 153
incidence, 56, 57, 74
incompatibility, 115
indentation, 137, 143
industries, 42, 106
inertia, 150
inflammation, 152
infrared spectroscopy, viii, 69
insertion, 14
insulation, 107
insulators, 127
integrity, 113
interface, 6, 7, 12, 13, 16, 17, 18, 19, 23, 113, 114,
115, 116, 117, 137, 144
interfacial bonding, 111
ions, 7, 23, 35, 45, 129
iron, 153
J
Japan, 25, 28, 43, 146, 147
joints, 113, 122, 152, 154
K
kinetics, vii, 1, 12, 13, 15, 18, 23, 102
L
laboratory studies, 115
laser radiation, 93
lattices, 126, 129
lead, viii, 15, 69, 80, 83, 86, 87, 90, 92, 93, 95, 98,
114, 117
leakage, 72
liberation, 116
lifetime, 5, 15
ligament, 45, 46
light, 70, 71, 72, 74, 75, 83, 98, 99, 126, 127, 135
light emitting diode, 70, 127
linear dependence, 132
liquid phase, viii, 2, 6, 7, 14, 15, 101, 102, 103, 104,
106, 108, 113, 117, 120, 134, 149, 151
low temperatures, 20, 93, 107
low thermal expansion coefficient, vii, 3, 33, 34, 52
low-frequency pulsed detonation spraying (LFPDS),
ix, 125, 133
M
magnesium, 12, 153
magnitude, 8, 12
majority, 12, 37, 48, 82, 110
manufacturing, 15, 102, 126
masking, 127
mass, vii, 1, 2, 4, 16, 17, 18, 105, 106, 113, 126, 133
mass loss, vii, 1, 4, 17, 18, 105
material dissociates, ix, 125, 132
materials, vii, viii, ix, 2, 3, 5, 13, 14, 15, 17, 19, 22,
23, 26, 33, 34, 37, 59, 63, 66, 101, 102, 104, 105,
Index
106, 107, 108, 112, 113, 114, 115, 116, 126, 127,
141, 143, 149, 151, 152, 153, 154
matrix, vii, viii, 14, 19, 23, 33, 34, 37, 41, 43, 48, 57,
70, 71, 72, 75, 86, 108, 110, 133, 143, 151
matter, 104
measurement, 76, 97
measurements, 67, 71, 74, 95
mechanical degradation, 119
mechanical properties, vii, viii, 2, 3, 4, 5, 6, 7, 13,
14, 33, 34, 40, 53, 54, 61, 101, 102, 103, 105,
106, 109, 112, 133, 151, 154
mechanical stress, 3, 81, 86, 89, 127
media, 51, 154
melt, 118
melting, 13, 14, 24, 55, 98, 108, 115, 116, 118, 150
melting temperature, 98, 115
membranes, 34
memory, 127
mercury, 71, 76
metal oxides, 118, 132
metals, 3, 114, 116, 118, 126, 128, 150
meteorites, 126, 129
Mexico, 1
microhardness, 127, 137
micropores, viii, 33, 36, 37, 48, 65
microscope, 72, 90
microscopy, 14, 83, 86, 87
microstructure, 2, 5, 15, 19, 34, 36, 37, 45, 47, 102,
103, 109, 111, 112, 114, 151
microstructures, 8, 45, 102, 103, 112
microwave heating, 107
microwave radiation, 107
microwaves, 107
migration, 79
military, 3
Ministry of Education, 33
mixing, 35, 42, 45, 133
models, 76
modifications, 109, 126
modulus, 41, 42, 126, 131
moisture, 30
molar ratios, 10
molecules, 35, 45
morphology, 13, 102, 109, 120, 151
N
nanocrystals, 70, 72, 76, 78, 81, 83, 84, 86, 90, 91,
97, 98
nanometer, 75
Nanostructures, 72
NATO, 30, 145
neodymium, 143
161
Netherlands, 25, 29, 30, 145, 146

