(Challenges and Advances in Computational Chemistry and Physics 26) Andrzej Koleżyński, Magdalena Król - Molecular Spectroscopy—Experiment and Theory_ From Molecules to Functional Materials-Springer I.pdf

Challenges and Advances
in Computational Chemistry and Physics 26

Series Editor: Jerzy Leszczynski
Andrzej Koleżyński
Magdalena Król Editors
Molecular
Spectroscopy—
Experiment and
Theory
From Molecules to Functional Materials
Challenges and Advances in Computational
Chemistry and Physics
Volume 26
Series editor
Jerzy Leszczynski
Department of Chemistry and Biochemistry
Jackson State University, Jackson, MS, USA
This book series provides reviews on the most recent developments in computa-
tional chemistry and physics. It covers both the method developments and their
applications. Each volume consists of chapters devoted to the one research area.
The series highlights the most notable advances in applications of the computa-
tional methods. The volumes include nanotechnology, material sciences, molecular
biology, structures and bonding in molecular complexes, and atmospheric
chemistry. The authors are recruited from among the most prominent researchers
in their research areas. As computational chemistry and physics is one of the most
rapidly advancing scientific areas such timely overviews are desired by chemists,
physicists, molecular biologists and material scientists. The books are intended for
graduate students and researchers.
More information about this series at http://www.springer.com/series/6918

Andrzej Koleżyński Magdalena Król
•
Editors
Molecular Spectroscopy—
Experiment and Theory
From Molecules to Functional Materials
123
Editors
Andrzej Koleżyński Magdalena Król
Faculty of Materials Science and Ceramics Faculty of Materials Science and Ceramics
AGH University of Science and Technology AGH University of Science and Technology
Kraków, Poland Kraków, Poland
ISSN 2542-4491 ISSN 2542-4483 (electronic)

Challenges and Advances in Computational Chemistry and Physics
ISBN 978-3-030-01354-7 ISBN 978-3-030-01355-4 (eBook)
https://doi.org/10.1007/978-3-030-01355-4
Library of Congress Control Number: 2018955928
© Springer Nature Switzerland AG 2019

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part
of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission
or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar
methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are exempt from
the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this
book are believed to be true and accurate at the date of publication. Neither the publisher nor the
authors or the editors give a warranty, express or implied, with respect to the material contained herein or
for any errors or omissions that may have been made. The publisher remains neutral with regard to
jurisdictional claims in published maps and institutional affiliations.
This Springer imprint is published by the registered company Springer Nature Switzerland AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Preface
The idea of this book originated in some discussions we carried out during one
of the conference we had the pleasure to organize, namely XIVth International
Conference on Molecular Spectroscopy, Kraków-Białka Tatrzańska, Poland in
2017, which was the last conference in a series. The history of these conferences
spans over 20 years. In 1991, Prof. Henryk Ratajczak from Wroclaw University,
Poland, the originator of the idea, organized the first one in a series (as a national
conference with international participation) with the intention to have the following
ones every two years. Similar conferences devoted to Vibrational Spectroscopy in
Materials Science were initiated in 1996 by Prof. Mirosław Handke from AGH
University of Science and Technology in Krakow, Poland. For a few years, these
conferences were organized independently and in 2009 were merged into a single
one, organized every two years, alternately by University of Wrocław and AGH
University of Science and Technology. The last conference was organized in 2017
in Białka Tatrzańska and gathered above three hundred participants (almost a third
from abroad) and was devoted to practical application of various spectroscopic
methods to study different types of materials and showed that modern spectroscopy
is an invaluable tool in recent scientific endeavor to understand the structure and
properties of various natural and synthetic materials as well as to create new or to
modify existing ones in order to obtain better suited (functionalized) materials
required by modern industry.
There are a lot of books devoted to various issues related to either spectroscopic
methods themselves or their use in the characterization of diverse classes of
materials exhibiting often very different properties. However, the books containing
comprehensive, up-to-date description of theory, experimentation, and simultane-
ous use of different spectroscopic methods to study the properties of a given type of
materials, provided by leading specialists in the field for various classes of materials
are still rare and sought after.
This book is intended to provide such comprehensive description of materials
characterization by means of a wide spectrum of modern spectroscopic methods
and is aimed at scientists interested in theory and practical application of various
spectroscopic methods in their work. It starts with a short introduction to theoretical
v
vi Preface
foundations and description of fundamentals of computational methods commonly

used in spectroscopy, followed by a series of chapters presenting the results
obtained using various spectroscopy methods combined with theoretical calcula-
tions. Consecutive chapters are devoted to the studies of a broad range of materials
(starting from organic molecules and biological systems, through quantum dots and
nanomaterials, to crystalline and amorphous inorganic solids) using various spec-
troscopic methods, like the most popular vibrational spectroscopy (IR and Raman),
Mossbauer spectroscopy, and UV–Vis electron spectroscopy.
Kraków, Poland Andrzej Koleżyński

Magdalena Król
Contents
1 Computational Methods in Spectroscopy . . . . . . . . . . . . . . . . . . . . 1

2 Scaling Procedures in Vibrational Spectroscopy . . . . . . . . . . . . . . . 49
Olga Bąk and Piotr Borowski
3 Quantum Dot and Fullerene with Organic Chromophores
as Electron-Donor-Acceptor Systems . . . . . . . . . . . . . . . . . . . . . . . 97
Danuta Wróbel and Bolesław Barszcz
4 Material Analysis Using Raman Spectroscopy . . . . . . . . . . . . . . . . 123
Animesh K. Ojha and H. Michael Heise
5 Ligand-Core NLO-Phores . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
Rodolphe Antoine
6 Small and Large Molecules Investigated by Raman
Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161
Krzysztof Czamara, Ewelina Szafraniec, Ewelina Wiercigroch,
Szymon Tott, Grzegorz Zając, Ewa Machalska, Monika Dudek,
Dominika Augustynska, Kamilla Malek, Agnieszka Kaczor
and Malgorzata Baranska
7 Hydantoins and Mercaptoimidazoles: Vibrational Spectroscopy
as a Probe of Structure and Reactivity in Different Environments,
from the Isolated Molecule to Polymorphs . . . . . . . . . . . . . . . . . . . 199
Rui Fausto, Gulce O. Ildiz, Elisa M. Brás and Bernardo A. Nogueira
8 Vibrational Spectroscopy in Analysis of Stimuli-Responsive
Polymer–Water Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223
Marcin Kozanecki, Marcin Pastorczak and Krzysztof Halagan
9 Mössbauer Spectroscopy of Magnetoelectric Perovskite
Oxides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273
Paweł Stoch and Agata Stoch
vii
viii Contents
10 Vibrational Spectroscopy of Zeolites . . . . . . . . . . . . . . . . . . . . . . . . 301

Magdalena Król, Andrzej Koleżyński, Andrzej Mikuła
and Włodzimierz Mozgawa
11 In Situ and Operando Techniques in Catalyst Characterisation
and Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 333
Przemysław Jodłowski and Joanna Łojewska
12 Application of Spectroscopic Methods in the Studies of
Polysiloxanes, Cubic Oligomeric Silsesquioxanes, and
Spherosilicates Modified by Organic Functional Groups
via Hydrosilylation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 361
Aleksandra Chechelska-Noworyta, Jan Mrówka, Maria Owińska
and Magdalena Hasik
13 Spectroscopic Aspects of Polydimethylsiloxane (PDMS) Used
for Optical Waveguides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 401
Dengke Cai and H. Michael Heise
14 The Luminescent Properties of Photonic Glasses and Optical
Fibers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 427
Jacek Zmojda, Marcin Kochanowicz, Piotr Miluski
and Dominik Dorosz
15 Spectroscopic Characterization of Silicate Amorphous
Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 457
Włodzimierz Mozgawa, Maciej Sitarz and Magdalena Król
16 Spectroscopy in the Analysis of Artworks . . . . . . . . . . . . . . . . . . . . 483
Tomasz Łojewski and Barbara Łydżba-Kopczyńska
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 519
Contributors
Rodolphe Antoine Institut Lumière Matière, UMR5306 Université Claude

Bernard Lyon1-CNRS, Université de Lyon, Villeurbanne Cedex, France
Dominika Augustynska Jagiellonian Centre for Experimental Therapeutics,
Jagiellonian University, Krakow, Poland
Olga Bąk Faculty of Chemistry, Maria Curie-Skłodowska University, Lublin,
Poland
Malgorzata Baranska Faculty of Chemistry, Jagiellonian University, Krakow,
Poland; Jagiellonian Centre for Experimental Therapeutics, Jagiellonian University,
Krakow, Poland
Bolesław Barszcz Faculty of Technical Physics, Institute of Physics, Poznan
University of Technology, Poznań, Poland; Institute of Molecular Physics Polish
Academy of Sciences, Poznań, Poland
Piotr Borowski Faculty of Chemistry, Maria Curie-Skłodowska University,
Lublin, Poland
Elisa M. Brás CQC, Department of Chemistry, University of Coimbra, Coimbra,
Portugal
Dengke Cai Laboratory of Micro-Structure Technology, Department of Electrical
Engineering and Information Technology, TU Dortmund University, Dortmund,
Germany
Aleksandra Chechelska-Noworyta Faculty of Materials Science and Ceramics,
AGH University of Science and Technology, Krakow, Poland
Krzysztof Czamara Faculty of Chemistry, Jagiellonian University, Krakow,
Krakow, Poland
ix
x Contributors
Dominik Dorosz Faculty of Materials Science and Ceramics, AGH University of

Science and Technology, Krakow, Poland
Monika Dudek Faculty of Chemistry, Jagiellonian University, Krakow, Poland
Rui Fausto CQC, Department of Chemistry, University of Coimbra, Coimbra,
Portugal
Krzysztof Halagan Faculty of Chemistry, Department of Molecular Physics, Lodz
University of Technology, Lodz, Poland
Magdalena Hasik Faculty of Materials Science and Ceramics, AGH University of
H. Michael Heise Interdisciplinary Center for Life Sciences, South-Westphalia
University of Applied Sciences, Iserlohn, Germany
Gulce O. Ildiz CQC, Department of Chemistry, University of Coimbra, Coimbra,
Portugal; Department of Physics, Faculty of Sciences and Letters, Istanbul Kultur
University, Istanbul, Turkey
Przemysław Jodłowski Faculty of Chemical Engineering and Technology,
Cracow University of Technology, Kraków, Poland
Agnieszka Kaczor Faculty of Chemistry, Jagiellonian University, Krakow,
Krakow, Poland
Marcin Kochanowicz Faculty of Electrical Engineering, Bialystok University of
Technology, Białystok, Poland
Andrzej Koleżyński Faculty of Materials Science and Ceramics, AGH University
of Science and Technology, Krakow, Poland
Marcin Kozanecki Faculty of Chemistry, Department of Molecular Physics, Lodz
University of Technology, Lodz, Poland
Magdalena Król Faculty of Materials Science and Ceramics, AGH University of
Joanna Łojewska Faculty of Chemistry, Jagiellonian University, Kraków, Poland
Tomasz Łojewski Faculty of Materials Science and Ceramics, AGH University of
Barbara Łydżba-Kopczyńska Faculty of Chemistry, University of Wroclaw,
Wroclaw, Poland
Ewa Machalska Faculty of Chemistry, Jagiellonian University, Krakow, Poland
Contributors xi
Kamilla Malek Faculty of Chemistry, Jagiellonian University, Krakow, Poland;

Jagiellonian Centre for Experimental Therapeutics, Jagiellonian University,
Krakow, Poland
Andrzej Mikuła Faculty of Materials Science and Ceramics, AGH University of
Piotr Miluski Faculty of Electrical Engineering, Bialystok University of
Włodzimierz Mozgawa Faculty of Materials Science and Ceramics, AGH
University of Science and Technology, Krakow, Poland
Jan Mrówka Faculty of Materials Science and Ceramics, AGH University of
Bernardo A. Nogueira CQC, Department of Chemistry, University of Coimbra,
Coimbra, Portugal
Animesh K. Ojha Department of Physics, Motilal Nehru National Institute of
Technology, Allahabad, India
Maria Owińska Faculty of Materials Science and Ceramics, AGH University of
Marcin Pastorczak Institute of Physical Chemistry, Polish Academy of Sciences,
Warsaw, Poland
Maciej Sitarz Faculty of Materials Science and Ceramics, AGH University of
Agata Stoch Institute of Electron Technology Krakow Division, Krakow, Poland
Paweł Stoch Faculty of Materials Science and Ceramics, AGH University of
Ewelina Szafraniec Faculty of Chemistry, Jagiellonian University, Krakow,
Poland
Szymon Tott Faculty of Chemistry, Jagiellonian University, Krakow, Poland
Ewelina Wiercigroch Faculty of Chemistry, Jagiellonian University, Krakow,
Poland
Danuta Wróbel Faculty of Technical Physics, Institute of Physics, Poznan
University of Technology, Poznań, Poland
Grzegorz Zając Faculty of Chemistry, Jagiellonian University, Krakow, Poland
Jacek Zmojda Faculty of Electrical Engineering, Bialystok University of
Chapter 1
Computational Methods in Spectroscopy
Abstract Spectroscopy investigates the interaction of electromagnetic radiation

with matter. Along with the development of theoretical methods, increasingly
effective numerical algorithms and computational methods as well as computer
technologies and resulting growing computer power available for scientists, the
so-called in silico experiments—computer simulations of materials and their prop-
erties in computer—have become an irreplaceable tool supporting experimental
research, often allowing a better understanding of phenomena taking place during
these interactions, and associated material properties. As a result, it becomes pos-
sible in growing number of cases to effectively design new materials with desired
properties and to modify existing ones, to improve their properties. This chapter is
devoted to a brief introduction to issues related to theoretical foundations of quantum
mechanics and density functional theory, both in stationary and time-dependent
form. The key assumptions of these theories are presented, together with the
description of various approximations and simplifications necessary for their prac-
tical application to the calculation of properties examined by spectroscopic methods.
The most important practical problems encountered during calculations, resulting
from the complexity of real materials and typical ways of dealing with these prob-
lems by means of various simplifications, idealizations, and abstractions in designed
structural models corresponding to real materials, are also presented.
1.1 Introduction
A broadly understood spectroscopy is devoted to the studies of the interactions of

electromagnetic radiation with matter. There are many different spectroscopic
techniques, but all of these techniques are based on the assumption that—when
certain conditions are fulfilled—each material emits or absorbs energy, usually in
A. Koleżyński (&)
Faculty of Materials Science and Ceramics, AGH University of Science
and Technology, 30 Mickiewicza Av., 30-059 Krakow, Poland
e-mail: andrzej.kolezynski@agh.edu.pl
© Springer Nature Switzerland AG 2019 1

A. Koleżyński and M. Król (eds.), Molecular Spectroscopy—Experiment
and Theory, Challenges and Advances in Computational Chemistry
and Physics 26, https://doi.org/10.1007/978-3-030-01355-4_1
2 A. Koleżyński
the form of electromagnetic radiation. The spectrum of electromagnetic radiation

extends in a very wide range of energy (and thus radiation frequencies). Various
spectroscopic techniques, when investigating the interaction of electromagnetic
radiation with a matter (absorption or emission, primary or secondary, by the matter
of a part of energy in the form of electromagnetic radiation), use different frequency
ranges of electromagnetic radiation, depending on the studied spectroscopic pro-
cesses and characteristic for them magnitude of the energy changes (in Fig. 1.1, a
schematic spectrum of electromagnetic radiation with the assignment of radiation
ranges to individual spectroscopic techniques is shown). For example, nuclear
magnetic resonance (NMR) technique [1–3] uses radio frequency radiation of about
108 Hz, recording changes in the spin of atomic nucleus (with very small energy
change) in atoms with a nonzero nuclear spin, e.g., 1H, 2H, 6Li, 7Li, 13C, 29Si.
Under the influence of the applied external magnetic field, there is a split of energy
levels in these atoms into two groups, depending on whether the nuclear spins are
directed parallel or antiparallel to the direction of the magnetic field. This level
separation is very small, of the order of 0.01 J/mol for the applied field of 1 T,
which corresponds to the radiation frequency of approximately 500 MHz. The
magnitude of energy changes and the associated frequency of electromagnetic
radiation strongly depends on the type of atom and its chemical environment, so
one can get a lot of relevant information about the structure of the material under
investigation.
The method closely related to the NMR is EPR/ESR method (electron param-
agnetic resonance/electron spin resonance) [4, 5]—the difference is that in this
method the change in the configuration of the electron spins, and not nuclei, is
studied. The EPR method is based on the presence of permanent magnetic dipoles,
i.e., unpaired electrons (as is the case, e.g., in many transition metals), and examines
the change in spins of such unpaired electrons under the influence of an external
magnetic field. Similarly to NMR, the observed energy changes are also small
(although about 2 orders of magnitude larger) and they are approximately 1 J/mol,
which corresponds to the microwave radiation frequency of approximately
3 1010 Hz. The obtained data allow analyzing direct environment of a given
atom and hence provide the information on the oxidation state, electron configu-
ration and coordination number of the paramagnetic ion, structural distortions due
Fig. 1.1 Main regions of electromagnetic spectrum with related spectroscopic techniques
1 Computational Methods in Spectroscopy 3
to, e.g., Jahn–Teller effect, or the degree of covalence of bonds formed by the ion
with surrounding atoms or ligands.
With the increase of radiation frequency, the energy of photons increases and
hence the possibility appears of studying phenomena in which energy changes are
higher and correspond to radiation with frequencies of about 1011–1013 Hz, e.g.,
molecular rotations and atomic vibrations in molecules and solids (the domain of IR
and Raman spectroscopy). While rotational spectra can only be obtained for
molecules (those with a permanent dipole moment) and thus provide the infor-
mation on the structure of such molecules (lengths of bonds and angles between
bonds), atomic vibrations are present in both, molecules and solids (amorphous and
crystalline), hence the vibrational spectroscopy (IR and Raman) [6] is a universal
method of studying the local structure of materials (force constants, bonds lengths,
angles, etc.) by analyzing normal modes of vibrations and changes in absorption
(IR) or scattering (Raman) of infrared radiation associated with structural changes
resulting, for example, from phase transitions or changes in chemical composition
(defects, admixtures, etc.).
The next group of spectroscopic methods is UV-Vis [7, 8] techniques related to
electronic transitions between different energy levels and the corresponding radi-
ation range, from near infrared, through visible region, to ultraviolet (about 1014–
1015 Hz). Various types of electronic transitions are possible, both between the low
lying localized levels of the internal shells of individual atoms and the levels of the
outermost valence shells (either localized on atoms or delocalized, forming bands).
Generally, there are four basic types of electron transitions:
• Excitation of an electron from a localized orbital to another localized orbital on
the same atom but with higher energy: The examples are d-d or f-f transitions in
compounds containing transition metals, transitions on outer shells in heavy
metals compounds (e.g., 6s–6p in lead(II) compounds), or transitions related to
defects (trapped electrons or holes, color centers);
• Excitation of an electron from a localized orbital on one atom to a higher energy
orbital localized on an adjacent atom (charge transfer processes, e.g., from
oxygen to chromium in tetrahedral [CrO4]2− anion);
• Excitation of an electron from a localized orbital on an atom to a delocalized
conduction band (e.g., photoconductive materials);
• Electron transitions from valence band to conduction band (e.g., in
semiconductors).
UV-Vis spectroscopy has a very wide range of applications related to the study
of the local structure of materials, because the positions of absorption bands are
very sensitive to the immediate surroundings and the nature of chemical bonding.
For example, by adding a small amount of spectroscopically active elements, such
as transition metals or p-block heavy metals (e.g., Pb2+), a number of information
on the local structure of amorphous materials can be obtained. Materials for lasers
often contain transition metal ions as active species, e.g., Al2O3 doped with Cr3+
(ruby laser) or Nd3+-doped glass (neodymium laser), and UV-Vis spectroscopy
4 A. Koleżyński
allows obtaining detailed knowledge about energy levels and possible transitions—
a key information from the point of view of effective laser working. UV-Vis
spectroscopy is also widely used in analytical chemistry, e.g., to determine the
concentration of ions, molecules, or functional groups responsible for individual
absorption peaks.
Electron spectroscopy methods [9–12] (XPS—X-ray photoelectron spec-
troscopy, UPS—ultraviolet photoelectron spectroscopy, or AES—Auger electron
spectroscopy) measure the kinetic energy of electrons emitted from the material as a
result of exposure to high energy ionizing radiation or high energy particles
bombardment. The main difference between XPS and UPS methods is in the atomic
shells which can be studied: The inner shells electrons can be ejected using higher
energy X-rays (frequencies around 1018 Hz), while valence shells electrons can be
ionized using lower energy ultraviolet light with the frequencies about 1016 Hz. The
kinetic energy of electrons measured by these methods is equal to the difference
between the energy ht of the used incident radiation and the binding energy
(ionization potential) of the emitted electron and depends strongly on the type of
atom from which the electron was emitted, its oxidation state, and its immediate
environment; whereby it is possible to use these techniques to identify the atoms of
a given element in the material, providing the basis for electron spectroscopy for
chemical analysis (ESCA) developed by Siegbahn et al. [13].
The method directly related to the XPS and UPS spectroscopy is the Auger
electron spectroscopy [11, 14], in which emitted electrons are not the electrons
primary ionized by X-ray or ultraviolet radiation, but the product of secondary
processes, i.e., decay of ionized atoms from excited states to lower energy states
associated with emission of a photon, which can be emitted (and measured by XPS
or UPS) or absorbed by the electron from outer shells, resulting in the emission of
such secondary ionized electron (called the Auger electron). Electron spectroscopy
methods are very important and useful methods to determine energy levels in atoms
and molecules. In the case of solids, they are particularly useful for studying the
local structure of the surface of materials, since due to the relatively low energy of
emitted electrons and their strong dissipation in the material, in practice it is pos-
sible to observe only electrons ejected from the surface layer with a thickness of up
to 5 nm.
Even higher energies than those used in electron spectroscopy require processes
of transitions between energy levels of atomic nuclei (around 1019 Hz). The method
that examines such processes is the Mössbauer spectroscopy [15, 16], based on the
Mössbauer effect (i.e., resonant and recoil-free emission and absorption of c-rays by
atomic nuclei in a solid). In Mössbauer spectroscopy, three types of nuclear
interactions are observed:
• Isomer shift (energy shift of the absorption peak resulting from the modification
of the energy levels of the nucleus by changes in the distribution of electron
density in the atom, due to the changes in the local environment of the atom; this
shift may, e.g., correlate with the electron density at atomic nucleus of the outer
s shells electrons);
• Quadrupole splitting (in the case of nuclei with nuclear spin I > ½, the distri-
bution of positive charge in the nucleus is not spherically symmetrical resulting
in quadrupole moment Q, which in turn results in the split of the nuclear energy
levels and hence the split of the corresponding absorption bands in the
Mössbauer spectrum);
• Magnetic hyperfine Zeeman splitting (when the atomic nucleus with spin I will
be subjected to the magnetic field, then each nuclear energy level splits into
2I + 1 levels).
The source of radiation in this method is a highly monochromatic gamma-ray
beam whose energy is changed using the Doppler effect: The sample is fixed and
the radiation source moves at a constant velocity toward or away from the sample,
and as a result, the energy of radiation incident on the sample increases or decreases
with respect to the primary energy of radiation. In this way, an absorption spectrum
of gamma radiation is created, the analysis of which allows obtaining information
about the local structure (oxidation state, coordination number, nature of binding,
etc.) and precisely measuring—due to a high energy and extremely narrow line
widths of c-rays—small energy changes in nuclei, atoms, and crystals induced by
electrical, magnetic, or gravitational field. The studies of hyperfine splitting,
especially as a function of temperature, allow additionally obtaining information on
the ordering of magnetic materials.
Theoretical studies on the above-mentioned spectroscopic methods are cur-
rently conducted in practice only with the help of methods based on the for-
malism of quantum mechanics. In practice, only in the case of vibrational
spectroscopy, it is possible to use classical methods (both stationary-molecular
mechanics and dynamical—molecular dynamics), in which atoms are treated as
charged particles, and their interactions are described by means of analytic two
and more center functions, the so-called interatomic potentials. It is then possible
to calculate the total potential energy of such a classical system, hence the forces
acting on individual atoms and their minimization (relaxation of the structure),
and then calculating (for the relaxed structure) the frequencies of normal modes
of vibrations and simulation of vibrational spectra. However, due to the lack of a
correct description of the electron density, such an approach does not allow to
analyze the processes dealt with by other spectroscopic methods mentioned
above, and therefore, in remaining parts of this chapter, classical methods will not
be further discussed.
1.2 Theoretical Foundations for Modeling of Real Systems

and Processes Studied by Spectroscopic Methods
The physical reality that surrounds us is extremely complex and impossible to be

accurately described using available physical theories. This applies in particular to
the micro-world and the structure and properties of materials of various sizes and
6 A. Koleżyński
level of complexity, starting with the simplest atoms, through molecules and
clusters, to amorphous and periodic solids. In the past, experimental studies were
the main and, in some cases, practically the only source of knowledge about the
properties of materials. The development of computer technology and theoretical
methods leads gradually to the increasing use of computer calculations and simu-
lations (the so-called in silico experiments) in materials research, which in many
cases becomes an equivalent (and sometimes also the first, preliminary) tool
allowing understanding of the properties of materials and their modification and
functionalization.
Among the theoretical methods used in material research, the so-called ab initio
methods (methods from the first principles), whose basic idea is to treat
multi-atomic systems as systems of many bodies composed of atomic nuclei and
electrons and to analyze them employing quantum mechanics—the arguably most
important contemporary physical theory, without using any empirical parameters.
In the following few sections, the most important ingredients of quantum
mechanics formalism as well as the density functional theory, rooted in this
formalism (and nowadays the dominant approximation used in theoretical cal-
culations), required for their practical application in modeling of the properties of
complex many-body systems will be presented. Apart from the above-mentioned
“static” methods, i.e., methods based on the solutions of appropriate equations
formulated for stationary states, there are also other, computationally much more
complex and demanding, dynamic (ab initio molecular dynamics allowing to
study the changes of system state with time) and statistical approaches (ab initio
quantum Monte Carlo, allowing going beyond the mean field theory and
independent electrons approximations and direct treatment and description of
complex interactions in many-body systems). However, due to the limited volume
of the chapter and still relatively rare use of these methods, they will not be
presented here and interested readers are referred to the rich literature devoted to
this topic [e.g., 17–21].
1.2.1 Ab Initio Methods
1.2.1.1 Schrödinger Equation
The physical theory describing the phenomena occurring in the micro-world, and
therefore also the processes studied by various spectroscopic methods, is quantum
mechanics. The most popular formulation of quantum theory is the formalism of
wave mechanics proposed by Erwin Schrödinger [22–27]. The starting point in
Schrödinger’s formulation is a number of postulates regarding the wave function of
complex values (representing the state of the system), linear Hermitian operators
and their eigenvalues (mathematical representations of observable quantities, the
so-called observables), a recipe enabling finding an operator associated with a given
observable, a description of the measurement process and a way of predicting its
possible results (measurement problem, collapse of the wave function), which

culminates in the formulation of the equation describing the evolution of the state of
the system, the so-called time-dependent Schrödinger equation (TDSE):
@Wðr; tÞ ^
ih ¼ HWðr; tÞ ð1:1Þ
@t
where W—wave function describing the state of the system and h—Planck’s
^
constant h, divided by 2p. The Hamiltonian H of the system (the total energy
operator) used in the Schrödinger equation is defined as:
2 2
^ ¼ h @ þ @ þ @ þ Vðr; tÞ ¼
h2 2
2 2
^ ¼ T^ þ V
H r þ Vðr; tÞ ð1:2Þ
2m @x2 @y2 @z2 2m
and the time-dependent Schrödinger equation takes its final form:
@Wðr; tÞ h2 2
ih ¼ r Wðr; tÞ þ V ðr; tÞWðr; tÞ ð1:3Þ
@t 2m
In the vast majority of cases, when we model a given system we are interested in
its properties at stationary state (ground or excited), which is why we are usually
interested in solving the so-called time-independent Schrödinger equation (TISE). If
we assume that in the stationary state the external potential is constant, independent
of time V ðr; tÞ ¼ VðrÞ, then we can perform a mathematical operation of separating
the variables and write the function of the state of the system as:
Wðr; tÞ ¼ wðrÞTðtÞ ð1:4Þ
The Schrödinger equation then takes the form:

h2 2 @
h ½wðrÞTðtÞ
r þ VðrÞ wðrÞTðtÞ ¼ i ð1:5Þ
2m @t
which after the transformation gives the equation:
h2 1 1 dTðtÞ
r2 wðrÞ þ VðrÞ ¼ i
h ð1:6Þ
2m wðrÞ TðtÞ dt
true for all r and t, so both sides must be equal to constant E, called the separation
constant. In this way, we obtain two independent equations describing the depen-
dence of the wave function, first on the position:
h2 1 ^
r2 wðrÞ þ VðrÞ ¼ E ) HwðrÞ ¼ EwðrÞ ð1:7Þ
2m wðrÞ
8 A. Koleżyński
and second on time:
1 dT ðtÞ
ih ¼E ð1:8Þ
T ðtÞ dt
The solution of the latter has the form T ðtÞ ¼ AeiEt=h , where E, is energy of
stationary state, one of the eigenvalues of the Hamiltonian operator defined for the
system in time-independent (stationary) state. Finally, the solution to the
Schrödinger equation has the form:
Wðx; tÞ ¼ wðxÞTðtÞ ¼ wðxÞeiEt=h ð1:9Þ
One can easily see that although the potential of V(r) is independent of time, the
wave function oscillates in time with the frequency depending on the energy cor-
responding to the current stationary state of the system (however, at the same time,
the probability distribution Pðx; tÞ ¼ jWðx; tÞj2 ¼ jwðxÞj2 , and therefore also the
distribution of total electron density, is constant in time).
Since the solution of the time-dependent part of Schrödinger equation is identical
in each case, in order to obtain full information about the system’s state and its
evolution over time, it is sufficient to find a solution for time-independent
Schrödinger equation, of the general form:
^
Hwðr 1 ; r2 ; . . .; rn ; Ra ; Rb ; . . .; Rm Þ ¼ Ewðr1 ; r2 ; . . .; ri ; Ra ; Rb ; . . .; Rm Þ ð1:10Þ
where ri and Ra—the coordinates of electrons and atomic nuclei, respectively, n—

number of electrons, m—number of atomic nuclei, and respective total energy
operator assumes the form:
X 1 h2 X 2 X X Za Zb e2
b ¼ h
2
b ¼T
H bþV r2a ri þ
2 a
ma 2me i a a[b
rab
ð1:11Þ
X X Za e2 X X e2
þ
a i
ria j
r
i [ j ij
Except for a few simplest cases, finding an exact, analytical solution of this
equation is impossible, and thus, some simplifications are usually necessary to be
used and solutions are obtained in numerical form.
1.2.1.2 Born–Oppenheimer Approximation, Potential Energy

Hypersurface
The first simplification, bringing us closer to the solution of the time-independent

Schrödinger equation for a multi-electron system, is called Born–Oppenheimer
approximation (BOA). This approximation is rooted directly in the simple observation
that since nuclei are much heavier then electrons (approximately 103–105 times,
depending on the number of protons and neutrons making up particular nucleus), they
move much slower than electrons which have a response time of the order of fem-
toseconds and can adjust themselves almost instantaneously to the new atomic con-
figuration (thus—from electrons’ point of view—nuclei seem to be clamped and the
idea goes, we can treat the motions of electrons and nuclei as independent). In this
case, the effective interactions between the atoms take place via electrons, and we
carry out the calculations looking for a solution for the ground state of the electron
cloud, for the assumed, fixed spatial configuration of motionless atoms. The
assumption of the independent movement of the atomic nuclei and electrons allows
separation of the variables describing their positions and writing the multi-electron
wave function as a productof two functions: one dependent only on nuclei positions
(nuclei wave function wN Ra ; Rb ; . . .; Rm ) and the other on positions of electrons

(electronic wave function wel r1 ; r2 ; . . .; rn ; Ra ; Rb ; . . .; Rm ) and current geometry
(nuclei positions):

w r1 ; r2 ; :::; rn ; Ra ; Rb ; :::; Rm ¼ wel r1 ; r2 ; :::; rn ; Ra ; Rb ; :::; Rm wN Ra ; Rb ; :::; Rm
ð1:12Þ
This enables us to solve the electron and nuclear Schrödinger equation sepa-
rately, starting with the one for electrons:
^ el wel ðr; RÞ ¼ Eel wel ðr; RÞ
H ð1:13Þ
where
2 X X X Za e2 X X e2
^ el ¼ h
H r2i þ ð1:14Þ
2me i a i
ria r
j i [ j ij
The Born–Oppenheimer approximation leads to a very important concept—the

potential energy hypersurface U(R):
X X Za Zb e2
UðRÞ ¼ Eel þ VNN ¼ Eel þ ð1:15Þ
a a[b
rab
Once we have the potential energy surface (PES), we can solve the nuclear
Schrödinger equation:
^ N wN ðRÞ ¼ E N wN ðRÞ
H ð1:16Þ
where
X 1
^ N ¼ h
2
H r2 þ U ðRÞ ð1:17Þ
2 a
ma a
10 A. Koleżyński
The solution of this equation allows obtaining information on the quality of the
initially adopted geometry of the studied system (spatial configuration of atoms).
Usually, the assumed geometry deviates from the equilibrium geometry and the
atoms are subjected to the forces that can be easily calculated (as a potential energy
gradient with respect to the atomic position). Information about forces acting on
atoms allows the geometry modification (optimization) by changing the positions of
all atoms according to the direction and magnitude of the respective forces and
repeating the total energy calculations, another forces estimation, geometry modi-
fication according to new forces calculated, and repeating the whole procedure until
the calculated forces on the atoms are smaller than the assumed precision. In
most cases, the relaxation of the structure (minimization of forces on atoms) is a
necessary step in the calculation, because the majority of physical quantities that are
of interest is defined for systems that are in the minimum of potential energy well
(relaxed, equilibrium geometry).
1.2.1.3 Hartree–Fock Method and Post-HF Extensions
The solution of the nuclear time-independent Schrödinger equation allows deter-

mination of a large variety of molecular properties (e.g., vibrational energy levels,
phonon dispersion in crystals). But this is just the first step, since by employing
Born–Oppenheimer approximation we have simplified our Hamiltonian a little; but
we still have to solve electronic TISE for our multi-electron system, which is
impossible to be done exactly (the reason being the last term in electronic
Hamiltonian, namely electron–electron interactions) and we need to make further
approximations. Since we cannot calculate exactly very complicated electron–
electron interactions, let us assume that electrons are independent from each other
and every single electron is moving in an effective single-particle potential due to
all fixed nuclei and the average electron density distribution of all remaining
electrons. This simple idea, called independent electrons approximation (IEA) , was
first formulated by Hartree [28] and led to the development of famous Hartree–
Fock SCF method, consisting of a set of self-consistent single-particle equations
with multi-electron wave function defined as a product of one-electron wave
functions (Hartree product).
wel ðr1 ; r2 ; . . .; rN Þ ¼ /1 ðr1 Þ/2 ðr2 Þ /N ðrN Þ ð1:18Þ
Since such simple product does not fulfill the requirement for electronic wave
function of the system to be antisymmetric (due to Pauli’s Exclusion Principle,
which electrons—as indistinguishable fermions—have to obey) while the linear
combination of different products does, the original Hartree product was replaced
later by antisymmetric-by-definition Slater determinant (with (spin)orbitals as its
elements) of the form:

/1 ðr1 Þ /1 ðr2 Þ /1 ðrN Þ

1 /2 ðr1 Þ /2 ðr2 Þ /2 ðrN Þ
wel ðr1 ; . . .ri ; . . .; rj ; . . .; rN Þ ¼ pffiffiffiffiffi . .. .. ð1:19Þ
N! .. . .

/ ðr1 Þ /N ðr2 Þ /N ðrN Þ
N
In the general case, the wave function for the N-electron system can be written in
the form of a linear combination of Slater determinants (including in the determi-
nants also partially occupied spinorbitals for excited states, empty in the ground
state):
X
wel ¼ di Ui ¼ d0 U0 þ d1 U1 þ d2 U2 þ ð1:20Þ
i
^ el of the form (1.14)

A more detailed analysis of the electron energy operator H
shows that it consists of two types of contributions:
• One-electron contribution (depending on the coordinates of only one electron):
P
n P
m
^hi where ^hi ¼ Tê þ V
^Ne ¼ h2 r2i Za e 2
ð1:21Þ
2me ria
i¼1 a¼1
• Two-electron contribution (depending on the coordinates of two electrons):
X X e2
^ ee ¼
V ð1:22Þ
j
r
i [ j ij
The former can be calculated accurately, whereas the two-electron contributions

can only be approximated, due to the complexity of interelectronic interactions. The
potential energy associated with two-electron interactions is equal to:
n X
X n
Vee ¼ Jee Kee ¼ ðJij Kij Þ ð1:23Þ
i j[i
where Jij is a Coulomb integral defined as:

ZZ
1
Jee ¼ /2i ð1Þ /2j ð2Þdr1 dr2 ð1:24Þ
r12
describing Coulomb repulsion between electron 1 on orbital i and electron 2 on

orbital j, while Kij is the exchange integral (nonzero only for like-spin electrons),
12 A. Koleżyński
with no classical interpretation—this integral is entirely a quantum mechanical

effect resulting from fermions’ antisymmetry.
ZZ
1
Kee ¼ /i ð1Þ/j ð2Þ /i ð2Þ/j ð1Þdr1 dr2 ð1:25Þ
r12
In order to be able to solve this Schrödinger equation, we must replace the

êe with an approximate, effective, one-electron operator
electron–electron operator V
êff :
V
X
n
êff ðiÞ ¼
V 2^Jj ðiÞ K
^ j ðiÞ ð1:26Þ
j
where
Z
^Jj ð1Þ ¼ 1
uj ð2Þ u ð2Þdr2 ð1:27Þ
r12 j
Z
^ j ð1Þui ð1Þ ¼ 1
K uj ð2Þ u ð2Þdr2 uj ð1Þ ð1:28Þ
r12 i
and as a result, the so-called one-electron Fock operator is obtained:
X
^fi ¼ h r2
2 m
Za e2 ^
þ V eff ðiÞ ð1:29Þ
2me i a¼1
ria
Using variational principle:

^ trial
wtrial Hw ^ trial
wtrial Hw
hEtrial i ¼ Eexact ; d ¼0 ð1:30Þ
wtrial wtrial wtrial wtrial
to the problem of energy minimization, we get a set of appropriate one-electron

Hartree–Fock equations:
^fi /ðiÞ ¼ ei /ðiÞ; i ¼ 1; 2; . . .; N ð1:31Þ
One-electron Fock operator ^fi depends, through effective potential V

êff ðiÞ, on the
one-electron wave functions /ðiÞ of all other electrons, and the Hartree–Fock
equations, despite the form resembling a classical eigenvalue problem, are not
possible to be solved directly; hence, it is necessary to perform calculations itera-
tively (using the so-called self-consistent field—SCF method):
• A set of one-electron initial functions (orbitals) is defined and respective Fock
operator is calculated;
• Next, the system of Hartree–Fock equations is solved;

• The set of new orbitals being the solution of the HF equations is used in the next
cycle of calculations to determine a new Fock operator;
• The system of Hartree–Fock equations with the new Fock operator is solved;
• The last two steps are repeated until the orbitals used to define the Fock operator
and the orbitals obtained as a result of solving the system of HF equations with
this operator (or total energies) differ by less than the predetermined, required
maximum error (i.e., until we get a self-consistent solution).
By applying these two approximations (BOA and IEA) to initial TISE, we have
made a big step toward the practical applicability of QM to real systems. But one
thing remains unresolved—we have assumed electrons to be independent particles,
but we have to define yet the functions describing (spin)orbitals in Slater deter-
minant—the state vectors of these electrons. Since we do not know the exact form
of such functions, we are forced to make some assumptions about their shape and
usually define these functions as a linear combination of some analytical functions
and search for the best form (best expansion coefficients for chosen set of analytical
functions) employing variational principle. In principle, one can choose any
functions, but in practice, due to the numerical efficiency, only limited set of various
analytical functions is used—depending on the system of interest, the functions
which “mimic” the electron distribution of real system best, are chosen, since this
can greatly decrease the number of calculation cycles and thus the time necessary to
carry out the calculations. For molecular system, the usual choice are atomic
orbitals (AO)—one can safely assume that when atoms approach each other and
start to form the molecule, the resulting electron density distribution will change,
but will resemble, even in case of valence electrons, the original atomic distribution
and thus the linear combination of such atomic orbitals, defining the shape of
molecular orbital should be the most efficient one in molecular system. But again—
since the exact form of atomic orbitals is not known, one can further approximate
AO by means of linear combination of some simple analytical functions. Originally
Slater-type orbitals were used, but later, due to their numerical inefficiency, they
were substituted in most quantum chemical codes by Gaussian-type orbitals,
functions poorly describing the density distribution (so we need more of them in
linear combination), but definitely more efficient numerically, which more than
enough compensates their inefficient shape. Such Gaussian-type orbitals can be
single Gaussian functions (primitives) or their linear combination (contracted).
Currently, many predefined basis sets (depending on a definition of atomic orbitals
by means of Gaussian functions) is available, and the user is responsible for a
choice of the basis set (theory level) best suited for the system being studied. In case
of periodic solids, where one can find both extreme cases (metals and molecular
crystals) and a whole lot of intermediate ones, electrons can be described by plane
waves (best suited for metals) or atomic orbitals (best for molecular crystal). One
can use either of them (having in mind that their efficiency will be quite different
depending on the crystal studied) or some kind of hybrid functions like (L)APW,
ASW, (L)ASO, (L)MTO, to name the few—see, e.g., [29] for more details.
14 A. Koleżyński
Since the original HF method does not (by definition) take into account electron
correlation effects, in order to get the results in better agreement with experiment, a
few extensions, the so-called post-HF methods, have been proposed in following
years, e.g., CI—Configuration interaction [30–32], MP2 [33, 34], MP3 [35, 36],
and MP4 [37] Møller–Plesset Perturbation Theory, CC—Coupled Cluster Theory
[38, 39], QCI—Quadratic Configuration Interaction [40], CASSCF—Complete
Active Space Self-Consistent Field calculation [41, 42], MRSDCI—
Multi-Reference Single and Double Configuration Interaction [43] or Quantum
Chemistry Composite methods, like Gaussian G1 [44, 45], G2 [46], G3 [47], and
G4 [48], FPD Feller–Peterson–Dixon approach [49, 50], ccCA—correlation con-
sistent Composite Approach [51], or CBS—Complete Basis Set methods [52]. Such
extensions improved considerably the quality of calculations, but at the expense of
significantly higher demand for computer resources and calculation time (in the
worst-case scenario, calculation time increases fivefold with the size of the system,
comparing to original HF method).
1.2.1.4 Density Functional Theory
HF method with post-HF extensions is still commonly used (especially by chemists

interested in molecular compounds), but in recent years the Density Functional
Theory (DFT) approach became to dominate in computational physics and in great
part of computational chemistry. The origins of density functional theory go back to
the first years of the twentieth century and the works of Thomas [53] and Fermi [54],
describing the properties of multi-electron systems using total electron density,
extended in following years by Dirac [55] (exchange energy term) and Weizsäcker
[56] (first-order correction for kinetic energy functional). These attempts turned out,
however, to be unsatisfactory, due to, inter alia, approximate character of the
expression for kinetic and exchange energy and complete neglect of electron corre-
lation resulting in fundamentally wrong description of chemical bonding: The cal-
culated total energy of any molecule was higher than the sum of the energies of
isolated atoms, and therefore, Thomas–Fermi–Dirac model predicted molecules to be
unstable [57]. The problem of inaccurate description of kinetic energy in the Thomas–
Fermi–Dirac–Weizsäcker model was circumvented in Kohn–Sham’s Density
Functional Theory (with electron density as a key variable in the description of the
properties of non-uniform electron gas in multi-electron systems) for which firm
mathematical foundation was given in 1964 in two Hohenberg–Kohn theorems [58].
The starting point in density functional theory is the Born–Oppenheimer
approximation, resulting in constant external Coulomb potential Vext(r), generated
by motionless atomic nuclei, in which all electrons move:
X Za e2
Vext ðrÞ ¼ ð1:32Þ
a
j r ra j
Hence, the system Hamiltonian is defined as the sum of two operators:

X 1XX 1
^ ¼ 1
F r2i þ ð1:33Þ
2 i
2 i j6¼i ri rj
And
X
êxt ¼
V Vext ðri Þ ð1:34Þ
i
Functional F^ is the same for each N-electron system; therefore, the Hamiltonian
and the ground state W are completely defined by N and Vext(r), and the W
eigenstate corresponds to the electron density n(r), expressed as:
Z
nðrÞ ¼ N jWðr; r2 ; . . .; rN j2 dr2 . . .drN ð1:35Þ
According to the first Hohenberg–Kohn theorem, the ground state properties of

such multi-electron system are uniquely determined by the total electron density
and there is an unequivocal mapping Vext(r) $ n(r) (one-to-one correspondence)
between the external potential Vext(r) and the total electron density of the ground
state n(r) (and therefore, the electron density of the ground state is uniquely
determined by the external potential and at the same time this potential is uniquely
determined by the total electron density distribution of the ground state).
The second Hohenberg–Kohn theorem states that the total energy of the system
of N interacting electrons, subjected to the external potential Vext(r), is a functional
of the total electron density n(r):
Z
E½nðrÞ ¼ F½nðrÞ þ nðrÞVext ðrÞdr ð1:36Þ
and reaches the minimum for the total N-electron density of the ground state in the
external potential Vext(r). The ground state energy of such a system can be calcu-
lated using the variational method, since for each density n′(r) that is not eigenstate
of the external potential Vext(r), Ev[n′(r)] > Ev[n(r)], where the electron density that
minimizes the total energy of the system is the exact density of the ground state
(however, this is true only for the exact functional Ev[n(r)]).
An extraordinary result of DFT is the demonstration of the existence of a uni-
versal functional F[n(r)], independent of external potential, which means that instead
of dealing with the N-electron wave function, we operate on the electron density, i.e.,
three-dimensional spatial coordinates function. The level of complexity of the
problem is thus dramatically reduced and scales linearly with the size of the system
(compare the HF method, where the problem scales with second– third power).
Unfortunately, the exact form of the functional F[n(r)] is unknown and the practical
16 A. Koleżyński
problem faced by DFT (finding an approximate function) results directly from the
complexity of the problem of interacting electrons in many-body system.
Kohn and Sham [59] have reformulated the above problem by replacing the
system of N interacting electrons with a special reference system—the system of
N non-interacting electrons, having the same ground electron density. In the case of
such a reference system, this electron density is equal to the sum over the
one-electron wi(r) orbitals (Kohn–Sham orbitals):
X
N=2
nðrÞ ¼ 2 jwi ðrÞj2 ð1:37Þ
i¼1
being a solution to the following Schrödinger equations:

1 2
r þ VKS ðrÞ wi ðrÞ ¼ ei wi ðrÞ ð1:38Þ
2
and satisfying the orthonormality condition:

Z

w ðrÞw ðrÞdr ¼ dij ð1:39Þ
i j
The existence of a uniquely determined VKS(r) potential for the ground state
electron density n(r) is a direct consequence of the first Hohenberg–Kohn theorem.
Therefore, we have a situation in which the problem of finding a universal func-
tional F[n(r)] was replaced by the problem of finding an effective VKS[n(r)] func-
tional. This problem can be solved using the variational principle. Lets write down
the energy in the form:
Z
E ¼ TS ½nðrÞ þ EH ½nðrÞ þ EXC ½nðrÞ þ nðrÞVext ðrÞdr ð1:40Þ
The first part is the kinetic energy of the system of non-interacting electrons:
Z
h2 X
N=2
TS ½nðrÞ ¼ 2 wi ðrÞr2 wi ðrÞdr ð1:41Þ
2m i¼1
The second is Hartree energy, classic electrostatic energy:

Z
e2 nðrÞnðr 0 Þ
EH ½nðrÞ ¼ drdr 0 ð1:42Þ
2 jr r 0 j
and the last term is the classical Coulomb energy of electron interactions with
external potential. The key element in the expression for the energy of the system is
the third term whose exact form is unknown—this is the exchange and correlation
energy EXC[n(r)], containing non-classical electrostatic interactions and the differ-

ence in kinetic energy of the interacting and non-interacting particles systems. The
purpose of such energy partitioning is to separate the first two and the last, which
can be calculated accurately, giving together most of the total energy and third term,
a small part of the total energy, which contains the complex quantum effects of
multi-electron interactions (here our entire ignorance about the system is hidden).
According to the second H-K theorem, the energy of the ground state can be
calculated by minimizing the energy functional. The appropriate variational prob-
lem for the H-K density functional, when imposing the constraints on the number of
electrons N, takes the form:
R R
d F½nðrÞ þ nðrÞVext ðrÞdr k nðrÞdr N ¼ 0;
ð1:43Þ
F½nðrÞ ¼ Ts ½nðrÞ þ EH ½nðrÞ þ EXC ½nðrÞ
where k is undetermined Lagrange multiplier associated with constraints on the

number of electrons N, and the corresponding Euler equation has the form:
dF½nðrÞ
l¼ þ Vext ðrÞ ð1:44Þ
dnðrÞ
In Kohn–Sham’s formulation, respective Euler’s equation transforms into:
dTS ½nðrÞ
l¼ þ VKS ðrÞ ð1:45Þ
dnðrÞ
where
VKS ðrÞ ¼ Vext ðrÞ þ VH ðrÞ þ Vxc ðrÞ ð1:46Þ
and
Z
dEH ½nðrÞ nðrÞ dEXC ½nðrÞ
VH ðrÞ ¼ ¼ dr 0 ; VXC ðrÞ ¼
dnðrÞ jr r 0 j dnðrÞ ð1:47Þ
Both Euler equations are equivalent, which means that the density obtained as a
solution to the variational problem for the KS reference system is identical to the
density obtained for the original system of N interacting electrons and can be in
practice obtained by solving a system of N one-electron Schrödinger equations
(Kohn–Sham equations):

1
r2 þ VKS ðrÞ wi ðrÞ ¼ ei wi ðrÞ ð1:48Þ
2
18 A. Koleżyński
Since the Kohn–Sham potential VKS(r) depends on the total electron density n
(r), dependent on the Kohn–Sham one-electron orbitals wi(r), which in turn depend
on the VKS(r) potential, such system of equations (in analogy to the Hartree–Fock
equation system) must be solved in a self-consistent way. Since we do not know the
exact form of the exchange–correlation potential, it is necessary to define it in an
approximate form. The exchange–correlation functional can be written down as:
Z
EXC ½nðrÞ ¼ nðrÞeXC ðrÞdr ð1:49Þ
where eXC ðr Þ is exchange–correlation energy density. Based on the way in which

the density surrounding each electron is sampled in order to define eXC ðr Þ, the
approximations of exchange–correlation functionals can be divided into five dis-
tinct classes: Local Density Approximation (LDA), Generalized Gradient
Approximation (GGA), Meta-Generalized Gradient Approximation (MGGA),
Hybrid Functionals and Non-Local Functionals.
Historically first and the simplest approximation of the exchange–correlation
functional was the Local Density Approximation proposed by Kohn and Sham in
their original work [59]. The exchange–correlation energy EXC ½nðr Þ can be
expressed within this approximation as:
Z
EXC ½nðrÞ ¼ nðrÞe0XC ½nðrÞdr ð1:50Þ
where e0XC ½nðr Þ is the exchange–correlation energy density of a homogeneous

electron gas, usually written as separate expressions for the exchange and corre-
lation part:
e0XC ½nðrÞ ¼ e0X ½nðrÞ þ e0C ½nðrÞ ð1:51Þ
The first term, exchange energy, was provided in 1928 in analytic form by Dirac
[55]:
Z
eX ½nðrÞ ¼ Cx nðrÞ ) EX ½nðrÞ ¼ Cx nðrÞ4=3 dr
0 1=3 LDA
ð1:52Þ
but for the second (correlation) term, the exact form of the functional is unknown.
There are, however, exact fittings to the results obtained by Ceperley and Alder
using the Monte Carlo method for a homogeneous electron gas [60], e.g., analytic
expression parameterized by Perdew and Zunger in a form [61]:
pffiffiffiffi
0:9164=rs 0:2846=ð1 þ 1:0529 rs þ 0:3334rs Þ; rs 1
eXC ½nðrÞ ¼
0:9164=rs 0:0960 þ 0:0622 ln rs 0:0232rs þ 0:0040rs ln rs ; rs \1
ð1:53Þ
where
sffiffiffiffiffiffiffiffiffiffiffiffiffiffi
3
rs ¼ ð1:54Þ
3
4pnðr Þ
Often, in order to increase the flexibility of the approximate functional, the

procedure (somewhat artificial, because the exact functional depends only on the
total density) of separation of the density into two parts, depending on the spin, is
employed:
nðrÞ ¼ na ðrÞ þ nb ðrÞ ð1:55Þ
In this case, the appropriate XC functional reads:

Z
LSDA
EXC ½na ðrÞ; nb ðrÞ ¼ nðrÞe0XC ½nðrÞ; fðrÞdr ð1:56Þ
where fðrÞ is a spin polarization parameter, defined as:
na ðrÞ nb ðrÞ
fðrÞ ¼ ð1:57Þ
na ðrÞ þ nb ðrÞ
The value of fðrÞ equal to 0 means paramagnetic state (full spin compensation),
while equal to ±1 ferromagnetic state (full polarization) when one of spin densities
vanishes completely. As before, the explicit form of the functional is known only
for the exchange part.
The accuracy of L(S)DA approximation is often considered sufficient in solids,
but much less so in molecular systems. L(S)DA gives to high energy, overestimates
cohesion energy (over 20%), underestimates unit cell parameters (and thus bonds
lengths), and in many cases wrongly predicts phase stability. These defects can be
largely removed by introducing gradient corrections. In this case, the XC functional
is defined as a function of the local density (information on the density at a given
point) and its gradients (information on how the density changes near this point):
Z
GGA
EXC ½nðrÞ ¼ XC ½nðrÞ; rnðrÞdr
nðrÞeGGA ð1:58Þ
Also in this case, we can split this functional into two parts, exchange and
correlation:
XC ½nðrÞ ¼ eX
eGGA ½nðrÞ þ eGGA ½nðrÞ ð1:59Þ
GGA
C
There is a number of different GGA functionals, but most often just a few are
used, e.g., PW91 [62], PBE [63], (with different LDA correlation functional pos-
sible, i.e., VVN [64], PZ [61] or PW [65]), PBESol [66] (optimized for solid state
20 A. Koleżyński
and surface systems), BLYP (with Becke exchange [67] and Lee–Yang–Parr cor-
relation [68]) or WC [69].
The next step (meta-GGA approximation) is the addition of consecutive
semi-local information (defined locally at the point r), e.g., higher density deriva-
tives. Meta-GGA functionals [70] can be written in the following general form:
Z
mGGA
EXC ½nðrÞ ¼ f ½nðrÞ; rnðrÞ; r2 nðrÞ; g1 ðrÞ; g2 ðrÞ; g3 ðrÞ; . . .dr ð1:60Þ
where g1(r), g2(r), g3(r), … are other semi-local quantities that can be used in the
MGGA functional construction. Such a quantity, often used is, for example, kinetic
energy density s(r), containing derivatives of the occupied Kohn–Sham orbitals:
1Xocc
sðrÞ ¼ jrwi ðrÞj2 ð1:61Þ
2 i
The integrated energy density s(r) is equal to the kinetic energy of the system of
non-interacting electrons TS[n(r)]:
Z Z
h2 X
N=2
TS ½nðrÞ ¼ sðrÞdr ¼ 2 wi ðrÞr2 wi ðrÞdr ð1:62Þ
2m i¼1
There are many meta-GGA functionals available currently, e.g., M06-L [71], TPSS
[70], SSB-D [72], MVS [73] as well as the entire group of approximate functional
based on meta-GGA approximation, with a common name Minnesota Functionals,
containing term dependent on kinetic energy density and based on complicated
functional forms parameterized on high-quality benchmark databases [74].
Although we do not know the exact form of the XC functional, it is known from
numerical simulations that the exchange energy definitely dominates over the
correlation energy, and therefore, looking for the best XC functional, we should
focus primarily on the exchange part. As is known, in the Hartree–Fock method, the
exchange energy is determined exactly (but unlike in DFT—non-locally), and the
correlation energy is completely neglected. Since in Kohn–Sham method non-local
effects are neglected, an idea was proposed, to mix non-local exact HF exchange
energy with the local GGA and in this way to construct the hybrid XC functional of
the form:
EXC ¼ aEXHF þ ð1 aÞEXlocal þ EClocal ð1:63Þ
where the consecutive terms are the exact non-local HF exchange energy (but
calculated for KS orbitals, not HF ones), local GGA exchange energy, and local
GGA correlation energy. Of the currently available hybrid functionals, the most
popular are B3LYP [75, 76], PBE0 [77, 78], and HSE [79]. For example, the first
one is defined as (a0 = 0.2; ax = 0.72; ac = 0.81):
B3LYP
EXC ¼ a0 EXHF þ ð1 a0 ÞEXLDA þ ax EXBecke þ ð1 ac ÞECVVN þ ac ECLYP ð1:64Þ
and includes additionally LDA exchange and correlation functionals.