nitrides, 30, 102, 114, 155
nitrogen, 16, 17, 18, 19, 21, 22, 35, 43, 45, 52, 71,
72, 80, 81, 89, 102, 105, 106, 108, 113, 114, 126,
128, 129, 130, 131, 133, 135, 136, 150, 151
nitrogen gas, 22, 105, 114
non-refractory substrates, viii, 69, 99
novel materials, 107
nuclei, 5, 126
nucleus, 126
O
OH, 22, 26, 27, 29, 145, 146
operations, 106
opportunities, 5, 24, 34
optical microscopy, 86
optical properties, 89, 90, 98
optimization, ix, 125, 144
ores, 43
oxidation, vii, viii, ix, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 12,
13, 14, 15, 16, 17, 18, 19, 20, 22, 23, 24, 29, 30,
31, 33, 34, 101, 105, 118, 119, 120, 125, 131, 132
oxidation products, 13, 15, 20
oxidation rate, 4, 5, 6, 8, 10, 12, 13, 15, 16, 17, 18,
19, 20, 22, 24
oxidizing atmospheres, vii, 1, 3
oxygen, 2, 4, 5, 6, 9, 10, 12, 15, 16, 17, 18, 19, 20,
22, 23, 35, 45, 80, 104, 118, 119, 131, 135, 150,
151
P
pain, 154
parallel, 56, 57, 59, 76, 129, 139
passivation, 9, 127
permit, 113
phase diagram, 134
phase transformation, vii, viii, 33, 34, 35, 37, 48,
101, 104, 105, 108, 109, 110, 120, 133, 150
phonons, 80, 81
phosphate, vii, viii, 33, 34, 37, 45, 48, 63
phosphate bonded silicon nitride ceramics, vii, viii,
33
phosphorous, 127
phosphorus, 63
photoluminescence, viii, 69, 70
photons, 74, 98
photovoltaic cells, 127
physical and mechanical properties, 102
physical properties, 2, 64, 65
physics, 113
Plank constant, 97
162
Index
plasma enhanced chemical vapor deposition

(PECVD), viii, 69
plasticity, 20, 70
Poland, 73
polarization, 56, 57, 59, 72, 79, 80
polymer, 34, 46
polystyrene, 153
population, 153
porosity, viii, 33, 34, 36, 37, 40, 41, 43, 44, 45, 46,
48, 49, 50, 54, 57, 59, 63, 106, 131, 134, 137, 151
porous materials, 37
praseodymium, 143
precipitation, 103, 151
preparation, vii, 33, 34, 43, 59, 60, 61, 62, 63, 64,
133, 151
primary products, 126
principles, 81, 89
probability, 79, 131
probe, 7, 74, 98, 118
propagation, 56, 111, 120, 141, 153
propylene, 135
prostheses, 150
prosthesis, 152, 154
prosthetic device, 150
protection, 127
protective coating, viii, 4, 15, 125, 132, 139
prototype, 154
pumps, 126
purity, 6, 19, 42, 63
Q
quantum dot, 70
quantum dots, 70
R
radiation, 93, 107, 129
radius, 74, 132
Raman spectra, 72, 80, 81, 82, 83, 91, 92, 93, 94, 95,
96
Raman spectroscopy, viii, 69, 72, 81
ramp, 45
raw materials, 63, 106
RE, 13, 106
reactants, 22
reaction time, 16
reactions, 2, 18, 105, 113, 130, 131, 133, 134
reality, 144
recession, 14
recombination, 86, 89
red shift, 76, 97
redistribution, 75
refraction index, 74, 75, 77