The last group of XC functionals is fully non-local functionals such as ADA [80]
(average density approximation) or WDA [80, 81] (weighted density approxima-
tion), proposed in mid-70s of the twentieth century, where the starting point is the
exact functional of the density defined as:
Z Z
1 nXC ðr; r 0 Þ 0
EXC ½nðrÞ ¼ nðrÞdr dr ð1:65Þ
2 jr r 0 j
where nXC ðr; r 0 Þ is the electron density of the exchange–correlation hole (density
depletion) resulting directly from the exchange–correlation effects related to
electron interactions and modeling of the exchange–correlation hole using
some analytic function. Such a fully non-local functional has finally the following
form:
Z Z 0
1 XC ðr; r Þ
nmodel
NL
EXC ½nðrÞ ¼ nðrÞdr 0
dr 0 ð1:66Þ
2 jr r j
Thus, the crucial point of this approach is the best possible selection of the
0
analytical function, modeling the electron density nmodelXC ðr; r Þ of the exchange–
correlation hole. Non-local functionals have a number of advantages over previous
approximations (the minimal number of simplifications), but they are very com-
putationally demanding and hence less frequently used.
Density Functional Theory quickly gained popularity in calculations in
solid-state physics, but in chemistry, it began to be applied to a broader range of
systems only in the 90s of the twentieth century, when better approximations of XC
potentials were proposed. Nowadays, due to significantly lower computational
power requirements compared to HF and post-HF methods, DFT has dominated the
electronic structure calculations for both solids and molecules/clusters. However,
due to the difficulty of using this formalism, for example, in the description of
intermolecular interactions (of key importance for the understanding of chemical
reactions), especially dispersive as well as of transition states and defects con-
taining or highly correlated systems, it is still being intensively developed.
One of the fundamental problems associated with DFT in Kohn–Sham’s for-
mulation, resulting from the lack of knowledge of the exact form of the exchange–
correlation functional and hence the necessity to use of approximate ones, is the
need to select the appropriate functional for the problem studied (no universal XC
functional); hence, a wide variety of functionals have been proposed exhibiting
different efficiency in calculations for a given type of system and/or physical
property. In the case of solid-state calculations, the simplest LDA and GGA
potentials are still quite often used, but in the case of strongly correlated systems,
e.g., 3d-block transition metal oxides, modified versions of these potentials are
22 A. Koleżyński
often used, introducing an additional Hubbard U parameter [82] (LDA + U,

GGA + U potentials). In the electronic structure calculations, carried out mainly by
physicists, the PBE potential (or its PBESol variant, optimized for semiconducting
materials) is probably the most widely used, whereas in typical calculations carried
out by chemists, high popularity gained the previously described hybrid functionals,
e.g., B3LYP, in which the parameters are usually fitted for the collection of
experimental data, so one can get very accurate results for similar systems.
Unfortunately, these approximate XC functionals cannot be systematically
improved (KS equations with approximate XC potential do not fulfill the variation
principle); hence, it is not possible to unequivocally estimate errors associated with
the use of a given XC potential, and each time the results of calculations must be
compared with the results obtained by other theoretical methods or experimental
data.
1.2.2 Practical Aspects of the Application of Computational

Methods in Spectroscopy
Computational spectroscopy is a rapidly growing field that provides comprehensive

tools for the simulation, analysis, and interpretation of spectra in the context of
related physical and chemical processes and phenomena. Applied theoretical
approaches can be divided into two groups: time-independent calculations for
stationary states in minimum of the potential energy (e.g., vibrational spectra,
phonon dispersion curves) and time-dependent calculations for electron transitions
and excited states and their electronic properties (electron, NMR, and Mössbauer
spectra). Vibrational spectra of molecules, clusters as well as amorphous and
periodic systems are routinely modeled today and interpreted on the basis of a
harmonic approximation for potential energy near the minimum, normally available
in many programs using very different theory levels, starting from classical
approach with classic interparticle potentials, through hybrid QM/MM approaches,
quantum HF, post-HF, or DFT, up to molecular dynamics [83–87]. Spectroscopic
studies of conformational spaces or reaction pathways in molecular systems are also
broadly supported by routine ab initio and DFT calculations of the potential energy
hypersurface for the ground state, allowing a better understanding of intra- and
intermolecular interactions [88, 89].
1.2.2.1 Harmonic Approximation—Vibrational Spectroscopy
Atoms in molecules, clusters, and solids oscillate with respect to the equilibrium
positions, for which the potential energy of the system is the smallest. Any such
atomic displacement from the equilibrium position is associated with an increase in
potential energy, and the magnitude of these displacements depends on the
temperature, the type of atoms (their mass), and the strength of interatomic inter-
actions. If we could observe the motions of atoms in real time, we could see that
they are completely random, and the maximum atomic displacements form a kind
of ellipsoid in the Cartesian system, inclined, in general, at some angle to the axes
of the coordination system and with different lengths of each ellipsoid axis. This
random atomic motion can be decomposed into separate, independent “normal
modes of vibration”, the vibrations of all atoms in phase, with the same frequency
but with different amplitudes (these vibrations are, the well-known ones from
vibrational spectrum analyzes, symmetric and antisymmetric stretching vibrations,
bending, out of plane vibrations, twisting, rocking, etc.). In general case, for the
N atomic system, we have 3N vibrations, of which three are rotations (absent in
periodic systems), three are translations and 3N − 6 are vibrations, and the most
general motion of a system is a superposition of its all normal modes. The concept
of normal modes of vibrations can be derived from the classical model of a
molecule made up of point masses connected by springs that satisfy the Hooke’s
law, vibrating with respect to the equilibrium positions. The dynamics of such a
system is described in the Lagrange equation [90]:
d @T @V
þ ¼0 ð1:67Þ
dt @ x_ i @xi
where T and V are the kinetic and potential energies, respectively, defined as (within
harmonic approximation):
!
X
T ¼ 1=2 mi x_ i
2
ð1:68Þ
i
and
V ¼ 1=2kx2i ð1:69Þ
and xi are the Cartesian displacement coordinates (the dot denotes time derivative).
pffiffiffiffiffi
When using the so-called mass-weighted displacement coordinates qi ¼ mi xi ,
Lagrange equations become:
X
3N
€qi þ fij qj ¼ 0 ð1:70Þ
j¼1
where €qi denotes second-time derivative, and:
@2V
fij ¼ ð1:71Þ
@qi @qj
are mass-weighted Cartesian force constants. The above equation represents a set of
3N simultaneous differential equations with the solutions:
24 A. Koleżyński
pffiffiffiffi
qi ¼ Ai sin ki t þ e ð1:72Þ
pffiffiffiffiffiffiffiffiffi
where Ai are amplitudes, e are phase angles, and k ¼ x2 , x ¼ 2pm ¼ k=m, m—
reduced mass. The solution for this system of equations is N normal modes of
vibrations (simultaneous in-phase displacements of all atoms) with the corre-
sponding frequencies k. Using the relation:
d2 q i
¼ kqi ð1:73Þ
dt2
we can transform the above system of 3N linear differential equations into the
system of 3N homogeneous linear equations, expressed in matrix form as:
ðF KÞ A ¼ 0 ð1:74Þ
where F is a force constants matrix, K is a diagonal matrix of ki values, and A is a

vector consisting of Ai amplitudes. The latter equation has two solutions: one trivial
for A ¼ 0, i.e., when all Ai amplitudes
equal
to zero, and one non-trivial, when
determinant jF Kj is zero, i.e., fij dij ki ¼ 0. To obtain vibrational frequencies
and to be able to visualize normal modes of vibrations, it is necessary to solve this
last system of linear equations, usually by numerically diagonalizing the matrix F,
according to the relation LT FL ¼ K, where L is a matrix composed of eigenvec-
tors, which simultaneously allows the transformation of mass-weighted Cartesian
coordinates to a new coordinate system, the so-called normal coordinates
Q: Q ¼ Lq. These new coordinates are defined in such a way that each of the
3N normal modes of vibrations corresponds only to one normal coordinate
Q, where six of these normal coordinates correspond to three rotations and three
translations with a vibration frequency equal to zero.
From the quantum mechanical point of view, the problem of finding the fre-
quencies of atomic vibrations is much more straightforward, although computa-
tionally more demanding—as a result of quantum mechanical calculations, we get
information about the total energy of the system and after geometry optimization
about the energy in equilibrium. In the general case, the shape of the potential
energy hypersurface near the minimum is unknown and the dependence of total
energy on the atomic displacement Dx ¼ x x0 can be expanded into the Taylor
series:

dE 1 2 d2 E 1 3 d3 E
EðxÞ ¼ Eðx0 Þ þ Dx þ Dx þ Dx þ
dx x¼x0 2! dx2 x¼x0 3! dx3 x¼x0
ð1:75Þ
For x x0 , the first derivative is equal to zero and hence the expression,
neglecting the terms with derivatives of the order higher than the second (harmonic
approximation), simplifies to the form:

1 2 d2 E
EðxÞ ¼ Eðx0 Þ þ Dx ð1:76Þ
2 dx2 x¼x0
Thus, the problem comes down to the calculation of the second derivatives of the
total energy of the system, with respect to the displacement of atoms in Cartesian
coordinates from equilibrium geometry. This allows creation of a matrix composed
of the energy second derivatives (the so-called Hessian matrix) and the formulation
of the eigenvalue problem equation, and by solving it, finding the respective
eigenfrequencies for a given vibrational system:
X
3N X
3N
ðHjk ki Mjk ÞXki ¼ 0 ð1:77Þ
k¼1 j¼1
where eigenvalues ki are associated with vibrational eigenfrequencies mi by relation:
ki ¼ 2pm2i ð1:78Þ
Mjk is diagonal matrix of atomic masses, Xki is a matrix composed of eigen-

vectors, transforming atomic displacements from Cartesian coordinates to normal
coordinates, and Hjk is the Hessian matrix of second-order partial derivatives of the
total energy of the system, with respect to the Cartesian displacement coordinates:
2 @2 E @2E @2E
3
@n21 @n1 @n2 ... @n1 @n3N
6 7
6 @2 E @2E
... @2E 7
6 @n2 @n1 @n22 @n2 @n3N 7
H¼6 .. .. .. .. 7; where ni ¼ fxi; yi ; zi g ð1:79Þ
6 7
4 . . . . 5
@2 E @2E @2E
@n3N @n1 @n3N @n2 ... @n23N
The calculated deformations of the equilibrium geometry caused by vibrations

and the resulting change in the distribution of the total electron density provide
information on changes in the dipole moments and polarizations of the electron
cloud and hence allow calculating the intensities of vibrations. There are a number
of different approximations to calculate IR intensity and Raman activity, e.g.,
localized molecular orbitals [91, 92], the finite perturbation method [93], IR
intensities through Berry phase [94, 95] or through maximally localized Wannier
functions scheme [96, 97], IR or Raman intensities through Coupled Perturbed
Hartree–Fock/Kohn–Sham approach [98–100]; however, due to the limited space,
and especially the fact that the user of a given program for ab initio calculations has
26 A. Koleżyński
usually a fairly limited choice of methods of calculating IR intensity and Raman

activity, their detailed description will not be presented here.
Harmonic approximation is very useful and in many cases allows obtaining
accurate data on vibrations in the system. In many cases, however, in systems in
which anharmonic interactions have a significant effect on properties and thus
cannot be neglected, this approximation is insufficient, and there is a need to
consider the higher-order terms of Taylor series expansion. In this case, the
vibrational spectrum calculations are significantly more complex and computa-
tionally demanding, hence limited to relatively small systems—in the case of
medium and large systems, such calculations are still in practice impossible and we
have to use harmonic approximation or classical methods based on interatomic
potentials.
1.2.2.2 Time-Dependent DFT
Calculations of the energy for excited states of the systems and their electronic
properties have also become an irreplaceable auxiliary tool in the interpretation of
electronic spectra and dynamics of excited systems. While the calculations with the
so-called chemical accuracy for systems in the ground state have become in last
years relatively simple, the accurate estimation of energy of excited states is still
quite a challenge, both from theoretical and computational complexity point of view
(excited states are often close to each other, hence very high precision of calcula-
tions is necessary, associated with a high level of theory, often also with additional
geometry optimization for excited states, which is associated with extremely
time-consuming calculations, impossible to do in the case of large systems).
Currently, there are a number of different quantum mechanical methods for cal-
culating different properties of excited states. A detailed discussion of such methods
goes far beyond the scope of this chapter and is available in many studies [e.g.,
101–105]. Generally, these methods can be divided into two groups: The first is the
previously mentioned classical, post-HF-based “configuration” methods (single
configuration CI, multi-configuration MCSCF, CASSCF, CASPT2, or
multi-reference MRCI, in which any electron state is defined in the form of a linear
combination of a number of Slater determinants corresponding to various electronic
configurations) or derived from the coupled cluster (CC) family. However, these
methods are very computationally demanding, and hence, they cannot be used for
larger molecules or more complex solids.
The second group consists of methods based on electron density. In classical
DFT, the effects of exchange and correlation are taken into account approximately
by means of XC functionals, usually obtained by fitting functional parameters to
experimental data or by imposing physical constraints on the shape of XC potential
well, based on the properties of the theoretical model system.
The extension of classical DFT approach for excited systems is the TD-DFT
method [105–109], which proved to be an extremely effective and useful tool, with
precision comparable to the most sophisticated post-HF methods, but with
considerably lower computational requirements, resulting in increasing number of

studies published in recent years devoted to TD-DFT application for the analysis of
various systems in excited states and related absorption and emission processes.
Despite some objections to the TD-DFT method, e.g., lack of equal precision for
electronic transitions of various nature, incorrect description of long-range disper-
sive forces by standard functionals, an over-polarization problem, limited precision
of the description of induced delocalized states (molecular valence states with
extended p systems, Rydberg states, and doubly excited states), a problem with the
selection of a suitable functional and basis function set (the results obtained with
their use should be verified based on available experimental data) [102, 107, 108,
110], due to very large computational requirements in the case of correlated wave
function-based post-HF methods, TD-DFT based methods (thanks to the ability to
deliver good quantitative results for medium-sized systems at reasonably low
computational costs) have become very popular in simulations of excited
properties. Given the above, it seems that currently TD-DFT is arguably the best
compromise between precision and computational costs for modeling and analyzing
the properties of excited states of medium and larger molecular and periodic sys-
tems. Therefore, in the following paragraphs, the basic assumptions of this method
will be presented.
The starting point for the formulation of TD-DFT is the time-dependent
Schrödinger equation:
^ ðtÞWðtÞ ¼ i @ WðtÞ
H ð1:80Þ
@t
for N-electron wave function WðtÞ ¼ Wðr1 ; r2 ; . . .; rN ; tÞ and the theorems formu-
lated by Runge and Gross [111]. The first Runge–Gross theorem states that the
time-dependent electron density qðr; tÞ, together with the initial wave function
W0 ¼ Wð0Þ, determines the external potential with the accuracy of up to an additive
function of time, and hence the wave function is determined with the accuracy up to
the phase:
WðtÞ ¼ ei/ðtÞ W½q; W0 ðtÞ ð1:81Þ
and therefore all the observables can be calculated based on the knowledge of the
time-dependent electron density. While in the case of the classical DFT formulated
for time-independent processes, the ground state can be determined by the varia-
tional principle and the minimization of total energy, this procedure is impossible
for time-dependent processes (the variation principle cannot be defined based on the
energy of the system, since the energy is not a saved quantity). According to
the third Runge–Gross theorem, the equivalent of the variational principle for
stationary state can be defined in the case of a time-dependent system using the
Frenkel–Dirac stationary action principle [112] from which follows that
the time-dependent Schrödinger equation is satisfied in the time interval (t0, t), if the
quantity (called quantum mechanical action):
28 A. Koleżyński
Zt
@
A½W½W0 ðt; t0 Þ ¼ ds W½W0 ðsÞi HðsÞ
^ W½W0 ðsÞ
ð1:82Þ
@s
t0
is made stationary with respect to wave function W½W0 ðsÞ variations within time
interval s 2 ðt0 ; tÞ, assuming the following constraints imposed on the system:
dW½W0 ðt0 Þ ¼ dW½W0 ðtÞ ¼ 0. Next step is to substitute instantaneous energy

E ½W0 ðsÞ ¼ W½W0 ðsÞH ^ ðsÞW½W0 ðsÞ in quantum mechanical action formula
with Kohn–Sham energy and the wave function with Kohn–Sham determinant. As
a result, the quantum mechanical action:
dA
¼0 ð1:83Þ
dqðr; sÞ
is stationary for the true time-dependent electron density. Based on the latter
condition, the time-dependent Kohn–Sham equation can be derived. Assuming, in
analogy to classical KS system, the existence of the external potential Veff ðr; sÞ for
the reference system of independent electrons whose one-electron wave functions
wi ðr; sÞ give the same total charge density qðr; sÞ as the original system of inter-
acting electrons, i.e.:
X
qðr; sÞ ¼ ci jwi ðr; sÞj2 ð1:84Þ
i
and minimizing the quantum mechanical action, we obtain the set of

N time-dependent Kohn–Sham equations:
Z
1 qðr; sÞ 0 @w ðr; sÞ
r2 þ Vext ðr; sÞ þ 0
dr þ Vxc ½qðr; sÞ wi ðr; sÞ ¼ i i ð1:85Þ
2 jr r j @s
where, as in classical KS equations, the first term on the left-hand side of the
equations is kinetic energy, next two are classical Coulomb electron–nuclei and
electron–electron (Hartree term) interactions, and the last one, Vxc ½qðr; sÞ, is the
time-dependent counterpart of the classical DFT, stationary exchange–correlation
functional, defined as:
Zt1
dAxc ½qðr; sÞ
Vxc ½qðr; sÞ ¼ ; and Axc ¼ Exc ½qðr; sÞds ð1:86Þ
dqðr; sÞ
t0
It is worth to note that unlike the exchange–correlation potential in ground state

DFT, the exchange–correlation potential in time-dependent Kohn–Sham theory is
formally dependent on the entire history of the density, along with both the initial
wave function W(0) of the real interacting system of N-electrons and the initial KS
wave function U(0) on reference system of N non-interacting particles.
Like in classical DFT, the time-dependent KS theory yields exact results,
assuming the knowledge of an exact form of the exchange–correlation potential.
Unfortunately, as in classical DFT counterpart, the exact formula for this potential
is unknown, so we are forced to use approximations. The simplest one is the
so-called adiabatic approximation, the local approximation in time, which assumes
instantaneous reaction of XC potential to any temporal changes in charge density
(and without memory of previous states, i.e., depending only on the density qðsÞ at
the same time) and therefore can be used reasonably for the systems with potential
which is slowly varying with time [113]:
dAxc ½qðr; sÞ dExc ½qs ðrÞ

Vxc ½qðr; sÞ ¼ ¼ Vxc ½qs ðrÞ ð1:87Þ
dqðr; sÞ dqs ðrÞ
The second approach, dominating in practice today in available TD-DFT

implementations in computer programs, is the approximation based on linear
response function theory. The detailed description of this approximation is far
beyond the scope of this chapter, but in short it consists of using perturbation theory
and treating the change in electron density under the influence of external potential
as a specific perturbation to the original electron density of stationary state at a
constant external potential [114]. Consider the system in the ground state, with
constant external potential V0 ðrÞ (usually nuclear attraction), to which at the time t0
time-dependent external potential V1 ðr; tÞ (e.g, electric or magnetic field) starts to
act, and the total external potential is equal Vext ðr; tÞ ¼ V0 ðrÞ þ V1 ðr; tÞ. For any
time t t0, the initial ground state electron density q0 ðrÞ can be defined as the
self-consistent solution of the classical system of KS equations. Time-dependent
electron density qðr; tÞ is, on the other hand, the functional of external potential
Vext ðr; tÞ only, i.e., qðr; tÞ ¼ q½Vext ðr; tÞ. Assuming that the potential V1 ðr; tÞ is
small at any time t > t0, we can treat within perturbation theory the density changes
related to this potential as a small perturbation and expand the electron density
qðr; tÞ using the Taylor expansion series:
qðr; tÞ ¼ q0 ðr; tÞ þ q1 ðr; tÞ þ q2 ðr; tÞ þ þ q3 ðr; tÞ. . . ð1:88Þ
where the superscript denotes the order of the perturbation (0 means the electron
density of the unperturbed ground state). Hence the first correction, the first-order
(linear) response can be defined as:
Z Z
q1 ðr; tÞ ¼ dt0 d3 r 0 vðr; r0 ; t; t0 ÞV1 ðr0 ; t0 Þ ð1:89Þ
where v is the total electron density first-order response to perturbing external

potential V1 ðr; tÞ [114]:
30 A. Koleżyński

0 dq½Vext ðr; tÞ
0
vðr; r ; t; t Þ ¼ ð1:90Þ
dVext ðr0 ; t0 Þ V0
Analogous Kohn–Sham electron density first-order response function for the

reference system of non-interacting particles with unperturbed density q0 ðr; tÞ,
moving in external potential VKS ðr; tÞ, is given by:

0 0 dq½VKS ðr; tÞ
vKS ðr; r ; t; t Þ ¼ ð1:91Þ
dVKS ðr0 ; t0 Þ VKS ½q0 ðr;tÞ
while the potential VKS ðr; tÞ of a system of non-interacting particles corresponding

to the potential Vext ðr; tÞ of the actual system is equal:
Z
qðr; tÞ 0
VKS ðr; tÞ ¼ Vext ðr; tÞ þ dr þ Vxc ½qðr; tÞ ð1:92Þ
j r r0 j
Using the functional chain rule and calculating the functional derivative of
VKS ðr; tÞ with respect to Vext ðr; tÞ, we get the necessary link between the electron
density linear response function of the real system of interacting particles and the
reference system of the Kohn–Sham system of non-interacting particles and hence
the expression for the density linear response function of the non-interacting par-
ticles system:
Z Z
q ðr; tÞ ¼
1
dt0 dr 0 vKS ðr; r0 ; t; t0 ÞVKS;1 ðr0 ; t0 Þ ð1:93Þ
with effective potential VKS;1 ðr0 ; t0 Þ equal to:

Z Z Z
q1 ðr; tÞ 0
VKS;1 ðr; tÞ ¼ V1 ðr; tÞ þ dr þ dr 0 dt0 fxc q0 ðr; r0 ; t; t0 Þq1 ðr0 ; t0 Þ
j r r0 j
ð1:94Þ
where time-independent kernel:

0 0 0 dVKS ½qðr; tÞ
fxc q ðr; r ; t; t Þ ¼ ð1:95Þ
dqðr0 ; t0 Þ q0 ðr;tÞ
is a functional of the initial ground state electron density. The obtained expression
for the electron density linear response function of the system of non-interacting
particles allows practical application of the linear response function approximation
within the TD-DFT formalism for the calculation of time-dependent properties of
multi-electron systems (analogous derivation can be carried out for higher-order
corrections). If the system is subjected to a small time-dependent external potential
V1 ðr; tÞ, then there is no need to solve the full time-dependent KS equation, and
most of the system’s properties of interest can be derived from first-order changes
of electron density q1 ðr; tÞ [107, 115, 116] (in some cases, however, it is also
necessary to take into account the corrections of higher orders). Despite the obvious
advantages of the discussed approximations (adiabatic and linear response function)
allowing the study of many processes related to excitations from the ground state of
the system, there are still a number of problems requiring solutions and phenomena
that are still impossible to study within TD-DFT formalism (e.g., electronic tran-
sitions with substantial contributions from double excitations or non-adiabatic
effects) [107]. Thus, there is ongoing work on the development of new, better and
better and more accurate exchange–correlation functionals, especially suitable for
excited states and enabling the description of more complex phenomena within the
TD-DFT formalism, which will allow in the near future for a correct description of
intermingled optical properties using more realistic models and thus significant
broadening of various complex spectroscopic processes possible to be described
and analyzed within TD-DFT approach.
1.3 The Problem of Computational Complexity

and the Resulting Necessary Simplifications
and Approximations
The above-mentioned, obviously very brief description of the most popular theo-
retical methods used to model material properties shows how much can be modeled
today, but at the same time indicates many limitations that prevent effective and
accurate study of many phenomena and processes, including those that are the
domain of spectroscopy. Nevertheless, the development of theoretical methods, the
emergence of new theoretical models of various phenomena, new approaches to
difficult theoretical issues as well as the continuous development of the existing
ones, together with the involvement of many research teams around the world in the
development of computational chemistry, physics, and materials science, allows
hoping that in the coming years it will be possible to conduct more and more
accurate computer simulations of a growing number of applicationally important
processes and physicochemical phenomena.
The second barrier, besides the need to develop theoretical methods, that inhibits
the effective and widespread use of computational methods in the analysis of the
properties and in materials characterization is the complexity of many physical
systems, which still prevents the ability to obtain the results which are correct
quantitatively, and in many cases even just qualitatively. These problems result
from the complexity on the two separate levels—the structural (non-stoichiometric
composition, point defects, admixtures, partial occupation of some crystal lattice
sites, local compositional disorder, etc.) and the microstructural, important in
polycrystalline solids (grain size distribution, structure and properties of grain
boundaries, segregation of admixtures and defects at grain and phase boundaries,
32 A. Koleżyński
porosity—open and closed, linear defects, various phase and chemical composition
of particular grains in polycrystalline material and the resulting phenomena and
processes of mass transport on grain boundaries, etc.). Both groups of problems
require completely different approaches, and while the problem of complexity at the
structural level is currently in many cases taken into account in theoretical modeling
in an approximate way, often allowing to obtain also quantitatively accurate results,
the complexity at the microstructural level is still a huge challenge and apart from
few cases of attempts to model these effects using ab inito methods (obviously with
very simplified models) requires the use of coarse classical methods that give at best
the acceptable quality results and is still waiting for the development of better
models and theories, as well as more efficient algorithms and significantly faster and
more powerful computers. Therefore, in the following paragraphs, only selected
problems related to structural complexity will be discussed and some most common
practical ways of solving them effectively will be presented.
1.3.1 The Size and Complexity of the System
The development of computer technologies, their increasing availability, and

growing computing power with simultaneous development of theoretical methods
and the increase in the efficiency of computational algorithms have caused signif-
icant increase in applications of theoretical calculations in scientific research. In
turn, this resulted in increasing importance of calculations, both in the first, early
stages of research, to reduce the number of compounds and compositions requiring
experimental study, by screening possible materials and narrowing potential can-
didates to those with the most promising properties, as well as later, during the
synthesis and experimental research, to modify and improve the physicochemical
properties of the materials of interest.
All this increases the interest in the application of theoretical methods and in
silico experiments in the study of materials with a very wide range of applications,
huge diversity, and structural complexity.
Unfortunately, despite the extremely rapid progress observed, still many mate-
rials are beyond the scope of modeling possibilities using the available tools. This is
mainly due to the fact that in general the ab inito calculations refer to ideal systems
(e.g., ideal crystals of infinite sizes, isolated molecules or clusters) at the absolute
zero temperature. However, real systems are much more complex–neutral or
charged particles and clusters are not isolated, but they interact (in the gas phase,
liquid or molecular crystals) with a various strength; solids are always—to a dif-
ferent extent—defective, and these defects can exhibit various types of structural
disorder or local ordering and clustering. Moreover, ab initio computational
methods for solids are, by definition—due to Bloch theorem—dedicated to periodic
systems exhibiting translational symmetry, which makes them unable to handle
amorphous solids. The mentioned factors cause that in practically every case when
structural complexity appears; it is necessary to use various types of simplifications,
idealization, or abstraction related to omitting some of the interactions, when

constructing model structure of material under investigation. The type and range of
simplifications depends heavily on the complexity of the system and hence the next
sections of this chapter will be devoted to this issue described from the viewpoint of
the most frequently used simplifications of structural models, followed in consec-
utive subchapter by discussion of the simplifications related to the level of theory
employed in calculations (applied formalism, basis functions used to define
molecular or crystal orbitals, type of crystalline potential, termination method used
for smaller structural fragments cut out of a complex structure for their stabilization,
etc.).
1.3.1.1 Size of the System: Molecules, Clusters, Amorphous

and Periodic Solids
The simplest systems, which in many cases do not require any additional model
simplifications, are molecules and charge neutral clusters. Obtained results are
usually in very good agreement with experimental data, especially those obtained in
the gas phase. In the case of molecules in solutions, the situation complicates
significantly, due to additional interactions of those molecules with surrounding
solvent molecules. In such cases, solvation models of varying complexity are
usually used, the broader discussion of which goes beyond the scope of this chapter.
These can be the so-called implicit (continuous), explicit, or hybrid solvation
models. In the implicit solvation models, the solvent interacting with the molecule
under the study is represented as a continuous medium devoid of the properties of
individual “explicit” solvent molecules [117]—the most commonly used implicit
models are polarizable continuum model (PCM) [118, 119] and COnductor-like
Screening MOdel (COSMO) [120, 121]. In contrast, explicit solvation models treat
the solvent molecules explicitly (i.e., in general, their coordinates and interactions
with the molecule studied are explicitly included) [122]. The problem is, however,
that due to the complexity of such explicit solvation approach, its use is usually
limited to classical MM/MD applications and the main effort is put to find appro-
priate force field allowing accurate and realistic modeling. The third group of
models, hybrid solvation models, is a kind of trade-off, taking the best from both
worlds—the efficiency of implicit solvation models and accuracy of explicit sol-
vation models and usually treats the solvent molecules closest to studied molecule
or cluster explicitly and the remaining part of the solvent implicitly as continuous
uniform solution [123].
A separate group of systems are charged ions and clusters of various sizes, which
require additional treatments related to the overall charge neutralization, because
such charge-possessing systems are computationally unstable and it is usually
impossible to perform calculations related to the structure relaxation and finding the
equilibrium geometry. Typically, the charge neutralization and stabilization of the
structure of the studied system is achieved by adding external, constant charge, or
placing additional cations or anions in the proximity of the ion or cluster (depending
34 A. Koleżyński
on the ion or cluster net charge sign). This allows performing successful calcula-
tions devoted to geometry optimization; however, in the analysis of the obtained
results, it is a must to take into account the influence of this additional neutralizing
charge (type of cations or anions, their position, valence, etc.) on the spatial con-
formation of the ion or cluster and hence its properties.
The most computationally complex are solids, both amorphous and crystalline
with translational symmetry. Crystalline solids, if they maintain stoichiometry, and all
Wyckoff’s positions are fully occupied, are not usually a challenge for modern pro-
grams (apart from the very long, sometimes prohibitive, calculations in the case of
structures with unit cells containing a very large number of atoms and hence some-
times computationally still unavailable for computer modeling). In the latter case, it is
necessary to apply structure simplifications (reduction of the size of the unit cell), so as
to maintain long-range structural properties with significantly shortened calculation
time. The example materials with a large number of atoms in the unit cell are zeolites
(see Chapter 10). The structure simplifications are necessary even for the most
structurally simple zeolites like zeolite A (LTA-type structure), whose unit cell is
shown in Fig. 1.2a—for the sake of clarity, the figure omits oxygen atoms and
non-tetrahedral, extra-framework cations (the latter constitute a separate computa-
tional problem because they partially occupy some of Wyckoff’s positions in unit cell
—more on this problem later in this section). The unit cell is composed of sodalite b
cages (Fig. 1.2c) connected to each other by D4R double four-membered rings
(Fig. 1.2b), forming a system of 8 a chambers (Fig. 1.2a). Thus, it is possible to
simplify this unit cell in such a way that the new and smaller unit cell consists of only
one sodalite chamber and eight single four-membered rings constituting the halves of
the respective double four-membered rings (Fig. 1.2d). In this way, significant sim-
plification of the unit cell is achieved and the computational complexity for this
structure reduced. Unfortunately, this gain is related to the appearance of additional
Al–O–Al and Si–O–Si bonds in the model LTA structure, which are much less
probable in real structure (Löwenstein’s rule of aluminum avoidance, which in fact is
not always fulfilled [124]). However, it is an acceptable cost—as shown in [86], this
simplified model allows the calculation of LTA vibrational spectra of high accuracy,
thanks to which it was possible to investigate the influence of the type and amount of
selected non-tetrahedral cations on the structural properties and vibrational spectra of
zeolite A and to compare the results with experimental data obtained for synthesized
LTA subjected to sorption processes. This approach can be an effective way to sig-
nificantly reduce computation time and is often the only (as in the described case of
LTA) possibility to model, within the framework of ab inito theories, the properties of
more complex structures using the currently available computing power of
supercomputers.
A separate problem to deal with often is the partial occupation of some of
Wyckoff’s position in unit cell and the presence of adsorbed molecules, like water,
carbon and nitrogen oxide or dioxide, methane. While the latter can in many cases
be omitted in calculations (which is often practiced), this cannot be done with the
atoms that partially occupy Wyckoff’s positions. This is a big challenge, because
the vast majority of quantum mechanical programs do not allow partial occupancy
of atoms in the cell and therefore simplifications are necessary: If occupation of a

given position is close to zero, it is often possible to omit such an atom altogether,
whereas if occupation is close to 1, then full occupation can be assumed.
The problem arises when a given position is occupied by considerable fraction
that its omission is not possible without significant undesirable influence on the
results.
One can then use one of two approaches: The first is to reduce the symmetry of
the system with full occupation of only part of the original incomplete Wyckoff’s
positions and leaving the remaining sites empty. This is, however, connected with
the additional problem of which position to choose as being occupied and the
influence of the introduced in the process artificial local order, on the structural and
physicochemical properties of the system of interest—for example, if the original
Wyckoff’s position had a multiplicity of 4, and its occupation number was about
0.25, then after symmetry reduction one can select just one of new symmetrically
non-equivalent sites, leaving remaining three empty. If, however, the multiplicity of
such a position is so large that for a given occupation, after reducing the symmetry,
it will be necessary to fill more than one new symmetrically non-equivalent posi-
tion; the problem is additionally complicated, because there is more than one
possible configuration of occupied positions (e.g., for multiplicity of 8 and frac-
tional occupation of 0.25, there are eight new positions to be filled with just two
atoms). This means that a series of calculations must be carried out for all different
local arrangements in order to find energetically optimal structure and reproduce the
experimental results best. Again, an example of such a complex structure may be a
natural zeolite A, whose unit cell in a hydrated form (Fig. 1.3a) [125] with the
chemical formula |Na92.7(H2O)6.95| [Si96.96Al95.04O384] contains 1840 atomic sites
(not counting hydrogen atoms from water molecules!), including part of the Na and
O sites partially occupied, and in the dehydrated form (Fig. 1.3b) [126] with the
chemical formula |Na91.78| [Si96Al96O384] 1072 atomic sites (in the case of sodium
partially occupied). In the latter case, sodium occupies three different Wyckoff’s
positions: 64 g, occ. 0.972; 96i, occ. 0.242; and 96 h, occ. 0.066. Therefore, we can
simplify the model structure considerably, assuming full occupation of the 64 g
position, and 0.25 partial occupation in position 96i (neglecting completely hardly
occupied position 96 h). However, the problem of a very large number of possible
configurations of the position 96i filling remains (96 sites to be filled with four
times smaller number of sodium atoms), which is connected with the necessity of
creating many model structures for all possible configurations and performing
calculations to find the best model structure. It is worth mentioning here that even if
such extensive calculations can be made, the resultant best structure will still be too
complex to be able to carry out even relatively simple calculations of energy
Hessian matrix and the vibrational spectrum simulation; therefore, further simpli-
fications are still needed, such as the reduction of the unit cell size discussed earlier,
at the cost of a small change in the structure (in mentioned case the appearance of
homonuclear aluminum and silicon bonds).
The second, much more computationally complex approach involves the use of
superstructures (creation of bigger unit cell built of many original unit cells) and
36 A. Koleżyński
Fig. 1.2 a Zeolite A unit cell, b the same structure shifted by (0.25, 0.25, 0.25), c sodalite cage,
d LTA framework rebuilt using simplified SOD-like unit cell, e a fragment of the simplified LTA
model structure consisting of two sodalite cages (Al–Al and Si–Si bonds are depicted using the
respective atomic colors)
simulation of statistical, incomplete occupancy of a given site in the structure. This

approach will be presented in more detail in the next section, when discussing the
problem of local disorder related to, inter alia, structure defects.
Fig. 1.3 Structure of zeolite A in a exemplary hydrated [125] and b dehydrated form [126]
1.3.1.2 Structure Disorder: Point Defects, Dopants, Vacancies
Another problem often encountered in the theoretical modeling of periodic systems

is the local disorder of the structure (substitutional or compositional disorder, in
contrast to positional, topological or structural disorder, typical for amorphous
solids, related to the loss of regular spatial periodicity), usually caused by structural
defects or admixtures of foreign atoms. In practice, two approaches are most often
used: the first ones being the mean field approximation method; here the most
popular are two approximations: the simpler one is the so-called Virtual Crystal
Approximation, VCA [127–129] basically consisting of creation at each potentially
disordered Wyckoff site a virtual atom interpolating the properties between the
actual atoms, host, and dopant, by means of weighted average of their site occu-
pation (and therefore neglecting local effects like surrounding distortion and
averaging the properties of the site, thus making it practically impossible to
reproduce finer details of the disordered structure very accurately) or more
sophisticated approach, namely Coherent Potential Approximation, CPA, proposed
initially by Soven [130] and developed further by Velicky et al. [131, 132] and
consisting of replacement of varying potential of disordered system with periodic
effective potential of ideal crystal composed of average atom.
Second approach is based on constructing larger, ordered supercells followed by
a series of calculations carried out for various (preferably all possible) configura-
tions and post-hoc configurational averaging. This approach is computationally
very expensive with practical application limited to bigger concentrations and thus
smaller supercells. The main advantage is, however, that this approach allows
studying local properties (structure deformation, electron density, or bonding
changes, etc., due to point defect—substitutional atom or vacancy), which is simply
not possible in mean field approximations [133, 134].
38 A. Koleżyński
The procedure employed to construct supercell is very straightforward: One

starts with original unit cell (in Fig. 1.4a the unit cell of SrTiO3 cubic perovskite
structure is shown) and extends the structure in one, two, or three dimensions, e.g.,
constructing 3 3 3 superstructure (Fig. 1.4b) one obtains new, three times
bigger cubic superstructure (Fig. 1.4c) with 27-fold increased number of atoms. In
this way, it is possible to substitute from 1 to 13 host atoms at B-cation position and
simulate the substitution with the ratios of 1:28, 2:27, …13:14, corresponding to the
dopant concentration wrt B-cation of approximately 3 at.%, 6 at.%, …, 48 at.%
(and since there are now 135 atoms in unit cell, i.e., 27 Sr, 27 Ti, and 81 O, to total
concentration of 0.75, 1.48, …, 9.63 at.%). Of course it should be pointed out here
that some of these concentrations can be also obtained for smaller superstructure,
e.g., 1:8 doping ratio can be simulated using 2 2 2 superstructure and just one
dopant atom in B-cation position, instead of using three dopants in 3 3 3
superstructure—in practice, it is almost always advisable to use the smallest pos-
sible superstructure in order to save computation time, which increases significantly
with the size of superstructure. The additional problem stems from the fact that
when there is more than one point defect in superstructure (be it dopant or vacancy,
or simultaneously both defect types), it is necessary to construct all possible con-
figurations, i.e., assuming different positions occupied by defects and run analogous
calculations for each of the model structure and then to average obtained properties
(in Fig. 1.4d the exemplary 3 3 3 superstructure with three substitutional
defects is depicted). Unfortunately, in some cases, the number of configurations can
be prohibitively large and it is necessary to constrain the calculations and analysis
to highest symmetry configurations only, which are usually energetically most
probable or—in order to study defects tendency to clusterization—select appro-
priate configurations allowing direct comparison of various local ordering of
defects.
The amorphous solid generates additional issues, related mainly to the lack of
long-range translational symmetry and thus similar to those met in big molecules.
The problem of amorphous solids is, however, more complicated than in case of big
molecules, since the bond lengths between atoms as well as local environment vary
significantly in amorphous structures, and thus the simulations are even more
time-consuming. In general, one can only model a small part of the system,
mimicking to some extent statistical ordering of the atoms and then either use it as a
cluster embedded in bigger environment or create unit cell filled with such statis-
tically distributed atoms and use translational symmetry to generate the bigger
amorphous solid. There are two different approaches to model such systems:
starting with small oligomers and create the cluster of increasing size randomly
attaching new oligomers to the structure (bottom-up approach) [83, 135] or starting
with periodic structure (e.g., in case of amorphous silica one can start with
cristobalite structure), construction of big superstructure, then removing randomly
part of atoms (or smallest polyhedra like [SiO4]4− tetrahedra in silica glass) and then
relax the obtained structure (top-down approach) [136–138]. Both approaches are
extremely time-consuming (even if instead of ab initio methods, classical approach
is used) and provide only selective results, highly dependent on the approach used,
Fig. 1.4 SrTiO3 unit cell (a); 3 33 extended unit cell (b); 3 33 superstructure with one
(1:28 doping ratio) (c); and three point defects (1:8 doping ratio) in one of the possible, exemplary
configuration (d). The defects are depicted in different color and size for better visibility
source of randomness (random number generator) selected, etc., so in case of such

structures there is still much to be done in the future, when significantly faster
computers will be available.
A separate group of computational problems, extremely important for catalysis,
concerns the method of describing the surface or space inside the catalytically
active mesoporous materials (zeolites, MOFs) or clusters and the catalytic prop-
erties of these materials. This is a very complex issue, worth a separate discussion
in itself, hence far beyond the scope of this chapter. It is worth mentioning, how-
ever, that also here there are a number of different approaches related to surface
simulation, catalytically active sites, thermodynamics, and kinetics of catalytic
reactions, etc., similar to those described above. These approaches are based on
standard ab initio methods, hybrid methods combining the accuracy of ab initio
methods with the speed of classical methods based on classic interparticle potentials
(e.g., ONIOM [139]), but also dynamical or statistical (both, classical and ab initio)
40 A. Koleżyński
methods. The calculations are often carried out for model surfaces (symmetric or
asymmetric slabs) or structural fragments containing catalytically active sites cut
from a larger structure and frozen to preserve the original structure or subjected to
the relaxation procedure. Detailed information on the methods and approximations
commonly used in the context of catalytic properties of materials can be found inter
alia in [140–143].
1.3.2 Model Structure Simplifications and Approximations

(Theory Level)
The simplifications and idealizations discussed above relate to the reduction of the
complexity of the structure and hence a significant reduction in the calculation time.
In most cases, it is a necessary, but insufficient step to be able to effectively simulate
the properties of the material of interest. Another, even more important step, con-
cerns the necessity to make a decision regarding the applied level of theory (type of
approximations), since this will strongly influence the ability to get the results of
acceptable quality in reasonable time, which strongly depends on the particular
choice of the formalism, theory level, and complexity of the system.
In general case, a careful decision should be made regarding:
• The choice of the formalism to be used in calculations—classical (with chosen
appropriate classical Force Field), semi-empirical (with approximate
Hamiltonian, e.g., HMO, PPP, EH, CNDO, INDO, MNDO, AM1, PM3, PM6),
Hartree–Fock, DFT, TD-DFT, etc.;
• Relativistic treatment of the electrons—non-relativistic, semi-relativistic, fully
relativistic;
• Spin treatment—non-spin-polarized or spin-polarized (with particular magnetic
ordering in periodic solids);
• Representation of 3D bulk—non-periodic (cluster) versus periodic (unit cell)
and 2D surfaces—asymmetric slab, symmetric slab, cluster;
• Form of external potential in periodic systems—muffin tin (MT), atomic sphere
approximation (ASA), full potential (FP), or pseudopotential (PP);
• In case of ab initio methods, additionally the choice of electronic interactions
treatment (described earlier) has to be done;
– For HF formalism, the level of post-HF approximation
Configuration Interaction (Full-CI, CIS, CISD, CISDT, MRCI, MRSDCI,
…);
Møller–Plesset Perturbation Theory (MP2, MP3, MP4, …);
Coupled Cluster Theory (CC, CCSD, CCSD(T), …);
Other (QCI, CASSCF, G1, G2, G3 or G4, FPD, ccCA, CBS, …);
– for DFT formalism, the type of exchange–correlation functional

LDA functionals (LDA, LSDA, LDA + U, LSDA + U, …)
GGA functionals (PW91, PBE, PBESol, BLYP, WC 2006, …),
Meta-GGA functionals (M06-L, TPSS, SSB-D, MVS, …),
Hybrid functionals (B3LYP, PBE0, HSE, …),
Non-local functionals (ADA, WDA, …).
• Function basis set used to represent molecular or crystal orbitals: (PW, (L)APW,
ASW, (L)ASO, (L)MTO, localized atomic basis functions). In case of the latter,
also the appropriate basis set (minimal, split valence, Pople style, polarized, dif-
fused, correlation consistent, polarization consistent, etc.) has to be additionally
chosen—this choice strongly influences the time of calculations, which scales as
NX, (x > 3, N—number of basis functions for various QM theory levels);
The choice of level theory is a crucial step in every calculations, and it has to be
done thoughtfully after careful consideration supported by literature reports on
similar systems, before the “production” calculations start. In many cases, due to the
lack of sufficient information on similar system studied from given properties
viewpoint, it is necessary to run preliminary calculations to properly determine
required theory level for given formalism to ensure the ability to simulate the
properties of the system under study in acceptable time, using reasonable computer
resources and simultaneously to obtain useful, physically meaningful results. It
should be, however, mentioned here that in many cases the available options to
choose from are strongly limited by the prior choice of the computer software, which
will be used for calculations, since in various programs only limited number of listed
above options is implemented (currently a huge number of quantum chemistry and
solid-state physics programs, both commercial and free is available [144]).
As final remark, it is worth to mention that whatever theory level is chosen for
practical calculations, when a series of calculations is carried out for a group of
model structures, in order to analyze the influence of the chemical composition,
defects, local order, etc, on the physic-chemical properties of interest, it is a must to
use for all calculations the same theory level and convergence criteria, since
otherwise, one will get the results which simply cannot be sensibly compared and
the whole endeavor of in silico experiment will simply end up in wasted time and
computer resources.
1.4 Conclusions
The use of theoretical methods as tools supporting experimental research is asso-

ciated with many crucial limitations resulting, on the one hand, from the limitations
of the methods themselves, and on the other from the complexity of the materials
studied. In the case of spectroscopy investigating the interaction of electromagnetic
radiation with matter, a number of practically available theoretical methods that
42 A. Koleżyński
allow modeling and studying these phenomena are in most cases limited to ab initio
methods derived from the quantum mechanics formalism. For this reason, this
chapter was devoted to those methods and the first part of the chapter discussed the
basic assumptions of quantum mechanics formalism in the Schrödinger’s wave
mechanics version and the Density Functional Theory, together with the most
commonly used approximations allowing practical application of these theories in
theoretical studies on the structure and properties of various materials (molecules,
clusters as well as crystalline and amorphous solids). The most important approa-
ches using quantum mechanics and DFT formalisms (both stationary and dynam-
ical) to study the spectroscopic properties of materials were briefly described.
Typical practical problems encountered during such studies, resulting from the
complexity of the structure (chemical composition, local order, defect, number of
symmetrically non-equivalent atoms, etc.) and the most usual ways to solving them
were also presented.
It follows from the presented discussion that the rapid development of theoretical
methods and the simultaneous increase in the computing power of modern com-
puters allowed for a significant increase of the role and importance of theoretical
modeling (the so-called in silico experiments) in the studies of the structure and
properties of various materials and currently, in many cases, such experiments are
carried out routinely, and their results serve to more effectively design new mate-
rials and modify existing ones. On the other hand, in many cases the computing
power of modern computers is still insufficient and there are a number of limitations
on the available theoretical tools, often critical, that prevent their practical appli-
cation to the study of more complex systems and phenomena. Nevertheless, looking
back on the speed of development of the so-called computational chemistry and
computational physics, one can expect that theoretical modeling of an increasing
number of more and more complex physical systems and processes will be possible
in the coming years, and in silico experiments will play gradually more important
role in scientific research, becoming an irreplaceable tool not only supporting
experimental research but also becoming a key tool in the research process,
probably on an equal footing with experimental methods (as is already the case, in,
e.g., the search for new drugs).
References
1. Lambert JB, Mazzola EP (2018) Nuclear magnetic resonance spectroscopy: an introduction

to principles, applications, and experimental methods, 2nd edn. Wiley
2. Chan JCC (ed) (2012) Solid state NMR. Topics in current chemistry, vol 316.
Springer-Verlag, Berlin
3. Buhl M, van Mourik T (2011) NMR spectroscopy: quantumchemical calculations. WIREs
Comput Mol Sci 1:634–647
4. Shukla AK (ed) (2017) EMR/ESR/EPR spectroscopy for characterization of nanomaterials.
Advanced structured materials, vol 62. Springer, India
5. van Doorslaer S, Murphy DM (2012) EPR spectroscopy: applications in chemistry and
biology. Topics in current chemistry, vol 321. Springer, Berlin
6. Diem M (2015) Modern vibrational spectroscopy and micro-spectroscopy: theory,

instrumentation, and biomedical applications. Wiley, Chichester
7. Perkampus HH (1992) UV-Vis spectroscopy and its applications. Springer, Berlin
8. Kumar C (ed) (2013) UV-Vis and photoluminescence spectroscopy for nanomaterials
characterization. Springer, Berlin
9. Hollas JM (2004) Modern spectroscopy, 4th edn. Wiley, Chichester, England
10. Kuzmany H (2009) Solid-state spectroscopy. An introduction, 2nd edn. Springer, Berlin
11. Briggs D, Seah MP (1990) Auger and X-ray photoelectron. Practical surface analysis
spectroscopy, vol 1. Wiley, Chichester
12. Eland JHD (1984) Photoelectron spectroscopy. An introduction to ultraviolet photoelectron
spectroscopy in the gas phase. Butterworth-Heinemann, Oxford
13. Siegbahn K (1973) Electron spectroscopy for chemical analysis. In: Smith SJ, Walters GK
(eds) Atomic physics, vol 3. Springer, Boston, MA
14. Wolstenholme J (2015) Auger electron spectroscopy: practical application to materials
analysis and characterization of surfaces, interfaces, and thin films. Momentum Press
15. Sharma VK, Klingelhofer G, Nishida T (eds) (2013) Mossbauer spectroscopy: applications
in chemistry, biology, industry, and nanotechnology. Wiley
16. Gütlich P, Bill E, Trautwein AX (2011) Mossbauer spectroscopy and transition metal
chemistry: fundamentals and applications. Springer, Heidelberg
17. Marx D, Hutter J (2009) Ab initio molecular dynamics: basic theory and advanced methods.
Cambridge University Press, Cambridge, UK
18. Gatti F (2014) Molecular quantum dynamics: from theory to applications. Springer, Berlin
19. Nightingale MP, Umrigar CJ (eds) (1998) Quantum Monte Carlo methods in physics and
chemistry. Kluwer Academic Publishers, Dordrecht
20. Mitas L, Roy PN, Tanaka S (eds) (2016) Recent progress in quantum Monte Carlo. In: ACS
symposium series, American Chemical Society, Washington, DC, USA
21. Anderson JB (2007) Quantum Monte-Carlo: origins, development, applications. Oxford
University Press, USA
22. Schrödinger E (1926) Quantisierung als eigenwertproblem. Erste mittteilung. Ann D Physik
79:361–376
23. Schrödinger E (1926) Quantisierung als eigenwertproblem. Zweite mittteilung. Ann D
Physik 79:489–527
24. Schrödinger E (1926) Quantisierung als eigenwertproblem. Dritte mittteilung. Ann D Physik
80:437–490
25. Schrödinger E (1926) Quantisierung als eigenwertproblem, Vierte mittteilung. Ann D Physik
81:109–139
26. Schrödinger E (1926) Der stetige übergang von der Mikro- zur Makromechanik.
Naturwissenschaften 14:664–666
27. Schrödinger E (1926) An undulatory theory of the mechanics of atoms and molecules. Phys
Rev 28:1049–1070
28. Hartree DR (1928) The wave mechanics of an atom with a non-Coulomb central field. Math
Proc Cambridge Phil Soc 24:89–110
29. Nemoškalenko VV, Antonov NV (1999) Computational methods in solid state physics. CRC
Press
30. Löwdin PO (1955) Quantum theory of many-particle systems. I. Physical interpretations by
means of density matrices, natural spin-orbitals, and convergence problems in the method of
configurational interaction. Phys Rev 97:1474–1489
31. Löwdin PO (1959) Correlation problem in many-electron quantum mechanics I. Review of
different approaches and discussion of some current ideas. Adv Chem Phys 2:207
32. Sherrill CD, Schaefer HF III (1999) The configuration interaction method: advances in
highly correlated approaches. In: Löwdin PO (ed) Advances in quantum chemistry, vol 34.
Academic Press, San Diego, pp 143–269
33. Møller C, Plesset MS (1934) Note on an approximation treatment for many-electron
systems. Phys Rev 46:618–622
44 A. Koleżyński
34. Head-Gordon M, Pople JA, Frisch MJ (1988) MP2 energy evaluation by direct methods.
Chem Phys Letters 153:503–506
35. Pople JA, Seeger R, Krishnan R (1977) Variational configuration interaction methods and
comparison with perturbation theory. Int J Quantum Chem 12:149–163
36. Pople JA, Binkley JS, Seeger R (1976) Theoretical models incorporating electron
correlation. Int J Quantum Chem 10:1–19
37. Raghavachari K, Pople JA (1978) Approximate fourth-order perturbation theory of the
electron correlation energy. Int J Quantum Chem 14:91–100
38. Purvis GD, Bartlett RJ (1982) A full coupled-cluster singles and doubles model: the
inclusion of disconnected triples. J Chem Phys 76:1910–1919
39. van Voorhis T, Head-Gordon M (2001) Two-body coupled cluster expansions. J Chem Phys
115:5033–5041
40. Pople JA, Head-Gordon M, Raghavachari K (1987) Quadratic configuration interaction.
A general technique for determining electron correlation energies. J Chem Phys 87:5968–
35975
41. Roos BO, Taylor PR, Siegbahn PEM (1980) A complete active space SCF method
(CASSCF) using a density matrix formulated super-CI approach. Chem Phys 48:157–173
42. Olsen J (2011) The CASSCF method: a perspective and commentary. Int J Quantum Chem
111:3267–3272
43. Buenker RJ, Peyerimhoff SD (1975) Energy extrapolation in CI calculations. Theor Chim
Acta 39:217–228
44. Pople JA, Head-Gordon M, Fox DJ, Raghavachari K, Curtiss LA (1989) Gaussian-1 theory:
a general procedure for prediction of molecular energies. J Chem Phys 90:5622–5629
45. Curtiss LA, Jones C, Trucks GW, Raghavachari K, Pople JA (1990) Gaussian-1 theory of
molecular energies for second-row compounds. J Chem Phys 4:2537–2545
46. Curtiss LA, Raghavachari K, Trucks GW, Pople JA (1991) Gaussian-2 theory for molecular
energies of first- and second-row compounds. J Chem Phys 94:7221–7230
47. Curtiss LA, Rachavachari K, Redfern PC, Rassolov V, Pople JA (1998) Gaussian-3 (G3)
theory for molecules containing first and second-row atoms. J Chem Phys 18:7764–7776
48. Curtiss LA, Redfern PC, Raghavachari K (2007) Gaussian-4 theory. J Chem Phys
126:084108
49. Feller D, Peterson KA, Dixon DA (2008) A survey of factors contributing to accurate
theoretical predictions of atomization energies and molecular structures. J Chem Phys
129:204105
50. Peterson KA, Feller D, Dixon DA (2012) Chemical accuracy in ab initio thermochemistry
and spectroscopy: current strategies and future challenges. Theor Chem Acc 131: 1079–5
51. Deyonker NJ, Cundari TR, Wilson AK (2006) The correlation consistent composite
approach (ccCA): an alternative to the Gaussian-n methods. J Chem Phys 124(11):114104
52. Petersson G (2002) Complete basis set models for chemical reactivity: from the helium atom
to enzyme kinetics. In: Cioslowski J (ed) Quantum-mechanical prediction of thermochemical
data, vol 22. Springer, Netherlands, pp 99–130
53. Thomas LH (1927) The calculation of atomic fields. Proc Camb Phil Soc 23:542–548
54. Fermi E (1928) Eine statistische Methode zur Bestimmung einiger Eigenschaften des Atoms
und ihre Anwendung auf die Theorie des periodischen Systems der Elemente. Z Phys 48:73–79
55. Dirac PAM (1930) Note on exchange phenomena in the thomas atom. Proc Camb Phil Soc
26:376–385
56. Weizsäcker CF (1935) Zur theorie der Kernmassen. Z Phys 96:431–458
57. Teller E (1962) On the stability of molecules in Thomas-Fermi theory. Rev Mod Phys
34:627–631
58. Hohenberg P, Kohn W (1964) Inhomogeneous electron gas. Phys Rev B 136:864–871
59. Kohn W, Sham LJ (1965) Self-consistent equations including exchange and correlation
effects. Phys Rev A 140:1133–1138
60. Ceperley DM, Alder BJ (1980) Ground state of the electron gas by a stochastic method. Phys
Rev Lett 45:566–569
61. Perdew JP, Zunger A (1981) Self-interaction correction to density-functional approximations

for many-electron systems. Phys Rev B 23:5048–5079
62. Perdew JP, Chevary JA, Vosko SH, Jackson KA, Pederson MR, Singh DJ, Fiolhais C (1992)
Atoms, molecules, solids, and surfaces: applications of the generalized gradient approxi-
mation for exchange and correlation. Phys Rev B 46:6671–6687
63. Perdew JP, Burke K, Ernzerhof M (1996) Generalized gradient approximation made simple.
Phys Rev Lett 77:3865–3868
64. Vosko SH, Wilk L, Nusair M (1980) Accurate spin-dependent electron liquid correlation
energies for local spin density calculations: a critical analysis. Can J Phys 58:1200–1211
65. Perdew JP, Wang Y (1992) Accurate and simple analytic representation of the electron-gas
correlation energy. Phys Rev B 45:13244–13249
66. Perdew JP, Ruzsinszky A, Csonka GI, Vydrov OA, Scuseria GE, Constantin LA, Zhou X,
Burke K (2008) Restoring the density-gradient expansion for exchange in solids and
surfaces. Phys Rev Lett 100:136406–136409
67. Becke AD (1988) Density-functional exchange-energy approximation with correct asymp-
totic behavior. Phys Rev A 38:3098–3100
68. Lee C, Yang W, Parr RG (1988) Development of the Colle-Salvetti correlation-energy
formula into a functional of the electron density. Phys Rev B 37:785–789
69. Wu Z, Cohen RE (2006) More accurate gradient approximation for solids. Phys Rev B
73:235116–235121
70. Tao J, Perdew JP, Staroverov VN, Scuseria GE (2003) Climbing the density functional
ladder: nonempirical meta-generalized gradient approximation designed for molecules and
solids. Phys Rev Lett 91:146401–146404
71. Zhao Y, Truhlar DG (2008) The M06 suite of density functionals for main group
thermochemistry, thermochemical kinetics, noncovalent interactions, excited states, and
transition elements: two new functionals and systematic testing of four M06-class
functionals and 12 other functional. Theor Chem Acc 120:215–241
72. Swart M, Solà M, Bickelhaupt FM (2009) A new all-round DFT functional based on spin
states and SN2 barriers. J Chem Phys 131:094103
73. Sun J, Perdew JP, Ruzsinszky A (2015) Semilocal density functional obeying a strongly
tightened bound for exchange. Proc Nat Acad Sci 112:685–689
74. Peverati R, Truhlar DG (2014) Quest for a universal density functional: the accuracy of
density functionals across a broad spectrum of databases in chemistry and physics. Phil
Trans R Soc A 372:20120476
75. Becke AD (1993) A new mixing of Hartree-Fock and local density-functional theories.
J Chem Phys 98:1372–1377
76. Becke AD (1993) Density-functional thermochemistry. III. The role of exact exchange.
J Chem Phys 98:5648–5652
77. Perdew JP, Ernzerhof M, Burke K (1996) Rationale for mixing exact exchange with density
functional approximations. J Chem Phys 105:9982–9985
78. Adamo C, Barone V (1999) Toward reliable density functional methods without adjustable
parameters: the PBE0 model. J Chem Phys 110:6158–6170
79. Heyd J, Scuseria GE, Ernzerhof M (2003) Hybrid functionals based on a screened Coulomb
potential. J Chem Phys 118:8207–8215
80. Gunnarsson O, Jonson M, Lundqvist BI (1976) Exchange and correlation in atoms,
molecules and solids. Phys Lett A 59:177–179
81. Alonso JA, Girifalco LA (1978) Nonlocal approximation to the exchange potential and
kinetic energy of an inhomogeneous electron gas. Phys Rev B 17:3735–3743
82. Anisimov VI, Zaanen J, Andersen OK (1991) Band theory and Mott insulators: Hubbard U
instead of Stoner I. Phys Rev B 44:943–954
83. Koleżyński A, Król M, Żychowicz M (2018) The structure of geopolymers—theoretical
studies. J Mol Struct 1163:465–471
46 A. Koleżyński
84. Groenhof G (2013) Introduction to QM/MM simulations. In: Monticelli L, Salonen E