reinforcement, 131
relaxation, 14, 79
reliability, 3, 22, 29, 57, 114, 126
requirements, viii, 61, 62, 101
researchers, 6, 8, 9, 102, 150
resistance, vii, 1, 2, 3, 5, 6, 7, 8, 12, 13, 14, 15, 17,
18, 19, 23, 24, 33, 34, 46, 105, 118, 126, 127,
132, 137, 138, 144, 149, 150, 152, 154
resolution, 14, 23, 40, 71, 72, 83, 84
rings, 126, 129
room temperature, vii, 33, 34, 35, 36, 37, 42, 70, 72,
97, 137, 139
root, 22
roughness, 152
routes, 2, 126, 130, 151
Russia, 69, 71
rutile, 6, 23, 126
S
safety, 131
sapphire, 98
scanning electron microscopy, 119
scattering, 67, 72, 80, 81, 83, 87, 91, 93, 95
segregation, 114
SEM micrographs, 20
semiconductor, viii, 69, 70, 90, 98, 99, 127, 149
semiconductor nanoclusters, viii, 69, 70, 90, 98, 99
semiconductors, 70, 87, 98
sensitivity, 71
sensors, 127
serum, 152
services, 113
shape, 7, 21, 34, 41, 45, 63, 66, 74, 76, 83, 91, 104,
105, 108, 112, 118, 120
shear, 150
shear strength, 150
sheep, 154
shock, 2, 34, 127, 132, 139, 140, 141, 144
showing, 13, 20, 40, 76, 134
silica, 6, 14, 15, 17, 19, 22, 24, 74, 102, 129, 151
silicon nanocrystals, 83, 86
Silicon nitride-based ceramics, vii, 1
silicon-rich nitride (SRN), viii, 69, 70
Simulated Body Fluid (SBF), ix, 149
simulation, 57, 66
SiO2, vii, 1, 2, 3, 4, 6, 7, 8, 10, 12, 13, 14, 16, 18, 19,
20, 21, 23, 34, 37, 39, 45, 46, 47, 48, 54, 67, 102,
103, 105, 106, 107, 109, 118, 133, 134, 135
SiO2 surface, 4
sodium, 127
solid phase, 18, 108, 113
Index
solid solutions, 135
solid state, 2, 37, 40, 55, 149
solidification, 24, 115, 118
solubility, 6, 108, 114, 118
solution, 103, 114, 132, 139, 152
Spain, 1, 28
spark plasma sintering, viii, 63, 101
species, 17, 22, 23, 113
specifications, 130
spectroscopy, viii, 69, 72, 81
sponge, 154
Spring, 27
stability, 3, 8, 14, 15, 20, 90, 102, 126, 127, 151, 154
starch, 43, 44, 45, 46, 66
stars, 126
state, 2, 37, 40, 55, 75, 108, 113, 141, 149, 151, 152
states, 71, 79
statistics, 131
steel, 126, 133, 137, 138, 139, 143, 144, 145
stoichiometry, viii, 69, 70, 74, 75, 76, 80, 81, 83, 84
storage, 139
stress, viii, 4, 8, 22, 89, 113, 114, 125, 127, 131, 137
stretching, 83, 84, 86
structural changes, 91, 95
structural characteristics, 79
structure, vii, 14, 33, 34, 35, 36, 37, 39, 43, 45, 47,
48, 54, 57, 58, 59, 64, 65, 68, 76, 80, 81, 90, 91,
95, 98, 102, 127, 129, 131, 150, 151, 154
substrate, 70, 72, 74, 80, 81, 93, 95, 139, 140, 144
substrates, viii, 17, 69, 70, 74, 91, 92, 93, 94, 95, 99,
125, 133, 137, 138, 139
Sun, 24
supernovae, 126
surface area, 84
surface layer, 4, 14
Switzerland, 25, 145
symmetry, 129, 150
synthesis, vii, 34, 56, 102, 126, 127, 129
T
target, 81
techniques, viii, 22, 24, 34, 69, 101, 105, 107, 108,
112, 113, 115, 127, 132, 149, 151, 154
technologies, 70, 112, 129, 132, 133
technology, ix, 14, 67, 70, 118, 125, 140, 144
TEM, 14, 19, 46, 84
temperature dependence, 71
testing, 15, 152
texture, 41
thermal analysis, 46
thermal decomposition, 130
thermal destruction, 127
163
thermal energy, 113, 141

thermal expansion, vii, 3, 6, 13, 20, 33, 34, 46, 52,
114, 126, 127, 132
thermal properties, 114, 127
thermal stability, 102
thermal treatment, 6
thermodynamics, 20, 89
thinning, 16
tibia, 154
time resolution, 71
titania, 23
titanium, 5, 6, 23, 153
TNF, 153
TNF-, 153
transformation, vii, viii, 13, 33, 34, 35, 37, 48, 80,
101, 104, 105, 108, 109, 110, 115, 120, 133, 150
transformations, 84, 110
transition temperature, 16, 23
translation, 74
transmission, 14, 23, 56, 57, 59, 60, 61, 62, 65, 67,
72
transmission electron microscopy, 14
Transmission Electron Microscopy, 72
transport, 16, 23, 117
transport processes, 23
treatment, 6, 7, 15, 22, 34, 102, 104, 105, 106, 113,
118, 119
tunneling, 79
U
UK, 146
uniform, 10, 23, 37, 43, 48, 65, 75, 106, 107, 140,
141
United, 14, 18
United States, 14
USA, 30
UV, 127
V
vacancies, 150
vacuum, 21, 115
valve, 126, 139
vapor, viii, 30, 69, 70, 131, 149, 151
variables, 114
variations, 151
velocity, 17, 22, 135, 140, 144
vibration, vii, viii, 69, 71, 80, 85, 86, 91, 97, 98
Vickers hardness, 105
Vickers indentation hardness, 143
viscosity, ix, 15, 104, 125, 132, 133, 140, 143, 151
volatilization, 22, 45, 46, 47, 48, 50, 55, 114
164
Index
W
waste, 126
waste heat, 126
water, vii, 1, 3, 21, 22, 23, 30, 135, 137, 144, 152,
153, 154
water vapor, 30
wave propagation, 56
wear, 106, 126, 137, 139, 144, 149, 150, 152, 154
Weibull equation, 131
weight changes, 6, 12
weight gain, 1, 2, 3, 4, 5, 6, 8, 9, 10, 11, 12, 13, 14,
19, 22
weight loss, 10, 22, 152
welding, 113, 126
wettability, 108, 153
wetting, 116, 154
windows, 66, 67
workers, 89
X
xenon, 71
X-ray diffraction, 7, 35, 39, 44, 47, 48, 110, 118
X-ray diffraction (XRD), 7
XRD, 7, 12, 16, 19, 37, 48, 54, 111, 118, 143
Y
yield, 153
ytterbium, 153
yttrium, 20, 131, 143, 153
Z
zirconia, 63, 127, 128, 150
zirconium, 34