(eds) Biomolecular simulations: methods and protocols, methods in molecular biology, vol
924. Springer Science+Business Media, New York
85. Morzan UN, Alonso de Armiño DJ, Foglia NO, Ramírez F, González Lebrero MC,
Scherlis DA, Estrin DA (2018) Spectroscopy in complex environments from QM–MM
simulations. Chem Rev 118:4071–4113
86. Koleżyński A, Mikuła A, Król M (2016) Periodic model of LTA framework containing
various non-tetrahedral cations. Spectrochim Acta A 157:17–25
87. Thomas M, Brehm M, Fligg R, Vöhringer P, Kirchner B (2013) Computing vibrational
spectra from ab initio molecular dynamics. Phys Chem Chem Phys 15:6608–6622
88. DickinsonJA Hockridge MR, Kroemer RT, Robertson EG, Simons JP, McCombie J,
Walker M (1998) Conformational choice, hydrogen bonding, and rotation of the S1 ← S0
electronic transition moment in 2-Phenylethyl Alcohol, 2-Phenylethylamine, and their water
clusters. J Am Chem Soc 120:2622–2632
89. Kubelka J, Keiderling TA (2001) Differentiation of b-sheet-forming structures: Ab
initio-based simulations of IR absorption and vibrational CD for model peptide and protein
b-sheets. J Am Chem Soc 123:12048–12058
90. Wilson EB, Decius JC, Cross PC (1955) Molecular vibrations: the theory of infrared and
Raman vibrational spectra. McGraw-Hill co., New York, USA
91. Nafie LA, Polavarapu PL (1981) Localized molecular orbital calculations of vibrational
circular dichroism. I. General theoretical formalism and CNDO results for the
carbon-deuterium stretching vibration in neopentyl-1-d-chloride. J Chem Phys 75:2935–
2944
92. Nafie LA, Freedman TB (1981) A unified approach to the determination of infrared and
Raman vibrational optical activity intensities using localized molecular orbitals. J Chem
Phys 75:4847–4851
93. Kormornicki A, McIver JW Jr (1979) An efficient ab initio method for computing infrared
and Raman intensities: application to ethylene. J Chem Phys 70:2014–2016
94. King-Smith RD, Vanderbilt D (1993) Theory of polarization of crystalline solids. Phys
Rev B 47: 1651–1564
95. Resta R (1994) Macroscopic polarization in crystalline dielectrics: the geometric phase
approach. Rev Mod Phys 66:899–915
96. Marzari N, Vanderbilt D (1997) Maximally localized generalized Wannier functions for
composite energy bands. Phys Rev B 56:12847–12865
97. Silvestrelli PL, Parrinello M (1999) Water molecule dipole in the gas and in the liquid phase.
Phys Rev Lett 82:3308–3311
98. Maschio L, Kirtman B, Orlando R, Rèrat M (2012) Ab initio analytical infrared intensities
for periodic systems through a coupled perturbed Hartree-Fock/Kohn-Sham method. J Chem
Phys 137:204113
99. Maschio L, Kirtman B, Rérat M, Orlando R, Dovesi R (2013) Ab initio analytical Raman
intensities for periodic systems through a coupled perturbed Hartree-Fock/Kohn-Sham
method in an atomic orbital basis. I. Theory. J Chem Phys 139:164101
100. Maschio L, Kirtman B, Rérat M, Orlando R, Dovesi R (2013) Ab initio analytical Raman
intensities for periodic systems through a coupled perturbed Hartree-Fock/Kohn-Sham
method in an atomic orbital basis. II. Validation and comparison with experiments. J Chem
Phys 139:164102
101. Grimme S (2004) Calculation of the electronic spectra of large molecules. In: Lipkowitz KB,
Larter R, Cundari TR (eds) Reviews in computational chemistry. Wiley
102. Dreuw A, Head-Gordon M (2005) Single-reference ab initio methods for the calculation of
excited states of large molecules. Chem Rev 105:4009–4037
103. Olivucci M, Sinicropi A (2005) Computational photochemistry. In: Olivucci M
(ed) Computational photochemistry, vol 16. Elsevier, Amsterdam
104. Gagliardia L, Roos BO (2007) Multiconfigurational quantum chemical methods for
molecular systems containing actinides. Chem Soc Rev 36:893–903
105. Burke K, Werschnik J, Gross EK (2005) Time-dependent density functional theory: past,
present, and future. J Chem Phys 123:62206
106. Adamo C, Jacquemin D (2013) The calculations of excited-state properties with
time-dependent density functional theory. Chem Soc Rev 42:845–856
107. Casida ME, Huix-Rotllant M (2012) Progress in time-dependent density-functional theory.
Annu Rev Phys Chem 63:287–323
108. Laurent AD, Jacquemin D (2013) TD-DFT benchmarks: a review. Int J Quantum Chem
113:2019–2039
109. Marques MAL, Nogueira FMS, Gross EKU, Rubio A (eds) (2012) Fundamentals of
time-dependent density functional theory, vol 837. Springer-Verlag, Heidelberg
110. Dreuw A, Head-Gordon M (2004) Failure of time-dependent density functional theory for
long-range charge-transfer excited states: the zincbacteriochlorin–bacteriochlorin and
bacteriochlorophyll–spheroidene complexes. J Am Chem Soc 126:4007–4016
111. Runge E, Gross EKU (1984) Density-functional theory for time-dependent systems. Phys
Rev Lett 52:997–1000
112. Langhoff PW, Epstein ST, Karplus M (1972) Aspects of time-dependent perturbation theory.
Rev Mod Phys 44:602–644
113. Bauernschmitt R, Ahlrichs R (1996) Treatment of electronic excitations within the adiabatic
approximation of time dependent density functional theory. Chem Phys Lett 256:454–464
114. Gross EKU, Dobson JF, Petersilka M (1996) Density functional theory of time-dependent
phenomena. Topics in chemistry. In: Nalewajski RF (ed) Density functional theory. Topics
in current chemistry, vol 181. Springer-Verlag, Berlin
115. Improta R (2012) UV-visible absorption and emission energies in condensed phase by PCM/
TD-DFT methods. In: Barone V (ed) Computational strategies for spectroscopy: from small
molecules to nano systems. Wiley, Hoboken, New Jersey
116. Stratmann RE, Scuseria GE, Frisch MJ (1998) An efficient implementation of
time-dependent density-functional theory for the calculation of excitation energies of large
molecules. J Chem Phys 109(19):8218–8224
117. Mennucci B, Cammi R (eds) (2008) Continuum solvation models in chemical physics: from
theory to applications. Wiley, Chichester, England
118. Cossi M, Rega N, Scalmani G, Barone V (2003) Energies, structures, and electronic
properties of molecules in solution with the C-PCM solvation model. J Comput Chem
24:669–681
119. Mennucci B, Tomasi J, Cammi R, Cheeseman JR, Frisch MJ, Devlin FJ, Gabriel S,
Stephens PJ (2002) Polarizable continuum model (PCM) calculations of solvent effects on
optical rotations of chiral molecules. J Phys Chem A 106:6102–6113
120. Klamt A, Schüürmann G (1993) COSMO: a new approach to dielectric screening in solvents
with explicit expressions for the screening energy and its gradient. J Chem Soc, Perkin Trans
2:799–805
121. Klamt A (2005) From quantum chemistry to fluid phase thermodynamics and drug design.
Elsevier, Boston, USA
122. Skyner RE, McDonagh JL, Groom CR, van Mourik T, Mitchell JBO (2015) A review of
methods for the calculation of solution free energies and the modelling of systems in
solution. Phys Chem Chem Phys 17:6174–6191
123. Kamerlin SCL, Haranczyk M, Warshel A (2009) Are mixed explicit/implicit solvation
models reliable for studying phosphate hydrolysis? A comparative study of continuum,
explicit and mixed solvation models. Chem Phys Chem 10:1125–1134
124. Fletcher RE, Ling S, Slater B (2017) Violations of Lowenstein{‘}s rule in zeolites. Chem Sci
8:7483–7491
125. Ikeda T, Izumi F, Kodaira T, Kamiyama T (1998) Structural study of sodium-type zeolite
LTA by combination of Rietveld and maximum-entropy methods. J Chem Mat 10:3996–
4004
48 A. Koleżyński
126. Pluth JJ, Smith JV (1980) Accurate redetermination of crystal structure of dehydrated zeolite
A. Absence of near zero coordination of sodium. Refinement of silicon, aluminum-ordered
superstructure. Am Chem Soc 102:4704–4708
127. Bellaiche L, Vanderbilt D (2000) Virtual crystal approximation revisited: application to
dielectric and piezoelectric properties of perovskites. Phys Rev B 61:7877–7882
128. Íñiguez J, Vanderbilt D, Bellaiche L (2003) First-principles study of (BiScO3)1-x–(PbTiO3)x
piezoelectric alloys. Phys Rev B 67:224107
129. Winkler B, Pickard C, Milman V (2002) Applicability of a quantum mechanical “virtual
crystal approximation” to study Al/Si-disorder. Chem Phys Lett 362:266–270
130. Soven P (1966) Coherent-potential model of substitutional disordered alloys. Phys Rev
156:809–813
131. Velický B, Kirkpatrick S, Ehrenreich H (1968) Single-site approximations in the electronic
theory of simple binary alloys. Phys Rev 175:747–766
132. Velický B (1969) Theory of electronic transport in disordered binary alloys:
coherent-potential approximation. Phys Rev 184:614–627
133. Drożdż E, Koleżyński A (2017) The structure, electrical properties and chemical stability of
porous Nb-doped SrTiO3—experimental and theoretical studies. RSC Adv 7:28898–28908
134. Mikuła A, Drożdż E, Koleżyński A (2018) Electronic structure and structural properties of
Cr-doped SrTiO3. Theoretical investigation. J Alloys Compd 749:931–938
135. Kupwade-Patil K, Soto F, Kunjumon A, Allouche E, Mainardi D (2013) Multi-scale
modeling and experimental investigations of geopolymeric gels at elevated temperatures.
Comp Struct 122:164–177
136. Kroll P (2003) Modelling and simulation of amorphous silicon oxycarbide. J Mater Chem
13:1657–1668
137. Tian KV, Mahmoud MZ, Cozza P, Licoccia S, Fang D, Di Tommaso D, Chass GA,
Greaves N (2016) Periodic vs. molecular cluster approaches to resolving glass structure and
properties: anorthite a case study. J Non-Cryst Solids 451:138–145
138. White CE, Provis JL, Proffen T, Riley DP, van Deventer JS (2010) Combining density
functional theory (DFT) and pair distribution function (PDF) analysis to solve the structure
of metastable materials: the case of metakaolin. Phys Chem Chem Phys 12:3239–3245
139. Chung LW, Sameera WMC, Ramozzi R, Page AJ, Hatanaka M, Petrova GP, Harris TV,
Li X, Ke Z, Liu F, Li HB, D L, Morokuma K (2015) The ONIOM method and its
applications. Chem Rev 115:5678–5796
140. Asthagiri A, Janik MJ (eds) (2013) Computational catalysis: RSC (Catalysis series). Roy
Soc Chem, UK
141. van Santen RA, Neurock M (2006) Molecular heterogeneous catalysis: a conceptual and
computational approach. Wiley-VCH Verlag GmbH & Co., Weinheim, Germany
142. van Santen RA, Sautet P (eds) (2009) Computational methods in catalysis and materials
science: an introduction for scientists and engineers. Wiley-VCH Verlag GmbH & Co.,
Weinheim, Germany
143. Thiel W (2014) Computational catalysis—past, present, and future. Angew Chem Int Ed
53:8605–8613
144. https://en.wikipedia.org/wiki/List_of_quantum_chemistry_and_solid-state_physics_
software
Chapter 2
Scaling Procedures in Vibrational
Spectroscopy
Olga Bąk and Piotr Borowski
Abstract This chapter contains a brief review of the up-to-date scaling procedures
that are used in the computational vibrational spectroscopy to improve agreement
between the calculated harmonic frequencies and the observed fundamentals.
Initially, the basics of vibrational spectroscopy are reminded. This includes the
concept of potential energy surface, harmonic approximation, and a basic quantum
chemistry treatment of the anharmonicity for a diatomic molecule. Brief description
of the Wilson–Decius–Cross method for polyatomic molecules is also presented.
Then the commonly used scaling procedures are discussed. The distinction between
single- and multi-parameter scaling procedures is made. Four scaling procedures are
reviewed. First, Pople’s uniform scaling is presented. Second, Yoshida’s
wavenumber linear scaling method is discussed. Both methods are simple
single-parameter frequency scaling methods. Then basics of two multi-parameter
scaling methods, which are much more accurate but less straightforward to use, are
given. Thus, Pulay’s scaled quantum mechanical force field method, in which
scaling factors are applied directly to the calculated force constants is reviewed.
Finally, introduction to quite recently proposed multi-parameter frequency scaling
method, called effective scaling frequency factor method, is provided. The relevant
sections start with a short description of the theory for a given method. Then a brief
literature review on the historical background of methodology development is
given.
List of Acronyms
ARPE Average relative percentage error
ESF Effective scaling factor
ESFF Effective scaling frequency factor method
FC Force constant
FF Force field
IC Internal coordinate
O. Bąk P. Borowski (&)

Faculty of Chemistry, Maria Curie-Skłodowska University,
Maria Curie-Skłodowska Sq. 3, 20-031 Lublin, Poland
e-mail: piotr.borowski@poczta.umcs.lublin.pl

50 O. Bąk and P. Borowski
LSF Local scaling (scale) factor

LSMF Least-squares merit function
NIC Natural internal coordinate
PEC Potential energy curve
PES Potential energy (hyper)surface
PIC Primitive internal coordinate
QC Quantum chemistry (methods, calculations…)
RMS Root-mean-square (deviation)
SF Scaling factors
SQM(FF) Scaled quantum mechanical (force field) method; simply SQM
US Uniform scaling
WDC Wilson–Decius–Cross method
WLS Wavenumber linear scaling
X A general second-row atom
Y A general third-row atom
ZPVE Zero-point vibrational energy correction
Symbols Used in Mathematical Formulas

A, B, … General indices used in summations over atoms
B Transformation matrix between internal and Cartesian
coordinates
f ¼ ðf1 ; f2 ; . . .; fNscl Þ A vector with scaling factors
eff eff
f ¼ f1 ; f2 ; . . .; fK
eff eff Effective scaling factors
opt opt opt A vector with optimal scaling factors
f ¼ f1 ; f2 ; . . .; fNscl
opt
fx Cartesian force constant matrix with elements fijx

F Force constant matrix in internal coordinate represen-
tation with elements Flm
i, j, … General indices used in summations over Cartesian
coordinates
I, J, … General indices used to indicate types of internal
coordinates in multi-parameter scaling procedures
K Number of vibrational degrees of freedom, K = 3N − 6
(or K = 3N − 5 in the case of linear molecules)
L Number of redundant primitive internal coordinates
(typically L > 3N)
L Transformation matrix between Cartesian displace-
ments and normal coordinates
N Number of atoms in a molecule
Nmol Number of molecules in a training set
Nscl Number of scaling factors
Ntyp Number of types (groups) of internal coordinates (=Nscl)
2 Scaling Procedures in Vibrational Spectroscopy 51
Nvib Number of vibrational modes used in optimization of

scaling factors; additional superscript indicates the
molecule the number of modes refers to
p, q, … General indices used in summations over frequencies of
a training set of molecules, basis functions, etc.
s = (s1, s2, …) Vector with internal coordinates; additional superscripts
refer to the type a given coordinate belongs to.
Additional subscript “e” denotes the equilibrium values.
When used in the potential energy expression the
symbols refer to the deviations of coordinates from their
equilibrium values
T Kinetic energy
U Potential energy
w Weight used for a given frequency in the scaling factor
optimization procedure
a Transformation matrix between internal coordinates
displacements and normal coordinates
hvib Vibrational temperature
l, m, … General indices used in summations over internal or
normal coordinates as well as normal modes of a
molecule
mh , mexpt , mexpt;h , mscl Harmonic, experimental, experimental harmonic, and
scaled frequencies, respectively (in cm−1)
whp Eigenfunctions of harmonic oscillator
2.1 Introduction
Computational chemistry turned out to be an excellent tool for experimentalists,

which helps to understand physicochemical phenomena at the molecular level,
starting from structure determination, reactions mechanisms investigation, etc., and
ending at the interpretation of various spectra. One of the most important appli-
cations is interpretation of the vibrational spectra of the molecules. Before the
advent of modern NMR, vibrational spectroscopy was the main spectroscopic
method for molecular characterization. There are two main disadvantages of
vibrational spectroscopy compared to NMR. First, most vibrations are strongly
delocalized, precluding an easy assignment of a structure based on the spectra, and
restricting vibrational spectroscopy to fingerprinting. Second, the resolution of
vibrational spectra is limited. However, its major advantage over NMR is much
higher sensitivity. Modern quantum chemistry methods are capable of predicting
vibrational frequencies with remarkable accuracy, and intensities fairly well with
modest computational effort. This effectively eliminates the first disadvantage if the
identity of the species can be narrowed down to a small number of candidates.
Thus, vibrational spectroscopy supported by quantum chemistry calculations
appears to be increasingly effective structural tool in modern chemistry.
Scaling procedures are now routinely used to improve agreement between the
calculated harmonic frequencies and the observed fundamentals. In this chapter, a
short review of the available scaling methods will be given. This includes the
simplest uniform scaling, wavenumber linear scaling, scaled quantum mechanical
force field, and quite recently proposed effective scaling frequency factor approa-
ches. This is not the authors’ intention to provide a detailed review of the up-to-date
literature. In the subsequent sections, we will briefly summarize the main “theory”
of each method and then describe methodology development from the historical
point of view, citing the methodological references. No papers with applications
only will be cited. In addition, we will also make no citations to quantum chemistry
methods we are referring to, basis sets, etc., to make the reference list as short as
possible. We will use a number of acronyms to keep the text concise. The first time
(apart from this section) a given term appears, the acronym will be given in
parentheses. Then it will be used throughout the chapter in most cases. A list of
acronyms extended by the explanation of mathematical symbols used is provided in
a separate section in the beginning of this chapter for the reader’s convenience.
2.2 Fundamentals
In this section, fundamentals of vibrational spectroscopy will be given. Basic ideas,

including the concept of the potential energy curve and (hyper)surface, harmonic
approximation, anharmonicity treatment, will be reminded. Finally, an outline of
the Wilson–Decius–Cross method of polyatomic molecules vibrations treatment
will be given.
2.2.1 Potential Energy Surface
Within Born–Oppenheimer approximation the potential energy curve (PEC) U(r) of

a diatomic molecule, where r is the internuclear distance, has a Morse-like shape
shown in Fig. 2.1a. It can be approximated using various quantum chemistry
(QC) methods, like the Hartree–Fock (HF) method, post-Hartree–Fock methods
(MCSCF, MP2, CI, CC…), DFT methods, etc. The accuracy of each approximation
is different with respect to the dissociation energy De, curvature of PEC at mini-
mum, etc. It can be found, say, in the pointwise way by calculating energies for a
number of
internuclear distances, which affords the curve as a set of points
rp ; U rp ; p ¼ 1; 2; . . . From the point of view of vibrational spectroscopy, the
most important point on PEC is re corresponding to minimum, called equilibrium

bond length. Alternatively, one can expand PEC in a Taylor series around re, i.e.,
1 1
U ðr Þ ¼ U ðre Þ þ U ð1Þ ðre Þðr re Þ þ U ð2Þ ðre Þðr re Þ2 þ U ð3Þ ðre Þðr re Þ3
2 6
1 ð4Þ 4
þ U ðre Þðr re Þ þ
24
ð2:1Þ
where

@U
U ð 1 Þ ðre Þ ¼ U 0 ðre Þ ¼ ; ð2:2Þ
@r r¼re
2
ð2Þ 00 @ U
U ðre Þ ¼ U ðre Þ ¼ ; etc: ð2:3Þ
@r 2 r¼re
The linear term vanishes for obvious reason. The derivatives, in particular third
and higher derivatives, can be calculated numerically; in the case of second
derivatives, most QC packages have analytic derivatives implemented. The
expansions of PEC through fourth order are shown in Fig. 2.1b.
In the case of polyatomic molecules (N-atomic; from now on N denotes number
of atoms; other symbols introduced in a similar way will be also used throughout
the entire text) possessing K ¼ 3N 6 (or K ¼ 3N 5 for linear molecules)
vibrational degrees of freedom the potential energy hypersurface, frequently called
potential energy surface (PES), is a function of K variables called internal coor-
dinates (ICs) which form a vector sT ¼ ðs1 ; s2 ; . . .; sK Þ (“T” means transpose; we
prefer to define vectors as column vectors). There is no unique choice of ICs. They
can be ordinary primitive ICs (PICs; bond lengths, valence and torsion angles),
natural ICs (NICs) [1], etc. PES exhibits a minimum at a point called equilibrium
geometry se. Expansion of PES in Taylor series around se gives
Fig. 2.1 a PEC of a diatomic molecule. b Expansions of PEC in Taylor series through fourth
order around minimum
1X K
ð2Þ

U ðsÞ ¼ U ðse Þ þ Ulm ðse Þ sl sl;e sm sm;e
2 l;m¼1
ð2:4Þ
1 X K
ð3Þ

þ Ulmj ðse Þ sl sl;e sm sm;e sj sj;e þ
6 l;m;j¼1
where

ð2Þ 00 @2U
Ulm ðse Þ ¼ Ulm ðse Þ ¼ ; etc: ð2:5Þ
@sl @sm s¼se
An example of hypothetical PES for a symmetric nonlinear diatomic molecule

(like water), or rather its cross section for a fixed angle, as well as the second-order
Taylor expansion is shown in Fig. 2.2.
2.2.2 One Dimensional Vibrational Problem
Harmonic oscillator (Fig. 2.3a), i.e., oscillator fulfilling Hooke’s law, has a quad-
ratic PEC
1
U ð xÞ ¼ f ð x xe Þ 2 ð2:6Þ
2
where f is the so-called (quadratic) force constant (FC). It can be shown that
vibration of two masses m1 and m2 connected by a weightless spring (Fig. 2.3b) is
Fig. 2.2 a Hypothetical cross section of PES for a water molecule at a fixed angle. b Expansions
of PES in Taylor series through second order around minimum
Fig. 2.3 One-dimensional oscillator. The descriptions of vibrations of: a mass l against infinitely
heavy wall, and b two masses m1 and m2 such that their reduced mass equals l are identical
equivalent to the vibration of a mas l relative to infinitely heavy “wall” (Fig. 2.3a)
provided
m1 m2
l¼ : ð2:7Þ
m1 þ m2
l is called a reduced mass of masses m1 and m2. Figure 2.3b is just a model for
the vibrating diatomic molecule—two heavy masses (atoms) connected via
weightless spring vibrate in a potential shown in Fig. 2.1a, which, for small dis-
placements from the equilibrium position, can be approximated by a parabola
(n = 2). In addition, comparing Eqs. (2.6) and (2.1) it can be concluded that the FC
is just a second energy derivative with respect to displacement calculated at equi-
librium, i.e., f ¼ U ð2Þ ðre Þ.
Solution of a Schrödinger equation with Hamiltonian
^ ¼ h d þ 1 f ðr re Þ2
2 2
H ð2:8Þ
2l dr 2 2
gives a well-known formula for the vibrational energy levels of harmonic oscillator,
namely

1
Et ¼ hm0 t þ ð2:9Þ
2
where t = 0, 1, 2, … is the vibrational quantum number, and m0 is the classical

vibrational frequency, i.e.,
sffiffiffi
1 f
m0 ¼ : ð2:10Þ
2p l
It follows that vibrational energy levels are equidistant with the energy gap equal
to DE ¼ hm0 . Assuming that molecule behaves like harmonic oscillator (which is a
reasonably good approximation taking into account the fact, that vibrational
amplitudes are rather low, at least for first two energy levels) one concludes that
Bohr condition, i.e., DE ¼ hm, where m is the frequency of radiation, is fulfilled when
m = m0, i.e., when radiation frequency is equal to the classical vibrational frequency
of a molecule. This is why we say that the band on a spectrum corresponds to a given
vibrational frequency. It should be remembered, however, that the band corresponds
to transitions between various energy levels, e.g., t = 0 ! t = 1 in the case of
fundamentals. Thus, in order to find the position of a spectral line on a (pure)
vibrational spectrum of a diatomic molecule, one has to calculate m0. This means that
at a given computational level one has to find re (geometry optimization), next
calculate f as a second energy derivative at re, and finally apply Eq. (2.10).
Consider the simple case of HCl molecule. It is well known that the fundamental
(experimental) vibrational frequency in gas phase is 2886 cm−1. In fact, this is the
wavenumber where the Q-branch (absent on the rotation–vibration spectrum) would
occur. The band’s position calculated according to Eq. (2.10) using CCSD/
aug-cc-pVTZ computational level affords the value 3014 cm−1 (experimental har-
monic frequency based on f ¼ 516 Nm [2] is 2989 cm−1; in addition, reCCSD ¼ 1:2766 Å
and reexpt ¼ 1:2746 Å [2]), which is nearly 130 cm−1 higher as compared with
experimental fundamental. This overestimation is obvious. Parabolic PEC is only an
approximation, and the real PEC of HCl is anharmonic. Apparently, harmonic
approximation overestimated the observed vibrational frequency, and therefore, more
sophisticated treatment is needed to obtain better agreement with the experimental
values. This can be accomplished by means of perturbation or variation approaches.
After neglecting fifth and higher-order terms in Eq. (2.1), the Hamiltonian for a
vibrating molecule is
^ ¼ h d þ 1 f ðr re Þ2 þ 1 f ð3Þ ðr re Þ3 þ 1 f ð4Þ ðr re Þ4
2 2
H ð2:11Þ
2l dr 2 2 6 24
where f ð3Þ and f ð4Þ are cubic and quartic FCs, respectively, i.e., third and fourth
energy derivatives calculated at re, cf. Equations (2.1)–(2.3). In the case of per-
turbation approach terms
^ ð1Þ ¼ 1 f ð3Þ ðr re Þ3 and H

H ^ ð2Þ ¼ 1 f ð4Þ ðr re Þ4 ð2:12Þ
6 24
are first- and second-order perturbations to the unperturbed (harmonic oscillator)

Hamiltonian (2.8), for which exact solutions, whp and Eph , are known (symbol “h”
stands for “harmonic”). One can easily calculate f ð3Þ and f ð4Þ by numerical energy
differentiation. Using central differences on energy method, we obtained f ð3Þ ¼
0:985 a.u. (step h ¼ 0:01 Å) and f ð4Þ ¼ 2:51 a.u. (step h ¼ 0:02 Å; in view of
differentiating energy, this value may not be very accurate, sufficient for our pur-
pose, though). These values can be used in standard perturbation theory formulas
for first-order correction to the wavefunction and first- and second-order corrections
to the energy (see, e.g., [3]) for states t = 0 and t = 1. The integrals that have to be
computed read as follows:
Zþ 1 Zþ 1
wht r 4 wht dr and whp r 3 wht dr where t ¼ 0; 1; p ¼ 0; 1; 2; . . . and
1 1
p 6¼ t:
ð2:13Þ
Skipping the details of such calculations (relevant integrals were calculated using
Wolfram Mathematica program for the purpose of this chapter; only some of them
were not equal to zero) we obtain second-order corrections to energy equal to
−12 cm−1 for t = 0, and −116 cm−1 for t = 1 (first-order corrections to energy
R þ1
proportional to 1 wht r 3 wht dr according to formulas given in [3] are zero for
symmetry reason). Since E0h 1507 and E1h 4521 cm−1, (DE h ¼ E1h E0h 3014
cm−1, as reported above), we conclude that the energy gap within the second-order
perturbation theory corresponds to 2910 cm−1, which is much closer to the exper-
imental value of 2886 cm−1.
Alternatively, one can use variation method within, say, Ritz framework, and
choose the trial wavefunction as a linear combination of harmonic oscillator
eigenfunctions whq which form a basis set, i.e.,
X
M
wtrial
t ¼ cqt whq ; t ¼ 0; 1: ð2:14Þ
q¼0
Assuming again that fifth and higher-order terms in the expansion (1) are neg-
ligible, i.e., the Hamiltonian (11) for anharmonic oscillator is exact (which is the
case for t = 0 to a very good approximation; for t = 1 higher-order terms are also
important) the Hamiltonian matrix can be easily set up by calculating integrals
(again, they were calculated using Wolfram Mathematica program)
Zþ 1 Zþ 1 Zþ 1 Zþ 1
d2 h
whp w dr; whp r 2 whq dr; whp r 3 whq dr and whp r 4 whq dr:
dr 2 q
1 1 1 1
ð2:15Þ
Straightforward diagonalization using, say, ordinary Jacobi routine, provides

eigenvalues and the corresponding eigenvectors (note that harmonic oscillator
eigenfunctions are orthonormal, so no overlap matrix needs to be calculated). The
lowest eigenvalue is an upper bound to the ground state (t = 0) energy. For t = 0
and 1, M = 6 turned out to be enough, i.e., practically no changes in two lowest
eigenvalues were observed upon further increasing of M, so the basis set for M = 6
can be regarded as (nearly) complete (unfortunately, such a low M value is not the
case in the electronic structure theory), and wavefunctions wtrial trial

0 and w1 as (nearly)
exact solutions. Since wtrial trial
1 is orthogonal to w0 , the second lowest eigenvalue is an
upper bound to the first excited state energy (t = 1). The two lowest eigenvalues
correspond to E0trial 1498, and E1trial 4434 cm−1, and thus, the energy gap cor-
responds to 2936 cm−1. The difference between theoretical and experimental
transition frequency is therefore 50 cm−1, which is again much lower than within
the harmonic oscillator framework (130 cm−1). Since, as mentioned above, the
calculated wavefunctions are (nearly) exact for Hamiltonian (2.11), and assuming
that the overall picture will not change upon inclusion of fifth and higher-order
terms, the difference as large as 50 cm−1 should be regarded as following from the
shortcomings in the treatment of correlation effects at the adopted computational
level.1 In view of this discussion, the better performance of the (approximate)
second-order perturbation approach (yielding the difference 24 cm−1) is due to
fortuitous cancelation of errors.
The relative error between the calculated harmonic and the experimental fun-
damental frequency is quite significant. In the case described above, it is equal to
4.4%. The values obtained within the perturbation or variation frameworks are also
not perfect. The reported differences of 24 and 50 cm−1 follow from the approxi-
mate nature of the calculations. First, we considered merely cubic and quartic terms
in the expansion of PEC. In addition, we used the second-order correction within
the perturbation framework. In the case of variation procedure, we used expansion
(2.14) which is long enough to assure (nearly) exact results for a given computa-
tional level and the assumed form of Hamiltonian (2.11). However, the quadratic,
cubic, and quartic FCs are approximate (recall that the CCSD/aug-cc-pVTZ method
was used), which also results in the above-mentioned deviations.
The energy levels of anharmonic oscillator are no longer equidistant. They
approach each other, which means that harmonic approximation overestimates the
observed fundamentals (in an overwhelming majority of cases). In addition, the
approximate nature of calculations (neglect of a part of correlation effects, and basis
set incompleteness) leads to errors in the predicted geometry and/or curvature of
PEC around the equilibrium position, which result in the error in the predicted FC,
and, consequently in the predicted harmonic frequency. It should be emphasized
that even in the case of highly accurate geometries (in the case considered above,
the deviation between the computed and experimental bond length is 0.16%), there
is no guarantee that the calculated curvature is correct. For these reasons, additional
endeavors are needed for the harmonic approximation to be conclusive when
applied to various physicochemical problems. One of them is scaling techniques
discussed in the present chapter.
1
In fact, this is not the case. Inclusion of fifth-order term to Hamiltonian (2.11) with the estimated
value of f(5) = -5.2 a.u. lowers the predicted by variation method transition frequency from
2936 cm−1 down to 2911 cm−1 (which is mostly due to substantial lowering of energy for t= 1).
This value is only 25 cm−1 higher as compared with the fundamental (25 cm−1 is also the dif-
ference between the calculated and experimental harmonic frequencies).
2.2.3 Wilson–Decius–Cross Method
As stated in the previous section within the harmonic oscillator approximation, a band
on the vibrational spectrum corresponding to t = 0 ! t = 1 transition appears at the
wavenumber corresponding to classical vibrational frequency. This can be general-
ized; finding classical vibrational frequencies of a polyatomic molecule will give us its
approximate vibrational spectrum. The detailed derivation of relevant formulas is
given elsewhere [4, 5] and will not be repeated here. We merely present the proper
practice and report formulas necessary for future discussion.
Finding classical vibrational frequencies of a polyatomic molecule consists in
solving Lagrange equations of motion. This requires the knowledge of the vibra-
tional kinetic and potential energies of a molecule. The most straightforward
coordinate system to treat molecular vibrations is the Cartesian system, the proper
choice of which is the following: Its origin is fixed at the center of mass of a
molecule, and it rotates with the molecule. The instantaneous position of the Ath
atom is rA ¼ ðxA ; yA ; zA Þ, while its equilibrium position is rA;e ¼ xA;e ; yA;e ; zA;e .
First it should be noted that change of any of the 3N Cartesian coordinates from
its equilibrium position results in translation and possibly rotation of a molecule,
P
since NA¼1 MA rA 6¼ 0 and, where MA denotes mass of the Ath atom. To a good
approximation kinetic energy T of a molecule is given by
1X N
T¼ MA jq_ A j2 ð2:16Þ
2 A¼1

where qA ¼ rA rA;e ¼ xA xA;e ; yA yA;e ; zA zA;e ¼ ðDxA ; DyA ; DzA Þ, and the
dot denotes time derivative, i.e., velocity of the Ath atom, provided relations
X
N
M A rA ¼ 0 ð2:17Þ
A¼1
and
X
N
MA rA;e rA ¼ 0 ð2:18Þ
A¼1
are satisfied. Equations (2.17) and (2.18) are called first and second Sayvetz con-
ditions, respectively (sometimes Eckart conditions) and denote that whenever
atomic vibration is to generate translation (generate a change in a position of a mass
center) and/or rotation (generate zero-order vibrational angular momentum; this can
be easily seen after differentiation of Eq. (2.18) with respect to time) the coordinate
system adjusts in such a way that both components of the motion are eliminated.
These conditions were introduced to eliminate to the best possible extent coupling
between translation and vibrations as well as translation and rotation in a general
kinetic energy expression. The term coupling molecular rotation and vibrations
cannot be eliminated though, but in low-resolution spectroscopy, it can be neglected
due to its insignificant contribution to the overall kinetic energy expression.
Introducing, for brevity, vectors dT ¼ ðd1 ; d2 ; d3 ; . . .; d3N Þ ¼ ðDx1 ; Dy1 ; Dz1 ; Dx2 ;
Dy2 ; Dz2 ; . . .; DxN ; DyN ; DzN Þ and mT ¼ ðM1 ; M1 ; M1 ; . . .; MN ; MN ; MN Þ we obtain
1X 3N
T¼ mi d_ 2i : ð2:19Þ
2 i¼1
Second, change of any of the 3N coordinates from its equilibrium position

results in change of the potential energy U of a molecule. Thus, to second-order
(harmonic approximation), U takes the form
3N 2
1X @ U 1X 3N
U ¼ Ue þ di dj ¼ Ue þ f x di dj ð2:20Þ
2 i;j¼1 @di @dj e 2 i;j¼1 ij
where the subscript “e” denotes,

that the values are computed for the equilibrium
coordinates, e.g., Ue ¼ U r1;e ; r2;e ; r3;e ; . . .; rN;e . Equation (2.20) defines matrix fx,
which is called Cartesian force constant matrix, or Cartesian force field (FF) of a
molecule. The elements of this matrix are now computed by QC packages at
various computational levels, in most cases in an analytic way.
Solving Lagrange equations of motion with T and U given by Eqs. (2.19) and
(2.20), respectively, leads after some manipulations (a tedious but straightforward
procedure) to the well-known Wilson–Decius–Cross (WDC) equations, which in
matrix notation (dimension 3N 3N) take the form
2
f x L ¼ ML mh ð2:21Þ
where M is a diagonal matrix with atomic masses on the main diagonal

ðMij ¼ mi dij Þ, m is a diagonal matrix mij ¼ mi dij with harmonic frequencies on
h h h
the main diagonal, and L is a transformation matrix between Cartesian displace-

ments d and normal coordinates QT ¼ Q1 ; Q2 ; . . .; QK ; Tx ; Ty ; Tz ; Rx ; Ry ; Rz , i.e.,
d ¼ LQ: ð2:22Þ
The L matrix defines normal coordinate vector which includes 6 (5 in the case of
linear molecules) additional coordinates describing translations and rotations. They
correspond to zero “frequencies”. The lth column of L provides (relative) ampli-
tudes for atomic displacements associated with lth frequency. Normal coordinates
satisfy the Sayvetz conditions in that when the molecule is distorted by adding
amplitudes corresponding to lth column (or its multiple), Eqs. (2.17) and (2.18) are
satisfied. Kinetic and potential energies of a molecule when expressed in terms of
normal coordinates are both diagonal, i.e.,
1 _T_ 1X K
1 2 1X K
Q_ 2
2
T¼ Q Q¼ and U ¼ Q T mh Q ¼ ml Q2l : ð2:23Þ
2 2 l¼1 l 2 2 l¼1
This is to be contrasted with Cartesian coordinates expressions (2.19) and (2.20)

which clearly show that Cartesian motions are coupled (the f x matrix is not diagonal).
For this reason, the total vibrational Hamiltonian expressed in terms of the normal
coordinates is a sum of independent terms, each one having a form of Eq. (2.8) with
r replaced by Q (the remaining modifications can be easily deduced). Thus, the
molecular vibrations can be resolved in terms of K independent normal vibrations:
Total vibrational wavefunction is a product of terms, each one associated with a given
normal coordinate, and total energy is a sum of the corresponding energies.
For the sake of discussion on the multi-parameter scaling procedures, it is
necessary to recall projection of equation (2.21) onto the space spanned by ICs of a
molecule. Given a set of ICs sT ¼ ðs1 ; s2 ; . . .; sK Þ the Cartesian FF f x can be
transformed to IC basis representation F according to
T
F ¼ B1 f x B1 : ð2:24Þ
B1 is a 3N K matrix containing the derivatives of Cartesian displacements

with respect to ICs, i.e.,
@di
B1
il ¼ ð2:25Þ
@sl
satisfying the Sayvetz conditions. It can be obtained, e.g., from the K 3N B matrix
@s
containing @dli derivatives (expressions for the B matrix elements are reported
elsewhere, see, e.g., [5]) using the so-called generalized Moore–Penrose matrix
inverse and projecting away translations and rotations from it. Alternatively, one
can define 3 translation and 3 (2 in the case of linear molecules) rotation coordinates
[5], differentiate them with respect to Cartesian displacements and extend the
B matrix to be of dimension 3N 3N. This matrix can be readily inverted using
“ordinary” inverse, and the vectors corresponding to translations and rotations
omitted in further calculations. The latter procedure is illustrated in Fig. 2.4.
Both procedures should give identical B−1 matrices, as we verified by coding

relevant routines in our lab. Note that both B−1 and F matrices can be also obtained
numerically in a very straightforward way. It can be shown that the set of Eq. (2.21)
of dimension now 3N 3N takes the (K K) form
2
Fa ¼ G1 a mh ð2:26Þ
Fig. 2.4 Construction of a B−1 matrix
T
where G1 ¼ B1 MB1 is the so-called mass tensor which is multidimensional
generalization of a reduced mass ðT ¼ 12 s_ T G1 s_ Þ and a is a transformation matrix
between internal and normal coordinates
s ¼ aQ for the 00 Q part00 of Q: ð2:27Þ
In addition, we have L ¼ B1 a and
1
U ¼ sT Fs: ð2:28Þ
2
In the potential energy expression s denotes deviations of internal coordinates

from their equilibrium values ðs se Þ, for simplicity. There are some advantages of
using Eq. (2.26) instead of (2.21) in particular, when a new QC method has to be
tested for the prediction of harmonic frequencies of small molecules. Apparently,
much fewer FCs need to be calculated numerically by central differences on energy,
which is always coded first prior to coding analytic gradient and hessian. Note that
Eq. (2.26) is a generalization of Eq. (2.10) as it reduces to the latter one in
one-dimensional case with l ¼ G1 11 .
Equations (2.21) and (2.26) provide harmonic vibrational frequencies along with
the atomic Cartesian/internal displacement amplitudes (L/a matrices), the latter
ones being used in the animation of normal modes by graphical interfaces to various
QC packages.
2.3 Scaling Procedures
Theoretical prediction of the vibrational spectra, i.e., calculations of the harmonic

frequencies, became affordable in the end of 1960s [6] when nuclear gradient
formulas become available, initially at the HF computational level with small basis
sets. Numerical differentiation was used to obtain the FC matrix. Somewhat later it
was recognized that the calculated harmonic frequencies significantly overestimate

the observed fundamentals [7]. In the case of the HF procedure, the typical error in
prediction of the vibrational frequencies was found to be up to 15%. Analytic FC
calculations at the HF level were presented in late 1970s [8]. Further development
of efficient algorithms for the correlated QC methods in conjunction with the
outburst of computational facilities made it possible to predict much more accurate
FCs. The calculated frequencies were still too high as compared with fundamentals.
The reasons are analogous to those described in Sect. 2.2.2 for a simple example of
a diatomic molecule:
• the harmonic approximation used in typical applications of the WDC method,
and
• approximations introduced when solving the Schrödinger equation to obtain the
FC matrix (consisting in the incomplete incorporation of the correlation effects
and incompleteness of the basis set used in the calculations). They have the
direct effect on both the quality of the obtained equilibrium geometry and the
curvature of the (approximate) PES with respect to the atomic displacements.
In most cases, harmonic approximation would lead to overestimation of the
observed fundamentals even if exact second energy derivatives at exact equilibrium
geometry were available (which is, of course, impossible). Second item can be
developed as follows. Consider, for example, simple Hartree–Fock approximation.
The approximate wavefunction, i.e., the Slater determinant, is regarded as uncor-
related wavefunction. The lack of bonding–antibonding orbital excitations (which
“remove” a part of bonding charge density from between the nuclei, like in the case
of post-HF methods) leads to too short and, consequently, to strong bonds. The
resultant overestimated FCs lead to overestimated frequencies even in the case of
perfectly harmonic vibrations.
Thus, the calculated frequencies have to be corrected to be applicable in the
interpretation process. The empirical scaling procedures came into being in the
1970s as a consequence of the above-mentioned observations. They constitute an
alternative to the purely QC treatment within variational or perturbation formalisms
analogous to the simple treatment described for diatomic molecules in Sect. 2.2.2.
The formal procedure of the determination of the vibrational spectrum beyond
harmonic approximation is based on solving the Schrödinger equation for the
nuclear motion in an approximate way with at least cubic potential. It is therefore
time-consuming; nowadays, it can be applied to fairly small systems at a reasonably
high computational level in spite of significant progress in computational chemistry.
The review of the relevant literature is beyond the scope of the present chapter; the
reader is referred to the original papers (see, e.g., [9, 10], and references therein). In
such cases, the scaling procedures, in particular multi-parameter ones, will be the
methods of choice for a long time, just as the methods of molecular mechanics are
the methods of choice in the determination of, e.g., the protein structure. The
additional computational time needed to determine the scaled frequencies is neg-
ligible as compared with that needed for the determination of the molecular
quadratic FF. Taking into account the high accuracy of the methods we may easily
predict their dominance in the nearest future, at least with respect to the systems of
the appreciable size. It should be noted at this stage that—as the pioneers of the
scaling procedures claim (see, e.g., [11])—none of them has a strict theoretical
basis. These procedures are empirical, and therefore, their validity is judged based
on the agreement of the calculations and experiment.
2.3.1 General Strategy
In all scaling procedures, regardless of their nature (frequency or FF scaling, single-

or multi-parameter scaling), the main strategy is the same: One uses scaling factors
(SFs) which depend on the computational level (defined as method/basis set) and
which—when determined for a well-defined set of Nmol molecules (called “a
training set” or “a calibration set”)—are assumed to be transferable to other
molecules. The choice of a training set depends on the problem being investigated.
One could consider, e.g., a set of small organic molecules vibrational spectra of
which are known and use the optimized factors to perform the relevant scaling for
other organic molecules, spectra of which are to be interpreted. Another problem is
associated with the choice of experimental frequencies. The proper assignment of
bands on IR and/or Raman spectra to vibrational (harmonic) modes of all molecules
of a training set is needed prior to optimization of SFs. In addition, bands of
imprecise experimental position as well as bands exhibiting excessively larger
deviations from theoretical frequencies as compared with the remaining ones should
be omitted.
Given the properly chosen set of experimental vibrational frequencies for a given
p , p ¼ 1; 2; . . .; Nvib , the so-called optimization (re-
training set of molecules mexpt
finement) of SFs f ¼ ðf1 ; f2 ; . . .; fNscl Þ has to be performed. Note that in case of
single-parameter scaling procedures Nscl ¼ 1. In general, the scaled frequencies are
p ¼ mp ðf Þ. In order to find the optimal SFs f , one has to
opt
functions of SFs, i.e., mscl scl
minimize the least-squares merit function (LSMF)
X
Nvib 2
F ðf Þ ¼ wp mscl
p ð f Þ m expt
p ð2:29Þ
p¼1
with respect to f, i.e., one has to solve a set of equations
@F
¼0 for q ¼ 1; 2; . . .; Nscl : ð2:30Þ
@fq
Note that the same symbol F is used for LSMF and the force constant matrix
elements in IC representation. Since the latter one has two indices, no ambiguity is
introduced in the notation. In the above expression, wp is a weight a particular
frequency p contributes to the overall LSMF. It can be used to change the

importance of specific frequencies in the fit. It is recommended that for
well-established frequencies, wp is inversely proportional to the frequency itself
[12], although in many applications presetting all weights to unity is a standard
procedure. In the case of multi-parameter scaling procedures, some of the SFs may
be given preset values. This is easily accomplished using, e.g., Lagrange method of
undetermined multipliers in conjunction with Eq. (2.29). Having found f opt , one
can calculate the root-mean-square (RMS) deviation as
vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
u Nvib 2
u 1 X
RMS ¼ t p ðf Þ mp
mscl opt expt
½cm1 ð2:31Þ
Nvib p¼1
where we used wp = 1, p = 1,2,…,Nvib, for simplicity, average relative percentage

error (ARPE) as

Nvib mscl ðf opt Þ mexpt
100 X p p
ARPE ¼ ½%; ð2:32Þ
Nvib p¼1 mexpt
p
and/or some other quantities determining the accuracy of a fit. In the above for-
mol
mulas, Nvib can be replaced by Nvib to obtain the corresponding values for a
particular molecule.
The main term used in all scaling procedures is transferability of Sfs. Sfs
meaning seems to be intuitively obvious. However, two aspects of transferability of
SFs: (i) transferability “within the molecule” and (ii) transferability “among (re-
lated) molecules” should be considered [13]. The first of these items can be for-
mulated as follows: The SFs are transferable within the molecule if, after
calculating them from rather localized modes (group vibrations), they are capable of
reproducing frequencies of delocalized modes, i.e., normal modes, involving a
number of local vibrations. This kind of transferability is particularly applicable in
the case of multi-parameter scaling procedures. The transferability among related
molecules naturally means that the factors computed for a training set of molecules
can be successfully applied to obtain accurate spectra of other molecules.
2.3.2 Uniform Scaling
2.3.2.1 Fundamentals of the Method
The simplest scaling procedure is single-parameter frequency scaling, frequently

called uniform scaling (US) . Its basic idea reads as follows: Scale harmonic fre-
quencies of a molecule obtained from solution of Eqs. (2.21) or (2.26) by a single
SF f opt obtained from minimization of the LSMF, Eq. (2.29), for a given training
set of molecules. In this case, the LSMF takes the form (assuming wp ¼ 1,
p ¼ 1; 2; . . .; Nvib )
Nvib
X 2
Fð f Þ ¼ f mhp mexpt
p ; ð2:33Þ
p¼1
which immediately gives the equation for the optimal SF

PNvib h expt
p¼1 mp mp
f opt
¼ P 2 ; ð2:34Þ
Nvib h
p¼1 mp
the same for all frequencies. When calculating some thermodynamic properties
from the vibrational partition function, in particular enthalpic and entropic effects, it
is necessary to obtain accurate frequencies on the red wing of a vibrational fre-
quency range. Equation (2.33) is the most appropriate for high frequencies. For low
frequencies, the following formula for LSMF:
!2
X
Nvib
1 1
Fð f Þ ¼ h
expt ; ð2:35Þ
p¼1
f mp mp
is recommended. Note that different authors use different criteria for the qualifi-
cation of a frequency as a “low frequency,” but the values 1000 and 1800 cm−1
appear the most frequently in the literature. Straightforward differentiation gives
PNvib h 2
p¼1 mp
f opt ¼P 1 : ð2:36Þ
Nvib
m h mexpt
p¼1 p p
US is easily adjusted for properties other than vibrational frequencies of a

molecule. They include zero-point vibrational energy (ZPVE) correction, as well as
vibrational component of the thermal contribution to enthalpy and entropy,
DHvib ðT Þ and Svib ðT Þ, respectively. The latter two quantities are very sensitive to
low-frequency vibrations. In the case of ZPVE correction, a theoretical value for a
given molecule based on harmonic frequencies is calculated as
N mol
1X vib
ZPVEtheor ¼ mh ½cm1 : ð2:37Þ

2 p¼1 p
This quantity is also available from experiment, in particular for small mole-
cules, provided all experimental harmonic frequencies mexpt;h and the corresponding
anharmonicity constants x are available. In this case, the relevant equation reads
mol
1X
Nvib
1
ZPVE expt
¼ mexpt;h 1 x p ½cm1 : ð2:38Þ
2 p¼1 p 2
The SF for ZPVE correction is then obtained by minimizing the LSMF of the
form
Nmol
X 2
Fð f Þ ¼ f ZPVEtheor
p ZPVEexpt
p : ð2:39Þ
p¼1
In the case of enthalpies and entropies, one uses (see, e.g., [14])
mol
X
Nvib
hvib
DHvib ðT Þ ¼ NA k p ½J=mol ð2:40Þ
p¼1 exp hp =T 1
vib
and
0 1
X
mol
Nvib
hvib =T
Svib ðT Þ ¼ R @ p
ln 1 exp hp =T A ½J=mol K;
vib
ð2:41Þ
p¼1 p =T 1
exp hvib
hcm
where k is the Boltzmann constant, NA is Avogadro’s number, and hvib p ¼ k is the
p
vibrational temperature associated with mode p of a given molecule, with mp (ex-

pressed in wavenumbers) being either mhp in the case of theoretical, or mexpt;h
p 1 12 xp
in the case of experimental quantities, respectively. In principle, to obtain the cor-
rected enthalpies and/or entropies from theoretical harmonic frequencies one could
use f mhp to calculate hvib
p , with f preferably optimized for the low-frequency range.
However, better agreement with experiment is expected when using formulas
Nvib 2
X expt
Fð f Þ ¼ DHvib
theor
ðT Þp DHvib ðT Þp ð2:42Þ
p¼1
and
Nvib 2
X expt
Fð f Þ ¼ vib ðT Þp Svib ðT Þp
Stheor ð2:43Þ
p¼1
hcmhp
p ¼f
with hvib k .
US is simple and straightforward, which makes it very attractive for routine

applications. The only thing one has to do after solving the vibrational problem in
harmonic approximation is to find the SF relevant for a given computational level in
the literature and scale frequencies. However, there are twofold disadvantages of
US as compared with the more sophisticated, multi-parameter treatments. First, the
quality of the scaled frequencies is not spectacular. The RMS values are typically
large. As will be discussed in the case of frequency scaling, they frequently exceed
30 cm−1 even with high-quality FFs (see, e.g., [15]). This is partially due to the fact
that the training set used in the optimization procedure in the cited work was very
diversified. It contains various types of molecules, including both organic and
inorganic ones, closed- and open-shell molecules, etc. Probably lower RMS would
be obtained if the training set consisted of molecules with rather similar structural
motifs (e.g., organic molecules only). Second, probably more important disad-
vantage, is associated with the fact that US does not provide the frequency swaps,
which are frequently necessary, in particular in the overcrowded spectral ranges.
Consider, for example, two close-lying vibrations, one being a simple stretch of
some bond, the other one—some out-of-plane bending vibration, with the harmonic
frequency of the former only somewhat greater than that of the latter one, and with
similar theoretical intensities. Assume, that they correspond to two well-resolved
close-lying bands on IR spectrum, also of similar intensities. Multiplying both
frequencies by the same SF would preserve their order, which may lead to
misassignment. The reason is that it may be expected, that the frequency lowering
for stretch due to anharmonicity is much greater than that for the out-of-plane
bending (the out-of-plane motion, like H motion out of the Ar plane in aromatic
compounds, has symmetric potential, like parabola, and for this reason it is more
harmonic than stretch). Having two SFs, lower one for the stretch, and larger one
for out-of-plane bending, would solve the problem. This is, in fact, what
multi-parameter scaling is handling quite well.
2.3.2.2 Development of US
Considerations presented in Sect. 2.3.2.1 exhaust the theory of US. In the fol-
lowing, we briefly report the literature on historical development of US to provide
the reader with the relevant references. For more details, the reader is referred to the
original papers. Probably, the first report on single frequency SF was given by
Pople and co-workers back in 1981 [16]. A value of 0.89 was derived for a training
set of 38 molecules (nearly 500 frequencies) and recommended for subsequent use
with HF/3-21G frequencies. This work initialized relentless progress in US pro-
cedure due to continuous QC methodology and basis set development as well as
advances in software and hardware manufacturing, which enable increasingly
accurate QC calculations on increasingly large molecules. Factors of 0.8929 and
0.9210 were derived one year later at more sophisticated HF/6-31G* and MP2-fu/
Table 2.1 Recommended SFs reported in [19] for HF, DFT/B3LYP and DFT/B3PW91 with
6-31G* basis set for various quantities (frequencies, ZPVE, DHvib(T), and Svib(T))
Method Frequencies Low ZPVE DHvib(T)d Sib(T)d
frequencies
HF 0.8953 0.9061 (15b) 0.9135 0.8905 0.8978
(50a) (0.71c) (0.044e) (0.28f)
DFT/ 0.9614 1.0013 (15b) 0.9806 0.9989 1.0015
a
B3LYP (34 ) (0.42c) (0.036e) (0.26f)
DFT/ 0.9573 0.9930 (14b) 0.9772 0.9885 0.9920
B3PW91 (34a) (0.37c) (0.036e) (0.24f)
The RMS values for the considered quantities are reported in parentheses
a
In [cm−1]
b
In units of 10−5 cm
c
In [kJ/mol]
d
At 25 °C
e
In [kJ/mol]
f
In [J/mol K]
6-31G* computational levels, respectively [17]. The calculations were carried out
for a set of 36 one- and two-heavy-atom molecules comprising second-row ele-
ments (first-row using terminology of Ref. [17]). The calculated frequencies were
recommended to obtain molecular entropies, from which entropies of reactions that
are in good agreement with experimental ones can be calculated. Similar calcula-
tions, this time using much more extended training set of molecules (122 molecules,
1066 frequencies) were reported in 1993 [18]. The recommended scaling factors are
0.8929 and 0.9427 for HF/6-31G* and MP2/6-31G* fundamentals, respectively,
and 0.9135 and 0.9646 for HF/6-31G* and MP2/6-31G* ZPVE corrections,
respectively. More comprehensive study came in 1996 [19]. The authors deter-
mined SFs for vibrational frequencies, ZPVE corrections as well as thermal con-
tributions to enthalpy and entropy for subsequent usage with 19 computational
levels. They include semiempirical AM1 and PM3 levels, ab initio HF, MP2, and
QCISD as well as a few DFT (BLYP, B3LYP, BP86, B3P86, and B3PW91)
computational levels with a few Pople basis sets. The calculations were based on
various training sets, including that of 122 molecules for frequencies, as before. It
was shown that the best DFT results were obtained with B3LYP and B3PW91
functionals. In addition, HF procedure turned out to provide better results as
compared with the post-HF MP2 method.2 The recommended SFs for the
above-mentioned three computational levels are reported in Table 2.1.
A number of other contributions to the development of US between 1996 and
2007 appeared. The recommended literature includes works by Wong [20], Truhlar
[21–26], Curtiss [27], Schlegel [28], Wilson [29], Uvdal [30], Csonka [31], and
Tantirungrotechai [32]. A brief summary is reported in Table 2.2. These works
2
This behavior was also observed in the case of multi-parameter scaling procedures. Although
MP2 is known to predict very accurate molecular geometries, it does not provide accurate
vibrational spectra.
Table 2.2 Significant contributions to the development of US procedure in the period of 1996–
2007
Author/year Factors for Computational levels Basis sets
Wong/1996 Frequencies, ZPVE MP2-fu, SVWN, 6-31G*
BLYP, B3LYP,
BVWN, B3P86
Truhlar/1999–2005 Frequencies, ZPVE inter alia MP2, MP4, The selected Pople and
CCD, CCSD, CCSD(T), Dunning basis sets
QCISD and a series of
density functionals as
specified in Table 2 of
[26]
Curtiss/2001 ZPVE B3LYP 6-31G(2df,p)
Schlegel/2001 Frequencies HF, SVWN, BLYP, Sadlej pVTZ
(>1800 cm−1), low B3LYP, B3PW91, MP2
frequencies
(<1800 cm−1)
Wilson/2004 Frequencies HF, B3LYP, MP2 cc-pVDZ, cc-pVTZ,
(>1000 m−1), low cc-pVQZ and the
frequencies augmented modifications
(<1000 cm−1),
ZPVE, DHvib(T),
Svib(T)
Uvdal/2005 frequencies B3LYP 6-311+G**
(>1000 cm−1), low
frequencies
(<1000 cm−1),
ZPVE
Csonka/2005 ZPVE B3LYP, B3PW91, PBE, 6-31G*, 6-31+G*, 6-31
PBE0, TPSS, TPSSh +G**, 6-31G(2df,p)
Tantirungrotechai/ Frequencies, ZPVE B972, B98, G96LYP, 3-21G, 6-31G*, 6-31+G*,
2006 HCTH, OLYP, O3LYP, 6-31G**, 6-311G**,
VSXC, PBE0 6-311G(df,p), 6-311+G(df,
p), cc-pVDZ, aug-cc-pVDZ
provide valuable results that can be applied to a variety of (thermo)chemical prob-

lems. Frequent citations, up to several hundred times in a few cases (as at July 2018),
are a clear evidence of the usefulness of the US approach. However, works by
Radom and co-workers [15, 19] appear to be the fundamental ones in the field of
routine applications of US. The first one (1996, [19]) has been cited nearly 6000
times, the second, more recent (2007, [15])—more than 1400 times (as at July 2018).
This is the reason for selecting the period of 1996–2007 in the present paragraph.
The above-mentioned work [15] includes extended ab initio and DFT (with a
variety of density functionals) calculations with commonly used Pople (and
Dunning) basis sets. Apart from the recommendations as to the SFs used for fre-
quencies, low frequencies, ZPVE corrections, DHvib(T), and Svib(T), a careful study
of the basis set, geometry convergence criteria, percentage of the exact exchange in
the modified B3LYP functional, etc., effects is also presented. The selected
Table 2.3 Recommended SFs reported in [15] for the selected ab initio and DFT methods with
6-311+G** basis set for various quantities (frequencies, ZPVE, DHvib(T), and Svib(T))
Method Frequencies Low frequencies ZPVE DHvib(T)a Sib(T)a
HF 0.9059 0.9146 0.9255 0.8967 0.9041
MP2 0.9523 1.0157 0.9768 1.0071 1.0158
QCISD 0.9560 1.0086 0.9812 0.9970 1.0049
QCISD(T) 0.9647 1.0429 0.9907 1.0274 1.0382
CCSD 0.9542 1.0034 0.9795 0.9918 0.9998
CCSD(T) 0.9639 1.0399 0.9897 1.0244 1.0351
BLYP 1.0001 1.0915 1.0189 1.0766 1.0870
B3LYP 0.9688 1.0189 0.9887 1.0102 1.0161
B971 0.9684 1.0162 0.9893 1.0072 1.0134
B972 0.9587 1.0027 0.9799 0.9922 0.9983
B98 0.9676 1.0132 0.9884 1.0046 1.0106
OLYP 0.9842 1.0609 1.0056 1.0455 1.0546
O3LYP 0.9690 1.0230 0.9904 1.0123 1.0186
PBE 0.9944 1.0714 1.0154 1.0534 1.0637
a
At 25 °C
recommended SFs are reported in Table 2.3. The post-2007 literature comprises
further contributions to the development of the US methodology. They include,
inter alia, recent works of Radom and Chan [33, 34], Wilson and co-workers [35],
and a number of others [36–43].
Specifying SFs using four significant figures has been a common practice in all
but the very first papers. Therefore, works on uncertainties of SFs, and—conse-
quently—uncertainties of the scaled frequencies, have to be mentioned [44–47].
There are a few reasons why SFs exhibit uncertainties. First, the uncertainty of a
computed frequency arises from systematic error with respect to the observed
fundamental [44]. Second, there is an uncertainty associated with the determination
of the experimental frequencies the SF optimization procedure is based on. It is
closely related to the resolution of the IR or Raman experiment, which in typical
applications is 4 cm−1, although it ranges between 1 and 15 cm−1 [44]. Third, there
is an uncertainty in the predicted theoretical harmonic frequencies. The calculated
frequencies depend to some extent on the threshold values (geometry convergence
thresholds, SCF convergence threshold, integral prescreening thresholds, etc.) used
in a given QC package. For a given package, the differences are typically not
significant provided threshold values are not looser than default. These errors are
random and are frequently regarded as negligible. However, some vibrational
modes (not necessarily the low-frequency ones) may exhibit frequencies that differ
by a few cm−1 depending on the package being used, even if tight convergence
criteria were adopted. The source of these errors is not clear. By performing careful
statistical considerations, Irikura et al. demonstrated [44] that vibrational frequen-
cies SFs have only two significant figures!!! The calculated uncertainties represent
the standard deviations of SFs state-of-knowledge probability distributions. Similar
calculations were carried out for SFs of ZPVE corrections [45, 46]. In addition, SFs
for a few dozen of common computational levels along with the calculated
uncertainties were reported in both works [44, 45]. High uncertainties of scaling
factors were also confirmed by Cordeiro et al. [47] who carried out calculations for
X3LYP functional. The considerations reported above opened a debate on the
uncertainties of SFs [48–50]. High uncertainties of SFs put US procedure in a little
bad light. However, as stated in the pioneering work [44] “…until the underlying
probability distributions have been characterized, it is not possible to obtain
probabilistic uncertainty intervals. Improved methods for classifying vibrational
frequencies will probably lead to distributions that are more nearly normal and to
smaller uncertainties.” It seems at the moment that the SFs uncertainty problem has
not been fully solved yet.
2.3.3 Wavenumber Linear Scaling
Another single-parameter scaling procedure was devised by Yoshida and

co-workers in 2000 [51]. It is called wavenumber linear scaling (WLS). To our best
knowledge, only two methodological, frequently cited works from early 2000s
appeared in the literature [51, 52] so far.
The authors considered 205 organic and inorganic molecules, for which they
considered as many as 1729 fundamentals [51]. Then, after calculating harmonic
frequencies at the B3LYP/6-311+G** computational level they noticed that the ratio
expt
f ¼ mmh , called wavenumber scale factor, is typically within 1.00 ± 0.05 range
(although larger deviations were also observed) and decreases linearly with increasing
frequency. In addition, they noticed that at high frequencies the spread in the factor is
fairly low and increases along with the decreasing wavenumber. After rejecting 28
modes from the set, for which deviations of f were higher than 10%, i.e., using 1701
fundamentals, they obtained the relation for the frequency-dependent optimal SF of
the form

f opt mh ¼ 1:0087ð9Þ 0:0000163ð6Þmh ð2:44Þ
where mh is expressed in [cm−1] and the value in parentheses are uncertainties at the
last significant figure. In fact, similar relation was provided at the same time [53],
but it seems to be not as general as Yoshida’s one, since only a few molecules and
merely 139 fundamentals were considered.
The idea of the WLS was extended by Yoshida et al. two years later [52]. The
authors performed additional calculations on 164 organic and inorganic compounds
using the same as before computational level and scaled them according to
Eq. (2.44). The average deviation from the experimental values was 3.4% for 1223
vibrational modes. In addition, they examined a set of 224 diatomic molecules and
ions, for which experimental harmonic frequencies were available. After calculating
mexpt;h
mh
, they observed that in more than 90% cases the ratio was 1.0 ± 0.1 and
expt
concluded that the decrease of mmh ratio for polyatomic molecules with increasing mh
value is attributed almost exclusively to the neglect of anharmonicity. Finally, they
proposed a simple method of calibrating the WLS method to computational levels
different from B3LYP/6-311+G**. They recommended use of liquid indane and its
20 fundamentals [52] instead of a large training set of small molecules which
apparently facilitates the calculations. The relation they obtained for the considered
computational level reads

f opt mh ¼ 1:0 0:00001356mh : ð2:45Þ
2.3.4 Scaled Quantum Mechanical Force Field Method
Scaled quantum mechanical (SQM) force field (FF) method, called selective scal-
ing, is a multi-parameter scaling method, in which SFs are applied directly to FCs
expressed in IC basis prior to solving the vibrational problem (2.26). As in
Sect. 2.3.2, the outline of the up-to-date theory will be presented first, saving for
later a brief literature review on the historical background of methodology
development.
The first mention of the SQM scaling procedure was given in 1981, when Pulay
and co-workers published their paper [11] on FFs, dipole moment derivatives, and
vibronic constants of benzene. The general idea of the theory itself survived until
now, the only significant modification introduced by Baker et al. [54], consisting in
adopting redundant set of PICs, was proposed. As was already mentioned, the
solution of Eqs. (2.21) or (2.26) usually gives too high frequencies for the reasons
discussed in the beginning of Sect. 2.3. However, the frequency lowering that
brings the calculated frequencies as close to experiment as possible should not be
the same for all vibrations. This means that different SFs should be used for dif-
ferent FCs. The basic idea behind the method is classification of ICs (initially
non-redundant NICs, later redundant PICs) into the limited set of chemically similar
types (groups, I ¼ 1; 2; . . .; Ntyp ), and those coordinates within each type share the
same SF. Such a division is usually based on chemical intuition unless some test
calculations, say, with extended set of SFs, show that some ICs can be put into one
group (the values of SFs are nearly identical). For example, all CH bonds in a given
molecule or in all molecules of a training set can be put into the first group, all XX
bonds, where X denotes a general second-row atom, into the second group, etc. We
will come back to this idea later on. In principle, one could calculate the force
constant matrix F in the non-redundant ICs representation numerically by setting
input in z-matrix format, vary the coordinates accordingly, and apply the central
differences method on energies (or gradient), but the more common procedure used
now is to calculate the Cartesian FF f x and transform it to F by calculating the B−1
Fig. 2.5 Structures of the B−1 and F matrices obtained by grouping internal coordinates into types
matrix, Eq. (2.25), for a given set of ICs followed by using Eq. (2.24). By grouping
the non-redundant ICs according to

N N N
s ¼ s11 ; s12 ; . . .; s1k1 ; s21 ; s22 ; . . .; s2k2 ; . . .; s1 typ ; s2 typ ; . . .; skNtyp ; ð2:46Þ
typ
where the superscript indicates the type number and the relation
X
Ntyp
kI ¼ K; ð2:47Þ
I¼1
is satisfied, the structures of the B−1 and F matrices are as those presented in
Fig. 2.5.
Then, for a given set of optimal SFs f opt ¼ f1opt ; f2opt ; . . .; fNopt
typ
, called SQM SFs
or simply FF SFs, the diagonal FCs are multiplied by the corresponding factor, and
off-diagonal ones—by the geometric mean of the corresponding SFs, i.e.,
qffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Flscal
I mJ
¼ fIopt fJopt Flm ð2:48Þ
where additional indices in lI and mJ are used to indicate that coordinates sl and sm
belong to types I and J, respectively. Obviously, this means that off-diagonal FCs in
a diagonal block bII corresponding to given type I (see Fig. 2.5) are multiplied by
the same factor as diagonal ones.
Optimal FF SFs are obtained from minimization of the LSMF, Eq. (2.29), for a
given training set of molecules. In this case, the LSMF is non-quadratic in contrast
to US procedure and has to be minimized
iteratively
using,
e.g., Newton–Raphson
@F @2 F
procedure. For that purpose, first @fI and second @fI @fJ derivatives (gradient and
hessian, respectively) of the LSMF F ðf Þ have to be evaluated in every iteration. It
can be shown [54] that gradient and hessian of the LSMF take the form
Nvib
X @mscl ðf Þ
@F ðf Þ p
¼ 2 mexpt m scl
ðf Þ ð2:49Þ
@fI p¼1
p p
@f I
and
!
X @mscl @ 2 mscl ðf Þ
p ðf Þ @mp ðf Þ
Nvib scl
@ 2 F ðf Þ p
¼2 mexpt mscl
p ðf Þ ð2:50Þ
@fI @fJ p¼1
@f I @f J
p
@f I @fJ
respectively, where we skipped the weights w, for simplicity. The procedure to

calculate first frequency derivatives with respect to FF SFs is straightforward [55].
More problematic are the second derivatives that enter Eq. (2.50) but, as discussed
in [54], their contribution to @@fFI @fðfJÞ vanishes when scaled frequencies approach the
2
experimental ones. Consequently, the simplified hessian based on the first deriva-
tives only can be computed in ith iteration, inverted and multiplied by the negative
gradient to provide step Df ðiÞ used to obtain f ði þ 1Þ ¼ f ðiÞ þ Df ðiÞ . Obviously, some
of the SFs may be given preset values in the optimization procedure. As usually, a
gradient norm (or maximum component) can be a basic criterion for convergence.
2.3.4.2 Development of SQM
The idea of dividing the ICs into groups came into being a few years before the first
paper on SQM [11] appeared and was due to Blom and Altona in a series of papers
from the mid-1970s [56–61]. The authors recognized that adopting different SF to
different FCs provides much better agreement with experimental frequencies than in
the case of using a single SF. They calculated ratios between diagonal FCs derived
from both experimental fundamental and experimental harmonic frequencies and
the calculated ones (HF/4-31G) for methane, ethane, propane, ethene, cyclo-
propane, and cyclopropene [56] and observed that they remain rather constant
regardless of the molecule for a given kind of IC. On the other hand, the ratios
(loosely speaking, SFs) differ for various coordinates. At that time, the idea of
training set was not introduced probably due to the computational cost of QC
calculations, but the term “transferability” occurred already in the first paper of the
series [56]. Next, the authors performed a series of calculations using limited set of
molecules and optimized (refined) SFs each time to match the experimental
fundamentals (and/or experimental harmonic frequencies, if available). In the case

of ethane, propane, and cyclopropane [57], they obtained scaled FFs that repro-
duced the experimental fundamentals of the molecules and their deuterated analogs
with remarkable accuracy (with RMS ca. 10 cm−1) using merely 5 (or 6 in the case
of cyclopropane) SFs. This was the case of refinement carried out for each molecule
separately (SFs remained roughly the same, regardless of molecule) and for all
molecules simultaneously, indicating good transferability of SFs. Then the calcu-
lations of the scaled FF for water (to match both harmonic and fundamental
experimental frequencies) and methanol (to match experimental fundamentals)
were carried out [58]. The average error for harmonic frequencies of water was very
low (1.6 cm−1) while for fundamentals much larger (14.8 cm−1) indicating that the
significant part of the difference may be ascribed to anharmonicity. Similar cal-
culations were performed on propene [59] using both precomputed [58] and opti-
mized (for the set of experimental fundamentals of propene and its deuterated
analogs) SFs. The RMS turned out to be 9.3 and 8.3 cm−1, respectively, for the set
84 fundamentals using 7 SFs. The FF SFs were then obtained for ethylene and its
deuterated analogs [60] were the harmonic and fundamental frequencies were
reproduced with remarkable accuracy. In the case of 50 harmonic frequencies, the
RMS value turned out to be 3.9 cm−1. This value reflects the neglect of the cor-
relation effects in QC treatment, which apparently cannot be accounted for by a few
SFs. In the case of 81 fundamental frequencies, the RMS was somewhat larger and
amounted to 7.1 cm−1. Further, calculations carried out on dimethylether [61]
proved, that the method can be generally applied to reproduce vibrational spectra
for more complex systems.
The research described above constitutes an embryo of the modern SQM
method. The problematic thing in Blom and Altona approach was the treatment of
off-diagonal (interaction or coupling) FCs. Apparently, they did not know how to
solve this problem, so they decided to attribute one of the SFs to all off-diagonal
FCs. In the pioneering work on SQM [11] Pulay et al. said: “It is difficult to see why
physically different off-diagonal elements should share the same scale factor…” and
decided to adopt the geometric mean of SFs as in Eq. (2.48). With this modifica-
tion, SQM survived until now.
Further development of the SQM procedure3 consisted in, inter alia, extending
its applicability to a wider range of systems. In 1983, Pulay et al. [12] considered a
series of molecules (glyoxal, acrolein, butadiene, formaldehyde, and ethylene) with
similar structural motifs treated at HF/4-21G computational level. They used six
SFs for 115 in-plane modes and three factors for 41 out-of-plane modes and
managed to reproduce the fundamentals with RMS of 12.3 and 6.7 cm−1, respec-
tively. It was also concluded that, in contrast to SFs, the FCs themselves are not
well transferable between molecules. This means that common FF SFs can be used
for a large group of molecules. Twelve years later [62, 63] calculations of the
3
As mentioned in the introductory section in the following we will not cite papers, where only
applications are presented.
recommended SFs for subsequent use with FFs obtained with BLYP and B3LYP
density functional with 6-31G* basis set were carried out. The authors used 20
small organic molecules containing C, H, N, and O atoms and selected a number of
securely assigned bands. First, they noticed that good overall performance of the
BLYP functional with regard to predicting fundamentals without scaling is due to
the fortuitous cancelation of errors due to overestimation of bond length which
results in a decrease of FCs. Second, they calculated single FF SF, which for BLYP
functional was equal to nearly unity (for the reasons described above) and for
pffiffiffiffiffiffiffiffiffiffiffi
B3LYP—0.928 (which corresponds to the value of 0:928 ¼ 0:963 for frequen-
cies). Third, they proved good transferability of SFs by using them in prediction of
vibrational spectra of 11 additional molecules of a test set. Fourth, they optimized
11 SFs for the selected types of NICs with the RMS value of 19.1 and 12.8 cm−1
for BLYP and B3LYP, respectively. These factors are included in Table 2.4. From
now on the following notation will be used: X denotes a general second-row atom
and Y—a general third-row atom. The two-letter symbols, e.g., XX, refer to bond
lengths, three-letter symbols, e.g., XXX—to valence angles, and four-letter sym-
bols, e.g., XXXX—to torsion angles. The above-mentioned work was extended,
and a new SF for CCl stretch equal to 1.017 (B3LYP/6-31G*) was proposed [64].
First applications of the SQM procedure described so far were based on the FFs
expressed in non-redundant NICs [1], called local symmetry coordinates, which are
just the appropriate linear combinations of PICs (simple stretches, bends, and tor-
sions). Numerical procedures to obtain FFs were adopted [11]. Now Cartesian
quadratic FCs f x can be obtained analytically at a number of computational levels,
the B matrix for NICs constructed from B matrix for PICs by taking the appropriate
linear combinations and inverted (as described in Sect. 2.2.3), and F matrix cal-
culated according to Eq. (2.24). The approach based on non-redundant NICs has a
few disadvantages, though. First, the constraints imposed on the weights of PICs in
Table 2.4 Recommended No. Type of NIC BLYP B3LYP

[62] SFs for FFs expressed in
non-redundant NICs basis 1 XX 1.007 0.922
computed using BLYP and 2 XH 0.977 0.920
B3LYP density functionals 3 XXX 1.052 0.990
with the 6-31G* basis set 4 XCH 1.005 0.950
5 XOH, XNH 0.980 0.876
6 HCHa 0.964 0.915
7 Linear deformations 0.986 0.913
8 NH2 wagging 0.834 0.806
9 Out-of-plane 1.072 0.976
10 Conjugated torsions 0.990 0.935b
11 Single-bonded torsions 0.869 0.831b
a
Probably includes HNH bending in 1995 work (there is a lack of
consistency in the symbols used between 1995 [62] and 1998 [54]
papers)
b
Corrected in Ref. [63]
the given NIC may not fully account for the contributions of PICs in a given mode
in some cases. This problem is of particular importance in the case of the
multi-parameter frequency scaling procedure discussed in the next section. Thus, it
appears that set of PICs is more flexible. Second, manual generation of NICs is
tedious. Although there are programs, e.g., [65] that automatically generate NICs,
they are not free of shortcomings, and—as stated in [54]—“… algorithms … may
fail occasionally for unusual topologies, e.g., cage compounds.” Third, NICs are
frequently delocalized, which is also a disadvantage. Fourth, some of the NICs are
linear combinations of PICs which should not share the same SF. For example, NIC
for the symmetric bending of the methyl group is a linear combination of XCH and
HCH primitive bends, which are now usually put into different types.
On the other hand, the choice of PICs that form a non-redundant set is not
unique. The only thing one has to take care of is their linear independence. As
frequencies rather than FCs are observables, they are the same for all possible
choices of ICs. It was thus decided [54] to modify the original SQM procedure and
transform the Cartesian FF to the set of redundant PICs. They are easily generated
based on atomic connectivities, and their number L is greater than K, and even
greater than 3N in an overwhelming majority of cases. Consider, e.g., ethylene
molecule (3N = 18, K = 12). The redundant set of PICs consists of 5 stretches, 4
CCH and 2 HCH bendings, and 4 HCCH torsions (15 primitives overall). Thus, the
B matrix augmented with translations and rotations is a 15 + 6 = 21 18 matrix,
inversion of which can be accomplished using the generalized Moore–Penrose
inverse. Introduction of the generalized inverse whenever necessary is the main
difference between the “classical” and “modified” SQM procedure. The scaling
idea, i.e., division of ICs into chemically similar groups, is retained.
This time [54] the authors used a new training set of 30 small organic molecules,
the fundamentals of which were taken from the post-1993 literature. The set
includes a wide variety of organic molecules, including saturated and unsaturated
compounds, both having chain, cyclic and heterocyclic structures with hydroxyl,
carbonyl, amino, etc. substituents. They carefully selected as many as 663 vibra-
tional fundamentals, eliminated misassignments and fundamentals of uncertain
positions. Again the B3LYP/6-31G* computational level was used. A new clas-
sification of ICs into chemically similar groups was also proposed (see Table 2.5).
In the following, it will be called “11-parameter scaling frame.” Optimization of
SFs for the set of 663 fundamentals gave the RMS value of merely 12.04 cm−1.
Comparison of both classifications along with the recommended values is given in
Table 2.5. As can be seen some factors, in particular for the CCl stretch, turned out
to have a value larger than unity. This is also the case of multi-parameter frequency
scaling procedure described in the next section. This is apparently a “geometric
effect”—CCl bonds predicted by B3LYP are too long and, consequently, too weak.
For this reason, the harmonic frequency drops below the fundamental one. SFs for
CCl calculated, say, at the MP2 computational level have values lower than unity
(see, e.g., [66]). As before the factors calculated for the training set were applied to
the test set, which also consisted of 30 molecules (843 fundamentals found in the
post-1993 literature). The calculations proved again excellent transferability of FF
Table 2.5 Comparison of 1995 [63] and 1998 [54] classifications of ICs into types for molecules
composed of C, H, N, O, and Cl, and the corresponding scaling factors
Types B3LYP/6-31G* values
PICs (1998) NICs (1995) PICs (1995) NICs (1995)
XX 0.9207 0.922
a
XY (here: CCl) CCl 1.0438 1.017a
CH XH 0.9164 0.920
NH 0.9242
OH 0.9527
XXX 1.0144 0.990
XXH XCH 0.9431 0.950
XOH, XNH 0.876
HCHb 0.9016 0.915
HNH – 0.8753 –
Linear deformations 0.8847 0.913
All torsions NH2 wagging 0.9523 0.806
Out-of-plane 0.976
Conjugated torsions 0.935
Single-bonded torsions 0.831
a
Not considered in [63], taken from Ref. [64]
b
Probably includes HNH bending in 1995 work (there is a lack of consistency in the symbols used
between 1995 and 1998 papers)
SFs—the RMS value turned out to be 11.77 cm−1. The presented in this paragraph
FF scaling procedure is regarded now as a standard one in obtaining accurate
vibrational spectra in spite of the fact that it was published 20 years ago.
Another contribution to the development of the SQM procedure came in 2002
[67]. The authors considered a series of 20 molecules (341 frequencies) containing
second- and third-row atoms (called first- and second-row atoms, respectively, in
the paper). The molecules included the already investigated structural motifs based
on C, H, N, and O second-row atoms. In addition, they contained fluorine as well as
sulfur and phosphorus (in various oxidation states), silicon, and chlorine atoms. In
this way, a number of additional (to the already investigated) structural motifs were
introduced. Surprisingly, the authors decided to use SQM scheme based on the
non-redundant set of NICs rather than more flexible scheme based on redundant set
of PICs. The disadvantages of the adopted approach are discussed in the paper,
though. The B3LYP/6-31G* computational level was adopted. In the case of the
types of NICs for which the SFs are known, they decided to use the precomputed
values [63]. The new SFs were attributed to new structural motifs; the overall set
consisted of 20 SFs. New SFs were optimized, and the overall RMS was found to
be merely 10.7 cm−1 for this very diversified set of molecules. It is worth to
mention that optimization of all SFs only slightly improved the least-squares fit
(RMS = 10.6 cm−1). It turned out that most of the SFs initially taken from [63] did
not change significantly: the typical relative deviation was well below 1% (it was
frequently of the order of 0.1–0.2%) unless the factors’ optimization was based on
merely a few fundamentals. In passing the authors noticed that numerous scaling
factors for types involving the third-row atoms are greater than unity.
Further development of the SQM scaling procedure consisted in continuous
development of the SFs’ database. One of the authors of this chapter
(PB) developed an alternative, multi-parameter frequency scaling method, the
so-called effective scaling frequency factor (ESFF) method. To properly investigate
its predictive capabilities, the development of ESFF was accompanied by the
analogous SQM calculations, which were always considered as benchmark. The
methodological aspects will be provided in the next section, in which the ESFF
procedure is described in detail. Here we merely mention the fundamental things
concerning the FF SFs’ database development. First, the SFs for subsequent use
with B3LYP/6-311G** FFs were reported in 2008 as a consequence of verifying
the transferability of the ESFF SFs [68]. Baker’s training set of 30 molecules [54]
was used. The calculations were based on 660 fundamentals (3 less than originally),
and the original 11-parameter scaling frame [54] based on redundant PICs was
adopted. Second, a new, reduced set of FF SFs was proposed in 2010 [69]. With
this set of factors, the scaling procedure will be referred to as a “9-parameter scaling
frame”. The main difference between the standard, 11-parameter set, and a new,
9-parameter set is the splitting of the type composed of all XX bonds into two types
corresponding to (i) XX single and CC conjugated bonds, the so-called XX(s,c)-
type, and (ii) the remaining double and triple XX bonds, the so-called XX(d,
t) + CN(c)-type, respectively, while joining a few distinct groups into one group
(CH, OH, and NH as well as HCH and HNH were combined into the common XH
and HXH groups, respectively). Note that the additional flexibility obtained by
proliferating of the factors in the range of XH vibrations is unnecessary, as the
interpretation of the spectra in the range above 2500 cm−1 is fairly straightforward.
Moreover, in this range the largest deviations between the scaled and the experi-
mental frequencies are observed anyway, inter alia on account of the frequency
shifts due to the Fermi resonance. As before B3LYP/6-311G**, computational
level and Baker’s training set were used. The proposed, new set of factors seems to
be more flexible in the middle range of a vibrational spectrum, in that it is capable
of reproducing the vibrational spectra with higher accuracy in the region where XX
stretching vibrations occur. The obvious advantage of using the 11-parameter frame
is that the user does not have to assign a formal bond type or distinguish between,
say, a CC double bond or a conjugated bond, which is not always obvious, as all
XX bonds form one group. Thus, it can be still recommended in a large number of
cases. Third, a series of FF SFs within the 9-parameter scaling frame for subsequent
use with B3LYP FFs computed with a variety of Pople’s basis sets were reported in
2012 [70]. In addition to structural motifs spanned by molecules of Baker’s training
set (i.e., motifs derived from C-, H-, O-, N-, and Cl-containing molecules), for
which the FF SFs were also calculated, motifs obtained for silicon- and sulfur(II)-
containing compounds were considered, and FF SFs for new types of redundant
PICs (HY, XY, YY, XXY, and XYX+YYX) were determined (VDZ basis sets
only). Roughly at the same time [66], an extended database of FF SFs for 11- and
9-parameter frameworks, with a variety of density functionals and a variety of
Pople’s and Dunning’s basis sets (overall 370 computational levels) was presented.
All the FF SFs reported in the previous paragraph (except those mentioned in
Ref. [70]) were determined to reproduce spectra obtained in gas-phase experiments.
It was shown, however, that SFs are well transferable to molecules in condensed
phases unless hydrogen bonding strongly affects the bands’ position (see e.g., [71]).
The problem of gas phase versus Ar-matrix SFs was thoroughly considered in
reference [72]. A set of 33 molecules (347 vibrational modes), whose vibrational
spectra in both the gas phase and argon matrix are known was used. The authors
carried out calculations using four popular density functionals (PBE, B3LYP,
B3PW91, and M06-2X), two basis sets (6-31++G** and aug-cc-pVTZ), and
11-parameter scaling frame (with two additional types, i.e., SH and {H,X}XXH)
within the redundant PICs formalism. Surprisingly, the largest differences between
gas phase and Ar-matrix FF SFs were observed for the XXXX type (up to 3%)
rather than types involved in hydrogen bonding, i.e., OH and NH (<1%). Matrix
shifts were also reasonably reproduced with the new set of FF SFs. Finally, the
authors declared to report more FF SFs for recent density functionals (including
those accounting for the dispersion interactions) but we did not find works on that
in the post-2011 literature.
2.3.5 Effective Scaling Frequency Factor Method
Effective Scaling Frequency Factor (ESFF) method is a multi-parameter frequency

rather than FF scaling method, in which SFs are applied directly to frequencies after
solving the vibrational problem (26). Here again the outline of the up-to-date theory
will be presented first, deferring a brief literature review on the historical back-
ground of methodology development.
Let us first consider the original formulation of the method, in which potential
energy U of a molecule is expressed in terms of non-redundant ICs, chosen to be
NICs. The potential energy in harmonic approximation, Eq. (2.28), reads
1X K
U ðsÞ ¼ Flm sl sm : ð2:51Þ
2 l;m¼1
As in Sect. 2.2.3, s denotes here deviations of internal coordinates from their

equilibrium values. Introducing expansion (2.27) of ICs in terms of normal coor-
P
dinates, i.e., sl ¼ Kj¼1 alj Qj , we obtain
1X K X K
U¼ Qj Qk alj Flm amk : ð2:52Þ
2 j;k¼1 l;m¼1
PK
Since l;m¼1 alj Flm amk ¼ m2j djk (see, e.g., [5]), we obtain
1X K X K
U¼ Q2j alj Flm amj : ð2:53Þ
2 j¼1 l;m¼1
The last equation resolves the potential energy associated with each normal
mode (index j) into local modes (indices l and m), i.e., modes associated with the
change of one IC. For a given value of j, the expression plm;j ¼ alj Flm amj
describes the contribution of the coupled local modes l and m to the jth normal
mode. plm;j , called Potential Energy Distribution (PED) coefficients [73], were
introduced by Morino and Kuchitsu in early 1950s. The common practice is to use
normalized to unity diagonal PED coefficients, i.e.,
a2lj Fll
pll;j ¼ PK : ð2:54Þ
2
m¼1 amj Fmm
Coefficients defined in this way provide the (percentage) contribution of the local
mode l to the normal mode j. Note that PEDs are not only dependent on ampli-
tudes a, and they are also proportional to the FCs.
The basic assumptions of the ESFF scaling read as follows:
• as in SQM the idea of classifications of ICs into chemically similar types is
preserved, and each type shares the same SF called ESFF SF, or simply local SF
(LSF). Let us assume that the set of optimal (in the usual sense) LSFs f opt ¼

f1opt ; f2opt ; . . .; fNopt
typ
is known;
• from the PED coefficients, pll;j , evaluated after solving the vibrational problem,
one may determine the contribution of the entire type I of local modes to the jth
normal mode as a sum of pll;j for the internal coordinates (index l) belonging
to a given type I ¼ 1; 2; . . .; Ntyp , i.e.,
X
PI;j ¼ pll;j : ð2:55Þ
l2typeI
Note that a capital “P” letter is used instead. Also note that the term “local
mode” is referred to the mode associated with the change of one IC in spite of the
fact that NICs we are currently referring to are sometimes delocalized;
• the effective SF (ESF) for the jth normal vibration is determined as a sum of the
LSFs weighted by the contributions PI;j , i.e.,
X
Ntyp
fjeff ¼ PI;j fIopt ; ð2:56Þ
I¼1
• the corrected frequency for the jth normal vibration can be found as a product
of the corresponding harmonic frequency mh and the ESF fjeff , i.e.,
j ¼ fj mj :
mscl ð2:57Þ
eff h
As always optimal SFs, i.e., optimal LSFs, are obtained from minimization of
the LSMF, Eq. (2.29), for a given training set of molecules. In this case, the LSMF
is quadratic with respect to LSFs, like in the case of US, and therefore no iterative
procedure to find LSFs needs to be employed. Skipping for simplicity the weights
w and factors that are preset in the optimization procedure as in original work [74],
we obtain
!2
X
Nvib X
Ntyp
F ðf Þ ¼ mhp PI;p fI mexpt
p ð2:58Þ
p¼1 I¼1
or alternatively
X
Nvib X
Ntyp 2 X Nvib X
Ntyp Nvib
X 2
F ðf Þ ¼ PI;p PJ;p fI fJ mhp 2 PI;p fI mhp mexpt
p þ mexpt
p : ð2:59Þ
p¼1 I;J¼1 p¼1 I¼1 p¼1
Straightforward differentiation gives that the stationary solution is obtained by

solving a set of Ntyp linear equations
Af opt ¼ b ð2:60Þ
where
X
Nvib 2 X
Nvib
AIJ ¼ PI;p PJ;p mhp and bI ¼ p :
PI;j mhp mexpt ð2:61Þ
p¼1 p¼1
Further development of the ESFF scaling scheme was similar to that of SQM.
The implementation of all computational procedures should be as user-friendly as
possible. Therefore, a modification of the original procedure [74] consisting in
using redundant set of PICs was proposed [75]. It involved utilization of general-
ized Moore–Penrose inverse of a matrix. As in the case of SQM, it is necessary to
invert the transformation matrix Br between the redundant PICs and the Cartesian
displacements (L 3N matrix) which is then used in the transformation of the
Cartesian FFs fx to obtain Fr and generation of a mass tensor G1 r . It was also
decided to use the canonical FFs (see, e.g., [76]) in order to ensure that the solution
is unique. Since both matrices: Fr and G1 r are now singular, Eq. (26) was projected
onto some set of non-redundant ICs snr but the final ar matrix used to calculate
PEDs must refer to redundant PICs, i.e., to the transformation sr ¼ ar Q (indices “r”
and “nr” refer to redundant and non-redundant, respectively). It can be shown that
the A transformation matrix from sr to snr ðsnr ¼ Asr Þ can be obtained by diago-
nalizing Br BTr , and taking the transpose of O eigenvector matrix corresponding to
T x
nonzero eigenvalues, i.e., A ¼ OT . Then we have Fnr ¼ B nr f Bnr and
ar ¼ Br B nr anr , where Bnr ¼ AB r , a nr is a transformation matrix obtained from
solution of WDC equations (2.26) with Fnr , and the superscript “−” denotes the
generalized Moore–Penrose inverse. Having unique Fr and ar matrices the PEDs
can be calculated according to Eq. (2.54).
Finally, one more extension of ESFF, called ESFF2, was made [75]. Namely,
when calculating contributions of local modes to a given normal mode (note, the
term “contribution” is used instead of “PED”), one can skip the FCs in Eq. (2.54)
and use only amplitudes, i.e.,
a2lj
pll;j ¼ PK : ð2:62Þ
m¼1 a2mj
This approach, which turned out to be equally efficient as PED-based ESFF, is

particularly well suited to calculations on really large systems. This problem will be
briefly discussed in Sect. 2.3.6.3.
2.3.5.2 Development of ESFF
First application of the ESFF procedure, application to toluene, was made in the
first, methodological paper [74]. As in the case of pioneering works of scaling
procedures, the least-squares fit was made to frequencies of toluene itself, i.e., no
training set was used. Six LSFs were optimized, and the obtained results were
highly encouraging. However, as already discussed, the applicability of the scaling
procedures is contingent upon the already mentioned transferability of the SFs
among the molecules. Thus, the determination of the LSFs using a very diversified
training set of molecules (Baker’s training set [54]) was made [68]. Using this set,
apart from obtaining the SFs for the routine applications, comparison of the ESFF
and the SQM benchmark calculations was possible. The determination of both
LSFs and FF SFs at the B3LYP/6-311G** level was based on 660 experimental
fundamentals observed in the gas-phase spectra. The FF SFs for the redundant
PICs-based SQM formalism were determined using the standard implementation of

the procedure; LSFs for non-redundant NICs-based ESFF procedure were found
using the homemade program. 11-parameter scaling frame (vide infra) was used.
The RMS value of 11.66 cm−1 for ESFF was observed and was slightly lower than
in the case of SQM, in spite of using less flexible, NICs basis. The ARPE was also
lower in the case of ESFF. The low value of the RMS deviation obtained for a very
diversified set of molecules allowed to conclude that LSFs are well transferable, at
least as well as FF SFs.
At the same time, transferability of the SFs among the related molecules (see
Sect. 2.3.1) was investigated [13]. A set of three related molecules: toluene, styrene,
and p-methylstyrene was used. As before, the FFs were computed at the B3LYP/
6-311G** level. Two sets of experimental frequencies were considered. The first
set included 66 well-resolved fundamentals in the range of 3000–400 cm−1 (the
ArH and =C–H stretching vibrations were not considered) unambiguously assigned
to the normal modes. The second comprised 39 fundamentals corresponding to the
“pure” modes (i.e., 60% contribution or more from one NIC). Very low RMS
(3.30 cm−1) was obtained for all 66 frequencies. This showed that local SFs are
well transferable among the related molecules. It has to be emphasized that such a
good performance of ESFF follows from classification of NICs into types specific
for the three molecules considered (eight LSFs overall, out of which three were
designed specifically to describe the vinyl group modes). The confirmation of good
transferability of LSFs “within the molecule” follows from the test, in which the
factors were first optimized based on a set of 39 “pure” vibrations, and then applied
to scale all 66 frequencies. Only minor differences in the values of LSFs were
observed, and, in addition, the RMS value increased only slightly, up to 3.35 cm−1.
The ESFF calculations described above revealed that the SF attributed to the
C=C stretching vibration in styrene and p-methylstyrene was significantly lower
than that obtained for the ring stretching. This may constitute an evidence of
different anharmonicity of the stretching vibrations of bonds of different order, or of
their different sensitivity to the neglect of a part of correlation effects in the FFs
determination. The other research carried out roughly at the same time but pub-
lished somewhat later (see [71]) confirmed this observation suggesting the necessity
of some modifications within the standard, 11-parameter scaling frame [54], at least
in some applications. Utilization of the 11-parameter set of the SFs may lead to
errors in the assignment of bands to the normal modes, in particular, in the con-
gested region of a spectrum. Thus, a new, 9-parameter set [69], described in more
details in Sect. 2.3.4.2, was proposed and the LSFs (as well as FF SFs) were found
for Baker’s training set of molecules and at the B3LYP/6-311G** computational
level. The results obtained with the aid of this set are of comparable quality to those,
obtained in the 11-parameter calculations (the RMS values are similar, i.e.,
11.77 cm−1 for SQM and 11.62 cm−1 for ESFF), in spite of reducing the number of
factors by 2. It follows directly from the significant lowering of the RMS value in
the middle range of the spectrum, i.e., 1000–2500 cm−1, from 10.29 to 9.20 cm−1,
when going from the 11- to 9-parameter calculations, which compensates for the
higher deviations on the blue wing of the spectrum. At the same time, the results
confirm the rightness of the assumptions the ESFF method is based on.
Then calculations for large systems, containing several dozen atoms, and for
which the infrared spectra contain very congested regions due to the presence of a
great variety of structural motifs were presented [71], and transferability of ESFF
(and SQM) SFs to solid-state systems was investigated. For that purpose
1,2,4-triazole derivatives were chosen; the largest one includes 67 atoms (195
normal modes). The complexity of the systems leads to the presence of a number of
bands in the range of 1400–1600 cm−1. In the case of both 11- and 9-parameter
sets, the RMS values using the precomputed LSFs [68, 69] turned out to be lower
than 9 cm−1 for the total number of 293 normal vibrations indicating good trans-
ferability of LSFs to molecules in condensed phase. It was also shown that the
newly designed, 9-parameter set of SFs is capable of predicting the correct
sequence of bands more often than the original, 11-parameter set. In addition, a
simple procedure (nearly as simple as US) of performing the ESFF calculations just
“by hand”, provided the contributions of the local vibrations to the normal vibration
are known, was reported. Those can be determined using a great variety of the
well-established programs. Consider, for example, the selected modes identified
in this work, that is
• ca. 1630 cm−1, corresponding roughly to: 69% (Ph stretching) + 17% (PhH
rocking) + 14% (other motions), and
• ca. 1635 cm−1, corresponding roughly to: 62% (C=N stretching) + 18% (XX(s,
c)) + 20% (other motions).
The LSFs for the 9-parameter ESFF calculations are f(Ph stretching) = f (XX(s,
c)) 0.9800, f(PhH rocking) = f(XXH) = 0.9813, f(C=N stretching) = f(XX(d,
t) + CN(c)) = 0.9618 [69]. Assuming, say, f (other motions) = 0.99 (why not to set
it to 0.99 for all these minor contributions, for simplicity), we obtain the approx-
imate ESFs as follows:
1. fjeff (1630 cm−1) = 0.69 0.9800 + 0.17 0.9813 + 0.14 0.99 = 0.9816,
and
2. fjeff (1635 cm−1) = 0.62 0.9618 + 0.18 0.9800 + 0.20 0.99 = 0.9707.
Thus, the approximate ESFF-scaled frequencies, being fjeff mhj , are 1630 cm−1
0.9816 = 1600 cm−1, and 1635 cm−1 0.9707 = 1587 cm−1. These values com-
pare well with the experimental values of 1588 and 1577 cm−1, respectively [71],
and deviate from those obtained with the ESFF program by a small margin. Note in
passing that the scaled frequencies changed order as compared with harmonic ones,
which is not achievable using US. This swap is consistent with the observed and
calculated intensities.
After modification of the ESFF method which enabled calculations using
redundant set of PICs, a few more extensions of the LSFs database (accompanied
by extension of FF SFs database) appeared in the literature. A training set of 8
organosilicon compounds: tetraethoxysilane and its functionalized derivatives,
some of which contain sulfur(II) atoms was chosen. The number of bands was
carefully selected from their IR and Raman spectra and assigned. In contrast to the
previous calculations, spectra were obtained for the compounds in the liquid phase;
as was already pointed out the already known SFs are expected to be highly
transferable from the gas to condensed phase, provided they refer to the structural
motifs exhibiting weak intermolecular interactions. The considered molecules
contain a great variety of new structural motifs. To increase the population of the
sulfur(II)-containing motifs, the so-called auxiliary molecules were added to the set.
Thus, the set of bands was assigned to an overall number of over 600 normal
modes. This time redundant PICs-based formalism of the ESFF procedure was
used. The B3LYP calculations using a series of basis sets: 6-31xxx (VDZ) and
6-311xxx (VTZ), where xxx=++G**, +G**, +G*, G**, G*, and G, were carried
out. In addition, the LSFs and FF SFs within the 9-parameter scaling frame using
Baker’s training set were calculated for all basis sets listed above. These are the
recommended factors; thus, extension of the applicability of the scaling procedures
to the subsequent use with the common basis sets and in conjunction with the
B3LYP density functional (already mentioned in Sect. 2.3.4.2) was made. Since the
B3LYP computational level is one of the most popular in the vibrational frequency
determinations, LSFs (ESFF) and FF SFs (SQM) for subsequent use within the
redundant PICs formalism are presented in Tables 2.6 and 2.7.
In the calculations of the new SFs, the “old” ones were given the precomputed
values as they turn out to be highly transferable to the present set of molecules, as
expected. Initial calculations suggested the necessity of introducing five new SFs,
corresponding to HY, XY, YY, XXY, and XYX+YYX types, respectively. The
optimized factors are reported in Table 2.8. It should be noted that for the types
XX, XY, and YY an increase of the ESFF and FF SFs is observed in the reported
sequence, i.e., along with the increasing number of the third-row atoms, in which
case they adopt values greater than unity. This is also the case of CCl type and, as
already discussed, is due to the geometric effect.
Finally, an extended database of the SFs for ESFF and SQM calculations, which
is available through the files attached to the paper (supporting information) was
created [66] and a detailed analysis of the selected, the most important results was
given. The SFs for 26 typical density functionals, in conjunction with 14 typical
basis sets were determined. In particular, Pople’s (as before, apart from the smallest
ones, i.e., 6-311G and 6-31G), and Dunning’s cc-pVDZ, cc-pVTZ, aug-cc-pVDZ
and aug-cc-pVTZ basis sets were used. This gives, along with the MP2 calculations
carried out using all Dunning’s and the largest Pople’s basis sets4 (i.e., 6-311++G**
and 6-31++G**), an overall number of 370 typical computational levels and 1480
sets of SFs, i.e., ESFF and SQM SFs within the 11- and 9-parameter scaling frames.
Again the calculations were carried out using Baker’s training set. The reason is
4
Note that the requirements of the MP2 and other correlated ab initio approaches with respect to
the basis set are higher as compared with DFT, due to the necessity of the adequate description of
the correlating orbitals, not only the electron density.
Table 2.6 LSFs and FF SFs of the 9-parameter redundant PICs-based scaling formalism obtained
at the B3LYP level with the selected basis sets using Baker’s training set of molecules
6-311++G** 6-311+G** 6-311G** 6-31G**
SQM ESFF SQM ESFF SQM ESFF SQM ESFF
XX(s,c) 0.9549 0.9815 0.9546 0.9814 0.9514 0.9794 0.9280 0.9690
XX(d,t) + CN(a) 0.9352 0.9692 0.9350 0.9690 0.9226 0.9625 0.9059 0.9527
CCl 1.0364 1.0090 1.0354 1.0103 1.0452 1.0153 1.0265 1.0027
XH 0.9263 0.9628 0.9261 0.9627 0.9268 0.9635 0.9175 0.9581
XXX 1.0079 0.9968 1.0076 0.9964 1.0065 0.9957 1.0192 1.0056
XXH 0.9674 0.9827 0.9676 0.9828 0.9648 0.9817 0.9544 0.9769
HXH 0.9507 0.9752 0.9510 0.9752 0.9497 0.9747 0.9285 0.9639
linear def. 0.8603 0.9452 0.8644 0.9465 0.8438 0.9350 0.8937 0.9504
torsions 0.9595 0.9811 0.9554 0.9786 0.9509 0.9771 0.9508 0.9781
Table 2.7 LSFs and FF SFs of the 11-parameter redundant PICs-based scaling formalism
obtained at the B3LYP level with the selected basis sets using Baker’s training set of molecules
6-311++G** 6-311+G** 6-311G** 6-31G**
SQM ESFF SQM ESFF SQM ESFF SQM ESFF
XX 0.9408 0.9758 0.9406 0.9757 0.9309 0.9717 0.9125 0.9616
CCl 1.0345 1.0084 1.0334 1.0097 1.0431 1.0154 1.0263 1.0019
CH 0.9289 0.9638 0.9288 0.9637 0.9291 0.9644 0.9183 0.9583
NH 0.9150 0.9565 0.9149 0.9565 0.9177 0.9584 0.9129 0.9555
OH 0.9103 0.9541 0.9103 0.9541 0.9118 0.9553 0.9176 0.9579
XXX 1.0221 0.9997 1.0218 0.9994 1.0272 0.9996 1.0320 1.0093
XXH 0.9706 0.9843 0.9707 0.9844 0.9692 0.9837 0.9579 0.9788
HCH 0.9503 0.9748 0.9508 0.9749 0.9486 0.9742 0.9268 0.9633
HNH 0.9295 0.9619 0.9283 0.9614 0.9265 0.9568 0.9225 0.9531
linear def. 0.8575 0.9457 0.8617 0.9461 0.8407 0.9353 0.8944 0.9497
torsions 0.9596 0.9813 0.9555 0.9786 0.9511 0.9773 0.9516 0.9784
Table 2.8 LSFs and FF SFs for new types of ICs in silicon- and sulfur(II)-containing compounds
obtained at the B3LYP level with the selected basis sets
6-311++G** 6-311+G** 6-311G** 6-31G**
SQMa ESFF SQMa ESFF SQMa ESFF SQM ESFF
HY – 0.9637 – 0.9636 – 0.9633 0.9176 0.9579
XY – 1.0101 – 1.0103 – 1.0088 1.0112 1.0036
YY – 1.0771 – 1.0776 – 1.0996 1.1912 1.0449
XXY – 0.9942 – 0.9969 – 0.9771 1.0410 0.9507
XYX+YYX – 1.0358 – 1.0365 – 1.0322 1.0333 1.0426
a
Not calculated due to the problems described in Sect. 2.3.6.2
that the obtained results were supposed to be as consistent with those available in
the literature as possible. It is difficult to summarize that paper and all the rec-
ommendations that follow from it in a few sentences, so the reader interested in
scaling procedures is encouraged to go through the details by reading the reference
in the free time. At this point, we only mention that we recalculated standard
deviations for LSFs5 which are closely related to uncertainties described in
Sect. 2.3.2.2. They turned out to be on average by an order of magnitude lower than
those reported for US [44, 45, 47].
2.3.6 Comparison of Multi-parameter Scaling Procedures
The US procedure is the simplest one of all described, and for this reason, it is the
most frequently used in various applications. However, its obvious disadvantage is
lowest quality of the scaled frequencies. In this section, comparison of the two most
powerful scaling techniques: SQM and ESFF will be presented.
2.3.6.1 Quality of the Scaled Frequencies
First applications of the ESFF technique [13, 68, 69, 71, 74], although successful
could not be directly compared to those of SQM. First, they were carried out using
non-redundant NICs-based algorithm, while the SQM calculations carried out at the
same time for comparison used more flexible, redundant PICs basis. Second,
slightly different classifications of ICs were adopted. For example, in the first
methodological paper [74] only five SFs were optimized in the case of ESFF, eight
FF SFs were used in the case of SQM. In addition, in the paper devoted to toluene,
styrene, and 4-methylstyrene [13] three specific ICs were used to treat the vinyl
group motions in the ESFF scaling. Third, in the case of ESFF weights used in the
optimization procedure were always equal to unity. Nevertheless in an over-
whelming majority of cases, ESFF turned out to provide slightly better results than
SQM in that the RMS and ARPE values were lower. Results before the modifi-
cation of the ESFF procedure consisting in utilization of the redundant PICs
appeared [75] are summarized in Table 2.9.
Full comparison of SQM and ESFF method was given in Ref. [66]. This time
both approaches used the same, redundant PICs-based algorithms. From the sta-
tistical point of view, the ESFF method is noticeably better than SQM—in most
(524–207) out of 740 cases (370 computational levels 2 scaling frame, i.e., 9-
and 11-parameter), the ESFF RMS values were lower than those obtained with
SQM. The SQM method predicts somewhat more accurate XH stretching
5
They were calculated in earlier works devoted to ESFF procedure, but due to a mistake in data
handling they were incorrect.
frequencies (could be accidental, due to the complications that follow from, e.g., the
Fermi resonance in this range). In the high-frequency range, even small increase in
the relative error in the calculated frequency leads to the significant increase in the
frequency deviations (which are transferred to the RMS value). This is compensated
by better description in the middle range of a spectrum by the ESFF method. It is
evident when the ARPE values are analyzed. In this test, the ESFF procedure
surpasses SQM even more (602–92).
The main shortcoming of the ESFF procedure, alike any other frequency scaling
ones (the Pople’s US, or the WLS), is the lack of consistency between the calcu-
lated frequencies and the final molecular FF. In the case of SQM, the FCs and the
final frequencies are obviously consistent, as the latter are obtained from the scaled
(effective) FFs. It has the advantage, in particular if further analysis on the FCs is
required. In addition, apart from obtaining reliable frequencies, improved eigen-
vectors (normal modes) can be used in the transformation of the dipole moment
derivatives to obtain more reliable IR and/or Raman intensities.
2.3.6.2 Numerical Stability
Optimization of the SFs within the ESFF procedure is numerically stable. It consists
in solving a set of linear equations. In contrast, a search for SQM optimal SFs is an
iterative procedure and may fail occasionally (which we found a few times). In
addition, SQM sometimes becomes problematic in the case of factors’ optimization
for the third-row elements [70]. Namely, the successive increase of the YY factor
value with the decreasing quality of the VTZ-type basis set was observed; in the
case of 6-311G*, the value above 1.6 was found for that factor, at the expense of
another, which was significantly lower than 1. The FF SFs significantly larger than
unity for the structural motifs including the third-row atoms have been already
found before in earlier works [67]. However, the main idea behind the scaling
procedures consists in using the SFs that are close to unity. Thus in spite of
providing reasonable frequencies, the conclusions related to the SQM-scaled FFs
may be wrong when one, or a few SFs significantly deviate from unity. This is not
the case of the ESFF factors. Regardless of the reason of such a behavior, it seems
that the ESFF scaling procedure is more stable, in particular in the case of the
systems for which strong coupling of local modes is observed.
2.3.6.3 Possible Future Applications
The ESFF method is well suited to the description of the macromolecular com-
pounds, better than SQM. Obviously, we do not have in mind the simulations of
spectra, like those presented in [77], as the SQM method can be used in conjunction
with the methodology presented there with equal success. However, when
attempting to describe really large systems in the restricted frequency range at the
levels of the mode-tracking [78, 79] and/or intensity-tracking [80, 81] procedures,
Table 2.9 Comparison of the SQM and ESFF methods