Silicon Nitride - Synthesis, Properties and Application (2012)

Hochgeladen von

Dokumentinformationen

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

Silicon Nitride - Synthesis, Properties and Application (2012)

Hochgeladen von

Copyright:

Verfügbare Formate

CHEMICAL ENGINEERING METHODS AND TECHNOLOGY

CHEMICAL ENGINEERING METHODS

Additional E-books in this series can be found on Novas website

Additional E-books in this series can be found on Novas website

CHEMICAL ENGINEERING METHODS AND TECHNOLOGY

EMILIANO JOSE HIERRA

JESUS ANJEL SALAZAR

Nova Science Publishers, Inc.

Copyright 2012 by Nova Science Publishers, Inc.

Published by Nova Science Publishers, Inc. New York

Synthesis and Properties of Silicon Nitride Coatings

Silicon Nitride as Biomaterial

Emiliano Jose Hierra and Jesus Anjel Salazar

In: Silicon Nitride

HIGH TEMPERATURE OXIDATION

Centro de Investigacin en Materiales Avanzados, CIMAV S.C, Chihuahua, Mexico

M. H. Bocanegra-Bernal, A. Aguilar-Elguzabal and J. Echeberria

High Temperature Oxidation of Silicon Nitride Based Ceramics

2.1. Passive Oxidation

M. H. Bocanegra-Bernal, A. Aguilar-Elguzabal and J. Echeberria

2.2. Active Oxidation

High Temperature Oxidation of Silicon Nitride Based Ceramics

2.3. Effect of Sintering Additives on the Oxidation of Si3N4

M. H. Bocanegra-Bernal, A. Aguilar-Elguzabal and J. Echeberria

High Temperature Oxidation of Silicon Nitride Based Ceramics

2SiC + Y2O3 + 3O2 Y2Si2O7 + CO(CO2) (g)

Y2O3 + 2 Si02 Y2Si2O7

M. H. Bocanegra-Bernal, A. Aguilar-Elguzabal and J. Echeberria

Weight Gain, W (mg.cm -2)

Weight Gain, W2 (mg2.cm-4)

Weight Gain, W (mg.cm -2)

eight Gain, W2 (mg2.cm-4)

High Temperature Oxidation of Silicon Nitride Based Ceramics

Weight Gain, W2 (mg2.cm-4)

Weight Gain, W (mg.cm-2)

Weight Gain, W2 (mg2.cm-4)

2.4. Effect of Hot Pressing on the Oxidation of Si3N4

M. H. Bocanegra-Bernal, A. Aguilar-Elguzabal and J.(a)

Weight Gain, W (mg.cm-2)

Weight Gain, W 2 (mg2.cm-4)

Weight Gain, W 2 (mg2.cm-4)

Weight Gain, W (mg.cm-2)

High Temperature Oxidation of Silicon Nitride Based Ceramics

Weight Gain, W (mg.cm-2)

Weight Gain, W2 (mg2.cm-4)

Weight Gain, W2 (mg2.cm-4)

KY64 (2) Q = 325 kJ.mol-1

1/T.104 (oK -1)

M. H. Bocanegra-Bernal, A. Aguilar-Elguzabal and J. Echeberria

Likewise, Cubicciotti and Lue [109], investigated the oxidation of a commercial HP

High Temperature Oxidation of Silicon Nitride Based Ceramics

M. H. Bocanegra-Bernal, A. Aguilar-Elguzabal and J. Echeberria

High Temperature Oxidation of Silicon Nitride Based Ceramics

2.5. Oxidation of Si3N4 under Different Atmospheres

M. H. Bocanegra-Bernal, A. Aguilar-Elguzabal and J. Echeberria

High Temperature Oxidation of Silicon Nitride Based Ceramics

M. H. Bocanegra-Bernal, A. Aguilar-Elguzabal and J. Echeberria

2.6. Oxidation of Hot Isostatically Pressed Silicon Nitride

High Temperature Oxidation of Silicon Nitride Based Ceramics

M. H. Bocanegra-Bernal, A. Aguilar-Elguzabal and J. Echeberria

Oxidation rates in atmospheres with lower oxygen activity, aO

High Temperature Oxidation of Silicon Nitride Based Ceramics

Si3N4+7 vol% Y2Si2O7

Hot pressing 1800 oC,