References Application SQM (redundant PICs, ESFF (NICs, weights equal
recommended weights were to unity)
used)
# # RMS/ # # RMS/
SF frequ. ARPE SF frequ. ARPE
[74] toluene 8 19 3.29/– 5 19 2.48/–
a
[13] toluene 8 66 4.31 /– 8 66 2.85a/–
styrene 4.13b/– 3.19b/–
4-methylstyrene 5.24c/– 3.67c/–
[68] Baker’s training 11 660 11.81/ 11 660 11.66/
set 1.06 1.02
[69] Baker’s training 9 660 11.77/ 9 660 11.62/
set 1.02 0.99
[71] Baker’s training 9d 293 8.8/0.58 9d 293 8.6/0.55
set
Number of SFs (# SF), number of frequencies the calculations were based on (# frequ.), RMS (in
cm−1) and ARPE (in %) are given
a
For toluene
b
For styrene
c
For 4-methylstyrene
d
Precomputed SFs were used. Additional factor for the C=O bond participating in hydrogen
bonding was introduced and optimized
utilization of the SQM scaling is not possible, on account of the determination of

only the frequencies and normal modes (eigenvectors), rather than the molecular
FFs. The ESFF scaling can be applied, though. Admittedly, the FCs enter the
expression for the PED coefficients, Eq. (2.54); however, this problem is easily
eliminated with the aid of the described ESFF2 procedure [75], according to which
the contributions of the local vibrations to the normal modes can be determined
only on the basis of the amplitudes of the atomic displacements from the equilib-
rium positions, Eq. (2.62). As described earlier, this procedure turned out to be as
efficient as the original one.
References
1. Pulay P, Fogarasi G, Pang F, Boggs JE (1979) Systematic ab initio gradient calculation of

molecular geometries, force constants, and dipole-moment derivatives. J Am Chem Soc
101:2550–2560
2. Lide DR, Frederikse HPR (eds) (1994) CRC handbook of chemistry and physics. CRC Press,
London
3. Atkins PW (1993) Molecular quantum mechanics. Oxford University Press, New York
4. Wilson EB Jr, Decius JC, Cross PC (1955) Molecular vibrations. The theory of infrared and
Raman vibrational spectra. Dover Publications Inc., New York
5. Califano S (1976) Vibrational states. Wiley, London
6. Pulay P (1969) Ab initio calculation of force constants and equilibrium geometries in

polyatomic molecules. I Theory Mol Phys 17:197–204
7. Pulay P, Meyer W (1974) Comparison of ab initio force constants of ethane, ethylene and
acetylene. Mol Phys 27:473–490
8. Pople JA, Krishnan R, Schlegel HB, Binkley JS (1979) Derivative studies in Hartree-Fock
and Moller-Plesset theories. Int J Quantum Chem 16:225–241
9. Willetts A, Handy NC, Green WH, Jayatilaka D (1990) Anharmonic corrections to vibrational
transition intensities. J Phys Chem 94:5608–5616
10. Barone V, Bloino J, Guido CA, Lipparini F (2010) A fully automated implementation of
VPT2 Infrared intensities. Chem Phys Lett 496:157–161
11. Pulay P, Fogarasi G, Boggs JE (1981) Force field, dipole moment derivatives, and vibronic
constants of benzene from a combination of experimental and ab initio quantum chemical
information. J Chem Phys 74:3999–4014
12. Pulay P, Fogarasi G, Pongor G, Boggs JE, Vargha A (1983) Combination of theoretical
ab initio and experimental information to obtain reliable harmonic force-constants—scaled
quantum-mechanical (SQM) force-fields for glyoxal, acrolein, butadiene, formaldehyde, and
ethylene. J Am Chem Soc 105:7037–7047
13. Borowski P, Fernández-Gómez M, Fernández-Liencres M-P, Peña Ruiz T, Quesada Rincón
M (2009) An effective scaling frequency factor method for scaling of harmonic vibrational
frequencies: application to toluene, styrene and its 4-methyl derivative. J Mol Struct 924:493–
503
14. McQuarrie DA (1976) Statistical mechanics. HarperCollinsPublishers, New York
15. Merrick JP, Moran D, Radom L (2007) An evaluation of harmonic vibrational frequency scale
factors. J Phys Chem A 111:11683–11700
16. Pople JA, Schlegel HB, Krishnan R, Defrees DJ, Binkley JS, Frisch MJ, Whiteside RA,
Hout RF, Hehre WJ (1981) Molecular-orbital studies of vibrational frequencies. Int J
Quantum Chem 15:269–278
17. Hout RF, Levi BA, Hehre WJ (1982) Effect of electron correlation on theoretical vibrational
frequencies. J Comput Chem 3:234–250
18. Pople JA, Scott AP, Wong MW, Radom L (1993) Scaling factors for obtaining fundamental
vibrational frequencies and zero-point energies from HF/6-31 g-asterisk and MP2/
6-31 g-asterisk harmonic frequencies. Isr J Chem 33:345–350
19. Scott AP, Radom L (1996) Harmonic vibrational frequencies: an evaluation of Hartree-Fock,
Moller-Plesset, quadratic configuration interaction, density functional theory, and semiem-
pirical scale factors. J Phys Chem 100:16502–16513
20. Wong MW (1996) Vibrational frequency prediction using density functional theory. Chem
Phys Lett 256:391–399
21. Patton LF, Corchado J, Sanchez ML, Truhlar DG (1999) Optimized parameters for scaling
correlation energy. J Phys Chem A 103:3139–3143
22. Lynch BJ, Truhlar DG (2001) How well can hybrid density functional methods predict
transition state geometries and barrier heights? J Phys Chem A 105:2936–2941
23. Zhao Y, Lynch BJ, Truhlar DG (2004) Doubly hybrid meta DFT: new multi-coefficient
correlation and density functional methods for thermochemistry and thermochemical kinetics.
J Phys Chem A 108:4786–4791
24. Zhao Y, Lynch BJ, Truhlar DG (2004) Development and assessment of a new hybrid density
functional model for thermochemical kinetics. J Phys Chem A 108:2715–2719
25. Zhao Y, Truhlar DG (2004) Hybrid meta density functional theory methods for
thermochemistry, thermochemical kinetics, and noncovalent interactions: the MPW1B95
and MPWB1K models and comparative assessments for hydrogen bonding and van der Waals
interactions. J Phys Chem A 108:6908–6918
26. Schultz NE, Zhao Y, Truhlar DG (2005) Databases for transition element bonding: metal–
metal bond energies and bond lengths and their use to test hybrid, hybrid meta, and meta
density functionals and generalized gradient approximations. J Phys Chem A 109:4388–4403
27. Curtiss LA, Redfern PC, Raghavachari K, Pople JA (2001) Gaussian-3X (G3X) theory: use of
improved geometries, zero-point energies, and Hartree-Fock basis sets. J Chem Phys
114:108–117
28. Halls M, Velkovski J, Schlegel H (2001) Harmonic frequency scaling factors for
Hartree-Fock, S-VWN, B-LYP, B3-LYP, B3-PW91 and MP2 with the Sadlej pVTZ electric
property basis set. Theor Chem Acc 105:413–421
29. Sinha P, Boesch SE, Gu C, Wheeler RA, Wilson AK (2004) Harmonic vibrational
frequencies: scaling factors for HF, B3LYP, and MP2 methods in combination with
correlation consistent basis sets. J Phys Chem A 108:9213–9217
30. Andersson MP, Uvdal P (2005) New scale factors for harmonic vibrational frequencies using
the B3LYP density functional method with the triple-f basis set 6–311 + G(d, p). J Phys
Chem A 109:2937–2941
31. Csonka GI, Ruzsinszky A, Perdew JP (2005) Estimation, computation, and experimental
correction of molecular zero-point vibrational energies. J Phys Chem A 109:6779–6789
32. Tantirungrotechai Y, Phanasant K, Roddecha S, Surawatanawong P, Sutthikhum V,
Limtrakul J (2006) Scaling factors for vibrational frequencies and zero-point vibrational
energies of some recently developed exchange-correlation functionals. J Mol
Struc-THEOCHEM 760:189–192
33. Chan B, Radom L (2016) Frequency scale factors for some double-hybrid density functional
theory procedures: accurate thermochemical components for high-level composite protocols.
J Chem Theory Comput 12:3774–3780
34. Chan B (2017) Use of low-cost quantum chemistry procedures for geometry optimization and
vibrational frequency calculations: determination of frequency scale factors and application to
reactions of large systems. J Chem Theory Comput 13:6052–6060
35. Laury ML, Carlson MJ, Wilson AK (2012) Vibrational frequency scale factors for density
functional theory and the polarization consistent basis sets. J Comput Chem 33:2380–2387
36. Krasnoshchekov SV, Stepanov NF (2007) Scale factors as effective parameters for correcting
nonempirical force fields. Russ J Phys Chem A 81:585–592
37. Andrade SG, Gonçalves LCS, Jorge FE (2008) Scaling factors for fundamental vibrational
frequencies and zero-point energies obtained from HF, MP2, and DFT/DZP and TZP
harmonic frequencies. J Mol Struc-THEOCHEM 864:20–25
38. Sierraalta A, Martorell G, Ehrmann E, Añez R (2008) Improvement of scale factors for
harmonic vibrational frequency calculations using new polarization functions. Int J Quantum
Chem 108:1036–1043
39. Alecu IM, Zheng J, Zhao Y, Truhlar DG (2010) Computational thermochemistry: scale factor
databases and scale factors for vibrational frequencies obtained from electronic model
chemistries. J Chem Theory Comput 6:2872–2887
40. Basire M, Parneix P, Calvo F (2010) Finite-temperature IR spectroscopy of polyatomic
molecules: a theoretical assessment of scaling factors. J Phys Chem A 114:3139–3146
41. Friese DH, Törk L, Hättig C (2014) Vibrational frequency scaling factors for correlation
consistent basis sets and the methods CC2 and MP2 and their spin-scaled SCS and SOS
variants. J Chem Phys 141:194106
42. Kesharwani MK, Brauer B, Martin JML (2015) Frequency and zero-point vibrational energy
scale factors for double-hybrid density functionals (and other selected methods): can
anharmonic force fields be avoided? J Phys Chem A 119:1701–1714
43. Kashinski DO, Chase GM, Nelson RG, Di Nallo OE, Scales AN, VanderLey DL, Byrd EFC
(2017) Harmonic vibrational frequencies: approximate global scaling factors for TPSS, M06,
and M11 functional families using several common basis sets. J Phys Chem A 121:2265–
2273
44. Irikura KK, Johnson RD III, Kacker RN (2005) Uncertainties in scaling factors for ab initio
vibrational frequencies. J Phys Chem A 109:8430–8437
45. Irikura KK, Johnson RD III, Kacker RN, Kessel R (2009) Uncertainties in scaling factors for
ab initio vibrational zero-point energies. J Chem Phys 130:114102
46. Irikura KK, Johnson RD III, Kacker RN, Kessel R (2009) Erratum: “Uncertainties in scaling
factors for ab initio vibrational zero-point energies” [J. Chem. Phys. 130, 114102 (2009)].
J Chem Phys 131:169902
47. Teixeira F, Melo A, Cordeiro MNDS (2010) Calibration sets and the accuracy of vibrational
scaling factors: a case study with the X3LYP hybrid functional. J Chem Phys 133:114109
48. Pernot P, Cailliez F (2011) Comment on “Uncertainties in scaling factors for ab initio
vibrational zero-point energies” [J. Chem. Phys. 130, 114102 (2009)] and “Calibration sets
and the accuracy of vibrational scaling factors: a case study with the X3LYP hybrid
functional” [J. Chem. Phys. 133, 114109 (2010)]. J Chem Phys 134:167101
49. Irikura KK, Johnson RD III, Kacker RN, Kessel R (2011) Response to “Comment on
‘Uncertainties in scaling factors for ab initio vibrational zero-point energies’ and ‘Calibration
sets and the accuracy of vibrational scaling factors: A case study with the X3LYP hybrid
functional’” [J. Chem. Phys. 134, 167101 (2011)]. J Chem Phys 134:167102
50. Teixeira F, Melo A, Cordeiro MNDS (2011) Response to “Comment on ‘Uncertainties in
scaling factors for ab initio vibrational zero-point energies’ and ‘Calibration sets and the
accuracy of vibrational scaling factors: A case study with the X3LYP hybrid functional’”
[J. Chem. Phys. 134, 167101 (2011)]. J Chem Phys 134:167103
51. Yoshida H, Ehara A, Matsuura H (2000) Density functional vibrational analysis using
wavenumber-linear scale factors. Chem Phys Lett 325:477–483
52. Yoshida H, Takeda K, Okamura J, Ehara A, Matsuura H (2002) A new approach to
vibrational analysis of large molecules by density functional theory: wavenumber-linear
scaling method. J Phys Chem A 106:3580–3586
53. Kudoh S, Takayanagi M, Nakata M (2000) Infrared spectra of Dewar 4-picoline in
low-temperature argon matrices and vibrational analysis by DFT calculation. Chem Phys Lett
322:363–370
54. Baker J, Jarzecki AA, Pulay P (1998) Direct scaling of primitive valence force constants: an
alternative approach to scaled quantum mechanical force fields. J Phys Chem A 102:1412–
1424
55. Mills IM (1960) Vibrational perturbation theory. J Mol Spectrosc 5:334–340
56. Blom CE, Slingerland PJ, Altona C (1976) Application of self-consistent-field ab-initio
calculations to organic-molecules. I. Equilibrium structure and force constants of hydrocar-
bons. Mol Phys 31:1359–1376
57. Blom CE, Altona C (1976) Application of self-consistent-field ab-initio calculations to
organic-molecules. II. Scale factor method for calculation of vibrational frequencies from
ab-initio force constants—ethane, propane and cyclopropane. Mol Phys 31:1377–1391
58. Blom CE, Otto LP, Altona C (1976) Application of self-consistent-field ab-initio calculations
to organic-molecules. III. Equilibrium structure of water, methanol and dimethyl ether,
general valence force-field of water and methanol scaled on experimental frequencies. Mol
Phys 32:1137–1149
organic-molecules. IV. Force constants of propene scaled on experimental frequencies. Mol
Phys 33:875–885
organic-molecules. V. Ethene—general valence force-field scaled on harmonic and anhar-
monic data, infra-red and Raman intensities. Mol Phys 34:177–192
61. Blom CE, Altona C, Oskam A (1977) Application of self-consistent-field ab-initio
calculations to organic-molecules. VI. Dimethylether—general valence force-field scaled on
experimental frequencies, infra-red and Raman intensities. Mol Phys 34:557–571
62. Rauhut G, Pulay P (1995) Transferable scaling factors for density-functional derived
vibrational force-fields. J Phys Chem 99:3093–3100
63. Rauhut G, Pulay P (1995) Transferable scaling factors for density-functional derived
vibrational force-fields. J Phys Chem 99:14572
64. Rauhut G, Pulay P (1995) Identification of isomers from calculated vibrational-spectra—a
density-functional study of tetrachlorinated dibenzodioxins. J Am Chem Soc 117:4167–4172
65. Fogarasi G, Zhou X, Taylor PW, Pulay P (1992) The calculation of ab initio molecular
geometries: efficient optimization by natural internal coordinates and empirical correction by
offset forces. J Am Chem Soc 114:8191–8201
66. Borowski P (2012) An evaluation of scaling factors for multiparameter scaling procedures
based on DFT force fields. J Phys Chem A 116:3866–3880
67. Kalincsák F, Pongor G (2002) Extension of the density functional derived scaled quantum
mechanical force field procedure. Spectrochim Acta A 58:999–1011
68. Borowski P, Drzewiecka A, Fernández-Gómez M, Fernández-Liencres M-P, Peña Ruiz T
(2008) An effective scaling frequency factor method for harmonic vibrational frequencies: the
factors’ transferability problem. Chem Phys Lett 465:290–294
69. Borowski P, Drzewiecka A, Fernández-Gómez M, Fernández-Liencres M-P, Peña Ruiz T
(2010) A new, reduced set of scaling factors for both SQM and ESFF calculations. Vib Spect
52:16–21
70. Borowski P, Peña Ruiz T, Barczak M, Pilorz K, Pasieczna-Patkowska S (2012) Application of
the multi-parameter SQM harmonic force field, and ESFF harmonic frequencies scaling
procedures to the determination of the vibrational spectra of silicon- and sulfur(II)-containing
compounds. Spectrochim Acta A 86:571–585
71. Borowski P, Pilorz K, Pitucha M (2010) An effective scaling frequency factor method for
scaling of harmonic vibrational frequencies: application to 1,2,4-triazole derivatives.
Spectrochim Acta A 75:1470–1475
72. Fábri C, Szidarovszky T, Magyarfalvi G, Tarczay G (2011) Gas-phase and Ar-matrix SQM
scaling factors for various DFT functionals with basis sets including polarization and diffuse
functions. J Phys Chem A 115:4640–4649
73. Morino Y, Kuchitsu K (1952) A note on the classification of normal vibrations of molecules.
J Chem Phys 20:1809–1810
74. Borowski P, Fernández-Gómez M, Fernández-Liencres M-P, Peña Ruiz T (2007) An effective
scaling frequency factor method for scaling of harmonic vibrational frequencies: theory and
preliminary application to toluene. Chem Phys Lett 446:191–198
75. Borowski P (2010) An effective scaling frequency factor method for scaling of harmonic
vibrational frequencies: the use of redundant primitive coordinates. J Mol Spectr 264:66–74
́
76. Martınez-Torres E (2000) Formulation of the vibrational theory in terms of redundant internal
coordinates. J Mol Struct 520:53–61
77. Borowski P, Pasieczna-Patkowska S, Barczak M, Pilorz K (2012) Theoretical determination
of the infrared spectra of amorphous polymers. J Phys Chem A 116:7424–7435
78. Reiher M, Neugebauer J (2003) A mode-selective quantum chemical method for tracking
molecular vibrations applied to functionalized carbon nanotubes. J Chem Phys 118:1634–
1641
79. Herrmann C, Neugebauer J, Reiher M (2007) Finding a needle in a haystack: direct
determination of vibrational signatures in complex systems. New J Chem 31:818–831
80. Luber S, Neugebauer J, Reiher M (2009) Intensity tracking for theoretical infrared
spectroscopy of large molecules. J Chem Phys 130:064105
81. Kiewisch K, Luber S, Neugebauer J, Reiher M (2009) Intensity tracking for vibrational
spectra of large molecules. Chimia 63:270–274
Chapter 3
Quantum Dot and Fullerene
with Organic Chromophores
as Electron-Donor-Acceptor Systems
Danuta Wróbel and Bolesław Barszcz
Abstract This review paper is focused on the research of molecular mechanisms

occurring in porphyrin-like systems such as porphyrins, phthalocyanines, and
corroles as well as in chromophore-semiconductor quantum dot (QD-CdSe/ZnS) or
corrole-fullerene (C60) as electron-donor-acceptor unites. The basic spectroscopic
investigations describe properties of materials in organic solutions in the ultraviolet,
visible, and infrared ranges and in a form of Langmuir and Langmuir–Blodgett
molecular nanolayers to get knowledge on photophysics of dyes and the influence
of QD and C60 on the electron redistribution within the molecular structures. The
studies also allowed to explain the impact of solvent on the spectroscopic properties
of corroles and on the redistribution of the p-electrons in the excited state. The
fluorescence studies very evidently showed strong interaction between chro-
mophores and C60 or QD and clearly demonstrated the strong donor-acceptor nature
of the phthalocyanines-quantum dot and the corrole-fullerene dyad. In addition,
spectroscopic studies in polarized light allowed determining molecular arrangement
of the chromophore molecules in the Langmuir–Blodgett layers with respect to
solid substrates. The computer calculations (TD-DFT theory) confirmed the
experimental results, in particular the redistribution of the p-electrons in the excited
state and the location of HOMO and LUMO levels. The DFT calculations let also to
evaluate the reorganization energy values for the set of free-base corroles and C60
fullerene. In this review, it was shown the electron-donor-acceptor character of the
systems composed of: porphyrin-quinone, phthalocyanines-QD, corroles-C60
dyads. It has been demonstrated potential capabilities of the photoactive organic
materials with QD and fullerene in the future applications in many areas of opto-
electronic and in the process of converting solar energy into electric energy in solar
cells.
D. Wróbel (&) B. Barszcz

Faculty of Technical Physics, Institute of Physics, Poznan University
of Technology, Poznań, Poland
e-mail: danuta.wrobel@put.poznan.pl
B. Barszcz
Institute of Molecular Physics Polish Academy of Sciences, Poznań, Poland

98 D. Wróbel and B. Barszcz
3.1 Introduction
One of the most important phenomena in the life of living organisms is the process
of electron/energy transfer in photosynthetic systems that convert solar energy into
chemical energy, as well as in photovoltaic devices that convert solar energy into
electricity [1]. In many technological activities, man mimics nature and acquired
knowledge allows him to follow the mechanisms and processes occurring in the
electron-donor-acceptor systems that are also extremely important in the molecular
photovoltaics which is sometime called “artificial photosynthesis.”
The turn of the twentieth century and the beginning of the twenty-first century is
a period of extremely rapid development of material engineering, especially in
nanotechnology. The reason for such a development of those sciences from the
borderline of physics, chemistry, and electronics is the search for new materials
with desirable properties and at the same time being highly efficient, cheap, and
environmentally friendly. On the other hand, in the face of the depletion of fossil
energy carriers, and above all because of the caused by them atmospheric pollution,
there is an urgent need to find alternative energy sources cooperating with new
organic materials.
Among the many more or less economically justified and socially acceptable
energy sources, the dominant role is played by the conversion of solar energy into
electricity using so-called solar batteries. This term is usually understood as pho-
tovoltaic devices capable of converting light energy emitted by the sun into elec-
trical energy. Such devices are already working in countries with high insolation
but, unfortunately, they are not cheap converters and their efficiency is not satis-
factory from an economic point of view. The consequence of low efficiency is the
huge sizes of solar batteries, which practically eliminates them from more densely
populated areas. That is why there is a constant struggle to find highly efficient,
cheap, and environmentally friendly converters of solar energy into electricity.
This review deals with the most important aspects of the experimental consid-
erations supported by quantum-mechanical calculations for some selected organic
and inorganic materials. Such a choice is justified by the huge variety of known
organic dyes and systems as well by the easiness of their modification, so they are
even more efficient, more durable, cheaper and they can be easily disposed. In the
chapter, the authors intend to put a special emphasis on characterization of
porphyrin-like dyes and their mixtures with quantum dots (phthalocyanine-QDs).
Moreover, the dyads of porphyrin chromophores with fullerene (corrole-C60) which
are promising materials to be useful for optoelectronics as photoactive materials are
also illustrated. Exemplary photoactive materials are investigated in solutions and in
a form of nanolayers. Comprehensive discussion will focus on properties of
selected organic dyes (the porphyrin family—porphyrins, phthalocyanines, cor-
roles) as well as their interaction with QD and C60. Up to date, the role of QD,
corroles, and fullerenes in organic photovoltaics is not fully known yet. In the
paper, we present the selected experimental and theoretical results, including the
basic photophysical properties like UV-VIS and IR absorption, molecular
3 Quantum Dot and Fullerene with Organic Chromophores as … 99
aggregation, molecular orientation, fluorescence quenching, reorganization energy,

HOMO and LUMO levels as well as processes of energy/electron transfer and
electron redistribution. These materials are gaining a lot of interest not only among
chemists, but also physicists and materials engineers.
As it appears from the above, it will be a chapter giving a brief outline of the
mechanisms and processes occurring in photoactive organic materials important for
their prospective applications and can serve as one of the basic sources of infor-
mation about new materials. The research methods used to study organic materials
were conventional and unconventional spectroscopic methods, supported by com-
puter simulations.
3.2 Brief History of Photocells Based

on Organic Materials
So far, the photoactive materials, which are used for solar cell manufacture, are
based on germanium, silicon, gallium arsenide, amorphous silicon and their alloys,
and others. However, efficiency of commercial solar cells based on such materials
does not exceed 12%, and their great disadvantage is high cost when compared with
conventional sources of electric power.
In the first dye-sensitized photovoltaics (DSPV), the photoactive molecules were
sensitive to the UV and blue light [1–10]. However, over the past few decades
intense research to improve DSPV have brought a great evolution in photocell
designs which stemmed from the use of a series of p-electron dyes and their
complexes working in DSPV as photosensitizers [4]. Among them there are cells
based on porphyrins, phthalocyanines and other macrocyclic dyes absorbing
intensively in the visible range [7–10]. The first successful solid hybrid
dye-sensitized solar cells were reported to improve electron transport. Numerous
research groups have experimented with a wide variety of a family of organic
porphyrin dyes; however, DSPV with such materials reached rather low efficiency
[5, 6]. Many new dye-sensitized solar cells based on organic functional dyes
comprised of porphyrins covalently linked with other moieties such as terthio-
phenes building blocks, rhodamines or ruthenium complexes have been also
extensively reported as photosensitizers for developing photoelectrochemical solar
cells [7–10]. Also porphyrin-sensitized TiO2 or systems with porphyrin-gold
nanoparticles [7, 11] were demonstrated as materials for a new cell that directly
converts light energy into electricity. Many groups of scientists have presented the
results of the chromophore-based DSPV devices with different working media:
from traditional organic liquid electrolytes, ionic liquid electrolytes, polymer gel
electrolytes, quasi-solid state to self-assembled monolayer, Schottky barrier cells,
and other systems [7–11]. In many cases, these DSPV systems offer high photo-
voltaic effect, chemical and thermal stability as well as opportunity for developing
long-term operational solar cells. The basic spectroscopic properties of photoactive
dyes and interrelation between the molecular structure of dyes and their ability to
photocurrent generation were also the subjects of the studies [7–21]. However, the
efficiency of the systems proposed in [2–10] is still too low to fulfill current
technical and economic as well as ecological requirements.
3.3 Porphyrins and Phthalocyanies
It is very essential to develop molecular materials, which will be characterized by

not only strong light absorption in the range of the maximal solar exposure but also
fast charge separation and slow charge recombination. Figure 3.1 shows the
molecular structures of the organic materials presented as examples.
Many metallic porphyrin derivatives and relative compounds occur in nature.
One of the best-known families of porphyrin complexes is heme—a component of
hemoglobin. Porphyrins play a special role not only in human body but also in
photosynthesizing organisms (photosynthesis), and together with phthalocyanines
they are also promising photo-agents in medicine (e.g, photodynamic therapy) and
optoelectronics (e.g., photovoltaics).
Porphine is the parent chemical compound for types of the essential groups of a
large family of porphyrins, phthalocyanines and other derivatives. Porphyrin is an
organic aromatic heterocyclic compound containing four pyrrole rings joined by
four methine bridges forming a large macrocycle ring [2]. The total number of
p-electrons in porphyrins is 26 and 18 p-electrons create a planar macrocycle. The
porphyrin indicating D4h symmetry characterizes porphyrin coordinated to a
metal ion (metallic porphyrin—Mporphyrin) and possesses two Q bands in the
absorption spectrum or D2h (free-base porphyrin—H2-porphyrin), characterized by
the four Q bands depending on the presence or absence of a metal ion in the central
macrocycle [3]. Their rigid, square-planar ring, aromaticity, strong visible light
absorption, not bad fluorescence, good solubility in many organic solvents and easy
to create thin layers, easy to create cation and anion radicals and rich coordination
Fig. 3.1 Molecular structures of unsubstituted phthalocyanine (Pc), tetraphenylporphyrin (TPP),

and corrole (corr). M—metal
chemistry are the main values of those materials and make the porphyrin family a
great interest.
Another very good example of porphyrin species are phthalocyanines that can be
potentially used in photovoltaics. Phthalocyanines were discovered at the beginning
of the last century (1907) [4] and in that time and now they have been studied as
functional materials. When compared to porphyrin their macrocycles are build of
four indole units connected with azomethine groups instead of methine bridges.
Phthalocyanines are porphyrin-like dyes, and they belong to heterocyclic groups
which are characterized by: high chemical and physical stability, effective inter-
molecular electron transfer, easiness of molecular structure modification.
Phthalocyanine derivatives can change the color from red to blue depending on
metal ion inside the macrocycle. Phthalocyanines like porphyrins are also of a great
interest as photosensitizers to many scientific and technological applications.
Non-substituted phthalocyanines and some of their metalized derivatives and
complexes show rather poor solubility. Moreover, the substitution influences a shift
of the absorption bands toward the lower wavelengths. However, thanks to great
possibilities of modifying phthalocyanines molecular structure as well their par-
ticular photophysical and photochemical properties, they give a wide range of
possibilities to be used in many fields of molecular optoelectronics. Besides, the
substitution, many Pc are characterized by good solubility. The dyes are charac-
terized not only by strong absorption in the visible light, but some of them show
fluorescence. Also they are characterized by very high resistance to light, weak
influence of additional rings and metal ions on some spectral properties.
Both porphyrins and phthalocyanines are characterized by strong absorption
bands; porphyrin strong band is observed in the Soret region (420–450 nm) and
much less intense Q bands in the long-wavelength region [5]. Contrary, in
phthalocyanines the Q bands in the long-wavelength region (650–680 nm) are
much intensive whereas the less intense bands in the range of the Soret bands are
present [5].
The studies of spectral and photoelectric properties of a series of porphyrins and
phthalocyanines in the simple or more complex systems [6–27] as potential pho-
tosensitizers are required to find a correlation between molecular structure of
chromophores and their ability in photocurrent creation. Any change in the
molecular structure of dyes causes dissimilarity of spectroscopic properties and in
consequence in modification of photoactive properties. Typical examples for basic
porphyrins are: metal-free porphyrin (free-base tetraphenylporphyrin—H2TPP),
metallic porphyrins (MTPP—e.g., MgTNP, ZnTNP, PbTNP, CuTNP), metallic
tetranaphenylporphyrin (MTNP). For example, magnesium and zinc porphyrins
(MgTPP and ZnTPP, respectively) have large molar absorption coefficients and
good effectiveness in light energy to electric energy conversion. In contrast to those
dyes, free-base and lead porphyrins (H2TPP and PbTPP, respectively) are rather not
efficient in photoconversion. Likewise, the representatives of the phthalocyanine
group can be the following dyes: metal-free phthalocyanine (free-base phthalo-

cyanine—H2Pc), metallic phthalocyanine (M-Pc—e.g., CuPc, ZnPc), and those
substituted with peripheral groups like zinc phthalocyanine substituted with 16
fluorine atoms (ZnPcF16), zinc one substituted with 12 chlorine atoms (ZnPcCl12),
zinc dye substituted with long alkyloxy chains (ZnPcAlkyl), and unsymmetrical
indium phthalocyanine (InPc) and gallium (GaPc) substituted with terminal alkynyl
group [18–21], methane groups at peripheral positions and many others moieties
[6–21].
A large energy gap between the porphyrin and phthalocyanine absorption bands
is unfavorable and can lead to reduction in the efficiency of energy transport
between the dyes. One of the good solutions to this problem is the use of other dyes
that fill the missing energy gap—e.g., merocyanine dyes which posses high
absorption coefficient in the range of 450–580 nm [23–25]. The merocyanine dye
itself is not rather good energy converter. As shown by spectroscopic studies, the
light energy absorbed by merocyanines completes the porphyrin-phthalocyanine
energy gap, and as a result it can lead to increase photovoltaic cell efficiency
[26, 27].
In many systems, photoinduced electron transport is studied by fluorescence and
electron spin resonance (ESR) [28]. The transparent and simple example could be a
mixture of porphyrin and quinone. In a mixture composed of porphyrin (H2TPP,
MgTPP, ZnTPP) and quinone in dimethyloxide (DMSO), energy transfer was
observed by quenching fluorescence supported by the ESR experiment. The
changes in the dye’s fluorescence quantum yield, declining dye’s lifetime in the
presence of quinone and the Stern–Volmer equation analysis confirm quinone as a
good quencher. Those results confirmed occurrence of the electron transfer process
between the dye (donor) and quinone (acceptor). Another spectacular example how
the changes in the substituent in porphyrins affect dye’s photoactivity is a series of
studies of meso-tetraphenylzincporphyrins sensitized in TiO2 cells [29–31]. In this
point, it is worth to underline that the spectroscopic and photovoltaic experiments
done for photoactive materials in the simple models are placed in, e.g., nematic
liquid crystal (MBBA-EBBA), DMSO or polyvinyl alcohol with DMSO
(PVA-DMSO) or in polymers [5–30].
In many approaches, the experiments are carried out for the systems in the
volume phase as well as for thin-film systems (Langmuir, Langmuir-Blodgett,
Langmuir-Schaefer layers, and others). The systems based on organic chro-
mophores absorb strongly light in the range of solar exposure. However, efficiency
of the systems based on organics is rather low. The essential confinements are:
• low mobility of electron carriers in organic materials—this effect is partially
compensated by large absorption coefficients (>105 cm−1),
• and rather small diffusion depth of photoexcitation—it limits the thickness of
organic layers to ca. 10 nm.
It is connected with an exciton short lifetime (order of picoseconds) and leads to

fast charge recombination. Thus, searching for new systems which are able to omit
or reduce above problems seems to be very essential. To get knowledge on physical
and photophysical properties, the following research techniques are mostly used:
UV-VIS, IR absorption, FT-IR spectroscopy, in situ absorption spectroscopy,
fluorescence spectroscopy, electron paramagnetic resonance (EPR) spectroscopy,
thin film, and computer simulations.
3.4 Phthalocyanine and Porphyrin with Quantum Dots
For many reasons, organic photoactive systems composed with organic dyes and
semiconductor quantum dot are of a wide interest among many researchers due to
their perspective applications [32–38]. Phthalocyanines with a quantum dot are
more complex systems than those discussed above.
The process responsible for that was described as the fluorescence resonance
energy transfer. There are some examples of that. In CdTe-sulfonated aluminum
phthalocyanines systems, the QD can be used as a perfect energy donor to
phthalocyanine after light excitation. In the QD-organic dye system, phthalocya-
nines play a role of photosensitizers [33, 34]. The studies in the paper [37] on the
cell with QD and semiconducting polymer (PCPDTBT) evidently showed the
enhancement of the short-circuit current and that the blend of spherical quantum
dots and elongated nanorods results in a good electron pathway in the p-polymer
matrix, giving efficiencies of 3.6%. What is worth to underline is how fabrication
methods of the quantum dots, morphology of the photoanode, the type of elec-
trolyte used, and the choice of the counter electrode can improve solar cell effi-
ciencies [36].
Spectroscopic and thermodynamic examinations of dyes are one of the steps
allowing determination of the mechanisms of interrelation, aggregation, and elec-
tron transfer processes. Models composed of QD and chromophores in solution as
well as in the Langmuir layer can be an example which gives a description of the
interaction between units.
The core/shell CdSe/ZnS QD and structurally different zinc(II) and copper(II)
phthalocyanines substituted with butyl or octakis(octyloxy) peripheral groups
(2,9,16,23-tetra-tert-butyl-29H,31H-phthalocyanine 2,3,9,10,16,17,23,24-octakis
(octyloxy)-29H,31H-phthalocyanine) and indium and gal phthalocyanines are
excellent models to be investigated. Figure 3.2 presents absorption spectra of zinc
phthalocyanine in solution (Fig. 3.2a), creation of aggregates with the use of in situ
Langmuir layer (Fig. 3.2b), p-A isotherm (Fig. 3.2c) and compressibility of the dye
layer (Fig. 3.2d). Compression modulus of the layers evaluated with the use of a
Langmuir trough lets to determine their thermodynamic properties. The following
parameters can be estimated: p-A compression isotherms (compression versus
molecule area), compression-collapse parameters, and stability of layers. These
Fig. 3.2 Zinc phthalocyanine absorption spectra in solution (a), in situ absorption spectra of the
Langmuir layer (b), p-A isotherm (c), and compressibility of the layer (d)
investigations are needed to determine the best conditions to get Langmuir–Blodgett

layers of good quality (transfer ratio close to 1). The in situ light absorption technique
lets to establish interaction among closely packed species and creation of molecular
aggregates. Formation of the LB layers is usually performed on mica, quartz, gold or
silicon or other substrates. The number of layers can be controlled to get optimal
conditions for spectroscopic and photoelectric studies. It is well known that efficiency
of electron and energy transfer depends on the layer ordering, and thus, the spec-
troscopic studies are very often done with polarized light (in the UV-VIS and IR
regions) to follow bonding responsible for interaction with the solid surface.
The isotherms of surface pressure as a function of the mean area per molecule are
often examined, and the results for In and Gl, Zn phthalocyanines are presented in [19].
In many cases, it is not possible to determine unambiguously formation of aggregates
and/or their type. Due to such uncertainty, absorption spectra are also measured with the
use of the in situ method for the dyes in the 2D Langmuir layer—the in situ spectra done
for zinc phthalocyanines dyes evidently show that the dyes create aggregates (Fig. 3.2a,
b). For example, in the in situ spectrum, a broadband in the range 800–500 nm is
evidently observed, as a result of overlapping of single narrow bands and one
can observe the shift of the band 699 nm of about 40 nm toward the shorter wavelength
of 658 nm. The essential influence of a kind of the substituent attached to the
phthalocyanine macroring on dye thermodynamic properties is presented in [19]; the

creation of the H aggregate in GaPc or oblige in InPc (the later not shown). The in situ
studies clearly confirmed the significant influence of metal in the indium, gallium, or
zinc phthalocyanine macroring and the kind of aggregates [19]. The absorption spectra
and the in situ study of the Langmuir monolayer were also done for
copper(II) 2,9,16,23-tetra-tert-butyl-29H,31H-phthalocyanine, copper(II) 2,3,9,10,
16,17,23,24-octakis(octyloxy)-29H31H-phthalocyanine, zinc(II))2,9,16,23-tetra-
tert-butyl-29H,31H-phthalocyanine, and zinc(II) 2,3,9,10,16,17,23,24-octakis(octy-
loxy)-29H31H-phthalocyanine with QD. The in situ absorption spectra in the Q region
are characterized by clear two maxima—the result indicates the exciton energy splitting
and the presence of aggregates. The models composed of QD and chromophores in
solution as well as in the 2D Langmuir monolayers can be an example which gives
description of the interaction between an organic chromophore and a semiconductor
dot.
The correlation of the in situ spectra and electronic absorption spectra confirms
(or excludes) formation of molecular aggregates. When one compares the electronic
spectrum in chloroform with the in situ curves, it is evident that the existence of dye
aggregates in the monolayer can be observed. Investigations of thin layers are
extremely useful for determining such values as the compressibility and stability of
the layers that are important for optoelectronics. Besides, it is interesting that
quantum dots in Langmuir’s two-dimensional layers can significantly reduce the
formation of molecular aggregates [39, 40]. Moreover, it is also worth emphasizing
that orientation of the molecules is one of the most important factors that affect the
efficiency of the energy transfer process. For this reason, investigations of the dye
layers are essential to get knowledge on arrangement of the molecular skeleton
which influences electron transfer [39].
The photoinduced single-step electron transfer in the systems based on organics
and QD is of a great interest and is presented in many papers [41–44]. As said
above, the systems based on metallic phthalocyanine with QD is one of interesting
examples of electro-donor-acceptor mixture. The spectroscopic experiments of zinc
and copper phthalocyanines substituted with tetrabutyl side groups (ZnPc-tetr,
Cu-tetr) and octyloxy groups (ZnPc-octy, CuPc-octy), their mixtures with CdSe/
ZnS quantum dots in chloroform are a proof of creation of the efficient
donor-acceptor systems. Interaction between the phthalocyanine dye and CdSe/ZnS
quantum dot observed as dot fluorescence quenching is shown in Fig. 3.3.
The spectra show fluorescence of QD (max 585 nm) and its quenching with the
rising concentration of Pc resulting in energy transfer from QD to Pc. The influence
of the substituent attached to the phthalocyanine macroring and of the presence of
the quantum dot fluorescence is evident—the value of energy transfer yield (UET)
can reach up to 70–90% [20]. The Pc-QD system fulfills one of the important
conditions of the non-radiative resonance energy transfer (Fӧrster model—FRED)
between isolated D and A—overlapping of the donor emission with absorption
spectra of acceptor [20]. Confirmation of the photoinduced single-step electron
transfer on the nanometer scale in systems like strongly coupled zinc
Fig. 3.3 Quenching of QD

fluorescence in the presence
of zinc phthalocyanine
porphyrin-gold dyads was also studied with the use of advanced methods, and it
showed dynamics of the processes [41–45].
3.5 Corroles and Fullerene
The great hopes are associated with corroles—an analogue of porphyrins. They are
very intriguing materials of specific physicochemical properties. Corroles are also
aromatic tetrapyrroles planar p-electron systems, and thus they are characterized by
intense absorption in the visible region. However, their molecular structure differs
significantly compared to porphyrins and phthalocyanines [46]. They are charac-
terized by three carbon meso-positions and one direct bond between the pyrrole
rings [47–51]. The direct linkage between the pyrrole rings leads to reduction of the
symmetry from D4h (as in porphyrins) to C2v. Corroles possess four nitrogen atoms
and three –NH groups in their macroring. Moreover, corroles are a typical example
of the porphyrin-derivative dyes that are able to stabilize metals in higher oxidation
states compared to porphyrins which have two ionizable NHs. The significant
advantage of corroles over porphyrins is stronger absorption in the 700–550 nm
region, higher fluorescence quantum yield, lower first oxidation potentials, and
entirely different coordination chemistry [52, 53]. Due to their specific properties
like lower symmetry (C2v) (one meso-carbon bridge is missed and in consequence it
can result in strong changes of spectroscopic properties), the ring of corroles is
smaller, and there is a slightly less space in the middle of it which results in sitting
of many metals slightly out of plane and they are very reactive and sensitive to
oxidation and more difficult to be reduced. The corroles exist in a form of tautomers
—there are two forms of tautomers—Tautomer 1 (T1) and Tautomer 2 (T2) with
different location of proton H. In this chapter, some results concerning selected
spectroscopic investigations of corroles fullerene covalent dyads are also presented.
Two corroles 10-(4-nitrophenyl)-5,15-bis(pentafluorophenyl)corrole and

5,10,15-Tris(pentafluorophenyl)corrole are taken into consideration. The nitro-
phenyl corrole belongs to the A2B type whereas the pentafluoride to the type A3.
The results performed with the quantum chemical calculations indicate their
absolutely different behavior. The nitrophenyl corrole is tautomer T1 and penta-
fluoro corrole—tautomer T2. However, the DFT calculations for both tautomers of
nitrophenyl corrole show that the energy difference between them is very low
(<0.001 eV) in comparison to kT (0.026 eV at 300 K). It means that both tautomers
exist at room temperature while at low temperature the tautomer with lower total
energy (T2) is preferred. Since it is rather difficult to recognize Tautomer T1 from
T2, the Gouterman four-orbital model lets to assign the electronic energies of T1
and T2 and electronic transitions S1–S0, S2–S0 [54].
Because of such specific photophysicochemical properties, corroles are very
different from those of porphyrins and phthalocyanines and not yet fully under-
standable. Apart from experimental studies whose results showed unexpected
properties of the corroles (UV-VIS absorption, also with in situ absorption tech-
nique, FT-IR spectroscopy, fluorescence emission, fluorescence kinetics), also
computer chemical calculations are very often used as a helpful approach to the
issue of determining the HOMO and LUMO states as well reorganization energy
[55–57]. Moreover, the basic properties allow recognizing the processes of tau-
tomerization, deprotonation, and induced mesomeric effects occurring in corroles
and their influence on singlet and triplet excited states [58–62].
The more and more advanced chemical procedure has opened a gate to receive
many types of new corroles [63, 64]. The significant advantage of corroles is their
quite simple synthesis. Recent developments of chemistry made it possible—
e.g., preparation of trans-A2B-corroles [63] and also self-organization of
meso-substituted corroles possessing secondary amide groups methods were
developed [65–67]. Also creation of corrole aggregates is of a great interest [63, 64]
for their applications in the future optoelectronics and in many other fields of science
and technology such as electronics, sensory, medicine, photovoltaic cell models as
well as in modeling of photophysical and photochemical processes occurring in
photosynthesis [68–70].
The infrared spectroscopy is a powerful tool in the investigation of molecular
systems. The key feature is the ability to identify even very small changes in the
molecular structure and/or interaction with the surroundings. Additionally, if we
support the infrared experimental techniques with density functional theory (DFT)
methods using Gaussian [71] or similar software, the assignment of the observed
bands to appropriate normal modes becomes more reliable. Below we will
demonstrate the use of this method for investigations of two free-base corroles as an
example. Despite the significant and still growing number of investigations focused
on the corroles, there is still little known about their basic vibrational properties and
characteristic infrared or Raman bands. The works of Ghosh [72, 73] and
Czernuszewicz [74] are focused on metallocorroles and discussing mainly the
resonance Raman spectra. There are also the DFT studies of iron and germanium
corroles [75] and transition metal corrolazine complexes (corrolazine is a corrole
derivative with the meso-carbons replaced by nitrogens) [76]. Nevertheless, the

vibrational properties of the free-base corroles have not been described yet, except
for a few complex corroles [77, 78]. The free-base corroles synthesized and pre-
sented in [62] are quite similar to each other—the difference lies in one of the
substituents while the main core of the molecules remain unchanged. This allows
predicting a lot of similarities in the vibrational spectra of the samples with some
differences related to the changed group. The previous vibrational study of the
corrole-fullerene dyad showed that the B3LYP functional with a 6-31G basis set is
effective in the interpretation of experimentally observed vibrational bands [77]. For
that reason, we decided to use the same method for calculations of the set of
free-base corroles. The corrole macroring of the DFT optimized free-base corroles
is not planar, especially the NH containing part. Beenken et al. [60] performed DFT
calculations on meso-pyrimidinylcorrole and also obtained the non-flat geometry of
the corrole macroring. Moreover, this is in agreement with the experimental X-ray
results obtained by Gross et al. for TpFPC-H3 corrole [79]. The surface of the C6F5
substituents of the samples 1–6 (respectively named as:
• 1–10-(4,7-Dimethoxynaphthalen-1-yl)-5,15-bis(pentafluorophenyl)-corrole,
• 2–10-(2,7-Dimethoxynaphthalen-1-yl)-5,15-bis(pentafluorophenyl)-corrole,
• 3–10-(8-Methoxyquinolin-4-yl)-5,15-bis(pentafluorophenyl)-corrole,
• 4–10-(7,8-Dimethoxycoumarin-4-yl)-5,15-bis(pentafluorophenyl)-corrole,
• 5–5,10,15-Tris(pentafluorophenyl)-corrole,
• 6–10-(4-nitrophenyl)-5,15-bis(pentafluorophenyl)-corrole [62])
is inclined with respect to the macroring surface; however, the angle between
them is about 45°. The third substituent in 5 and 6 has similar inclination angle. The
remaining molecules (1–4) have the double ring structure as the third substituent
which is oriented more or less perpendicularly to the macroring surface. Beenken
et al. [60] suggested that massive groups attached to the aryl substituents are
responsible for perpendicular orientation between the substituent and the macroring
surface. Our results seem to confirm that suggestion. In Fig. 3.4, the exemplary
infrared absorption spectra of two investigated corroles recorded in KBr matrix are
presented. For comparison, an appropriate calculated spectrum of each molecule is
added in every panel. The relatively good agreement between the calculated data
and experimental spectra is clearly visible. A small shift of the calculated bands
toward higher wave numbers is typical and originates from the approximations used
for calculations (harmonic vibrations, isolated molecule). The most characteristic
feature is the double band observed in the spectra of both samples at about
1500 cm–1 (1497 and 1522 cm–1).
This band is related to the symmetrical (lower wave number) and asymmetrical
(higher wave number) stretching vibrations of the C–F bonds and stretching of the
C–C bonds in the corrole macroring. In the calculated spectra, this band appears as
a single band; however, it contains two normal modes mentioned above. The
similar band has been observed in the spectra of previously investigated corrole at
1521 cm–1 [77]. Wasbotten et al. also observed those two bands in the spectra of
Fig. 3.4 Infrared absorption spectra of two investigated corroles recorded in KBr matrix (red/
lower spectra) together with calculated ones (black/upper spectra). A3 corrole (C6F5 substituents)
(a) and trans-A2B-corrole (NO2C6H4 substituent) (b). Theory level: B3LYP/6-31G. Spectra are
vertically scaled and offset for clarity reasons
some copper(III) corroles, usually at slightly lower wave number [72]. The extra
sample-independent band is also observed at about 929 cm–1. It was also observed
earlier [77], and it is assigned to the in-plane deformation of the corrole macroring.
Moreover, the band at about 986 cm−1 related mainly to the stretching vibration of
the CF bonds in one of the C6F5 rings is also observed in all samples at almost the
same wave number. The group of bands at about 754, 770, and 794 cm−1 is related
to some in-plane bending of the C–C bonds in the corrole macroring and mostly to
the out-of-plane CH bending in the macroring. Since the common part of the
molecules is involved in those vibrations, the resulting bands also appear in the
spectra of both corroles. The differences related to the presence of the changed
substituent are easy to be noticed besides many similarities in the IR spectra
mentioned above. The NO2 group in trans-A2B-corrole is responsible for the
presence of the band at 1345 cm−1 related to the stretching of the C–NO2 bond. It is
quite interesting that this band appears in the calculated spectrum at lower wave
number than that in the experiment (1287 cm−1). Most probably, the reason of this
effect is that the interaction of the NO2 group with environment is stronger than in
the case of other terminal groups. The spectrum of A3-corrole (Fig. 3.4) is domi-
nated by the bands common for both samples and mentioned earlier. One can say
that the absence of the other infrared bands is characteristic feature of that sample. It
is somehow natural if one takes into account that this sample has three identical
substituents (C6F5 rings). The DFT calculations also reveal much simpler spectrum
than for the second sample.
Fullerene is a unique molecule with the specific structure and very interesting
properties. Fullerenes are molecules of carbons in the form of a sphere, ellip-
soid, tube, or many other shapes. The most popular fullerene is a C60 molecule with
a spherical structure—they are molecules composed of an even number of carbon
atoms with a spherical or ellipsoidal closed spatial structure. Their specific
molecular structure (3D system of conjugated p-electrons) causes C60 to attach
electron-donor chromophores and form supermolecular complexes. They are per-

fect electron acceptors, chemically and temperature very stable. Thus, fullerenes are
materials with special properties, they are currently on the top of the interest of
many researchers due to their specific spectroscopic properties, and it is material of
very low reorganization energy. The more detailed description of fullerene prop-
erties and applications is widely described in [80].
3.6 Energy and Electron Transfer
Extensive investigations of photoprocesses and their mechanisms by determination

of the ability of corrole-fullerene dyads to fast charge separation, slow charges
recombination, excitation energy, and electron transfer are being carried out. Thus,
the ability of the system to photocurrent generation after light absorption will be
better recognized. It is also important to get some knowledge what requirements
have to be fulfilled by proposed donor-acceptor units as effective agents in light
energy conversion into electric energy. Up to date, the role of corroles, fullerenes,
corroles-fullerene dyads in organic photovoltaics is not fully known yet. Thus, in
this chapter we have focused our attention on a family of meso-substituted corrole
dyes and their dyads with fullerene as an electron-donor-acceptor model.
One of the most important keys in tailoring new optoelectronic devices is the
knowledge of energy and electron transfer processes between electron donor
(D) and electron acceptor (A) as well as their mechanisms. It is very important and
useful in developing and proper designing of photovoltaic devices based on organic
materials. Participation of organic dyes and their dyads in energy and electron
transfer is the goal of many experimental and theoretical investigations [81–88].
Phenomena and processes essential for high effectiveness of optoelectronic devices
based on molecular systems include: strong p–p light absorption, efficient energy
transfer between D and A, fast electron transfer in a D–A chain followed by charge
separation and slow charge recombination [81–88].
The model of the excitation energy transfer process was proposed for the first
time by Förster [89, 90]—Förster long-range resonance excitation energy transfer
(FRET). The model is useful in studying interaction in a D–A unit distanced in the
nanoscale range. The energy transfer process described by the Förster model takes
place when the distances between D and A ranges from about 10 Å to about 100 Å
and requires overlapping of the D and A wave functions. The Förster mechanism is
approximated by the dipole–dipole interaction. The rate of energy transfer (kEnT)
depends very strongly on donor lifetime (s0) and the A–D distance R, and it is
expressed as follows:
6
1 R0
kEnT ¼ ð3:1Þ
s0 R
where
Z
Cj2 F ðmÞeðmÞ
R0 ¼ dm ð3:2Þ
n4 s 0 m4
R0 is the Förster distance (named Förster radius) of the D–A system which
energy transfer efficiency is 50%, j is the dipole factor that describes the mutual
orientation of D emission dipole moment and the A absorption dipole moment, F(m)
and e(m) are fluorescence and absorption over the m light frequency, respectively.
In the case of strong coupling interaction (the separation in the D–A pair is much
smaller; R 5 Å), the energy transfer process can be described with the use of the
exchange mechanism proposed by Dexter [91]. The strong coupling interaction also
requires overlapping of D and A wave functions. The excitation is delocalized over
D and A, and oscillates back and forth between them. The fluorescence lifetime of
D is of the order 10−8–10−9 s, and the energy transfer rate constants can range from
109 to 1014 s−1 depending strongly on a type of system and mechanism responsible
for the process.
In this chapter, selected results comprising corroles and fullerene C60 are pre-
sented to follow and characterize processes occurring in the corrole-C60 systems.
The presented results concern the experimental data with the use of modern com-
plementary experimental methods supported by quantum-mechanical calculations
[78, 92–95]. Exemplary absorption and fluorescence spectra of corrole and
corrole-fullerene dyad in solution are shown in Fig. 3.5. The absorption band at
about 260–270 nm originates from fullerene and the bands in the long-wave region
can be assigned to the corrole.
In the corrole-fullerene dyad the fluorescence intensity is dropping down com-
paring with fluorescence of the corrole itself. The quenching of corrole fluorescence
in the presence of C60 is caused by interaction between the corrole and C60 that is
strongly changed after excitation and leads to efficient energy transfer from the
chromophore to the fullerene. The evaluated values of the fluorescence quantum
yield of the corrole versus the corrole-fullerene unit decrease from 10% to about
3.5% and fluorescence lifetime from 3.80 to about 1.00 ns (90% weight), and they
are confirmed by fluorescence kinetics. The yield of the electron/electron transfer
was evaluated as about 60–70% [92].
The electron transfer (ET) theory was originally proposed by Marcus in 1956
[96, 97]. The simple model of ET can be presented as D + A ! D+ + A−. The
Marcus model explains the outer ET from D to A in two individual and not directly
linked species. The theory was then extended to the inner ET model [97], in which
two chemical species are linked chemically. In the Marcus theory, the rate constant
for electron transfer is expressed as:
!
2p 1 ðk þ DGÞ2
kET ¼ jHAB j2 pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi exp ð3:3Þ
h 4pkKb T 4kKb T
Fig. 3.5 Exemplary

absorption (a) and
fluorescence (b) spectra of
corrole (2) and
corrole-fullerene dyad (3) in
chloroform together with
room temperature
fluorescence decays (c).
Reprinted with permission
from [92], copyright (2013)
Elsevier
where HAB is the electronic coupling between the initial and final states, k is the
reorganization energy, DG is the total Gibbs free energy change for the transfer
reaction, Kb is the Boltzmann constant, and T is absolute temperature. The energy
reorganization k is a sum of internal energy ki (related to changes in molecular
geometry at the transition from the neutral molecule to its ion) and energy of outer
reorganization (related to the interaction with the solvent molecules). The latter part
is usually significantly smaller than the former and can be neglected. The transfer of
the charge in molecular systems can be described as the hopping of the electron (or
hole) between the adjacent molecules. It means that the electron (hole) is transferred
from the anionic (cationic) form of the molecule A to the neutral as follows:
Að þ Þ A0 ! A0 Að þ Þ ð3:4Þ
Considering such process, one must remember that conformation and energy of
the ionized molecule are different from the neutral species values. A molecule
which receives the charge (in a form of electron or a hole) is at the conformation
(and energy) of the neutral molecule and must relax to the geometry and energy of
the ionic species. Also the ionized molecule which loses its charge is not at its
equilibrium state and must relax toward the neutral conformation. This process
repeats resulting in the transfer of charge, but the efficiency of it depends on the
type of a molecule and a type of charge carriers. The mentioned differences in
conformation and energy of the neutral and ionic forms of the molecule are the
source of reorganization energy characteristic for every molecule. The electronic
coupling matrix element between a donor and an acceptor presented in the Eq. 3.3
(HAB) is expected to vary over the limited range of values [98], so we can assume
that difference for hole and electron transfer is relatively low. Thus, neglecting the
difference in the HAB values, the relative hopping rates are given by [98]:
12
ketðhÞ ke ke kh
e 4kT ð3:5Þ
ketðeÞ kh
The reorganization energies of the investigated molecules can be calculated with

DFT methods by the following procedure. First of all, the geometries of the
investigated molecules must be optimized and normal mode frequencies should be
calculated to be sure that we have reached an energetic minimum. These opti-
mizations must be of course performed for neutral and ionic species. The calcu-
lation of the normal mode frequencies is required because obtaining n imaginary
frequencies with such calculations inform us that molecule is not at the minimum
but at the nth-order saddle point. Next, the single-point energy calculations must be
performed for neutral molecule at neutral geometry, an ion at the neutral geometry,
an ion at the ion geometry, and neutral molecule at the ion geometry. Finally, the
reorganization energies are calculated for the hole (kh) and electron transfer (ke)
using following formulas:
kh ¼ E0þ E þ
þ
þ E0þ E00 ð3:6Þ
ke ¼ E0 E

þ E
0
E00 ð3:7Þ
0= þ =
where E0= þ = —energy of the neutral molecule/cation/anion at geometry of the
þ =
neutral molecule/cation/anion; E0 —energy of the cation/anion at geometry of
the neutral molecule; E0þ = —energy of the neutral molecule at geometry of the
cation/anion.
The values of calculated reorganization energies together with the relative
hopping rates of all investigated corroles are gathered in Table 3.1. The values for
C60 fullerene are also included. Calculations were done using B3LYP functional
and 6-31+G basis set. The Gaussian 09 [72] software package was used for all
above calculations. The hybrid functional B3LYP (3-parameter Becke exchange
functional combined with Lee–Yang–Parr correlation functional) is frequently used
because of its low cost (in a sense of calculation time) and quite good accuracy of
the results in comparison with the experiment as it was shown earlier. The basis set
was also chosen as a compromise between the speed and accuracy of the calcu-
lations. The 6-31+G basis set is large enough to obtain reliable results and at the
same time is significantly less time consuming than larger and more accurate basis
sets like 6-311+G or augmented correlation consistent basis sets aug-cc-pVDZ).
The diffuse functions have been included in the basis set because calculated
molecules are mostly ionized and in such case basis sets augmented with diffuse
functions are more accurate.
The obtained reorganization energy values show rather small differences
between various corroles. The exception is the corrole 6 (nitrophenyl) for which
both the ke and the kh values are significantly lower than for the other molecules.
This result suggests that corrole 6 is the most effective donor—very low kh value
means that it is “easy” for the molecule to get the positive charge, i.e., to donate
electron. The ke is higher than kh for most of the investigated corroles showing that
they prefer to act as a donor rather than as an acceptor. The only exception is the
corrole 1 for which ke < kh, so this molecule rather prefers to act as an acceptor. Of
course for the C60 fullerene, one also observes the reverse dependence because the
fullerene usually prefers to accept the electrons. The calculated reorganization
energy values for fullerene are in agreement with the works of Tokunaga [99, 100].
Table 3.1 DFT calculated reorganization energies and relative hopping rates for the set of
free-base corroles and C60 fullerene
Molecule C60 1 2 3 4 5 6
kh (eV) 0.164 0.359 0.248 0.209 0.226 0.185 0.058
ke (eV) 0.129 0.286 0.288 0.283 0.306 0.275 0.169
ket(h)/kaet(e) 0.640 0.440 1.600 2.350 2.530 2.880 4.990
a
kT = 0.026 eV (at 300 K)
It is easy to notice that the differences between ke and kh for investigated corroles
(except 6) are not very distinct—this suggests that it should be also possible to force
them to desired (donor or acceptor) behavior. This is in agreement with the results
of Brizet et al. who showed that direction of the energy transfer in the gallium
corrole-BODIPY dyad can be reversed by putting the appropriate substituent to the
BODIPY moiety [101]. The estimated relative hopping rates of holes versus
electrons ket(h)/ket(e) more or less reflect the values of reorganization energies and
show that mobility of holes in all corroles (except 1) is higher at least twice than
that of electrons. Again, the most distinct value was obtained for corrole 6 (4.99).
The results for 1 and fullerene show rather acceptor behavior (ket(h)/ket(e)< 1).
In addition to the intermolecular charge transfer, one should also consider the
possibility of the intramolecular charge transfer. Such transfer can be identified by
the position of the frontier molecular orbitals (HOMO and LUMO) when the
electron moves from one part of the molecule to another. Moreover, the transfer
depends on the molecular surroundings and on the attached groups. Figure 3.6
shows a good example—in the trans-A2B-corrole one observes a partial transfer
from HOMO to LUMO (Fig. 3.6a). The transfer becomes more complete if we take
into account the presence of a solvent (Fig. 3.6b)—in this case LUMO is located
almost only on the electron withdrawing group. The use of the fullerene instead of a
simpler substituent increases this effect as it can be seen in Fig. 3.6c.
Research groups devote time and attention to gaining knowledge on interactions
in covalent dyads of corrole-fullerene units. In many papers, it was plainly shown
that the corrole-fullerene dyads possess relatively long lifetime of the
charge-separated state. A very good proof are the results described in Hasabe et al.
[10], Gryko [48], Kadish [52], Imahori et al. [81, 82, 85], Ohkubo et al. [84],
Fig. 3.6 Frontier molecular orbitals of trans-A2B-corrole (a), trans-A2B-corrole in DMSO (b),
and corrole-fullerene dyad (c)
D’Souza et al. [65] (and references therein) papers. They present the long lifetime
of the charge-separated state in chlorine porphyrin and fullerene units and in fer-
rocene–porphyrin–fullerene systems. The papers [81, 82, 84] unquestionably
pointed out that the charge separation is viewed of electron transfer described in a
Marcus inverted region. It is closely connected with a small distance between
interacting units [84] Thus, super- and supramolecular covalent and non-covalent
systems consisting of organic chromophores and fullerene [102 and references
therein] are still under investigations. In some corrole-fullerene dyads investigated
by D’Souza et al., the reorganization energy values are estimated as 0.51–0.71 eV
depending on solvent polarity [65] and the charge separation rate ranges from 1010
to 1011 s−1. Also the charge transfer process and energy reorganization of por-
phyrin–fullerene complexes were investigated [103]; the reorganization energy
value in a porphyrin–fullerene dyad in benzene is about 0.23 eV.
3.7 Summarizing
Energy transfer in the systems containing corroles and fullerene has been presented
in several cited papers. However, there is still an open question concerning the
electron transfer and charge recombination processes. Thus, knowledge of basic
spectroscopic properties of corroles and fullerenes both in solutions and in a form of
ultrathin layers is needed to determine the conditions for receiving the best ultrathin
layers on solid substrates which could be applied in devices based on photoactive
corrole-fullerene.
In this chapter, we have concentrated on spectroscopic properties and interaction
of selected organic chromophores and their interaction with QD and fullerene C60.
Our attention has focused on three different electron/energy transfer models: from
the simplest models—dye–dye and porphyrin–quinone, through the metal
phthalocyanines with QD model, and more complex model—the corrole-fullerene
covalent dyad.
New organic chromophores like corroles and also fullerenes are an emerging
field in solar cell research as the photoactive materials. Since carbon-derivative
units are new extremely promising electronic materials in many scientific and life
science areas, in this chapter, we indicated how much studies are still extending
knowledge on the interaction of phthalocyanines with QD as well corroles with
fullerene covalent dyads and to confirm their donor-acceptor character. The final
effect of such necessary research should be finding suitable materials that will be
able to improve the photoefficiency of photoactive devices.
The fluorescence study (emission spectra and kinetics) presented in many papers
showed strong interaction between chromophores and C60 or QD and clearly
demonstrated a strong donor-acceptor nature of the metal phthalocyanines-QD
system and of the corrole-fullerene C60 dyad. The computer calculation (DFT)
confirmed the experimental results, in particular the redistribution of the p-electrons
in the excited state and the location of the HOMO and LUMO levels and the
reorganization energy values. Thus, there is a lot of evidence that selected organic
chromophores (porphyrins, phthalocyanines, corroles) with QD or C60 can be
promising materials for the composition of photoactive systems in the future.
Acknowledgements The paper is supported by Poznan University of Technology, the grant DS

06/62/DSPB/2181. The author is much grateful to Prof. D. T. Gryko (Institute of Organic
Chemistry, Polish Academy of Sciences, Warsaw, Poland) for the gift of the corrole and
corrole-fullerene samples. We also thank M.Sc. Eng. Kamil Kędzieski for the help of drawings.
References
1. Wróbel D (2016) From natural photosynthesis to molecular photovoltaics. Mol Cryst Liq
Cryst 627:4–22
2. Gong X, Milic T, Xu C, Batteas JD, Drain CM (2002) Preparation and characterization of
porphyrin nanoparticles. J Am Chem Soc 124:14290–14291
3. Huang X, Nakanishi K, Berova N (2000) Porphyrins and metalloporphyrins: versatile
circular dichroic reporter groups for structural studies. Chirality 12:237–255
4. Braun A, Tcherniac J (1907) Über die Produkte der Einwirkung von Acetanhydrid auf
Phthalamid. Ber Dtsch Chem Ges 40:2709–2714
5. Dolphin D (1978) The porphyrins, vol III. Academic Press, Cambridge
6. Grätzel M (2005) Solar energy conversion by dye-sensitized photovoltaic cells. Inorg Chem
44:6841–6851
7. Morandeira A, López-Duarte I, Martínez-Díaz MV, O’Regan B, Shuttle C, Haji-
Zainulabidin NA, Torres T, Palomares E, Durrant RJ (2007) Slow electron injection on
Ru–phthalocyanine sensitized TiO2. J Am Chem Soc 129:9250–9251
8. Luo L, Lin CJ, Tsai CY, Wu HP, Li LL, Lo CF, Lin CY, Diau EW (2010) Effects of
aggregation and electron injection on photovoltaic performance of porphyrin-based solar
cells with oligo(phenylethynyl) links inside TiO2 and Al2O3 nanotube arrays. Phys Chem
Chem Phys 12:1064–1071
9. Heimer TA, Heilweil EJ (1997) Direct time-resolved infrared measurement of electron
injection in dye-sensitized titanium dioxide films. J Phys Chem 101:10990–10993
10. Hasobe T, Imahori H, Kamat PY, Ahn TK, Kim SK, Kim D, Fujimoto A, Hirakawa T,
Fukuzumi S (2005) Photovoltaic cells using composite nanoclusters of porphyrins and
fullerenes with gold nanoparticles. J Am Chem Soc 127:1216–1228
11. Schmidt-Mende L, Campbell WM, Wang Q, Jolley KW, Officer DL, Nazeeruddin KM,
Grätzel M (2005) Zn-porphyrin-sensitized nanocrystalline TiO2 heterojunction photovoltaic
cells. Chem Phys 6:1253–1258
12. Lee MW, Lee DL, Yen WN, Yeh CY (2009) Synthesis, optical and photovoltaic properties
of porphyrin dyes. J Macromol Sci Part A 46:730–737
13. Smertenkov PS, Kostylev VP, Kislyuk VV, Syngaevsky AF, Zynio SA, Dimitriev OP
(2008) Photovoltaic cells based on cadmium sulphide–phthalocyanine heterojunction. Sol
Energy Mater Sol Cells 92:976–979
14. Wróbel D, Goc J, Ion RM (1998) Photovoltaic and spectral properties of tetraphenylopor-
phyrin and metallotetraphenyloporphyrin dyes. J Mol Struct 450:239–246
15. Wróbel D, Siejak A, Siejak P (2010) Photovoltaic and spectroscopic studies of selected
halogenated porphyrins for their application in organic solar cells. Sol Energy Mater Sol
Cells 94:492–500
16. Siejak A, Wróbel D, Ion RM (2006) Study of resonance effects in copper phthalocyanines.
J Photochem Photobiol A Chem 181:180–187
17. Siejak A, Wróbel D, Olejarz B, Ion RM (2009) Spectroscopic and photoelectric

investigations of resonance effects in selected sulfonated phthalocyanines. Dyes Pigm
83:281–290
18. Karimi AR, Khodadadi A (2012) Synthesis and solution properties of new metal-free and
metallo-phthalocyanines containing four bis(indol-3-yl)methane groups. Tetrahedron Lett
53:5223–5226
19. Bursa B, Wróbel D, Biadasz A, Kędzierski K, Lewandowska K, Graja A, Szybowicz M,
Durmuş M (2014) Indium-chlorine and gallium-chlorine tetrasubstituted phthalocyanines in
a bulk system, Langmuir monolayers and Langmuir-Blodgett nanolayers—spectroscopic
investigations. Spectrochim Acta A 128:489–496
20. Bursa B, Biadasz A, Kędzierski K, Wróbel D (2014) Quantum dot with zinc and copper
substituted phthalocyanines. 1. Energy transfer in solution and in-situ light absorption in
Langmuir monolayers. J Lumin 145:779–786
21. Kędzierski K, Barszcz B, Kotkowiak M, Bursa B, Goc J, Dinçer H, Wróbel D (2016)
Photophysics of an unsymmetrical Zn(II) phthalocyanine substituted with terminal alkynyl
group. J Lumin 180:132–139
22. Meyer T, Ogermann D, Pankrath A, Kleinermanns K, Müller TJ (2012) Phenothiazinyl
rhodanylidene merocyanines for dye-sensitized solar cells. J Org Chem 77:3704–3715
23. Wróbel D, Boguta A, Ion RM (2001) Mixtures of synthetic organic dyes in a
photoelectrochemical cell. J Photochem Photobiol 138:7–22
24. Chamberlain GA, Cooney PJ, Dennison S (1981) Photovoltaic properties of merocyanine
solid-state photocells. Nature 289:45–47
25. Steinmann V, Kronenberg NM, Lenze MR, Graf SM, Hertel D, Meerholz K, Bürckstümmer
H, Tulyakova EV, Würthner F (2011) Simple, highly efficient vacuum-processed bulk
heterojunction solar cells based on merocyanine dyes. Adv Energy Mater 1(5):888–893
26. Abdou EM, Hafez HS, Bakir E, Abdel-Mottaleb MS (2013) Photostability of low cost
dye-sensitized solar cells based on natural and synthetic dyes. Spectrochim Acta A 115:
202–207
27. Arjona-Esteban A, Lenze MR, Meerholz K, Würthner F (2017) Donor-acceptor dyes for
organic photovoltaics. In: Leo K (ed) Elementary processes in organic photovoltaics.
Advances in polymer science 272. Springer, Berlin
28. Wróbel D, Łukasiewicz J, Manikowski H (2003) Fluorescence quenching and ESR
spectroscopy of metallic porphyrins in the presence of an electron acceptor. Dyes Pigm
58:7–18
29. Pace NA, Reid OG, Rumbles G (2018) Delocalization drives free charge generation in
conjugated polymer films. ACS Energy Lett 3:735–741
30. Li L, Kang S-W, Harden J, Sun Q, Zhou X, Dai L, Jakli A, Kumar S, Li Q (2008)
Nature-inspired light-harvesting liquid crystalline porphyrins for organic photovoltaics. Liq
Cryst 35:233–239
31. Imahori H, Hayashi S, Hayashi H, Oguro A, Eu S, Umeyama T, Matano Y (2009) Effects of
porphyrin substituents and adsorption conditions on photovoltaic properties of
porphyrin-sensitized TiO2 cells. J Phys Chem C 113:18406–18413
32. Idowu M, Chen J-Y, Nyokong T (2008) Photoinduced energy transfer between
water-soluble CdTe quantum dots and aluminium tetrasulfonated phthalocyanine. New J
Chem 32:290–296
33. Britton J, Antunes E, Nyokong T (2010) Fluorescence quenching and energy transfer in
conjugates of quantum dots with zinc and indium tetraamino phthalocyanines. J Photochem
Photobiol A 210:1–7
34. Li L, Zhao J-F, Won N, Jin H, Kim S, Chen J-Y (2012) Quantum dot-aluminum
phthalocyanine conjugates perform photodynamic reactions to kill cancer cells via
fluorescence resonance energy transfer. Nanoscale Res Lett 7:386
35. Bursa B, Rytel K, Skrzypiec M, Prochaska K, Wróbel D (2018) Thin film of CdTeSe/ZnS
quantum dots on water subphase: thermodynamics and morphology studies. Dyes Pigm
155:36–41
36. Jun HK, Careem MA, Arof AK (2013) Quantum dot-sensitized solar cells—perspective and
recent developments: a review of Cd chalcogenide quantum dots as sensitizers. Renew Sust
Energ Rev 22:148–167
37. Jeltsch KF, Schädel M, Bonekamp J-B, Niyamakom P, Rauscher F, Lademann HWA,
Dumsch I, Allard S, Scherf U, Meerholz K (2012) Efficiency enhanced hybrid solar cells
using a blend of quantum dots and nanorods. Adv Funct Mater 22:397–404
38. Ma J, Chen J-Y, Idowu M, Nyokong T (2008) Generation of singlet oxygen via the
composites of water-soluble thiol-capped CdTe quantum dots-sulfonated aluminum
phthalocyanines. J Phys Chem B 112:4465–4469
39. Biadasz A, Bursa B, Barszcz B, Bogucki A, Laskowska B, Graja A, Wróbel D (2011)
Thermodynamics and in-situ absorption of Langmuir monolayers of selected copper
phthalocyanine substituted with different peripheral groups. Dyes Pigm 89:86–92
40. Martynenko IV, Orlova AO, Maslov VG, Fedorov AV, Berwick K, Baranov AV (2016) The
influence of phthalocyanine aggregation in complexes with CdSe/ZnS quantum dots on the
photophysical properties of the complexes. Beilstein J Nanotechnol 7:1018–1027
41. Fortage J, Boixel J, Blart E, Hammarström L, Becker HC, Odobel F (2008) Single-step
electron transfer on the nanometer scale: ultra-fast charge shift in strongly coupled zinc
porphyrin-gold porphyrin dyads. Chemistry 14:3467–3480
42. Leng H, Loy J, Amin V, Weiss EA, Pelton M (2016) Electron transfer from single
semiconductor nanocrystals to individual acceptor molecules. ACS Energy Lett 1:9–15
43. Claessens CG, Hahn U, Torres T (2008) Phthalocyanines: From outstanding electronic
properties to emerging applications. Chem Rec 8:75–97
44. Bae WK, Char K, Hur H, Lee S (2008) Single-step synthesis of quantum dots with chemical
composition gradients. Chem Mater 20:531–539
45. Toyoda T, Yindeesuk W, Kamiyama K, Katayama K, Kobayashi H, Hayase S, Shen Q
(2016) The electronic structure and photoinduced electron transfer rate of CdSe quantum
dots on single crystal rutile TiO2: dependence on the crystal orientation of the substrate.
J Phys Chem C 120:2047–2057
46. Aviv I, Gross Z (2007) Corrole-based applications. Chem Commun 20:1987–1999
47. Flamigni L, Gryko DT (2009) Photoactive corrole-based arrays. Chem Soc Rev 38:1635–
1646
48. Gryko DT (2008) Adventures in the synthesis of meso-substituted corroles. Porphyrins
Phthalocyanines 12:906
49. Harris RLN, Johnson AW, Kay IT (1966) The synthesis of porphins and related
macrocycles. Q Rev Chem Soc 20:211–244
50. Roberts JD, Streitwieser A, Regan CM (1952) Small-ring compounds. X. Molecular orbital
calculations of properties of some small-ring hydrocarbons and free radicals. J Am Chem
Soc 18:4579–4582
51. Ventura B, Esposti AD, Koszarna B, Gryko DT, Flamigni L (2005) Photophysical
characterization of free-base corroles, promising chromophores for light energy conversion
and singlet oxygen generation. New J Chem 29:1559–1566
52. Kadish KM, Shen J, Frémond L, Chen P, El Ojaimi M, Chkounda M, Gros CP, Barbe J-M,
Ohkubo K, Fukuzumi S, Guilard R (2008) Clarification of the oxidation state of cobalt
corroles in heterogeneous and homogeneous catalytic reduction of dioxygen. Inorg Chem 47
(15):6726–6737
53. Palmer JH (2011) Transition metal corrole coordination chemistry. In: Mingos D, Day P,
Dahl J (eds) Molecular electronic structures of transition metal complexes I. structure and
bonding, vol 142. Springer, Berlin, Heidelberg
54. Gouterman M, Wagnière GH, Snyder LC (1963) Spectra of porphyrins: Part II. Four orbital
model. J Mol Spectrosc 11:108–127
55. Lei H, Han A, Li F, Zhang M, Han Y, Du P, Lai W, Cao R (2014) Electrochemical,
spectroscopic and theoretical studies of a simple bifunctional cobalt corrole catalyst for
oxygen evolution and hydrogen production. Phys Chem Chem Phys 16:1883–1893
56. Kobayashi T, Mao K, Paluch P, Nowak-Król A, Sniechowska J, Nishiyama Y, Gryko DT,

Potrzebowski MJ, Pruski M (2013) Study of intermolecular interactions in the corrole matrix
by solid-state NMR under 100 kHz MAS and theoretical calculations. Angew Chem Int Ed
52:14108–14111
57. McNicholas BJ, Blumenfeld C, Kramer WW, Grubbs RH, Winkler JR, Gray HB (2017)
Electrochemistry in ionic liquids: case study of a manganese corrole. Rus J Electrochem
53:1189–1193
58. Ding T, Harvey JD, Ziegler CJ (2005) N-H tautomerization in triaryl corroles. J Porphyrins
Phthalocyanines 9:22–27
59. Konarev DV, Kariminov DR, Khasanov SS, Shestakov AF, Otsuka A, Yamochi H,
Kitagawa H, Lyubovskaya RN (2017) Solid state structures and properties of free-base
5,10,15-triphenylcorrole (TPCor) anions obtained by deprotonation and reduction. Effective
magnetic coupling of spins in (Cp*2Cr+)(H+)(H2TPCor˙2−) C6H4Cl2. Dalton Trans
46:13994
60. Beenken W, Presselt M, Ngo TH, Dehaen W, Maes W, Kruk M (2014) Molecular structures
and absorption spectra assignment of corrole NH tautomers. J Phys Chem A 118:862–871
61. Bursa B, Wróbel D, Barszcz B, Kotkowiak M, Vakuliuk O, Gryko DT, Kolanowski Ł,
Baraniak M, Lota G (2016) The impact of solvents on the singlet and triplet states of selected
fluorine corroles—absorption, fluorescence, and optoacoustic studies. Phys Chem Chem
Phys 18:7216–7228
62. Bursa B, Barszcz B, Bednarski W, Lewtak JP, Koszelewski D, Vakulyuk O, Gryko DT,
Wróbel D (2015) New meso-substituted corroles possessing pentafluorophenyl groups—
synthesis and spectroscopic characterization. Phys Chem Chem Phys 17:7411–7423
63. Kandala LVK, Kaur T, Ravikanth M (2017) One pot synthesis of unusual meso-dipyrrinyl
corrole. RSC Adv 7:19878–19884
64. Ooi S, Tanaka T, Park KH, Kim D, Osuka A (2016) Triply linked corrole dimers. Angew
Chem Int Ed 55:6535–6539
65. D’Souza F, Chitta R, Ohkubo K, Tasior M, Subbaiyan NK, Zandler ME, Rogacki MK,
Gryko DT, Fukuzumi S (2008) Corrole-fullerene dyads: formation of long-lived
charge-separated states in nonpolar solvents. J Am Chem Soc 130:14263–14272
66. Paolesse R, Nardis S, Sagone F, Khoury RG (2001) Synthesis and functionalization of
meso-aryl-substituted corroles. J Org Chem 66(2):550–556
67. Orłowski R, Gryko D, Gryko DT (2017) Synthesis of corroles and their heteroanalogs.
Chem Rev 117:3102–3137
68. D’Urso A, Nardis S, Pomarico G, Fragalà ME, Paolesse R, Purrello R (2013) Interaction of
tricationic corroles with single/double helix of homopolymeric nucleic acids and DNA. J Am
Chem Soc 135:8632–8638
69. Sheng X, Zhao H, Du L (2017) Selectivity of cobalt corrole for CO vs. O2 and N2 in indoor
pollution. Sci Rep 7:14536
70. Santos CIM (2014) Corroles: synthesis, functionalization and application as chemosensors.
Chem Open 3:88–92
71. Gaussian 09, Revision E.01 Frisch MJ, Trucks GW, Schlegel HB, Scuseria GE, Robb MA,
Cheeseman JR, Scalmani G, Barone V, Mennucci B, Petersson GA, Nakatsuji H,
Caricato M, Li X, Hratchian HP, Izmaylov AF, Bloino J, Zheng G, Sonnenberg JL, Hada M,
Ehara M, Toyota K, Fukuda R, Hasegawa J, Ishida M, Nakajima T, Honda Y, Kitao O,
Nakai H, Vreven T, Montgomery JAJr, Peralta JE, Ogliaro F, Bearpark M, Heyd JJ,
Brothers E, Kudin KN, Staroverov VN, Kobayashi R, Normand J, Raghavachari K,
Rendell A, Burant JC, Iyengar SS, Tomasi J, Cossi M, Rega N, Millam JM, Klene M,
Knox JE, Cross JB, Bakken V, Adamo C, Jaramillo J, Gomperts R, Stratmann RE,
Yazyev O, Austin AJ, Cammi R, Pomelli C, Ochterski JW, Martin RL, Morokuma K,
Zakrzewski VG, Voth GA, Salvador P, Dannenberg JJ, Dapprich S, Daniels AD, Farkas Ö,
Foresman JB, Ortiz JV, Cioslowski J, Fox DJ (2013) Gaussian, Inc., Wallingford CT
72. Wasbotten IH, Wondimagegn T, Ghosh A (2002) Electronic absorption, resonance raman,
and electrochemical studies of planar and saddled copper(iii) meso-triarylcorroles. highly
substituent-sensitive soret bands as a distinctive feature of high-valent transition metal

corroles. J Am Chem Soc 124:8104–8116
73. Steene E, Wondimagegn T, Ghosh A (2002) Resonance Raman spectroscopy and density
functional theoretical calculations of manganese corroles. A parallelism between high-valent
metallocorroles and metalloporphyrins, relevant to horseradish peroxidase and chloroper-
oxidase compound I and II intermediates. J Inorg Biochem 88:113–118
74. Czernuszewicz RS, Mody V, Zareba AA, Zaczek MB, Gałęziowski M, Sashuk V, Grela K,
Gryko DT (2007) Solvent-dependent resonance Raman spectra of high-valent oxomolyb-
denum(v) tris[3,5-bis(trifluoromethyl)phenyl]corrolate. Inorg Chem 46:5616–5624
75. Zakharieva O, Veeger C (2005) DFT normal coordinate analysis of the vibrational spectra of
iron and germanium corroles. J Mol Struct: THEOCHEM 723:171–182
76. Wang H, Yang C, Zhang Z, Wang M, Han K (2006) The molecular structure and vibrational
spectra of corrolazine metal complexes (CzM) by density functional theory. Spectrochim
Acta A 64:795–800
77. Lewandowska K, Barszcz B, Wolak J, Graja A, Grzybowski M, Gryko DT (2013)
Vibrational properties of new corrole–fullerene dyad and its components. Dyes Pigm
96:249–255
78. Bursa B, Wróbel D, Lewandowska K, Graja A, Grzybowski M, Gryko DT (2013) Spectral
studies of molecular orientation in corrole-fullerene thin films. Synth Met 176:18–25
79. Gross Z, Galili N, Simkhovich L, Saltsman I, Botoshansky M, Bläser D, Boese R,
Goldberg I (1999) Solvent-free condensation of pyrrole and pentafluorobenzaldehyde: a
novel synthetic pathway to corrole and oligopyrromethenes. Org Lett 1:599–602
80. Langa F, Nierenganter JF (eds) (2007) Fullerenes and applications. The Royal Society of
Chemistry (and references citated therein)
81. Imahori H, Sakata Y (1999) Fullerenes as novel acceptors in photosynthetic electron
transfer. Eur J Org Chem 1999:2445–2457
82. Imahori H, Mori Y, Matano Y (2003) Nanostructured artificial photosynthesis. J Photochem
Photobiol C 4:51–83
83. Łapiński A, Graja A, Olejniczak I, Bogucki A, Imahori H (2004) Supramolecular porphyrin/
fullerene interactions studied by spectral methods. Chem Phys 305:277–284
84. Ohkubo K, Kotani H, Shao J, Ou Z, Kadish KM, Li G, Pandey RK, Fujitsuka M, Ito O,
Imahori H, Fukuzumi S (2004) Production of an ultra-long-lived charge-separated state in a
zinc chlorine-C60 dyad by one-step photoinduced electron transfer. Angew Chem Int Ed
43:853–856
85. Imahori H, Guldi DM, Tamaki K, Yoshida Y, Luo C, Sakata Y, Fukuzumi S (2001) Charge
separation in a novel artificial photosynthetic reaction center lives 380 ms. J Am Chem Soc
123:6617–6628
86. Guldi DM (2002) Fullerene-porphyrin architectures; photosynthetic antenna and reaction
center models. Chem Soc Rev 31:22–36
87. Imahori H, El-Khouly ME, Fujitsuka M, Ito O, Sakata Y, Fukuzumi S (2001) Solvent
dependence of charge separation and charge recombination rates in porphyrin-fullerene
dyad. J Phys Chem A 105:325–332
88. Mizuseki H, Igarashi N, Belosludov RV, Farajian AA, Kawazoe Y (2003) Theoretical study
of phthalocyanine–fullerene complex for a high efficiency photovoltaic device using ab initio
electronic structure calculation. Synth Met 138:281–283
89. Förster Th (1949) Experimentelle und theoretische untersuchung des zwischenmolekularen
übergangs von elektronenanregungsenergie. Z Naturforsch A 4(5):321–327
90. Förster T et al (1965) Delocalized excitation and excitation transfer. In: Sinanoghi O
(ed) Modern quantum chemistry. Academic Press, New York, p 93
91. Dexter DL (1953) A theory of sensitized luminescence in solids. J Chem Phys 21:836
92. Lewandowska K, Barszcz B, Graja A, Bursa B, Biadasz A, Wróbel D, Bednarski W,
Waplak S, Grzybowski M, Gryko DT (2013) Absorption and emission properties of the
corrole-fullerene dyad. Synth Met 166:70–76
93. Lewandowska K, Bednarski W, Milczarek G, Waplak S, Graja A, Park EY, Kim T-D, Lee
K-S (2011) Photoelectrochemical cells based on LB films of fullerene-thiophene derived
dyads. Synth Met 161:1640–1645
94. Graja A (2012) Corrole-fullerene dyads: Will they place porphyrin-fullerene systems? Mol
Cryst Liq Cryst 554:31–42
95. Wróbel D, Lewandowska K (2011) Covalent dyads of porphyrin–fullerene and perylene–
fullerene for organic photovoltaics: spectroscopic and photocurrent studies. Opt Mater
33:1424–1428
96. Marcus RA (1956) On the theory of oxidation-reduction reactions involving electron
transfer. I. J Chem Phys 24:966
97. Marcus RA, Sutin N (1985) Electron transfers in chemistry and biology. Biochim Biophys
Acta 811:265–322
98. Lin BC, Cheng CP, Lao ZPM (2003) Reorganization energies in the transports of holes and
electrons in organic amines in organic electroluminescence studied by density functional
theory. J Phys Chem A 107:5241–5251
99. Tokunaga K (2009) On the difference in electronic properties between fullerene C60 and
C60X2. Chem Phys Lett 476:253–257
100. Tokunaga K (2012) Hydrogenation of fullerene C60: material design of organic semicon-
ductors by computation. In: Karamé I (ed) Hydrogenation InTech. https://doi.org/10.5772/
48534
101. Brizet B, Desbois N, Bonnot A, Langlois A, Dubois A, Barbe J-M, Gros CP, Goze C,
Denat F, Harvey PD (2014) Slow and fast singlet energy transfers in BODIPY-gallium
(iii)corrole dyads linked by flexible chains. Inorg Chem 53:3392–3403
102. Wróbel D, Graja A (2011) Photoinduced electron transfer processes in fullerene–organic
chromophore systems. Coord Chem Rev 255:2555–2577
103. Imahori H, Tkachenko NV, Vehmanen V, Tamaki K, Lemmetyien H, Sakata Y, Fukuzumi S
(2001) An extremely small reorganization energy of electron transfer in porphyrin-fullerene
dyad. J Phys Chem 105:1750–1756
Chapter 4
Material Analysis Using Raman
Spectroscopy
A Comparative Study of Graphite,
Single- and Multi-walled Carbon Nanotubes
Animesh K. Ojha and H. Michael Heise
Abstract A comparative Raman spectroscopic study of single-walled carbon

nanotubes (SWCNT), special multi-walled carbon nanotube (MWCNT) material,
and graphite is presented. Their Raman spectra have been recorded using different
excitation wavelengths, 532, 785, and 1064 nm, in the region of 1800–1200 cm−1.
The G-bands of SWCNTs observed at all excitation wavelengths were fitted taking
bands into account with Breit-Wigner-Fano (BWF) and Lorentzian line shape
functions. The contribution of both BWF and Lorentzian line shapes to the
asymmetric G-band shows a mixture of semiconducting as well as metallic CNTs in
the SWCNTs. For graphite and MWCNTs, only Lorentzian line shape functions
were used for band deconvolution. The variation in wavenumber position of
component bands of G-band with laser lines may be due to the resonance Raman
effect, where the energy of laser lines matches with the electronic transition energy
of CNTs with different diameters and chirality. The apparent Young’s modulus of
SWCNT and MWCNT materials was determined using the integrated band inten-
sity ratio of G- and D-bands, ID/IG, and it was found that the SWCNTs have a larger
value of the apparent Young’s modulus compared to that of the highly aligned
MWCNT material.
A. K. Ojha
Department of Physics, Motilal Nehru National Institute of Technology,
Allahabad 211004, India
H. M. Heise (&)
Interdisciplinary Center for Life Sciences, South-Westphalia
University of Applied Sciences, Frauenstuhlweg 31, 58644 Iserlohn, Germany
e-mail: heise.h@fh-swf.de

124 A. K. Ojha and H. M. Heise
4.1 Introduction
Carbon is one of the most abundant elements in nature and has long been known to
exist in three forms: amorphous carbon, graphite, and diamond. However, after the
discovery of buckminsterfullerene C60 in 1985 an entirely new branch of carbon
chemistry has been created opening up a new era in materials science and nan-
otechnology. The discovery of fullerenes [1] and carbon nanotubes (CNTs) [2] has
attracted considerable attention in recent years because of their unique properties
such as small dimensions, relative low density, high strength-to-weight ratio, and
high electrical and thermal conductivity. Since then, carbon nanomaterials have
been demonstrated to be useful, for example, as energy materials, in the develop-
ment of sensors [3, 4] or for drug delivery in medicine [5]. These exceptional
properties of CNTs have opened a wide range of promising applications like their
use in composite materials [6–9], where CNTs are expected to be one of the most
useful materials for improving the various properties of polymer matrix composites.
Another use of it has been suggested for hydrogen energy storage materials and
others [10–12]. Tremendous progress has been achieved in developing carbon
nanomaterials for high-performance energy conversion and storage systems [13].
Other aspects are concerned with their characterization by several analytical
methods when CNTs are found in environmental and biological samples; see recent
review [14]. Besides Raman, optical absorption spectroscopy within the ultraviolet,
visible and near-infrared ranges can be listed for their analysis. For morphological
studies, different microscopy techniques have been utilized like scanning and
transmission electron microscopy or atomic force microscopy. CNTs have seen
diverse commercial applications, where such materials have also been released into
the environment, posing a safety threat to those people who are exposed to these
nanomaterials. Therefore, monitoring and identification in particulate matter com-
prising CNTs become an important task, where toxicology aspects have also been
discussed [15]. On the other hand, therapeutic applications against cancer cells have
been suggested [16]. Other applications of CNTs can be found as sorbents for
solid-phase extraction and microextraction [17, 18] or as filters [19].
Nanotubes are one-dimensional wires having either metallic or semiconducting
nature, depending upon the orientation and chirality of the tube [20, 21]. A good
characterization of their structural and electronic properties is therefore of great
importance in material science. Raman spectroscopy has proven to be a nonde-
structive and nondissipative tool to characterize lower dimensional materials such
as nanostructures that have electronic and phonon properties different from the bulk
[22–24]. Raman spectroscopic measurements of CNTs provide a quick and simple
way to characterize them in terms of purity, diameter, chirality, conductivity, and
defects [25–29]. The first-order Raman spectra of multi-walled (MWCNT) and
single-walled CNTs (SWCNT) show a strong G-band corresponding to the E2g
symmetry and two defect dependent D- and D’-bands, appearing due to the double
resonance effect, which is associated with elastic phonon scattering close to the
4 Material Analysis Using Raman Spectroscopy 125
K and C points of the graphite Brillouin zone, respectively [30–33]. Also

second-order Raman spectra have been observed in MWCNTs and SWCNTs
resulting in a G’-band, which is basically the overtone of the D-band. Some other
overtone bands have also been observed in MWCNTs and SWCNTs. The G’-band
appears also in highly ordered graphite without defects. While elastic phonon
scattering results in the D-band, inelastic phonon scattering is responsible for the
G’-band. The metallic or semiconducting nature of CNTs is decided by the chirality
of the individual tube, and the chirality followed by the structure determines the
electron-phonon coupling [34, 35]. Therefore, the Raman spectra help to learn more
about the conductivity of CNTs by analysis of distinguishable G-band spectral
features. Also the diameter influences the Raman spectra of the CNTs. Some studies
[20, 21] have shown that the G-band line shape and the D-band wavenumber
positions depend on the diameter of the tubes.
In a more recent study [36], a relationship was proposed for the calculation of the
apparent Young’s Modulus, which is also called modulus of elasticity defining the
tensile elasticity mathematically. The estimate is based on the crystallinity of the
CNTs as described by the crystallinity parameter ID/IG. The latter can be derived
from the Raman line intensities of the D- and G-bands. The G-band is essentially
nondispersive, while the D-band wavenumber and intensity depend on the excita-
tion wavelengths. Heise et al. presented a series of studies [37, 38] on Raman
investigation of carbonaceous materials, i.e., SWCNTs, MWCNTs, graphitized
porous carbon, and graphite, using different excitation wavelengths. In those
studies, Lorentzian lineshapes were used only for band deconvolution. In their
recent study [38], they found that the ID/IG ratio can be used for the measurement of
degree of alignment for CNTs materials, which was well supported by SEM results.
The authors also performed a dispersion study on both D- and G’-bands of the
mentioned carbonaceous materials taking four different excitation wavelengths,
namely 488, 532, 785, and 1064 nm, into account. The dispersion of D- and G’-
bands evaluated from the wavenumber positions obtained through line shape
analysis of the experimental profiles turned out to be 52.7 ± 1.3 and
109.9 ± 4.0 cm−1/eV, respectively. However, the G-band position was only
slightly affected under the different excitation wavelengths. In CNTs, the
wavenumber features between 1500 and 1600 cm−1 are associated with the tan-
gential vibrational modes of the CNTs. The multiple nature of G-band is only seen
in CNTs, and it has not been observed in any sp2-bonded carbon materials [34]. The
unique optical properties of CNTs are observed due to the quantum confinement of
their electronic states in the two directions perpendicular to the nanotube axis.
When the energy of incident photons matches with the energy of an optical tran-
sition, the intensity of the Raman signal gets enhanced, known as resonance Raman
effect [39]. The CNTs with different diameter and chirality have different values of
resonance energy.
In view of the above discussion, we have performed a more extended compar-
ative resonance Raman study of CNT materials such as graphite, SWCNTs, and
MWCNTs with different excitation wavelengths. In the following, the results of our
analysis of the experimental Raman spectral data will be discussed. The spectral
line shape parameters were used to explore the physical properties, particularly, the
metallic or semiconducting nature of the CNTs in different CNT bundles. A detailed
Raman line shape analysis of the G-band has been performed using two functions,
Breit-Wigner-Fano (BWF) and Lorentzian band shapes. The parameters obtained
from these band deconvolutions were then used to calculate the Young’s Modulus
of the different CNTs.
4.2 Experimental
The MWCNT material with tube diameters ranging from 20 to 50 nm thickness

consisted of macroscale hollow carbon cylinders up to centimeters in diameter and
several centimeters long. For its synthesis, a continuous spray pyrolysis method was
used with ferrocene-derived iron particles acting as catalyst [19]. The well-aligned
nanotubes grew in radial directions on the walls of removable silica tube templates
leading to the formation of freestanding and continuous hollow cylindrical carbon
tubes. In this reaction, ferrocene breaks and gives nanoclusters of free iron particles,
which catalyze the growth of CNTs. Due to the high tube alignment, this material
has been suggested for nanopore filters; thus, in our previous publications [37, 38]
the term of “carbon nanotube filter” was used for this material. Furthermore, pure
graphite was also investigated.
For comparison, SWCNT materials were purchased from Aldrich (Milwaukee,
WI, U.S.A.; see also Ref. [37]). One of the charges was produced from CarboLex,
Inc. of Lexington (Kentucky, U.S.A.). The batch quantity (product# 519308) was of
AP-grade, consisting of bundles of single-walled CNTs with 10–200 individual
nanotubes per bundle with an average diameter of 1.2 to 1.5 nm and a bundle
length of approximately 20 µm and an individual tube length of 2–5 µm. The
purity of the AP-grade product ranged from 50 to 70% by volume. Major impurities
were carbon nanospheres and carbon-encapsulated catalyst nanoparticles. A second
single-walled CNT material (product # 636797), specified as SWCNT (Aldrich),
had diameters of 1–2 nm as measured by high-resolution transmission electron
microscopy (HRTEM) with the following purity statement: amorphous car-
bon: *3%, other nanotubes: *40%, single-walled CNT > 50% with an average
diameter of 1.1 nm and a length of 0.5–100 µm.
The dispersive Raman measurements on the CNT materials were carried out at
two of the three excitation wavelengths, 532, 785, and 1064 nm, using two different
Raman instruments, a XY-Dilor Raman spectrometer (HORIBA Jobin Yvon
GmbH) and a HoloSpec Raman Spectrometer from Kaiser Optical Systems,
respectively. For 532 nm wavelength excitation, the spectra were recorded using a
slit width of 100 µm with an integration time of 1000 s and 25 spectral frame
accumulations. An integration time of 25 s with 10 spectral frame accumulations
was used to record the Raman spectra at 785 nm wavelength with a laser power of
14 mW. The FT-NIR-Raman spectrometer model PE-2000 (PerkinElmer,
Überlingen, Germany) included a 2 W diode-pumped Nd:YAG laser (IE Optomech
Ltd, Newnham, England) for NIR excitation (1064 nm) and a GaAs detector. The
following experimental conditions were chosen: Spectral resolution of 16 cm−1
with 1000 interferogram scans coadded at 100 mW of laser power. Spectra of
reasonable signal-to-noise ratio could only be obtained for graphite and SWCNT
samples. For recording spectra of other materials, a rotating sample holder was
employed to reduce excessive sample heating from laser irradiation.
The nonlinear band fitting was done with either OPUS software from Bruker
Optics, Ettlingen (Germany) (Lorentzian bandshapes only) or ORIGIN version 6.0
from Microcal Software, Inc., Northampton, MA (including BWF functions; for
their definition, see below) using appropriate starting function estimates and the
Levenberg–Marquardt algorithm.
4.3 Results and Discussion
The Raman spectra of graphite, SWCNTs (from CarboLex and Aldrich), and
MWCNT materials between 1800 and 1200 cm−1, recorded using the three laser
excitation wavelengths of 532, 785, and 1064 nm, are presented in panels (a), (b),
and (c) of Fig. 4.1, respectively. The Raman spectra along with the fitted peaks are
presented in respective Figs. 4.2, 4.3, and 4.4. The spectral parameters obtained
from the deconvolution of the G-bands for the excitation wavelengths 532, 785, and
1064 nm are listed in Tables 4.1, 4.2, and 4.3, respectively.
The experimentally recorded G-bands of both types of SWCNTs (CarboLex)
and SWCNTs (Aldrich) have a noticeable asymmetry for all three excitation
wavelengths, and, therefore, the G-bands were fitted assuming the contribution of
three or four peaks. In case of the G-band of SWCNT (CarboLex) and SWCNTs
(Aldrich) recorded with 532 nm excitation wavelength, the best fit results were
obtained using two different line shape functions, BWF for the band component on
the low wavenumber side and Lorentzian functions for the other two bands.
The BWF line shape function has been defined as [40–42]:
½1 þ ðx xBWF Þ=qC2
I ðx Þ ¼ I 0 ð4:1Þ
1 þ ½ðx xBWF Þ=C2
The experimentally observed bands along with the fitted individual band com-
ponents are presented in Fig. 4.2a, b. However, for the G-band of MWCNTs, the
best fit results were obtained by fitting the band with two Lorentzian bands, and the
experimental as well as fitted spectra are presented in Fig. 4.2c. The low
wavenumber components of the G-bands in the Raman spectra of SWCNTs
(CarboLex) and SWCNTs (Aldrich) recorded using 532 nm excitation wavelength
can be found at approximately 1566 and 1572 cm−1, respectively. The value of the
fitting parameter, 1/q, was found to be around −0.12 for both types of SWCNT
G-bands. The wavenumber positions, symmetry, C (half of the full width at half
Fig. 4.1 Raman spectra of graphite, SWCNTs (CarboLex), SWCNTs (Aldrich), and MWCNT
materials in the range from 1800 to 1200 cm−1 recorded with 532 nm (a), 785 nm (b) and
1064 nm (c) excitation wavelengths
maximum (FWHM) for the Lorentzian bands) of the fitted band components of the
G-band of SWCNTs, MWCNTs, and graphite, recorded at 532 nm excitation
wavelength, are presented in Table 4.1. As mentioned above, the other two com-
ponents of the G-bands of the SWCNTs (CarboLex) and SWCNTs (Aldrich) were
fitted assuming a Lorentzian function, and the fitted peaks appeared at 1590 and
1606/1607 cm−1, respectively. The band at 1590 cm−1 belongs to the A(A1g) and
Fig. 4.2 Experimentally measured as well as fitted Raman spectra of the G-bands of SWCNTs
(CarboLex) (a), SWCNTs (Aldrich) (b), MWCNT (c) and graphite (d) recorded with 532 nm
excitation wavelength
Table 4.1 Results of a line shape analysis of the Raman G-bands observed in the Raman spectra
of SWCNT (CarboLex), SWCNT (Aldrich), MWCNT, and graphite measured using an excitation
wavelength of 532 nm. In the case of SWCNTs, the low wavenumber component is best fitted
assuming a BWF line shape function, while the other lines contributing to the G-band are fitted
using a Lorentzian line shape function with C the half full—width at half maximum
SWCNTs (CarboLex) SWCNTs (Aldrich) MWCNTs Graphite
ɷ (cm−1) C (cm−1) 1/q ɷ (cm−1) C (cm−1) 1/q ɷ (cm−1) C (cm−1) ɷ (cm−1) C (cm−1)
1566 (A1g) 8.2 −0.12 1572 16.9 −0.11 – – – –
1590 (A1g) 9.1 1590 11.5 1585 16.0 1587 9.4
1606 (E2g) 8.8 1609 9.4 1617 15.5 – –
the other bands at 1606/1607 cm−1 to the E(E2g) symmetry species of the semi-
conducting CNTs [30].
The G-band of graphite is centered at 1587 cm−1 for the 532 and 1064 nm
excitation wavelengths. It was fitted taking a Lorentzian function, and the experi-
mentally obtained as well as the fitted bands are shown in Figs. 4.2d and 4.4d,
respectively. The G-band of graphite recorded with 785 nm excitation wavelength
shows a small shoulder towards higher wavenumber (main band position at
1581.5 cm−1, satellite band at 1614.3 cm−1) and, therefore, was fitted with two
Lorentzian functions. Both experimental and fitted peaks are shown in Fig. 4.3a.
The calculated center of gravity for the doublet bands, based on their individual
band areas, was 1588 cm−1, very close to the band positions found under the other
excitation wavelengths. The G-band in graphite is due to the doubly degenerate
zone center E2g mode [43].
For the Raman G-band of the SWCNTs (CarboLex) and SWCNTs (Aldrich)
recorded with 785 nm excitation wavelength, the best fit results were obtained
again using two different line shape functions, BWF for the peak on the low
wavenumber side and Lorentzian functions for the residual components. In order to
get the best fit result, the G-band was decomposed by taking account of four peaks
assuming a BWF line shape for one and Lorentzian line shapes for the other three
band components. The experimentally observed bands along with the fitted peaks
are presented in Fig. 4.3a, b. The wavenumber positions, symmetry and C of the
fitted individual bands are reported in Table 4.2. The Raman G-band of MWCNT,
recorded at 785 nm excitation wavelength, is fitted with two Lorentzian functions,
and we obtained as best fit result peaks at 1584 and 1615 cm−1.
According to [42], in addition to the 1578 cm−1 mode of SWCNTs recorded
using a 632 nm HeNe laser, two low wavenumber modes were reported at
with a Lorentzian line shape function
1555 (A1g) 18.8 −0.30 1554 25.9 −0.27 – – – –
1569 (A1g) 6.6 1563 15.5 – – – –
1593 (A1g) 6.7 1585 9.1 1584 21 1582 13.5
1603 (E2g) 14.9 1593 14.1 1615 13 1614 12.0
approximately 1545 and 1561 cm−1. The mode at 1545 cm−1 was found to have a
BWF line shape and was assigned to the metallic nature of the SWCNTs. However,
the 1561 cm−1 mode was best described by a Lorentzian line shape and was
attributed to semiconducting SWCNTs. The peak at approximately 1578 cm−1,
which was also assigned to the metallic nature of the SWCNTs, showed a
Lorentzian line shape. The line shapes of the G-band components depend mostly on
the coupling of the discrete phonons to the electron continuum. While the line at
approximately 1578 cm−1 reported in [42] is best presented by a Lorentzian line
shape due to the weak coupling of phonons and electron continuum, in the present
study, this peak has a line shape coming closest to a BWF function, which points to
a stronger coupling. The second peak at 1593 cm−1 or closer to this value, observed
for all three excitation wavelengths, is associated with semiconducting SWCNTs.
This line is believed to appear because of the laser resonance with the energy gap
between 1D singularity in the 3d valence and 3d conduction bands of SWCNTs
with certain diameter [44].
The G-band seen in the Raman spectra of the SWCNTs, (CarboLex) recorded at
the excitation wavelength of 1064 nm, is best fitted assuming four components, one
having a BWF line shape and the other three having Lorentzian line shape func-
tions. However, for the SWCNT (Aldrich) species the G-band could only be
resolved into two components. The value of the fitting parameter, 1/q for both
SWCNTs (CarboLex) and SWCNTs (Aldrich) turns out to be −0.12/−0.11 for the
spectra obtained with 532 nm excitation wavelength. For 785 and 1064 nm, the 1/
q value is found to be −0.30/−0.27 and −0.10/−0.12 for SWCNTs (CarboLex) and
SWCNTs (Aldrich), respectively. The peak positions of the middle components of
the G-bands of both MWCNTs (*1585 cm−1) and SWCNTs (*1590 cm−1) are
changing only slightly with excitation wavelengths. For SWCNTs (CarboLex), the
peak at 1590 cm−1 for 532 nm is upshifted by 3 cm−1 for 785 and 1064 nm.
However, the low wavenumber components observed for the SWCNTs show a
downshift of 11 and 10 cm−1, respectively, when the excitation wavelength is
changed from 532 nm to 785 and 1064 nm. Similarly, the high wavenumber
components observed for the SWCNTs show a downshift of 3 cm−1 each when the
excitation wavelength is changed from 532 nm to 785 and 1064 nm.
with a Lorentzian line shape function
1556 (A1g) 12 −0.10 1548 15 −0.12 – – – –
1572 (A1g) 6.3 – – – – – –
1593 (A1g) 7.4 1589 14.7 1587 20 1586 20
1600 (E2g) 10.4 – – 1609 12 – –
The peak at approximately 1585 cm−1 in the Raman spectra taken from the
MWCNTs is downshifted by 1 cm−1 compared to the peak position found for the
MWCNTs for 785 nm excitation wavelength; see also Fig. 4.3c for the experi-
mental spectrum and fitted band components. Contrary, the 1585 cm−1 line position
is upshifted by 2 cm−1 when the wavelength is changed from 532 to 1064 nm. The
higher wavenumber peak at 1617 cm−1 for the MWCNTs recorded at 532 nm
shows a downshift of 5 and 8 cm−1 when the excitation wavelength is changed
from 532 nm to 785 and 1064 nm, respectively. The lower wavenumber peak in the
MWCNT Raman bands recorded at all three excitation wavelengths appeared at
almost the same position in graphite for corresponding excitation wavelengths. The
peak positions of the two fitted bands in the Raman spectra of MWCNTs recorded
with 1064 nm excitation are found to be at 1587 and 1609 cm−1. The peak position
and C (FWHM/2) for the G-band of the MWCNTs and graphite recorded at
1064 nm are presented in Table 4.3; see also Fig. 4.4c, d for the experimental
spectra and fitted individual bands. The laser energy dependence of the Stokes and
anti-Stokes Raman spectra of CNTs dispersed in aqueous solution and within the
solid bundles was thoroughly investigated by Fantini et al. in the energy range
1.52–2.71 eV using 76 laser lines, and the (n, m) assignment of nanotube is done by
the variation of resonance energy versus frequency of the ring-breathing mode
(RBM) [39]. The slight variation in wavenumber positions of component bands of
G bands with different laser lines may be due to the resonance Raman effect and
inhomogeneous nature of CNTs bundles such as different values of diameters and
chirality (n, m) [39].
Also the line widths are important when comparing the SWCNTs and the
MWCNTs. The low wavenumber components in the MWCNT spectra recorded
using the different excitation wavelengths show broader line widths compared to
those of the SWCNTs. A similar statement can be made comparing the two different
SWCNT varieties studied. The Aldrich species contains more impurities than the
CarboLex batch, so that band components of the SWCNT (Aldrich) material show
significantly broader line widths. However, the line widths of the low wavenumber
peaks for all excitation wavelengths behave differently. As was mentioned above,
this line is tentatively assigned to CNTs of metallic nature with a stronger coupling
between phonons and electron continuum; see also review by Dresselhaus et al.
[45]. The other two or three components of the G-bands show a Lorentzian line
shape and point to a semiconducting nature. The peak positions of the low
wavenumber lines of the MWCNTs appeared at higher wavenumbers compared to
the SWCNTs for all three excitation wavelengths. The contribution of both BWF
and Lorentzian line shape functions to the profile of the G-band shows that a
mixture of semiconducting and metallic CNTs can be found in SWCNTs studied
here. The positions and shapes of the different components clearly point to this
change of the electronic structure. A critical comment must be allowed with regard
to our Raman study. Recently, an optical, i.e., visible/near-infrared spectroscopic
evaluation of the metal-to-semiconductor ratio of SWCNTs was carried out by
Miyata et al. [46], who produced different materials by laser ablation and a sepa-
ration of both types by density gradient centrifugation. This is probably an effective
alternative method for material characterization.
4.4 Calculation of Young’s Modulus
Enomoto et al. [36] have performed a combined transmission electron microscopy

(TEM) and Raman spectroscopy study in order to measure the apparent Young’s
modulus of different CNT samples synthesized using various methods. An
empirical relationship between the apparent Young’s modulus and the crystallinity
of the different samples was reported there. For this, the authors used the parameter,
ID/IG obtained from the spectroscopic data. The values of the experimentally
measured apparent Young’s modulus of different CNT samples using TEM were
plotted versus the ID/IG values obtained from the Raman spectra. A nonlinear fitting
was done to determine the relation between the Young’s modulus and ID/IG values.
As a result, for the apparent Young’s modulus Ea, the empirical relation,
Ea = 2.7 104 (ID/IG)−1.6 MPa with the crystallinity factor ID/IG was obtained. It is
obvious from the above equation that the apparent Young’s modulus of the CNTs is
inversely depending on the order of crystallinity.
For the present analysis, the ID/IG values of the SWCNT and MWCNT bundles
were calculated using the integrated areas under the D (ID) and G-bands (IG) found
in the Raman spectra recorded at 532 nm excitation wavelength. The ID/IG values
determined for the CarboLex and Aldrich SWCNT varieties and MWCNTs are
0.085, 0.097 and 0.34, respectively. Thus, the value of ID/IG is much smaller for the
SWCNTs compared to the MWCNTs. This finding points to an increased number
of defects in the MWCNTs. Hence, the SWCNTs with the smaller value of ID/IG
have a high crystallinity sp2 network and, therefore, this structure also is the better
model system for a calculation of the Young’s modulus of CNTs theoretically.
Yakobson et al. [47] performed such a theoretical study using realistic many-body
potentials for the calculation. These authors suggested that the CNTs are resilient,
sustaining at extreme strain with no signs of brittleness, plasticity or atomic rear-
rangement. They predicted a theoretical value of the Young’s modulus, Ea = 5.5
TPa, for properly chosen tube parameters. In our case, the values of the apparent
Young’s modulus calculated using the expression, Ea = 2.7 104 (ID/IG)−1.6 MPa,
for SWCNTs (CarboLex), SWCNT (Aldrich) and MWCNTs turn out to be 1.39,
1.13 and 0.15 TPa, respectively. The order of magnitude of these values is in good
agreement with the theoretically calculated value given above [47].
Quite recently, a purity and defect density characterization of SWCNTs based on
the IG/ID ratio (inverse to our application) obtained from their Raman spectra was
investigated by Miyata et al. [48], which underlines the complexity of spectral
interpretation. In addition, different nanocarbon contributions to MWCNTs were
used by DiLeo et al. [49] for a calibration experiment aiming at a purity assessment
based on the ID/IG ratio of the corresponding Raman bands as a complement to
electron microscopy. Probably, several analytical methods are required for a unique
material characterization.
4.5 Conclusions
Raman spectra of graphite, single-walled, and multi-walled carbon nanotube

(SWCNT and MWCNT, respectively) materials were recorded using three different
excitation wavelengths. The G-bands of SWCNTs were found to have a noticeable
asymmetry at each wavelength and, therefore, were fitted taking three/four peaks
into account using two different line shape functions. The low wavenumber com-
ponent of the G-band was fitted with a Breit-Wigner-Fano (BWF) and the other
two/three components with a Lorentzian line shape function. The presence of both
BWF and Lorentzian features in the Raman profiles of the G bands of the SWCNTs
together with the observed changes of line positions and shapes point to a mixture
of semiconducting and metallic CNTs in the CNT bundles. The values of the
apparent Young’s modulus of the SWCNT and MWCNT materials were calculated
using an empirical relationship recently published [36]. These calculations yielded
the values 1.39, 1.13 and 0.15 TPa for the SWCNTs and MWCNTs, respectively.
Thus, the SWCNTs materials have a larger apparent Young’s modulus, which goes
along with a higher crystallinity compared to the MWCNTs.
Acknowledgements AKO is grateful to the Alexander von Humboldt Foundation for the award
of a research fellowship. The authors would like to express sincere thanks to Prof. B. P. Asthana
and Dr. A. Srivastava for the MWCNT material and suggestions to initiate this work.
References
1. Kroto HW, Heath JR, O’Brien SC, Curl RF, Smalley RE (1985) C60: Buckminsterfullerene.
Nature 318:162–163
2. Ijima S (1991) Synthesis of carbon nanotubes. Nature 354:56–58
3. Kruss S, Hilmer AJ, Zhang J, Reuel NF, Mu B, Strano MS (2013) Carbon nanotubes as
biomedical sensors. Adv Drug Deliv Rev 65:1933–1950
4. Zaporotskova IV, Boroznina NP, Parkhomenko YN, Kozhito LV (2016) Carbon nanotubes:
sensor properties. A review. Mod Electron Mater 2:95–105
5. Iannazzo D, Piperno A, Pistone A, Grassi G, Galvagno S (2013) Recent advances in carbon
nanotubes as delivery systems for anticancer drugs. Curr Med Chem 20:1333–1354
6. Dalton AB, Collins S, Munoz E, Razal JM, Ebron VH, Ferraris JP, Coleman JN, Kim BG,
Baughman RH (2003) Super-tough carbon-nanotube fibres. Nature 423:703–703
7. Bokobza L (2012) A Raman investigation of carbon nanotubes embedded in a soft polymeric
matrix. J Inorg Organometallic Polym 22(3):629–635
8. Baibarac M, Baltog I, Lefrant S (2011) Recent progress in synthesis, vibrational
characterization and applications trend of conjugated polymers/Carbon nanotubes composites.
Curr Org Chem 15:1160–1196
9. Bose S, Khare RA, Moldenaers P (2010) Assessing the strengths and weaknesses of various
types of pre-treatments of carbon nanotubes on the properties of polymer/carbon nanotubes
composites: a critical review. Polymer 51:975–993
10. Dillon AC, Jones KM, Bekkedahl TA, Kiang CH, Heben MJ (1997) Storage of hydrogen in
single-walled carbon nanotubes. Nature 386:377–379
11. Britto PJ, Santhanam KSV, Ajayan PM (1996) Carbon nanotube electrode for oxidation of
dopamine. Bioelectrochem Bioenerg 41:121–125
12. Che GL, Lakshmi BB, Fisher ER, Martin CR (1998) Carbon nanotubule membranes for
electrochemical energy storage and production. Nature 393:346–349
13. Lu W, Baek JB, Dai L (eds) (2015) Carbon nanomaterials for advanced energy systems:
advances in materials synthesis and device applications. Wiley, Hoboken, NJ
14. Herrero-Latorre C, Álvarez-Méndez J, Barciela-García J, García-Martín S, Peña-Crecente RM
(2015) Characterization of carbon nanotubes and analytical methods for their determination in
environmental and biological samples: a review. Anal Chim Acta 853:77–94
15. Braun EI, Huang A, Tusa CA, Yukica MA, Pantano P (2016) Use of Raman spectroscopy to
identify carbon nanotube contamination at an analytical balance workstation. J Occup Environ
Hyg 13:915–923
16. Rogers-Nieman GM, Dinu CZ (2014) Therapeutic applications of carbon nanotubes:
opportunities and challenges. Wiley Interdiscip Rev Nanomed Nanobiotechnol 6:327–337
17. Herrera-Herrera AV, Gonzalez-Curbelo MA, Hernández-Borges J (2012) Carbon nanotubes
applications in separation science: a review. Anal Chim Acta 734:1–30
18. Socas-Rodríguez B, Herrera-Herrera AV, Asensio-Ramos M, Hernández-Borges J (2014)
Recent applications of carbon nanotube sorbents in analytical chemistry. J Chromatogr A
1357:110–146
19. Srivastava A, Srivastava ON, Talapatra S, Vajtai R, Ajayan PM (2004) Carbon nanotube
filters. Nat Mat 3:610–614
20. Ebbesen TW, Lezec HJ, Hiura H, Bennett JW, Ghaemi HF, Thio T (1996) Electrical
conductivity of individual carbon nanotubes. Nature 382:54–56
21. Bockrath M, Cobden DH, McEuen PL, Chopra NG, Zettle A, Thess A, Smalley RE (1997)
Single-electron transport in ropes of carbon nanotubes. Science 275:1922–1925
22. Krauss TD, Wise FW, Tanner DB (1996) Observation of coupled vibrational modes of a
semiconductor nanocrystal. Phys Rev Lett 76:1376–1379
23. Bischof T, Lermann G, Schreder B, Materny A, Kiefer W, Ivanda M (1997)
Intensity-dependent micro-Raman and photoluminescence investigations of CdSxSe1−x
nanocrystallites. J Opt Soc Am B 14:3334–3338
24. Schreder B, Materny A, Kiefer W, Kümmell T, Bacher G, Forchel A, Landwehr G (2000)
Raman investigation of CdxZn1−xSe/ZnSe quantum wires: Strain relaxation and excitation
profile. J Appl Phys 88:764–771
25. Jorio A, Dresselhaus G, Dresselhaus MS, Souza M, Dantas MSS, Pimenta MA, Rao AM,
Saito R, Liu C, Cheng HM (2000) Polarized Raman study of single-wall semiconducting
carbon nanotubes. Phys Rev Lett 85:2617–2620
26. Kasuya A, Sasaki Y, Saito Y, Tohji K, Nishina Y (1997) Evidence for size-dependent discrete
dispersions in single-wall nanotubes. Phys Rev Lett 78:4434–4441
27. Dresselhaus MS, Jorio A, Hofmann M, Dresselhaus G, Saito R (2010) Perspectives on carbon
nanotubes and graphene Raman spectroscopy. Nano Lett 10:751–758
28. Jorio A, Kauppinen E, Hassanien A (2008) Carbon-nanotube metrology. In: Jorio A,
Dresselhaus G, Dresselhaus MS (eds) Carbon nanotubes. Topics applied in physics, vol 111.
Springer, Berlin, pp 63–100
29. Thomsen C, Reich S (2007) Raman scattering in carbon nanotubes. In: Cardona M, Merlin R
(eds) Light scattering in solid IX. Topics applied in physics, vol 108. Springer, Berlin,
pp 115–232
30. Saito R, Grüneis A, Samsonidze GG, Brar VW, Dresselhaus G, Dresselhaus MS, Jario A,
Canҫado LG, Fantini C, Pimenta MA, Filho AGS (2003) Double resonance Raman
spectroscopy of single-wall carbon nanotubes. New J Phys 5:157
31. Saito R, Dresselhaus G, Dresselhaus MS (1998) Physical properties of carbon nanotube.
Imperial College Press, London
32. Dresselhaus MS, Endo M (2001) Carbon nanotubes: synthesis, structure, properties and
applications. In: Dresselhaus MS, Dresselhaus G, Avouris P (eds) Topics Applied in Physics,
vol 80. Springer, Berlin, pp 11–28
33. Sadezky A, Muckenhuber H, Grothe H, Niessner R, Poeschl U (2005) Raman microspec-
troscopy of soot and related carbonaceous materials: spectral analysis and structural
information. Carbon 43:1731–1742
34. Dresselhaus MS, Dresselhaus G, Jorio A (2007) Raman Spectroscopy of Carbon Nanotubes
in 1997 and 2007. J Phys Chem C 111:17887–17893
35. Jorio A, Pimenta MA, Souza Filho AG, Saito R, Dresselhaus G, Dresselhaus MS (2003)
Characterizing carbon nanotube samples with resonance Raman scattering. New J Phys 5:139
36. Enomoto K, Kitakata S, Yasuhara T, Ohtake N (2006) Measurement of Young’s modulus of
carbon nanotubes by nanoprobe manipulation in a transmission electron microscope. Appl
Phys Lett 88:153115–153117
37. Heise HM, Kuckuk R, Ojha AK, Srivastava A, Srivastava V, Asthana BP (2009)
Characterisation of carbonaceous materials using Raman spectroscopy: a comparison of
carbon nanotube filters, single- and multi-walled nanotubes, graphitised porous carbon and
graphite. J Raman Spectrosc 40:344–353
38. Heise HM, Kuckuk R, Srivastava A, Asthana BP (2011) Characterization of carbon nanotube
filters and other carbonaceous materials by Raman spectroscopy—II: study on dispersion and
disorder parameters. J Raman Spectrosc 42:294–302
39. Fantini C, Jorio A, Souza M, Strano MS, Dresselhaus MS, Pimenta MA (2004) Optical
transition energies for carbon nanotubes from resonant Raman spectroscopy: environment and
temperature effects. Phys Rev Lett 93:147406–147410
40. Kataura H, Kumazawa Y, Maniwa Y, Umezu I, Suzuki S, Ohtsuka Y, Achiba Y (1999)
Diameter control of single-walled carbon nanotubes. Synth Met 103:2555–2558
41. Brown SDM, Jorio A, Corio P, Dresselhaus MS, Dresselhaus G, Saito R, Kneipp K (2001)
Origin of the Breit-Wigner-Fano lineshape of the tangential G-band feature of metallic carbon
nanotubes. Phys Rev B 63:155414–155421
42. Puretzky AA, Schittenhelm H, Fan X, Lance JM Jr, Allard LF, Geohegan DB (2002)
Investigations of single-wall carbon nanotube growth by time-restricted laser vaporization.
Phys Rev B 65:245425–245433
43. Tuinstra F, Foenig J (1970) Raman spectrum of graphite. J Chem Phys 53:1126–1130
44. Brown SDM, Corio P, Marucci A, Dresselhaus MS, Pimenta MA, Kneipp K (2000)
Anti-Stokes Raman spectra of single-walled carbon nanotubes. Phys Rev B 64:R5137
45. Dresselhaus MS, Dresselhaus G, Hofmann M (2007) The big picture of Raman scattering in
carbon nanotubes. Vib Spectrosc 45:71–81
46. Miyata Y, Yanagi K, Maniwa Y, Kataura H (2008) Optical evaluation of the metal-to-
semiconductor ratio of single-wall carbon nanotubes. J Phys Chem 112:13187–13191
47. Yakobson BI, Brabec CJ, Bernholc J (1996) Nanomechanics of carbon tubes: instabilities
beyond linear response. Phys Rev Lett 76:2511–2514
48. Miyata Y, Mizuno K, Kataura H (2011) Purity and defect characterization of single-wall
carbon nanotubes using raman spectroscopy. J Nanomater 786763
49. DiLeo RA, Landi BJ, Raffaelle RP (2007) Purity assessment of multiwalled carbon nanotubes
by Raman spectroscopy. J Appl Phys 101:064307–64311
Chapter 5
Ligand-Core NLO-Phores
Two-Photon Absorption and Two-Photon
Excited Emission Properties of Atomically
Precise Clusters of Gold and Silver
Rodolphe Antoine
Abstract Functional ligand-protected noble metal cluster nanomaterials with

enhanced two-photon absorption and two-photon excited emission may lead to new
technologies for bio-imaging applications. In this article, I review experimental and
theoretical methodologies allowing detailed investigation of two-photon absorption/
emission properties of ligand-protected silver and gold metal clusters. This includes
femtosecond two-photon excited fluorescence experimental setups and quantum
chemical methodologies based on time-dependent density functional theory.
I thoroughly analyze physical phenomena and trends leading to large two-photon
absorption/emission responses of model nanoclusters focusing on the effects of the
relaxation pathways in the linear and nonlinear optical regime, as well as strategies
aiming at enhancing their two-photon emission responses.
5.1 Introduction
Recent developments in optical imaging techniques, in particular multi-photon

excitation (MPE) microscopy that allows studies of biological interactions at a deep
cellular level, have motivated intensive research in developing multi-photon
absorption fluorophores. Biological tissues are optically transparent in the
near-infrared region. Therefore, fluorophores that can absorb light in the NIR region
by multi-photon absorption are particularly useful in bio-imaging. MPE microscopy
enables molecular imaging by using either exogenous markers or endogenous
signals. Endogenous nonlinear optics (NLO) signals are generally weak, unspecific
and usually require to improve the instruments detection threshold and to develop
R. Antoine (&)
Institut Lumière Matière, UMR5306 Université Claude Bernard Lyon1-CNRS,
Université de Lyon, 69622 Villeurbanne Cedex, France
e-mail: rodolphe.antoine@univ-lyon1.fr

140 R. Antoine
scientific innovations, such as endogenous fluorescent proteins [1] to be applicable

to MPE microscopy. Likewise, exogenous markers like organic dyes have been
synthesized for MPE microscopy, but their use remains limited by their relatively
low two-photon excited fluorescence cross sections in aqueous environments and
their rapid photo-bleaching [2]. In addition, other probes have been proposed
including quantum and polymer nanodots with broad absorption and discrete,
tunable emission wavelengths withstanding photo-bleaching [3]. Preliminary
studies have clearly demonstrated that noble metal nanoclusters (NCs), nanoma-
terials made of few to hundred gold or silver atoms protected by ligand molecules
exhibiting molecular-like properties [4], could also provide a valuable route in the
nonlinear optical regime and in particular MPE microscopy [5].
Multi-photon optics stems from the nonlinear light–matter interaction and can be
described by a medium polarization P, induced by an intense optical electric field
E as:
P ¼ vð1Þ E þ vð2Þ EE þ vð3Þ EEE þ . . . ð5:1Þ
where vð1Þ is the linear susceptibility tensor representing effects such as linear
absorption and refraction, vð2Þ is the second-order nonlinear optical susceptibility,
vð3Þ is the third-order nonlinear susceptibility, and so on. Second-harmonic gen-
eration (SHG) is a second-order process, whereas two-photon excitation fluores-
cence (TPEF) and third harmonic generation (THG) are both third-order processes.
The combination of the three processes intrinsically provides different contrasts for
living matter microscopy (Fig. 5.1; left) [6]. Of note, the use of THG in the higher
NIR region (1.3–1.7 µm) is one of the strategies of choice for nonlinear imaging in
scattering tissues [7].
Fig. 5.1 (left) Multicontrast images of an unperturbed human mammary tumor with various
extracellular vesicles. TPEF (2PF) is used to image fluorescent dyes or endogenous molecules.
SHG is used to image non-centrosymmetric structures such as collagen fibers, and THG is used to
visualize the refractive index differences such as the interface of lipid droplets. Adapted from
http://biophotonics.illinois.edu/imaging-technology/imaging-techniques/multiphoton-microscopy.
(right) Selected protected metal clusters that may display different NLO signals
5 Ligand-Core NLO-Phores 141
Designing highly efficient second-order v(2) and third-order v(3) NLO chro-
mophores is largely a matter of finely combining a high density of delocalized
electrons in a symmetrical or unsymmetrical environment. Gold and silver NCs
constitute therefore good candidates (Fig. 5.1; right), although the reported v(2)
values are still weaker than those of push–pull dyes [8] and are not currently
competitive as contrast agents for MPE microscopy. Our theoretical and experi-
mental joint investigation on ligand-protected silver and gold clusters [9–17] has
shown that the structure of the metal atom core, its charge, and symmetry, dra-
matically influencing the NLO cross sections and the core stabilizing ligands, play a
major role in NLO efficiencies. This new class of NLO materials is coined as
“ligand-core” NLO-phores [18, 19].
This feature article overviews mainly experimental methodologies used for
determination and analysis of two-photon absorption (TPA) and two-photon excited
(TPEF) emission properties of atomically precise clusters of gold and silver. Using
results obtained on atomically precise clusters of silver, I will show that both
experimental and theoretical data can be used to get an extensive comprehension of
the physics underlying the two-photon absorption process and its amplitude, as well
as to suggest an exploratory root for novel molecular engineering for further
enhancement of TPA [20]. Also, I will describe some strategies that have been
suggested recently to enhance TPEF properties. This includes the possibility to
increase the rigidity of the protective shell [21] as well as metal core-doping [22], a
promising strategy as it distorts the metallic atom core.
The paper is organized as follows. In Sect. 5.2, I define the atomically precise
clusters of gold and silver, from their synthesis and characterization toward their
(linear) optical properties. In Sect. 5.3, I summarize the main theoretical and
computational models used for calculation of the nonlinear responses and describe
the leading factors responsible for enhanced NLO efficiencies. Section 5.4 deals
with the principal experimental techniques and challenges for measurement of the
TPA/TPEF cross section. A variety of representative “ligand-core” NLO-phores are
investigated in Sect. 5.5, allowing for rationalization of different structural effects
on the TPA/TPEF cross section. Finally, the main conclusions as well as some
future perspectives are drawn in Sect. 5.6.
5.2 Atomically Precise Clusters of Gold and Silver:

Synthesis, Characterization, and Optical Properties
5.2.1 Atomically Precise Clusters of Gold and Silver
For small noble metal clusters—in the size range where each atom counts—
Mie-Drude-like model [23] that predicts the optical response of free-electron metals
in the bulk state is no longer appropriate to discuss absorption spectra in details.
142 R. Antoine
The geometry of the clusters must be determined by quantum chemistry methods

that often use group theory and the optic response is described in terms of molecular
transitions whose positions and intensities are predicted by sophisticated calcula-
tions of quantum mechanics [20]. As a pioneering work, the absorption spectra
computed by equation-of-motion coupled-cluster singles and doubles
(EOM-CCSD) and similarity transformed equation-of-motion coupled-cluster sin-
gles and doubles (STEOM-CCSD) methods with all-electron basis sets, for the most
stable structures of Ag5–8, nicely illustrated the molecular-like behavior of nan-
oclusters leading to an electronic energy quantization and the changes in the leading
features of the patterns as a function of the cluster size [24]. While the optical
properties of such small metal clusters have been largely investigated in the gas
phase, their study in the solid and liquid phase requires some “stabilization” that
prevents them from fragmentation or degradation. The use of solid gas or inorganic
matrices permits to protect gold or silver clusters from photodissociation [25–28].
Organic scaffolds allow the formation and stabilization of metal clusters in solution.
The use of organic scaffolds for fluorescent metal nanoclusters is relatively new,
first reported by the Zheng and Dickson [29] for silver nanoclusters. These organic
scaffolds have tremendous potentials, as the interaction between the ligands and
metal clusters can be adjusted, leading to tunability in their spectroscopic proper-
ties. For example, by using DNA oligomers as organic scaffolds and by playing
with the nucleotide sequence of DNA oligomers, it is possible to synthesize silver
nanoclusters that emit from the blue to near-infrared region [30].
Ligands play very important roles in the formation of NCs as protective agents,
which can prevent the metal clusters from aggregation and then keep the
size-dependent fluorescence property. The formation and stabilization of gold or
silver nanoclusters in solution have been accomplished in various ways and with
different scaffolds (Figs. 5.2 and 5.3). The proper choice of parameters for the
reaction, including the temperature, the reducing method, the stabilizers, and the
initial ratio of metal salt and stabilizer, plays a crucial role in the successful syn-
thesis of nanoclusters and to limit the size to few-atom nanoclusters. In addition to
the ultrasmall size, the ligands used for NCs preparation also have impacts on their
fluorescence properties. Wu and Jin [31] demonstrated that for gold NCs, the
surface ligands of NCs not only can be used as capping agent but also largely affect
the fluorescence of NCs through charge transfer from surface ligand to the gold
core. When the surface ligands have strong electron donation capability, the
fluorescence can be enhanced. And the ligands with electron-rich atoms or groups
have been found as a very effective choice for the promising surface ligand of NCs
to enhance the fluorescence.
Fig. 5.2 Protected metal nanoclusters with possible scaffolds
Fig. 5.3 (top) General route to produce thiolate-protected nanoclusters by a chemical reductant.
(bottom) PAGE for Ag:SG clusters by (left) Bigioni synthesis (see [36]) and (right) our focused
Ag:SG synthesis (see [9] and [18])
144 R. Antoine
5.2.2 Atomically Precise Clusters of Gold and Silver—

Synthetic Routes
The production of silver and gold nanoclusters can be performed following several
routes. The metal ions from dissolved metal salts can be reduced, either by a
chemical reductant (e.g., sodium borohydride), (Fig. 5.3), or by light (photore-
duction with near-ultraviolet light), or by c-rays (by radiolysis of water). The
chemical reduction and the photoreduction are the most commonly used methods
[32]. The specific properties of metal nanoclusters, such as the composition, sta-
bility, and fluorescence quantum yield, depend largely on the scaffold used during
reduction.
Thiols are frequently used on noble metal substrates because of the strong affinity
of sulfur for these metals. And thiolated ligands (-SR) have appeared to be extremely
good candidates to produce ultrasmall nanocluster sizes, in particular for gold [33].
Following the pioneering work of Brust et al. [34] based on the reduction of the
metal precursors and the formation of metal core, thiol-containing small molecules
were extensively used to stabilize gold and silver nanoclusters in the aqueous
solution [4, 35]. The use of thiol-containing small molecules as stabilizers permits to
better control the production of gold NCs (AuNCs) than phosphine-capped ones,
contributing to the stronger Au-S covalent bonding. Generally, the method of syn-
thesizing thiolate-capped AuNCs proceeds as follows. Gold salts [AuCl4]− are
dissolved in water and then transferred to an organic solvent by phase transfer agent;
the thiols are added to the mixture inducing reduction of Au3+ ions into Au+ ions and
form Au+–SR complexes or polymers; then, the Au+ polymers are reduced by
adding the reducing agent leading to thiolate-protected gold nanoclusters.
Glutathione (GSH), a ubiquitous low-molecular-weight thiol, played a signifi-
cant role in producing gold NCs which showed good water solubility, bioactive
surface, and high stability. Whetten and coworkers have proposed an unprecedented
thiol-protective AuNC route by using the GSH (N-c-glutamyl-cysteinyl-glycine) as
the stabilizer. The as-synthesized AuNCs were fractionated by using polyacry-
lamide gel electrophoresis (PAGE) and characterized by mass spectrometry (MS)
[37]. Tsukuda and colleagues have also reported the characterization of fractionated
AuNCs protected by GSH monolayers. The as-prepared AuNCs were isolated into
single-sized Aun(SR)m clusters by the PAGE method and characterized using
electrospray mass spectrometry [38, 39]. Then, Kumar et al. [36] bridged the gap to
use glutathione as a ligand to produce protected AgNCs that were isolated into
single-sized Agn(SR)m clusters by the PAGE method and characterized using
electrospray mass spectrometry.
While the routes for producing metal NCs lead to a mixture of Aun(SR)m
(Agn(SR)m) [40] cluster size, achieving atomic precision and molecular purity is
challenging because the nanocluster growth is extremely complicated and remains
poorly understood. Nevertheless, a systematic methodology called “size focusing”
for achieving atomically precise clusters of gold and silver with molecular purity
has been proposed. This methodology consists of two primary steps [41]:
• Kinetically controlled synthesis of an Mn(SR)m mixture with a properly con-

trolled size range and
• Thermodynamically dictated size focusing of the mixture to single-sized
nanoclusters.
In parallel, a new approach, which is to utilize ligand exchange to induce size
and structure transformation and, hence, to attain new Mn(SR)m nanoclusters, was
also proposed. In parallel to these wide developed synthesis methods, Yu et al. [42]
proposed a kinetically controlled route utilizing carbon monoxide (CO) as a mild
reducing agent for the synthesis of size-selected Au NCs.
Concerning the characterization techniques, X-ray crystallography is of course
the “holy-grail” technique to solve the crystal structure of nanoclusters and reveals
the nature of bonding and the packing of atoms [4, 43–45]. Recently, complete
structures have been experimentally resolved for many gold and silver nanoclusters,
including the most popular Au25(SR)18, Au38(SR)24, and Au102(SR)44 NCs by
X-ray crystallography [4]. Many other methods are used to characterize noble metal
NCs. X-ray photoelectron spectroscopy (XPS) is a valuable tool for confirming the
oxidation state of gold in the sample, based on the Au–Au and Au–S binding
energies. Various imaging techniques are widely employed to determine nanopar-
ticle size and sample dispersity, such as transmission electron microscopy
(TEM) and atomic force microscopy (AFM). The composition of NCs is best
resolved by high-resolution mass spectrometry (MS). The most widely used ion-
ization methods for NCs characterization by MS are matrix-assisted laser desorp-
tion ionization (MALDI) and electrospray ionization (ESI) [46–48].
5.2.3 Atomically Precise Clusters of Gold

and Silver—Optical Properties
One-photon excited fluorescence (OPEF) of metal nanoparticles is of considerable

interest due to their potential applications, for instance, in biomedicine. However,
the origin and underlying mechanism of OPEF in these clusters are still poorly
understood, although some recent theoretical investigations by Weerawardene and
Aikens [49–51], using time-dependent density functional theory (TDDFT), shed
some new lights on the key processes for photon emission in such nanosystems.
Unraveling the crystal structure has permitted to better evaluate the structure–
absorption relationships of some thiolate-protected gold clusters. Time-dependent
density functional theory (TDDFT) calculations of the electronic structure of Au25
clusters showed that the HOMO and the lowest three LUMOs are mainly composed
of 6 sp atomic orbitals of gold and a certain degree of the S (3p) [52]. The other
higher HOMO orbitals are mainly constructed from the 5 d10 atomic orbitals of
gold and hence constitute the d-band. The calculated absorption transitions are in
agreement with the experimental observations. Recent ultrafast spectroscopic
studies on relaxation of higher excited states have provided more insights into
146 R. Antoine
Fig. 5.4 Cartoon diagram

showing the relaxation
pathways in gold nanoclusters
following one-photon
absorption
understanding the photoluminescence mechanism of Au25 clusters. Devadas et al.

[53] found that 500 nm emission fundamentally arises from the electron–hole
recombination in the Au13 core with little perturbation from surface ligands, but
NIR emission at 700 nm originates from the recombination of holes in the ground
core state and electron decay from core-excited states to S–Au–S–Au–S semi-rings.
Wu and Jin in a seminal work [31] found that charge state and surface ligands also
have a significant influence on the NIR emission wavelength and quantum yields of
Au25 clusters. Clearly, more detailed photophysical studies are required for eval-
uating how surface ligands influence the photoluminescence mechanisms of such
systems. Over the last ten years, some general trends have been figured out con-
cerning the de-excitation pathways following a visible or near-UV absorption. The
following experimental and theoretical findings were assembled from the present
work and literature to derive the (very simplistic) energy diagram in Fig. 5.4 [54].
Near-ultraviolet and visible absorbance may arise from transitions between
molecular orbitals with high ligand contribution to orbitals with high metal char-
acter (LMCT) and from metal−metal electronic transitions (LMMCT). A rapid
(<1 ps lifetime) decay pathway for clusters which have a core of metal atoms may
lead to an emission in the visible. A long-lived (>100 ns lifetime), charge-transfer
component is exhibited for all clusters. NIR emission in the clusters is related to the
surface states and originates from the charge-transfer excited state (Fig. 5.4).
5.3 Atomically Precise Clusters of Gold and Silver as New

NLO Chromophores—Background and Design
The understanding of the structure–property relationship of molecular TPA is of

great importance for the rational design of optimized two-photon chromophores.
We will use the analogy with push–pull molecules in order to describe how
Fig. 5.5 Schematic illustration of (left) a nonlinear optical (NLO), push–pull chromophore and
(right) a “ligand-core” NLO-phore with ligand-protected silver and gold clusters
theoretical data can be used to get an extensive comprehension of the physics

underlying the two-photon process and its amplitude, as well as to suggest an
exploratory root for novel chemical engineering for further enhancement of TPA in
atomically precise clusters of silver and gold.
Push–pull dipolar molecules are characterized by a low-lying, high-intensity
absorption band, related to the intramolecular charge transfer (ICT) between the
electron donor (D) and acceptor (A) groups (Fig. 5.5).
The TPA cross section of such molecules is considered to be governed basically
by two factors: transition dipole moments and transition energies of the molecule.
As seen below, the theoretical expression of TPA cross section ðdTPA ðxÞÞ based on
the perturbation expansion is comprised of the numerator including transition
dipole moments and the denominator including the transition energies and the
incident photon energy. The energy term governs the wavelength dispersion of
dTPA ðxÞ, whereas the dipole moment term governs the overall magnitude of
dTPA ðxÞ. Thus, the structure–property relationship for the molecules with large
dTPA ðxÞ can be reduced by optimizing the transition dipole moments and fre-
quencies involved in the TPA process.
It is commonly accepted that from the perturbation theory, the TPA cross section
at the laser frequency of x is given as [55]:
ð2pÞ2 x2 2
dTPA ðxÞ ¼ 2
gðxÞSeg ð5:2Þ
ðchÞ
where c and h are the speed of light and the Planck constant, respectively.
gðxÞ
denotes the normalized lineshape function of the TPA transition; Seg is so-called
two-photon tensor. According to Przhonska et al. [56], the TPA cross section can be
written in SI units as:
148 R. Antoine
1 ðhxÞ2
dTPA ðxÞ ¼
5c2 hn2 e20 hxeg hx 2 þ C2
" 2 eg
2 #
leg jDlj2 Ceg leg jlee0 j2 Ce0 g
2 þ 2 ð5:3Þ
hxeg 2hx þ C2eg hxe0 g 2hx þ C2e0 g
The main parameters responsible for dTPA ðxÞ are: change in the permanent
dipole moment Dl; transition dipole moments leg ; lee0 ; angles between dipole
C; and detuning
moments; linewidth
energies from intermediate and final states,
hxeg hx and hxe0 g 2hx .
Enhancement of dTPA ðxÞ can be obtained by playing with the following factors:
• Increasing the transition dipole moments. In molecular design, this can be
realized by increasing the p-conjugation length, or by introducing electron
donor/acceptor groups. In non-centrosymmetric molecules, increasing the dif-
ference of the ground and excited state permanent dipole moments can also
increase dTPA ðxÞ.
• Maximizing resonance terms. Decreasing the detuning energy between inter-
mediate and ground states can significantly enhance dTPA ðxÞ. If the intermediate
state is located halfway between ground state and final state, a “double reso-
nance” condition can be achieved, which can lead to a dramatic enhancement of
dTPA ðxÞ [57].
• Reducing the linewidth of the lowest energy one-photon transition.
Such factors have been figured out in details for design strategies and structure–
property relations of cyanine and cyanine-like molecular structures with the goal of
enhancing TPA in the near-IR for multi-photon fluorescence sensing applications
[56].
Atomically precise nanoclusters of silver or gold can be viewed as a “multi-shell
system” composed by a metallic core, a metal–ligand interface, in particular with
staple motifs leading to metal–sulfur bonds, and the surface ligand molecules.
These three shells may communicate in two different ways: charge transfer from
ligand to metal core (analogy with ligand-to-metal charge transfer (LMCT) or
ligand-to-metal–metal charge transfer (LMMCT) observed in metal complexes) and
through direct bonding or direct donation of delocalized electrons of electron-rich
groups of the ligands [19]. Such “communications” between ligands and metal core
may increase the transition dipole moments leading to enhanced dTPA ðxÞ.
Density functional theory (DFT) and its time-dependent version (TDDFT) have
been used for determination of the structural and optical properties of
ligand-protected silver and gold clusters. There are two approaches for addressing
two-photon absorption within analytic response method: first, the calculation of
third-order frequency-dependent response function from the second hyperpolariz-
ability in which imaginary part is related to TPA cross section; the second approach
involves the single residue of the second-order response function or the first
hyperpolarizability. The latter represents more practical way of computing the TPA
cross section. In spite of the success, it is necessary to notice that in standard

response theory, the response functions can diverge, since the response function has
poles whenever one or more of the optical frequencies equal excitation energy. This
can lead to nonphysical behavior for molecular properties in the resonance region.
However, introducing damping terms in different ways the singularities of the
response functions can be corrected or effectively removed [17].
The two-photon absorption cross section for an excitation from the ground
state|
0> to a final state |f> is defined in terms of normalized shape function g xm þ xl
and the two-photon absorption transition amplitude tensor T xm xl ;f .
ð2peÞ4 xm xl x x ;f 2
rTPA ¼ g xm þ x l
T m l ð5:4Þ
c2
In order to obtain this tensor, the quadratic density functional response theory
was applied. Two-photon absorption transition amplitude tensor T xm xl ;f between
the ground and the excited state is defined as:
2x;f
X h0jl ^b jf i h0jl
â jkihkjl â jf i
^b jkihkjl

Tab ¼ þ ð5:5Þ
k
xk xf =2 xk xf =2
where it is assumed that frequency of incident radiation is equal to half of excitation

energy from ground to excited state, i.e., x ¼ xf =2. In the above equation, l â and
^b are Cartesian components of dipole moment operator l, and xk and xf are the
l
frequencies of excitation from |0> to |k> and |f>, respectively. Thus, application of
this formula includes explicit summation over-excited states and requires compu-
tation of dipole moment operator l between excited states. Of note, in order to
prevent the TPA cross sections from blowing up near the one-photon resonances,
the sum-over-states approach uses a damping factor C. More details about com-
putational details can be found in [17] and [19].
Russier-Antoine et al. [18] conducted a theoretical investigation of the nonlinear
optical properties for the lowest energy structures of the Ag11L7, Ag15L11, and
Ag31L19 nanoclusters, where L stands for the SCH3 group (Fig. 5.6). They contain,
respectively, 4, 8, and 12 delocalized electrons within the core. Several factors
influencing the TPA cross sections have been figured out:
• The excitation between ligands and the metal core are characteristic of the
nonlinear transitions;
• The “double resonance” between states involved in the OPA and TPA processes
is required to obtain giant TPA cross sections;
• Large dipole transition moments are related to a non-uniform electron distri-
bution within the metal core.
The role of the structural properties, i.e., of the geometry of the metal core in
determining this electron distribution, is therefore crucial.
150 R. Antoine
Fig. 5.6 Comparison of TDDFT TPA spectra (left) of Ag11(SCH3)7, Ag15(SCH3)11, and
Ag31(SCH3)19 nanoclusters for the lowest energy structures involving 4, 8, and 12 delocalized
electrons in the metal core. Adapted from [18], copyright (2018) The Royal Society of Chemistry
In the case of the TPA cross sections, interplay between a resonance effect and
large transition dipole moments is essential. As the size of the silver nanoclusters
increases, transition dipole moments should continue to increase due to larger core
size (and thus larger core-to-ligand distances) and/or non-uniform electronic dis-
tribution in the metal core. Also, the optical band gap decreases as the size of the
nanoclusters increases [43], shifting the spectra toward the NIR spectral region, a
feature useful for bio-imaging applications.
5.4 Measurement Techniques of Optical Nonlinearities—

Two-Photon Absorption/Fluorescence
The light source used in a typical experimental device is a titanium–sapphire

femtosecond laser tunable between 700 and 1000 nm. At the exit of this laser
cavity, an optical system was inserted, composed of a half-wave plate, a polarizer
cube allowing the control of the laser power.
The power-controlled laser beam is then directed to a second half-wave plate for
polarization control thereof. A low-pass filter placed immediately after this second
blade makes it possible to suppress the light generated at the harmonic frequency
before the cell. The laser beam at the fundamental frequency is then focused in the cell
whose walls are made of quartz in order to obtain an excellent transmission at the
harmonic frequency around 400 nm by a microscope objective X10 with numerical
aperture NA = 0.25. The scattered intensity is then collected at right angles using a
lens with a focal length of 50 mm and separated from the fundamental intensity
scattered by a high-pass filter. A polarization analysis system, consisting of a
half-wave plate and a polarizing cube adapted to the wavelength of 400 nm, is placed
on the detection line in order to select a polarization state defined by the diffused
harmonic wave. The intensity at the harmonic frequency is then detected, using a
photomultiplier tube placed at the output of a spectrometer.
A spectrometer makes it possible to choose the range of length of the collected
signal and to obtain the TPEF spectrum or the SHG signal.
Recently, we developed a “tout-en-un” experimental setup able to record TPA,
TPEF cross sections along with the first hyperpolarizability b values for liganded
silver and gold quantum clusters [10–12, 58, 59].
The light source for the present TPA experiments was the same as the above.
The beam was gently focused by a 5 cm focal lens and sent in transmission into a
0.5 cm path length spectrophotometric cuvette. The transmitted light was detected
with a large aperture photodiode. The incident power was controlled with a
half-wave plate and a polarizing cube. The sample absorption was then determined
as a function of the incident power. Fluorescein was used as a reference to deter-
mine the beam waist at focus in particular. The calibration of the photodiode signal
was obtained with a neat water cuvette through a variation of the incident power. As
expected with a distilled water cuvette, the plot of the transmitted light versus the
incident light exhibits a linear behavior, whereas in the presence of the nanoclus-
ters, a decrease of the transmitted intensity is observed as the absorption increases
due to the nonlinear contribution (Fig. 5.7).
The transmission factor T of the sample can then be expressed as a function of
intensity using the standard Beer–Lambert law:
T ¼ CeaL ¼ Ceða þ dIÞL ð5:6Þ
where C is a constant. It is possible to extract d from the slope of the T(Pm) versus
Pm dependence.
The TPE cross sections of the silver or gold NC samples were calculated using
fluorescein in aqueous solution as a reference (Fig. 5.7):
gref rref
2 c
ref
I
rTPEF ¼ ð5:7Þ
c I ref
Here, the index ref denotes values related to the reference measurements, where
r is the integrated area of the two-photon emission intensity of the sample and the
reference ref, cref is the concentration of fluorescein, and c the concentration of the
NCs. For example, the TPE cross section of reference (fluorescein) at 780 nm is 24
GM according to Makarov et al. [60].
152 R. Antoine
Fig. 5.7 Schematic view and photograph of the classical 90° angle hyper-Rayleigh scattering
geometry. The setup was recently upgraded in a “tout-en-un” experimental setup able to record
TPA, TPEF cross sections along with the first hyperpolarizability b values for liganded silver and
gold quantum clusters
5.5 Case Studies
5.5.1 NLO Emission of Ag29(DHLA)12—Playing

with Photons and Colors
As a first example, we investigated in details the nonlinear optical response of

Ag29(DHLA)12. The two-photon excited fluorescence spectra were recorded, and
their first hyperpolarizability is obtained. A general observation is these NCs have
huge two-photon absorption cross sections typically 5 104 GM for Ag29 while
their corresponding two-photon excited fluorescence cross section is only 0.5 GM
[11, 12]. For NLO applications, they are still less effective than dyes, but usually
these clusters present a good biocompatibility and low toxicity. We have also
observed for the Ag29 an interesting behavior for potential applications, which
concerns its emission properties following the absorption of 1 or 2 photons, as
shown in Fig. 5.8. With an excitation with one photon at 400 nm or by two photons
at 800 nm, the resulting emission spectrum is either blue or red (Fig. 5.8).
Interestingly, this unique Ag29 nanocluster displays two colors, playing with the
number of photons involved in the excitation.
Fig. 5.8 (left) Spectroscopic characterization of the synthesized Ag29(DHLA)12 clusters dispersed
in water. Absorption spectrum (black), OPEF spectrum (top) and TPEF spectrum (bottom). The
arrows represent the wavelength of excitation by OPE and TPE processes. (right) Schematics
showing the proposed excited state relaxation dynamics in the Ag29(DHLA)12 clusters in TPEF
experiments
This behavior is still poorly understood and probably reflects the complexity of
the relaxation of excited electronic state characteristic of the quantum
regime.
For organic molecules, the emission typically occurs from its lowest excited
energy level for a given spin multiplicity (referred to as Kasha’s rule [61]).
However, nanoclusters, as well as certain organic molecules, tend to break this rule
and emit at more than one possible wavelength depending on the electronic
structure, density of states, and excited state dynamics involved. Regardless of
whether a molecule/nanocluster obeys the Kasha’s rule or not, in order to gain more
detailed understanding of the excited state emission dynamics of a chromophore,
one may carry out the time-resolved experiments. Note that the photochemistry of
gold nanoclusters is more complex than the de-excitation of p–p* states of organic
molecules, where excitations within the gold core as well as couplings with surface
states (through LMCT and LMMCT), singlet to triplet states conversion, may
occur.
154 R. Antoine
5.5.2 Bulky Counter-Ions—a Simple Route to Enhance

the TPEF Efficiencies
At this stage, we reckon that protected gold quantum clusters are excellent two-photon
absorbers but rather poor two-photon excited emitters [58]. To enhance emission effi-
ciencies, the ligand shell rigidity is an interesting strategy that would allow for enhanced
photon emission as compared to non-radiative relaxation upon photo-excitation. Pyo
et al. [21] have shown that it is possible to achieve OPEF quantum yield >60% by
rigidifying the metal–sulfur interface with the binding of bulky groups. We recently
pushed forward this concept to the nonlinear optical regime.
The strategy developed recently is to use bulky ammonium counter-ions, a
nitrogen cation surrounded by four alkyl chains. In solution, the luminescence of
the NCs is often affected by the solvent (in particular water). Counter-ions will
interact with the NC surface by electrostatic interaction (between the counter-ions
and the carboxylate groups of the glutathione ligands) to stick to the surface of the
NCs which will have the effect of protecting it from the environment and also to
“rigidify” its surface. The effect is spectacular on the one-photon excited fluores-
cence spectra, as shown on Au15 in water. When such NCs are complexed with
tetrabutylammonium ions, they become extremely fluorescent (Fig. 5.9). We
Fig. 5.9 (left) Schematic and structure of bulky counter-ions used to bind to glutathione-protected
gold clusters. (right) One-photon excited fluorescence spectra of Au15(SG)13 in water and
TBA-Au15(SG)13 in methanol (top); two-photon excited fluorescence spectra at excitation
wavelength 780 nm of Au15(SG)13 in aqueous solution compared to different bulky cations—
Au15(SG)13 in methanol (bottom)
Fig. 5.10 (left) TPEF intensity image of Au NCs containing methanol droplets in heptane with
(right) the corresponding optical image (size of the image: 637 637 lm); kexc = 780 nm
extended this study to the NLO regime and were able to show that the same
dramatic effect is observed for TPEF and with a variety of bulky counter-ions; the
best candidate to date is tetrabutylammonium (TBA). The TPEF cross sections of
these objects then become interesting for multi-photon optics, and we have bridged
the gap toward two-photon confocal microscopy.
TPEF measurements were made with a confocal microscope with 780 nm
excitation from a focused fs laser in a small well containing the NCs solution. We
carried out an XY mapping of the sample which allowed us to locate an interesting
area and then a Z-scan with a motorized microscope objective. The emitted signal
was collected in epifluorescence mode. As a proof of concept and before going to
cell imaging, we made an emulsion from a methanol–heptane mixture; it gives rise
to methanol droplets trapped in heptane of micrometric size. Au NCs in the pres-
ence of TBA counter-ions are soluble in methanol and were used to image the
droplets. The topographic and TPEF images are given in Fig. 5.10, where we
recognize the droplet containing the NCs.
5.6 Conclusions and Outlooks
The aim of the present study was to gain a fundamental knowledge on the mech-
anism involved in multi-photon processes in atomically precise gold and silver
nanoclusters. By measuring two-photon absorption, two-photon excited fluores-
cence cross sections, and first hyperpolarizability for such NCs, we have provided
unique benchmarks for theoretical modeling of the origin of enhanced NLO
properties of ultrasmall-ligated metal clusters, as well as the interplay between the
cluster core and the interface between the ligand shell and the metallic part
(ligand-core NLO-phores).
It will also serve as a basis for further developments in the design of
high-efficiency NLO-phores with good stability and low toxicity for in vitro (and
156 R. Antoine
possibly ultimately in vivo) bio-imaging applications. One of the technological

objectives of this study will be to bridge the gap between new cell biomarkers and
two-photon NLO-phores design.
Designing highly efficient second-order and third-order nonlinear optical
(NLO) chromophores is largely a matter of finely combining a high density of
delocalized electrons in a symmetrical or unsymmetrical environment. For
enhancing NLO efficiencies, innovative strategies should be explored in the future.
To enhance second-order v(2) efficiencies, metal core-doping is a promising strategy
as it distorts the metallic atom core [62]. Also, enhanced fluorescence of AuNCs by
silver doping was reported by Guevel et al. [63] and Brach et al. [64]. Therefore, to
enhance third-order v(3) efficiencies, silver doping can be coupled to ligand shell
rigidity as the latter favors the radiative properties of NCs. This route was recently
explored by Brach et al. [64] who showed that AuAgNCs exhibit two-photon
excited luminescence (2PL) emission and second-harmonic generation (SHG) and
that these properties remain the same in the liquid crystalline matrix. An alternative
route to enhance third-order v(3) efficiencies is to use the concept of “rigidity.” We
recently pushed forward this strategy consisting in using bulky counter-ions to
enhance the luminescence in the nonlinear optical regime. We showed that by an
appropriate choice of bulky counter-ions and of solvent, a 30-fold increase in TPEF
signal in the red for glutathione-protected gold clusters could be obtained. A simple
way to increase the rigidity of the protective shell is to replace thiolated ligands by
proteins instead. Bovine serum albumin-Au25NCs indeed exhibit an efficient
two-photon absorption followed by blue [65] to red to near-infrared photolumi-
nescence [66]. Using (bio)organic thiolate templates, further enhancement can be
achieved by increasing the rigidity of the metal–sulfur interface. The precise tai-
loring of the hydrophilicity/hydrophobicity balance on the gold or silver NC sur-
faces can be a way to increase the fluorescence signal [67, 68].
Acknowledgements Most of the work presented would not have been possible without the
fruitful collaborations with Isabelle Russier-Antoine, Franck Bertorelle, Željka Sanader, Marjan
Krstić, Philippe Dugourd, Pierre-François Brevet, and Vlasta Bonačić-Koutecký. Therefore, I wish
to express my deepest gratitude to them. Furthermore, I would like to acknowledge financial
support of the French-Croatian project “International Laboratory for Nano Clusters and Biological
Aging, LIA NCBA.”
References
1. de Meulenaere E, Nguyen Bich N, de Wergifosse M, van Hecke K, van Meervelt L,

Vanderleyden J, Champagne B, Clays K (2013) Improving the second-order nonlinear optical
response of fluorescent proteins: the symmetry argument. J Am Chem Soc 135:4061–4069
2. Terenziani F, Katan C, Badaeva E, Tretiak S, Blanchard-Desce M (2008) Enhanced
two-photon absorption of organic chromophores: theoretical and experimental assessments.
Adv Mater 20:4641–4678
3. Medintz IL, Uyeda HT, Goldman ER, Mattoussi H (2005) Quantum dot bioconjugates for
imaging, labelling and sensing. Nat Mater 4:435–446
4. Jin R, Zeng C, Zhou M, Chen Y (2016) Atomically precise colloidal metal nanoclusters and
nanoparticles: fundamentals and opportunities. Chem Rev 116:10346–10413
5. Polavarapu L, Manna M, Xu Q-H (2011) Biocompatible glutathione capped gold clusters as
one- and two-photon excitation fluorescence contrast agents for live cells imaging. Nanoscale
3:429–434
6. Tu H, Liu Y, Marjanovic M, Chaney EJ, You S, Zhao Y, Boppart SA (2017) Concurrence of
extracellular vesicle enrichment and metabolic switch visualized label-free in the tumor
microenvironment. Sci Adv 3:e1600675
7. Yue S, Slipchenko MN, Cheng JX (2011) Multimodal nonlinear optical microscopy. Laser
Photonics Rev 5:496–512
8. Blanchard-Desce M (2002) Molecular engineering of NLO-phores for new NLO micro-
scopies. CR Phys 3:439–448
9. Bertorelle F, Hamouda R, Rayane D, Broyer M, Antoine R, Dugourd P, Gell L, Kulesza A,
Mitric R, Bonacic-Koutecky V (2013) Synthesis, characterization and optical properties of
low nuclearity liganded silver clusters: Ag31(SG)19 and Ag15(SG)11. Nanoscale 5:5637–5643
10. Bertorelle F, Russier-Antoine I, Calin N, Comby-Zerbino C, Bensalah-Ledoux A, Guy S,
Dugourd P, Brevet P-F, Sanader Ž, Krstić M, Bonačić-Koutecký V, Antoine R (2017)
Au10(SG)10: a chiral gold catenane nanocluster with zero confined electrons. Optical
properties and first-principles theoretical analysis. J Phys Chem Lett 8:1979–1985
11. Russier-Antoine I, Bertorelle F, Kulesza A, Soleilhac A, Bensalah-Ledoux A, Guy S,
Dugourd P, Brevet P-F, Antoine R (2016) Chiral supramolecular gold-cysteine nanoparticles:
Chiroptical and nonlinear optical properties. Prog Nat Sci-Mater Int 26:455–460
12. Russier-Antoine I, Bertorelle F, Hamouda R, Rayane D, Dugourd P, Sanader Z,
Bonacic-Koutecky V, Brevet P-F, Antoine R (2016) Tuning Ag29 nanocluster light emission
from red to blue with one and two-photon excitation. Nanoscale 8:2892–2898
13. Bellina B, Compagnon I, Bertorelle F, Broyer M, Antoine R, Dugourd P, Gell L, Kulesza A,
Mitric R, Bonacic-Koutecky V (2011) Structural and optical properties of isolated noble
metal-glutathione complexes: insight into the chemistry of liganded nanoclusters. J Phys
Chem C 115:24549–24554
14. Bellina B, Antoine R, Broyer M, Gell L, Sanader Z, Mitric R, Bonacic-Koutecky V,
Dugourd P (2013) Formation and characterization of thioglycolic acid-silver cluster
complexes. Dalton T 42:8328–8333
15. Bonačić-Koutecký V, Kulesza A, Gell L, Mitric R, Antoine R, Bertorelle F, Hamouda R,
Rayane D, Broyer M, Tabarin T, Dugourd P (2012) Silver cluster-biomolecule hybrids: from
basics towards sensors/ perspective article. Phys Chem Chem Phys 14:9282–9290
16. Sanader Z, Mitric R, Bonacic-Koutecky V, Bellina B, Antoine R, Dugourd P (2014) Nature of
electronic excitations at the metal-bioorganic interface illustrated on histidine-silver hybrids.
Phys Chem Chem Phys 16:1257–1261
17. Sanader Z, Krstic M, Russier-Antoine I, Bertorelle F, Dugourd P, Brevet P-F, Antoine R,
Bonacic-Koutecky V (2016) Two-photon absorption of ligand-protected Ag15 nanoclusters.
Towards a new class of nonlinear optics nanomaterials. Phys Chem Chem Phys 18:12404–
12408
18. Russier-Antoine I, Bertorelle F, Calin N, Sanader Z, Krstic M, Comby-Zerbino C, Dugourd P,
Brevet P-F, Bonacic-Koutecky V, Antoine R (2017) Ligand-core NLO-phores: a combined
experimental and theoretical approach to the two-photon absorption and two-photon excited
emission properties of small-ligated silver nanoclusters. Nanoscale 9:1221–1228
19. Antoine R, Bonacic-Koutecky V (2018) Liganded silver and gold quantum clusters. Towards
a new class of nonlinear optical nanomaterials. Springer, Cham
20. Bonačić-Koutecký V (2017) Theoretical design of new class of optical materials based on
small noble metal nanocluster-biomolecule hybrids and its potential for medical applications.
Adv Phys X 2:695–716
158 R. Antoine
21. Pyo K, Thanthirige VD, Kwak K, Pandurangan P, Ramakrishna G, Lee D (2015) Ultrabright
luminescence from gold nanoclusters: rigidifying the Au(I)–thiolate shell. J Am Chem Soc
137:8244–8250
22. Jin R, Nobusada K (2014) Doping and alloying in atomically precise gold nanoparticles.
Nano Res 7:285–300
23. Kreibig U, Vollmer M (1995) Optical properties of metal clusters. Springer Verlag, Berlin
24. Bonačić-Koutecky V, Veyret V, Mitrić R (2001) Ab initio study of the absorption spectra of
Agn (n = 5–8) clusters. J Chem Phys 115:10450–10460
25. Harbich W, Fedrigo S, Meyer F, Lindsay DM, Lignieres J, Rivoal JC, Kreisle D (1990)
Deposition of mass selected silver clusters in rare gas matrices. J Chem Phys 93:8535–8543
26. Harbich W, Fedrigo S, Buttet J, Lindsay DM (1991) Optical spectroscopy on size selected
gold clusters deposited in rare gas matrices. Zeitschrift für Physik D 19:157–159
27. Harbich W, Fedrigo S, Buttet J (1992) The optical-absorption spectra of small silver clusters
(N = 5–11) embedded in argon matrices. Chem Phys Lett 195:613–617
28. Fedrigo S, Harbich W, Buttet J (1993) Collective dipole oscillations in small silver clusters
embedded in rare-gas matrices. Phys Rev B 47:10706–10715
29. Zheng J, Dickson RM (2002) Individual water-soluble dendrimer-encapsulated silver nanodot
fluorescence. J Am Chem Soc 124:13982–13983
30. Richards CI, Choi S, Hsiang J-C, Antoku Y, Vosch T, Bongiorno A, Tzeng Y-L, Dickson RM
(2008) Oligonucleotide-stabilized Ag nanocluster fluorophores. J Am Chem Soc 130:5038–
5039
31. Wu ZK, Jin RC (2010) On the ligand’s role in the fluorescence of gold nanoclusters. Nano
Lett 10:2568–2573
32. Diez I, Ras RHA (2011) Fluorescent silver nanoclusters. Nanoscale 3:1963–1970
33. Jin R (2010) Quantum sized, thiolate-protected gold nanoclusters. Nanoscale 2:343–362
34. Brust M, Walker M, Bethell D, Schiffrin DJ, Whyman R (1994) Synthesis of thiol-derivatised
gold nanoparticles in a two-phase liquid-liquid system. J Chem Soc, Chem Commun: 801–802
35. Udayabhaskararao T, Pradeep T (2013) New protocols for the synthesis of stable Ag and Au
nanocluster molecules. J Phys Chem Lett 4:1553–1564
36. Kumar S, Bolan MD, Bigioni TP (2010) Glutathione-stabilized magic-number silver cluster
compounds. J Am Chem Soc 132:13141–13143
37. Schaaff TG, Knight G, Shafigullin MN, Borkman RF, Whetten RL (1998) Isolation and
selected properties of a 10.4 kDa Gold: Glutathione cluster compound. J Phys Chem B
102:10643–10646
38. Negishi Y, Takasugi Y, Sato S, Yao H, Kimura K, Tsukuda T (2004) Magic-numbered Au-n
clusters protected by glutathione monolayers (n = 18, 21, 25, 28, 32, 39): Isolation and
spectroscopic characterization. J Am Chem Soc 126:6518–6519
39. Negishi Y, Nobusada K, Tsukuda T (2005) Glutathione-protected gold clusters revisited:
Bridging the gap between gold(I)-thiolate complexes and thiolate-protected gold nanocrystals.
J Am Chem Soc 127:5261–5270
40. Zheng K, Yuan X, Goswami N, Zhang Q, Xie J (2014) Recent advances in the synthesis,
characterization, and biomedical applications of ultrasmall thiolated silver nanoclusters. RSC
Adv 4:60581–60596
41. Zeng C, Chen Y, Das A, Jin R (2015) Transformation chemistry of gold nanoclusters: from
one stable size to another. J Phys Chem Lett 6:2976–2986
42. Yu Y, Chen X, Yao Q, Yu Y, Yan N, Xie J (2013) Scalable and precise synthesis of thiolated
Au10–12, Au15, Au18, and Au25 nanoclusters via pH controlled CO reduction. Chem Mater
25:946–952
43. Jin R (2015) Atomically precise metal nanoclusters: stable sizes and optical properties.
Nanoscale 7:1549–1565
44. Zhu M, Aikens CM, Hollander FJ, Schatz GC, Jin R (2008) Correlating the crystal structure
of a thiol-protected Au25 cluster and optical properties. J Am Chem Soc 130:5883–5885
45. Qian H, Eckenhoff WT, Zhu Y, Pintauer T, Jin R (2010) Total structure determination of
thiolate-protected Au-38 nanoparticles. J Am Chem Soc 132:8280–8281
46. Hamouda R, Bertorelle F, Rayane D, Antoine R, Broyer M, Dugourd P (2013) Glutathione

capped gold AuN(SG)M clusters studied by isotope-resolved mass spectrometry. Int J Mass
Spectrom 335:1–6
47. Lu Y, Chen W (2015) Application of mass spectrometry in the synthesis and characterization
of metal nanoclusters. Anal Chem 87:10659–10667
48. Chakraborty I, Pradeep T (2017) Atomically precise clusters of noble metals: emerging link
between atoms and nanoparticles. Chem Rev 117:8208–8271
49. Weerawardene KLDM, Aikens CM (2016) Theoretical insights into the origin of
photoluminescence of Au25(SR)18 nanoparticles. J Am Chem Soc 138:11202–11210
50. Weerawardene KLDM, Guidez EB, Aikens CM (2017) Photoluminescence origin of
Au38(SR)24 and Au22(SR)18 nanoparticles: a theoretical perspective. J Phys Chem C
121:15416–15423
51. Weerawardene KLDM, Aikens CM (2018) Origin of photoluminescence of Ag25(SR)18
nanoparticles: ligand and doping effect. J Phys Chem C 122:2440–2447
52. Fernando A, Weerawardene KLDM, Karimova NV, Aikens CM (2015) Quantum mechanical
studies of large metal, metal oxide, and metal chalcogenide nanoparticles and clusters. Chem
Rev 115:6112–6216
53. Devadas MS, Kim J, Sinn E, Lee D, Goodson T, Ramakrishna G (2010) Unique ultrafast
visible luminescence in monolayer-protected Au25 clusters. J Phys Chem C 114:22417–22423
54. Zheng J, Zhou C, Yu M, Liu J (2012) Different sized luminescent gold nanoparticles.
55. Birge RR, Pierce BM (1979) A theoretical analysis of the two-photon properties of linear
polyenes and the visual chromophores. J Chem Phys 70:165–178
56. Przhonska OV, Webster S, Padilha LA, Hu H, Kachkovski AD, Hagan DJ, Van Stryland EW
(2010) Two-Photon absorption in near-IR conjugated molecules: design strategy and
structure-property relations. In: Demchenko AP (ed) Advanced fluorescence reporters in
chemistry and biology I: fundamentals and molecular design. Springer, Berlin, pp 105–147
57. Hu Z, Jensen L (2017) Importance of double-resonance effects in two-photon absorption
properties of Au25(SR)18. Chem Sci 8:4595–4601
58. Russier-Antoine I, Bertorelle F, Vojkovic M, Rayane D, Salmon E, Jonin C, Dugourd P,
Antoine R, Brevet P-F (2014) Non-linear optical properties of gold quantum clusters. The
smaller the better. Nanoscale 6:13572–13578
59. Bertorelle F, Moulin C, Soleilhac A, Comby-Zerbino C, Dugourd P, Russier-Antoine I,
Brevet P-F, Antoine R (2018) Bulky counterions: enhancing the two-photon excited
fluorescence of gold nanoclusters. Chem Phys Chem 19:165–168
60. Makarov NS, Drobizhev M, Rebane A (2008) Two-photon absorption standards in the 550–
1600 nm excitation wavelength range. Opt Express 16:4029–4047
61. Kasha M (1950) Characterization of electronic transitions in complex molecules. Discuss
Faraday Soc 9:14–19
62. van Steerteghem N, van Cleuvenbergen S, Deckers S, Kumara C, Dass A, Hakkinen H,
Clays K, Verbiest T, Knoppe S (2016) Symmetry breaking in ligand-protected gold clusters
probed by nonlinear optics. Nanoscale 8:12123–12127
63. le Guevel X, Trouillet V, Spies C, Li K, Laaksonen T, Auerbach D, Jung G, Schneider M
(2012) High photostability and enhanced fluorescence of gold nanoclusters by silver doping.
64. Brach K, Waszkielewicz M, Olesiak-Banska J, Samoc M, Matczyszyn K (2017) Two-photon
imaging of 3D organization of bimetallic AuAg nanoclusters in DNA matrix. Langmuir
33:8993–8999
65. Kindi HA, Mohamed A, Kajimoto S, Zhanpeisov N, Horino H, Shibata Y, Rzeznicka II,
Fukumura H (2018) Single bovine serum albumin molecule can hold plural blue-emissive
gold nanoclusters: a quantitative study with two-photon excitation. J Photoch Photobio A
357:168–174
160 R. Antoine
66. Raut SL, Shumilov D, Chib R, Rich R, Gryczynski Z, Gryczynski I (2013) Two photon
induced luminescence of BSA protected gold clusters. Chem Phys Lett 561:74–76
67. Porret E, Sancey L, Martín-Serrano A, Montañez MI, Seeman R, Yahia-Ammar A, Okuno H,
Gomez F, Ariza A, Hildebrandt N, Fleury J-B, Coll J-L, Le Guével X (2017) Hydrophobicity
of gold nanoclusters influences their interactions with biological barriers. Chem Mater
29:7497–7506
68. Shen D, Henry M, Trouillet V, Comby-Zerbino C, Bertorelle F, Sancey L, Antoine R, Coll
J-L, Josserand V, Guével XL (2017) Zwitterion functionalized gold nanoclusters for
multimodal near infrared fluorescence and photoacoustic imaging. APL Materials 5:053404
Chapter 6
Small and Large Molecules
Investigated by Raman Spectroscopy
From Conformational Study
to Biomedical Applications
Krzysztof Czamara, Ewelina Szafraniec, Ewelina Wiercigroch,

Szymon Tott, Grzegorz Zając, Ewa Machalska, Monika Dudek,
Dominika Augustynska, Kamilla Malek, Agnieszka Kaczor
and Malgorzata Baranska
Abstract This chapter presents selected techniques of Raman spectroscopy, i.e.

Raman imaging, Raman optical activity (ROA), and surface-enhanced Raman
spectroscopy (SERS), and gives an overview on their biomedical applications. The
current state of the art in the research on chiroptical compounds of biomedical
importance, as well as the study on early apoptosis and inflammation processes
occuring in the endothelium, is presented. The pathophysiology of the endothelium
is discussed based on the example of Raman imaging results for primary cells and
cell cultures. Moreover, the comparison of classical Raman imaging, application of
optical fiber probes, and immuno-SERS nanosensors in detection of marker proteins
in ex vivo studies is discussed.
6.1 Molecules of Biomedical Importance—Structure

and Interactions
In this section, we focus on principles of two Raman techniques which rely on

registration of light scattering generated due to interactions with optically active
chemical individuals (ROA, Raman optical activity) and with roughened metal
surfaces (SERS, surface-enhanced Raman spectroscopy). Active modes observed in
K. Czamara E. Szafraniec E. Wiercigroch S. Tott G. Zając

E. Machalska M. Dudek K. Malek A. Kaczor M. Baranska (&)
Faculty of Chemistry, Jagiellonian University, Krakow, Poland
e-mail: baranska@chemia.uj.edu.pl
K. Czamara D. Augustynska K. Malek A. Kaczor M. Baranska
Jagiellonian Centre for Experimental Therapeutics, Jagiellonian University,
Krakow, Poland

162 K. Czamara et al.
their spectra provide an additional insight into molecular structure when compared
to conventional Raman spectroscopy. We also summarize the latest applications of
ROA and SERS for modelling of inter- and intramolecular interactions of selected
biomedical systems.
6.1.1 The Basis of the Raman Optical Activity (ROA)
Raman optical activity (ROA) is one of two fundamental forms of vibrational

optical activity (VOA), complementary to vibrational circular dichroism (VCD),
and the most powerful structural probe of molecular chirality for the study of
various types of molecules in their natural environment. ROA spectroscopy can be
effectively used in a stereochemical analysis and in determining the absolute con-
figuration of molecules. This chiroptical method is based on the small intensity
difference in Raman scattering of right and left circularly polarized light (from
near-infrared to ultraviolet region) by optically active molecules. The primary
scattering mechanism accountable for ROA phenomenon was discovered in 1969
by Atkins and Barron [1], and then, two years later, Barron and Buckingham
provided the theoretical foundation for this original effect by means of dimen-
sionless circular intensity difference (CID), D, which was defined as [2]:
IR IL
CID ¼ D ¼ ð6:1Þ
IR þ IL
where IR and IL are the scattered intensities in right and left circularly polarized
incident light.
There are four circular polarized ROA strategies, depending on the polarization
of the incident and the scattered beams, including incident circular polarization
(ICP), scattered circular polarization (SCP), and in-phase or out-phase dual circular
polarization (DCPI and DCPII, respectively). The first ROA spectrum was obtained
for (+)-a-phenylethylamine by Barron et al. [3]. In the years 2004–2005, majority
of announced ROA spectra of biomolecules were measured using ICP-ROA in
backscattering instruments [4]. Nowadays, the backscattering SCP-ROA is the most
common form of ROA, measured in a commercially available chiral Raman
spectrometer from BioTools, Inc. [5]. It is worth to note that numerous laboratories
possess homemade ROA spectrometers committed for particular aims [6–13].
Moreover, due to fact that the ROA spectrum is a spectrum of intensity differences,
the ROA effect is a fairly weak process; hence, ROA intensities are small, usually
three or four orders of magnitude smaller than in a conventional Raman spectrum.
Therefore, long acquisition times, highly concentrated samples, and adequately
selected laser powers are needed to record a ROA spectrum [14, 15].
6 Small and Large Molecules Investigated by Raman Spectroscopy 163
6.1.2 ROA Research on Chiral Compounds and Its

Applications, Now and Over the Years
ROA spectroscopy provides information about structure, dynamics, and behavior of

chiral compounds. Therefore, it allows determining an absolute configuration as
well as studying conformational equilibrium of chiral molecules [16]. ROA spectra
of two enantiomers are mirror reflections like their structures. Moreover, ROA
enables to measure an enantiomeric excess in samples containing unequal amounts
of mirror image enantiomers. Hecht et al. [17] described a method for determination
of the enantiomeric composition with the accuracy of 0.1%, using a-pinene
enantiomers mixture as a test compound. On account of the exquisite sensitivity to
chirality, ROA is also applied to study conformational changes of important ‘life’
molecules like carbohydrates, proteins and nucleic acids, all in the aqueous solution
to reflect their natural environment (Fig. 6.1) [16]. ROA spectroscopy also appears
to be useful even in the structural studies of viruses. A few virus ROA spectra were
recorded, and the first one was reported for filamentous bacteriophages by Blanch
et al. in 1999 [18].
A wide range of carbohydrates was studied by ROA [19–21] providing a solid
base of structural information, often inaccessible in other spectroscopic methods.
ROA spectra reveal in detail the stereochemical arrangement of carbohydrates in
the aqueous solution, including the anomeric configuration, the ring conformation,
the relative orientation of the substituents around the ring, the absolute configura-
tion at anomeric centers and the glycosidic link configuration.
Studying protein structure, the ROA technique has much more to offer in
comparison with standard X-ray diffraction (XRD) and nuclear magnetic resonance
(NMR) spectroscopy. Various degrees of protein folding can make crystallization
difficult, or the structures are too large to be accessible to mentioned methods.
Protein ROA spectra reveal characteristic spectral patterns, sensitive to secondary
and tertiary structures of the polypeptide backbones, backbone hydration, and side
chain conformations, as well as structural elements in denaturation states [22].
Recently, Kessler et al. [23] shed a new light on ROA as a convenient optical tool to
monitor folding of proteins into amyloid fibrils, involved in serious medical dis-
orders as neurodegenerative diseases or diabetes type 2. ROA spectra of amyloidal
insulin, supported for the first time by the theoretical model, were presented. So far,
ROA has not been commonly used to study such structural variations due to
experimental artifacts.
Another valuable application of ROA is the structural evaluation of nucleic
acids, especially when it comes to study interaction of DNA with drugs.
Gąsior-Głogowska et al. [24] demonstrated that ROA spectroscopy is a sensitive
technique to follow the structural alteration of the DNA molecule upon binding
with anticancer drug cisplatin. In general, ROA spectra of nucleic acid provide
information on the sugar ring conformation, the base-stacking arrangement of base
rings, and mutual orientation of the sugar and base rings.
Fig. 6.1 Raman and ROA

spectra of selected
biomolecules. Representative
spectra of D-(+)-xylose, DNA
from herring sperm and
human lysozyme with
assigned marker bands
The computational approach plays a relevant role in complete interpretation of

experimental ROA spectra. The necessity of comparison of experimental data with
theoretical simulations results from the fact that there are no empirical rules to
predict the sign and intensity of recorded ROA signals. Currently, density func-
tional theory (DFT) has become the substantial scheme for theoretical calculations
of ROA spectra. The first entire ab initio calculations of ROA effect were suc-
cessfully performed by Polavarapu et al. [25, 26], while foremost calculations of
ROA CIDs, by means of DFT theory, were introduced by Ruud et al. [27].
Recently, sophisticated attempts of ROA calculations with the help of Cartesian
coordinate transfer (CCT) or molecules in molecules (MIM) methodology enable to

predict ROA spectra of much bigger structures, for instance cyclodextrins [28],
globular proteins [29] or even protein fibrils [23].
6.1.3 Sophisticated Applications of the ROA Techniques
The development of new ROA techniques and applications has been recently
entering a higher and higher level. New techniques, such as surface-enhanced
Raman optical activity (SEROA) [30] and coherent anti-Stokes Raman spec-
troscopy–Raman optical activity (CARS-ROA) [31], were performed in biomed-
ical applications, e.g., cancer detection [32] and sophisticated measurements like
chirality induction of solvent, [33] ROA of gas phase samples [34], diamagnetic
ROA of iodine and bromine gases [35], paramagnetic ROA of NOx gases [36], and
much more, shedding a new light on the analysis of molecular systems and their
interactions. Moreover, ROA spectrometers recently become useful also for cir-
cularly polarized light (CPL) measurements [37].
Another sophisticated methods are related to the attempts of ROA signal
enhancement. As ROA signal is very weak, measurements of low concentrated or
unstable chiral samples are very demanding, and therefore ROA signal enhance-
ment methods are of interest. One of the attempts to enhance the ROA signal is
resonance Raman optical activity (RROA). Similar to resonance Raman (RR),
strong enhancement of the RROA intensity is observed when the incident laser
wavelength coincides with one of the electronic transitions of a molecule. In case of
RROA, an absorbing chromophore needs to be chiral or in chiral environment;
therefore, the presence of rotatory strength of related electronic transition, charac-
terized by electronic circular dichroism (ECD) intensity, is necessary. Single
electronic state (SES) limit theory of Nafie [38] predicts monosigned RROA spectra
and with the opposite sign to the corresponding ECD band. What is more, relative
intensity of RROA bands is the same as in the parent RR spectrum and CID value
equals to −g/2 in the ICP and SCP-ROA strategy, where g is UV–Vis/ECD ratio of
relevant electronic transition [38]. First, RROA spectra were recorded for small
organic molecules [39] and metal complexes [40], however, the newest attempts
demonstrated that it is also possible for proteins [41] and carbon nanotubes [42].
Another technique of RROA is aggregation-induced resonance Raman optical
activity (AIRROA) observed recently by Zajac et al. in induction of strong RROA
signal after aggregation of carotenoids [43]. AIRROA spectra are monosigned and
opposite in sign to the relevant ECD band, in agreement with the SES theory—
however, CID values are rather comparable to the UV–Vis/ECD ratio; hence, twice
the value is predicted by the SES theory [38]. As carotenoid monomers do not
possess pronounced ECD signal in the spectral region of polyene chain chro-
mophore absorption, measurements of RROA in use of a 532-nm laser line are
rather impossible. After aggregation, in mixed water and organic solvent solutions,
strong induction of ECD signal in the discussed region is observed due to the chiral
Fig. 6.2 Raman and AIRROA spectra of H-aggregates formed by astaxanthin enantiomers.
Adapted with permission from [44], copyright (2016) American Chemical Society
arrangement of polyene chains from neighboring molecules. This enables to collect

RROA spectra after few-minute acquisition with a relatively low laser power (*10
mW). AIRROA was measured for the first time for astaxanthin enantiomers
(Fig. 6.2) [43, 44] and was confirmed for several other compounds [45, 46].
6.1.4 The Basis of Surface-Enhanced Raman Spectroscopy

(SERS)
Since 1974 when Fleishmann and co-workers observed enhanced Raman scattering
of pyridine molecules adsorbed on electrochemically roughened silver electrodes,
surface-enhanced Raman spectroscopy has developed into a mature and advanced
field with a special interest in biomedical applications. The reader can find several
extensive books and reviews discussing in detail the SERS phenomena [47–49];
here, we only briefly introduce its mechanism. To enhance the Raman signal, the
interaction of light with silver or gold nanoparticles (NPs) is required to generate
localized surface plasmon resonances (LSPRs). When a small metal nanoparticle of
size between 10 and 100 nm is irradiated by electromagnetic field, the conduction
electrons oscillate coherently at a plane of metal–dielectric interface, producing

huge electromagnetic field and consequently enhancing Raman scattering (elec-
tromagnetic mechanism). The oscillation frequency depends on dielectric function
of Ag/Au nanoplasmonic structures and is also associated with their shape and size.
The site of electromagnetic enhancement is ‘hot spots’ formed upon aggregation of
nanoparticles (Fig. 6.3a). According to calculations of local field enhancement for
nanospheres, a gap of ca. 2 nm between particles contributes to the largest gen-
eration of electromagnetic field. In this situation, molecules present at very short
distances (d) from a surface (a few nanometers) undergo Raman processes in an
environment where the laser excitation light is enhanced, cf. Fig. 6.3a. This
interaction relies on physisorption and has a d−3 dependence. A SERS spectrum
derived from electromagnetic mechanism is usually similar to a conventional
Raman spectrum of a neat substance; active SERS bands determine the orientation
of a molecule on NPs according to surface selection rules [49]. Since chemical
species (e.g., thiols) can exhibit a high affinity to Ag or Au surfaces, the process of
chemisorption can also occur yielding to an additional increase of enhancement
factor (EF), order of *102. This mechanism is accompanied by the transfer of an
electron between energy levels of the adsorbate and NPs (see Fig. 6.3b). The
contribution from resonance Raman scattering of the adsorbate–metal complex is
also possible [49]. A SERS spectrum in this case usually differs from the con-
ventional Raman spectrum.
Fig. 6.3 A schematic of electromagnetic (a) and chemical (b) mechanisms of SERS phenomenon;
(c) SEM images of Ag hydrogel (up), roughened Ag electrode (middle), and honeycomb
nanostructure covered with 5-nm-thick Ag layer (down)
There is a plethora of SERS substrates: hydrosols (colloids) easily synthesized

by wet chemistry methods, roughened electrodes, metal islands or periodic
nanostructures (Fig. 6.3c). Advances in nanoscience and nanotechnology allow
designing of various sizes and shapes of nanoparticles, from common nanospheres
to nanoflowers or nanopillars, and composed of single metals or their alloys. The
aim is engineering of LSPRs providing stable, reproducing, and high EF as *1010
[48].
6.1.5 Applications of SERS in Structural Studies—A Case

of Pharmaceuticals
The fundamental information gathered from SERS spectra gives an insight into
interactions between the adsorbate and enhancing metallic nanoparticles, including
spatial orientation and polarization of the local electric field. This fact has opened
new perspectives in studies on behavior of molecules of a pharmaceutical interest
under different physiological conditions (e.g., pH, concentration) located close to a
surface. Quantum-chemical calculations support the interpretation of SERS spectra
of drugs, however, they can be reliably applied when the chemical mechanism
strongly contributes to surface enhancement of the adsorbate and must include a
model of metal clusters [50].
SERS studies have been carried out on a wide diversity of pharmaceutical
compounds, e.g., analgesics, antidepressants, antibiotics, vitamins as well as on
potential drugs whose SERS characteristics are compared with their precursors [47–
49]. Here, we present briefly two groups of pharmaceuticals.
A series of tricyclic antidepressants TCA (imipramine, desipramine, clomipra-
mine, amitriptyline, nortriptyline, and doxepine) known as surface-active drugs has
been used as a model for evaluation of their adsorption mechanism on the metal
substrate and its relationship to mechanism of interaction with receptors, cf. Fig. 6.4
[51, 52]. SERS spectra clearly showed that the main site of interaction of all the
drugs is the p-electron system and the methyl-aminopropyl side chain, similarly to
docking of these ligands in serotonin and leucine transporters, where the ring
system is located almost perpendicularly in the outer vestibule of the receptor.
However, orientations of each drug are slightly different, so specific for each of
them (Fig. 6.4). Integral intensity ratios of bands assigned to the breathing mode of
the dibenzazepine ring and the 8a mode of the phenyl ring were found to be
different for particular groups of the drugs indicating changes in the orientation of
the tricyclic ring. For instance, the comparison of SERS features of clomipramine
(Clo) with imipramine (Imi) and desipramine (Des) indicated that the ring of Clo is
less tilted than for Imi and Des due to the substitution of the tricyclic ring by the
chlorine atom (Fig. 6.4). Živanovic et al. [53] continued SERS study on selected
TCAs observing them in vitro after the delivery by a nanoparticles-drug system.
Another example of the application of SERS in the examination of drug-target
Fig. 6.4 RS and SERS spectra of imipramine (Imi), desipramine (Des), and clomipramine (Clo)
and the proposed model of the adsorption mechanism of these antidepressants (colors of atoms
correspond to: gray—carbon, white—hydrogen, blue—nitrogen, green—chlorine, red—oxygen).
Adapted with permission from [51], copyright (2014) Elsevier
mechanism is a report of Cîntǎ-Pînzaru et al. [54], who probed interactions of

antiomalarials (chloroquine and mefloquine) with hematin by colloidal AuNPs.
Jaworska et al. [55–57] have investigated in detail SERS features of nicoti-
namide and its seven derivatives, i.e., trigonelline, nicorandil, 1-methylnicorandil,
1-methylpyridinium, dimethylpyridinium chlorides proposed as potential thera-
peutics for treatment of diseases resulted from endothelial dysfunction. These works
have shown SERS applicability in a rapid examination of similarities in surface
behavior of a group of potential drugs possessing the identical core structure.
Sensitivity of surface-enhanced Raman spectroscopy to determine differences in the
adsorption mechanism due to exchange of functional groups, variation of pH, and
anions in the pyridinium cation has been clearly proven. Each precursor adopts a
specific orientation with respect to the metal which is strongly dependent on the
functional groups attached to the pyridinium ring rather than on its aromatic system.
Both presented here Raman-based techniques are demanding but very promising
and useful in the research of molecular structure and intramolecular interactions.
Over a dozen attempts to combine SERS and ROA or RROA (SEROA or
SERROA) are presented in the literature. SEROA is even more demanding than
RROA but constitutes a perfect probe of chirality in diluted samples. Most of
SEROA studies were focused on interactions between a chiral analyte and metal
nanoparticles [30]. Recently, Blanch et al. [58] showed a new chirality induction
mechanism; i.e., the SERROA signal was collected for an achiral benzotriazole dye
reporter placed in the vicinity of achiral silver nanoparticles and induced by a chiral
analyte.
6.2 Application of Raman Spectroscopy to In Vitro

Endothelial Cell Cultures
From small molecules, through simple systems to highly sophisticated ones such as
single cells or tissues, Raman spectroscopy enables deciphering the structural and
chemical composition. A strong potential of this technique is in providing infor-
mation from various samples important in biology and medicine noninvasively and
in a label-free manner. Identifying, interpreting, and understanding of the studied
samples are possible based on basic knowledge about the structure of compounds
and analysis of characteristic bands in their Raman spectra. For example, investi-
gation of individual proteins [59] can be used to determine and define alterations in
the secondary structures of proteins in tissues [60] and spectroscopic analysis of
plant metabolites enables discrimination of different plant species and even
chemotypes [61]. Some biologically active compounds, for example hemoproteins
and lipids, exhibit very intense Raman signals due to resonance enhancement
(hemoproteins if irradiated with the wavelength in the suitable range) or the
molecular structure (the presence of long, nonpolar hydrocarbon chains in lipids).
Therefore, lipids are ideal specimens for Raman studies due to the large Raman
scattering cross section [62]. In case of biomedical applications, their analysis is
particularly important because changes in the lipid concentration and composition
are associated with many diseases, i.e., atherosclerosis or diabetes. On the whole,
the knowledge of the most characteristic marker features for groups of different
biocompounds facilitates analysis of complex Raman spectra of cells and tissues.
There are many analytical techniques for in vitro imaging. Historically, the first
attempts to visualize cells were based on fixed preparations initially using optical
microscopy and later electron microscopy. The latter technique enables imaging of
intracellular structures and organelles with a very high resolving power, however,
provide only morphological information. Similarly, development of microscopic
techniques in a bright field, based on phase contrast and differential interference
field, made possible observation of samples without fixation and even staining, and
imaging of living cells. The use of these techniques enables real-time imaging, but
they do not provide information about the chemical composition of cells. Modern
techniques of fluorescence microscopy yield highly resolved images, nevertheless,
only a limited number of components can be studied in one experiment, and
application of these methods requires long-lasting and laborious preparations of
samples. These problems do not concern Raman microscopy. The information
contained in each Raman spectrum provides knowledge about the structure and
chemical composition of the studied specimens. A number of advantages of this
analytical label-free technique make it more and more often used for studying
biological and biomedical systems. Due to the small scattering cross section for
water, the aqueous environment, convenient for tissue and cell measurements, can
be used. The high resolving power of Raman microscopy makes possible moni-
toring changes in the content of biocomponents in the subcellular scale and
determining their distribution in the organelles and the main cellular compartments.
All above-mentioned advantages make Raman microscopy an excellent tool for
investigating cellular processes and for future diagnostics of diseases [63].
Nowadays, the prevention and treatment of lifestyle diseases, such as
atherosclerosis or diabetes, are extremely hot topics and challenges for modern
medicine. A large amount of work is devoted to new, innovative diagnostic
methods and treatment regimens. However, future progress seems impossible
without deeper knowledge about mechanisms related to onset and development of
these pathologies. It is postulated that most of the cardiovascular diseases start with
endothelial dysfunction. The endothelium is a monolayer of cells, a barrier between
the blood and the inner part of the vessel and a very active regulative organ.
Considering its location, it is an integral and important structure in maintaining the
proper functioning of the vessel wall and the entire circulatory system [64]. At the
same time, it is extremely susceptible to influence pathological factors, and thus
endothelial cells are a very convenient and suitable model for in vitro studies of
development of circulatory system pathologies.
6.2.1 Characteristics of Endothelial Cell Cultures
Endothelial cells in vitro enable studying various physiological and pathological

processes and can be divided into three groups: primary cells, prolonged life cells,
and cells from immortalized cell lines [65]. The first one refers to cells directly
isolated from the tissue using enzymatic or mechanical methods. However,
endothelial cells are diverse. They exhibit morphological, physiological, and phe-
notypic differences resulting from the location in the arterial tree as well as due to
their origin, i.e., arteries or veins [64]. Their heterogeneous nature is observed within
the living organism in vivo as well as in in vitro cultures [66]. It is generally
accepted that primary cells are most suitable for studying biological processes that
are related to specific in vivo systems; however, the lifespan of such cells is very
limited. Individual variability, short lifetime, and time-consuming primary cell iso-
lation have prompted researchers to modify primary cultures. A successfully used
approach is introduction of foreign DNA/RNA, i.e., large T antigen of SV40 virus
[67] into cells by genetic transfection or mutation, or artificial modifications, i.e.,
fusion with cancer cells. Resulting cells have a stable, homogenous phenotype and
genotype, and can also be cultured almost indefinitely [65]. In the group constituting
the immortalized, commercially available cells, subgroups according to their
origin may be separated. The first subgroup is made up of cells from large blood
vessels, among which the best-defined and the most commonly used in endothelial
research is the EA.hy926 cell line [68]. The second subgroup is composed of
endothelial cell lines for the study of microvasculature with the first characterized
HMEC-1 (human microvascular endothelial cells) line from the foreskin microcir-
culation [69]. Compared with other culture lines, they proved to be the best model
for endothelial research, because they most closely reflected the processes observed
in primary cells [70].
6.2.2 In Vitro Cell Models of Pathophysiology

of the Endothelium
The number of published papers on endothelial studies using microscopic and

spectroscopic methods is rather small, and the majority of these works deal with the
endothelium in the blood vessel wall [71]. Thematically, they can be divided into
works addressing the general characterization of the endothelium and referring to
the disease states, mainly in murine models [60, 71–76]. The Raman imaging
combined with atomic force microscopy (AFM) and immunohistochemical staining
was used to characterize the chemical composition and topography of the
endothelium [75]. In vivo Raman studies on the endothelium in the aorta wall of the
sheep and in vitro in the human aorta have also been published [77]. In addition,
spectroscopic methods were used in studies of endothelial pathologies:
atherosclerosis [76, 78], hypertension [60, 73], diabetes type 2 [74] and cancer
metastasis [72]. In these studies, changes in the protein and/or lipid content
observed in the endothelial layer were treated as a marker of endothelial condition
in diseases and, therefore, could be used in diagnostic context. Analysis of alter-
ations arising from disease development requires knowledge about normal,
non-disease-affected samples. Raman microscopy was used for comparison of the
control EA.hy926 endothelial cells in reference to the endothelium of the murine
aortic wall [79]. This work involved Raman 3D profiling combined with the cluster
analysis (CA) to analyze the biochemical composition and endothelial morphology,
as well as to define the size, shape and composition of the main cell compartments
and organelles.
Many works on endothelial cells explore the topic of infecting the endothelium
with bacteria or parasites [80–82]. Studies on the interaction of Staphylococcus
aureus bacteria with EA.hy926 cells enabled to visualize the location and spectral
differences in the Raman spectral profile of cells after the bacteria invasion [81].
Presented research provide the chance to develop new preventive and therapeutic
strategies against intracellular bacteria. Other studies focus on the protozoan
Neospora caninum that attacks vertebrates and remains inside the brain and other
tissues for whole lifetime. In the study by Elsheikha and Kong [80, 82], Raman and
fluorescence imaging supported by biochemical tests were applied to analyze the
location of parasites and define metabolic and cellular changes associated with their
presence in human endothelial cells of cerebral microvessels. The chemometric
approach enabled to distinguish healthy cells from infected ones on the basis of
alterations in the lipids and nucleic acids composition. These research show that
endothelial cells are the target of the attack by the protozoa.
The largest number of endothelial cell research focus on the analysis of changes
in their lipid profiles. Two original papers deal with the formation of lipid droplets
(LDs) in human aortic endothelial cells (HAoECs) under the influence of polyun-
saturated fatty acids [83, 84]. These studies are based on Raman and fluorescence
microscopies. Due to the high resolving power of the methods, the size and dis-
tribution of LDs in the cytoplasm of cells in three-dimensional projections were
determined. The use of CA and comparison with Raman spectra of lipid standards
enabled to define the chemical composition of LDs. The appearance of character-
istic bands at 1660 and 3015 cm–1 originating from the stretching vibrations of the
C=C and =C–H groups, respectively, and the calculated values of the degree of
unsaturation confirmed the presence of unsaturated fatty acids and their storage in
the form of LDs. In addition, the effect of 1-methylnicotinamide (MNA) on the
uptake of arachidonic acid (AA) sodium salt was checked [84]. The evidence has
been provided that MNA may act as a cationic carrier to facilitate the uptake of
negatively charged AA ions, which may have physiological and pharmacological
significance. The topic of LDs was also investigated by Astantina et al. [85], who
provided a further evidence of the active LDs function, concluding that LDs are
involved as carriers in intercellular communication between HMEC-1 cells.
The formation of LDs in the cytoplasm of cells is associated also with the
inflammatory process. Inflammation in cellular models has been studied using many
spectroscopic and microscopic methods. A review on methods of imaging LDs,
starting from classical methods involving staining with fluorescence detection to
modern imaging methods, e.g., coherent anti-Stokes Raman spectroscopy (CARS),
is provided in the work by Melo et al. [86]. Some reports on the use of AFM
technique for the study of endothelial cells and macrophages can be also found [87,
88]. The inflammation of endothelial cells is the first symptom of blood vessel
pathology in cardiovascular diseases. The body’s aim is to reduce pathogen inva-
sion and damage after tissue disruption. On the cellular scale, it is the response to
proinflammatory factors, i.e., tumor necrosis factor alpha (TNF-a) or lipopolysac-
charides (LPS) followed by various mechanisms activated via connection of
inflammatory factors to the specific membrane receptors [89]. All these factors
trigger activation of NF-jB protein pathway, which transmits information about the
pathogen to the cell nucleus, from which the inflammatory response occurs.
Endothelial cells also show increased synthesis of PGI2 prostacyclin, which triggers
the arachidonic acid pathway through a series of prostaglandins. PGI2 is considered
a marker of inflammation. Spectroscopic studies shed the new light on this process.
Czamara et al. [90] applied 3D Raman profiling combined with chemometric
analysis to study the in vitro model of inflammation. In this work, HMEC-1 cells
exposed to TNF-a exhibited formation of two types of LDs: more saturated,
composed mainly of phosphatidylcholine and saturated cholesteryl esters and much
Fig. 6.5 Confocal 3D imaging of control and TNF-a-stimulated HMEC-1 cells. Raman images of
distribution for control (a) and stimulated with TNF-a (b) cells were obtained from layers every
1 lm step in the z-direction by integration in the region of 3030–2830 cm−1 (organic matter), 810–
760 cm−1 (DNA and RNA), 3030–3000 cm−1 (unsaturated lipids), and 715–695 cm−1 (choles-
terols). Intensities of bands between layers were not normalized. Adapted from [90], open access:
Creative Commons Attribution (CC BY) license
more numerous LDs containing lipids of unsaturated characteristics and choles-

terols (Fig. 6.5) [90]. The highly unsaturated characteristics of lipids in LDs was
confirmed by calculations of the degree of unsaturation [62] what strongly sup-
ported the thesis that formation of LDs was related to the increase in production of
prostacyclins, polyunsaturated lipids being endogenous mediators of inflammation.
Further research was performed on HMEC-1 cells by incubation with another
proinflammatory factor—lipopolysaccharides, endotoxins extracted from the cell
wall of gram-negative bacteria Escherichia Coli. Inside those cells, two types of
LDs were also found, however, in an inverted abundance. In conclusion, the for-
mation of LDs composed of unsaturated lipids is proposed as a universal marker of
inflammation.
Raman spectroscopy has been also successfully used to study the process of
apoptosis in various types of cells and under the influence of various factors, e.g.,
chemical compounds [91], including cytotoxic drugs [92, 93] or harmful ultraviolet
radiation [94]. A large part of the reports focuses on investigation of the effect of
anticancer drugs and their chemotherapeutic response in anticancer therapy [92, 95,
96] or is dedicated to the classification and differentiation of cells in apoptosis [97–
99], as well as a detailed description of biochemical changes in this process [100–
103]. Studies on programmed cell death using Raman spectroscopy enabled
defining a number of characteristic changes in the intensity of bands resulting from
changes in the level of bioconstituents. Fazio et al. [104] analyzed the molecular
changes associated with cell aging in the in vitro leukemia model in relation to the
control cell line. Using Raman spectroscopy in correlation with fluorescence
microscopy, markers of the cell aging process and subcellular changes in cellular
organization were found. Okada et al. [105], based on Raman and fluorescence
imaging, studied the cytochrome c alterations in the process of apoptosis. An
extremely interesting work by Brauche et al. [99] raises the problem of distin-
guishing the processes of apoptosis and necrosis. Although both processes have
different characteristics, the identification of apoptotic and necrotic cells in in vitro
cultures is a challenge. In this study, cells of two different cell lines for early and
advanced apoptosis and necrosis were analyzed using a combination of Raman
spectroscopy and fluorescence microscopy. Progressive apoptosis was diagnosed
by analyzing small band shifts in Raman spectra, e.g., for the amide I band, which
was observed for both cell lines. However, necrotic cells were characterized by the
increased intensity of bands originating from protein vibrations, in contrast to
apoptotic cells in which the reverse effect was observed. In general, the process of
apoptosis was associated with the decrease in the content of proteins and DNA in
cells, which was seen in Raman spectra as the decrease in the intensity of bands at
1007 and 785 cm–1, originating from the breathing mode of the phenylalanine ring
and the deformation vibrations of the nucleic acid bases, respectively [96, 98, 102].
It was also shown that the concentration of DNA in the cell nuclei depended on the
stage of apoptosis [99, 100]. The early stages of apoptosis were investigated in EA.
hy926 endothelial cells after activation by Fas ligand and cycloheximide [106].
Both agents are well-known pro-apoptotic inducers triggering programmed death
cell via binding to the membrane receptors and inhibition of protein synthesis,
respectively [107, 108]. The morphological and chemical changes were determined
with the most important conclusion from this work demonstrating that the early
apoptosis resulted in the statistically significant decrease of the protein level inside
all studied cellular structures with the simultaneous increase of the nucleic acids
content. Raman spectroscopy enabled to define these spectroscopic markers of
early-stage apoptosis independently on the activation pathway [106].
6.2.3 In Vitro Cell Models in Drug Screening
A significant chapter of Raman-based protocols for cellular investigations is in vitro

drug screening. The potential of Raman spectroscopy in this field has been well
established over the last decade, especially in understanding of chemotherapeutic
drugs’ actions. Raman spectroscopy gives an opportunity to determine the cellular
response to a given compound and also to follow the uptake and intracellular
accumulation of drugs and potentially identify cell-resistant strains [109]. Raman
spectroscopy together with the protocol of cell cloning also proved to be especially
useful for clinicians to better tailor the treatment to each individual patient [110].
Among a number of compounds tested with the use of Raman-based methodology
such as cisplatin, vincristine, and nilotinib, a substantial interest has been put on
doxorubicin (DOX) and its analogues [111, 112]. DOX belongs to the group of
compounds named anthracyclines, which exhibit cytostatic activity and are widely
used in chemotherapy. Despite the relatively well-established usage of DOX in
anti-tumor treatment, its mechanism of action is not fully elucidated, but it is known
that it includes such events as intercalation into nuclear DNA, DNA alkylation,
inhibition of DNA synthesis or formation of free radicals, altogether leading to cell
apoptosis [113, 114]. DOX-based therapy is complicated by its cardiotoxicity
which can lead to severe cardiac complications even years after administration. This
duality of DOX activity leads to a large number of investigations that aim on
assessing both the activity and cardiotoxicity effects but also synthesis of analogues
and derivatives of DOX [115]. Structurally, anthracyclines possess a tetracycline
ring structure with a daunosamine group attached by a glycosidic linkage. This
tetracycle is a chromophore group giving the absorption band at 480 nm (p−p*
transition). This feature gives rise to the Raman spectrum of DOX acquired at the
488-nm excitation wavelength exhibiting a very strong fluorescence background
(Fig. 6.6) [116]. The non-resonant spectra of DOX can be recorded at the 785-nm
excitation wavelength, where the characteristic spectral profile is present with
several marker bands (Fig. 6.6). Considering the above facts, spectral features at
both wavelengths are suitable for investigations focused on DOX–cell interactions.
The potential of imaging in assessment of nuclear accumulation of anthracyclines in
endothelial cells was demonstrated with the use of the 488-nm excitation [117]. The
maximum of DOX fluorescence coincides with the range in which the increased
background in DOX-treated cells is elevated. This enables for the analysis of
changes in cell spectral profile that occurs after treatment with DOX but also for
imaging of its distribution inside the cell (Fig. 6.6). Investigations performed at the
785-nm excitation reveal similar information on the distribution of the drug based,
however, on the characteristic doxorubicin Raman marker bands [112]. However,
spectra collected with the use of the 785-nm excitation enable for more specific
investigations of the modes of interaction of the drug, as, in the absence of the
strong fluorescence background, the drug itself as well as the biochemical cell
environment can be monitored in parallel. On the other hand, for fast imaging with
better spatial resolution, application of the 488-nm excitation is preferable.
Fig. 6.6 DOX uptake by endothelial cells. Representative Raman images (a) of an endothelial
cell (EA.hy926) treated with 10 µM concentration of DOX representing distribution of organic
matter and nuclear accumulation of DOX. Raman spectra of DOX recorded with 488 nm (b) and
785 nm (c) excitation wavelengths. Scale bars equal 5 lm
6.2.4 Probing of Intracellular Environment by SERS

Imaging
Although Raman microscopy has been widely employed for imaging of cells, it is
limited because of a weak signal in comparison with the background noise or
autofluorescence. Therefore, as data collection process is time-consuming, reliable
statistical analysis of weakly scattering biosamples is hindered [118]. This obstacle
can be overcome by using NPs as a SERS agent to enhance Raman scattering
(Sect. 6.1.4) and to gather information on specific cellular processes as well as type
of molecules and changes in the intracellular environment. Intracellular SERS
detection is usually accomplished by two strategies, i.e., label (indirect) and
label-free (direct) methods. In the latter, bare (‘naked’) Au or Ag nanoparticles or
nanoaggregates are used to interact with cellular biomolecules [119, 120]. An
effective and successful utilization of NPs for in vitro SERS analysis requires
understanding of pathways and mechanisms of their uptake and distribution.
Endocytosis and passive diffusion are known to be a major pathway for AuNPs
[121, 122]. Metallic nanoparticles after uptake by cells are mostly localized in
endosomes; therefore, the SERS intracellular spectra represent localization of these
structures, their composition, and local environment [47, 122]. Kneipp et al. [123]
have showed that a label-free method of probing living macrophages and
endothelial cells differentiates cell lines and characterizes changes in cellular
environments by detection of physiologically relevant molecules. This method has
been successfully employed in elucidating biomolecular dynamics of stress
response upon UV-C irradiation at a cellular level in healthy and cancerous cells.
Functionalization of Au nanocubes with nuclear localization signal peptide and
cell-penetrating peptide enabled observing physiochemical damages of cytosolic
proteins containing sulfur and aromatic amino acids by changes in their secondary
structures showing that SERS was a powerful approach to understand mechanisms
of UV-light-triggered cellular biochemical changes [124]. Another example of an

advanced application of label-free SERS imaging is a work of Huang et al. [125] on
tracking intracellular drug release. For this purpose, graphene oxide nanoplatforms
were loaded with DOX and AgNPs. SERS spectra of DOX were used to monitor
the real-time process of drug release inside living cells [125].
An indirect SERS probing of cells relies on decoration of NPs by a Raman
reporter exhibiting a specific Raman signature and/or is sensitive to particular
chemical changes/reactions [119]. This approach has been extensively developed in
response to bioanalytical needs to obtain selective and ultrasensitive quantitative
analysis of target molecules present in complex chemical environment. SERS
biosensors can be additionally conjugated with antibodies, aptamers, or peptides as
recognition elements [47, 119]. Such SERS tags have labelling functions similar to
fluorophores but overcome limitations of fluorescence microscopy due to their high
photostability and multiplexing quantification [119, 126].
The indirect SERS technology has found a great application in monitoring of
intracellular pH and redox potential changes. Local pH of intracellular compart-
ments can be measured by using a Raman reporter present on the surface of NPs
that exhibits a strong SERS signal and is sensitive to the pH response. The probe
molecules undergo deprotonation/protonation due to pH variation, resulting in
changes in SERS spectra [127]. Such pH-sensitive Raman reporters are
4-mercaptobenzoic acid (4-MBA) [128–131], 4-ethynylpyridine [132], and
aminotiophenols [130, 133]. For example, intensities of SERS bands assigned to the
deprotonated and protonated 4-MBA carboxylic group at *1720 and *1420 cm–
1
, respectively, correspond to changes of intracellular pH from 5 to 9 [128]. A great
importance for the determination of subcellular events is also employing SERS tags
for tracking targeted delivery of pH-sensitive drugs [134, 135]. pH-controllable
drug carriers proposed by Chen et al. [134] are bifunctional. They provide
pH-sensitive delivery of the DOX-AuNPs complex and monitor the release of DOX
inside living cells [134].
Chemical reactions of Raman reporters with a target analyte have been used in
designing of SERS sensors detecting the intracellular level of carbon monoxide and
nitric oxide [136, 137]. Carbonylation of the palladacycle moiety was employed to
trap CO molecules [137], whereas monitoring of the NO concentration was possible
due to reaction of NO with a SERS probe producing a benzotriazole derivative
which exhibits in turn an intense Raman band of the –N=N– vibration at 1440 cm–1
[136].
An effective and selective delivery of anticancer drugs to cancer cells was
proposed by introducing to the design of SERS biosensors conjugation with an
antibody against a target molecule. Song et al. [135] employed SERS probes
conjugated with human epidermal growth factor receptor 2 (HER2) antibody and
observed simultaneously and specifically drug delivery to the SKBR-3 cancer cells
[135]. This approach has been employed in the fabrication of SERS tags for
detection of inflammatory molecules in endothelial cells. Gold nanorods decorated
with 4-MBA and ICAM-1 antibody indicated an increased level of inflammatory

proteins after short-time (1–5 h) treatment of cells with lipopolysaccharides [138].
An important issue in designing SERS sensors is to construct organelle-specific
nanoprobes which recognize particular compartments of cells as nuclei or mito-
chondria. Three different subcellular organelle-targeting SERS nanotags have been
employed by Kang et al. [139] for multiplex cell imaging and monitor cell mor-
phologies during cell death. They attached cell-penetrating, mitochondria-targeting,
and nucleus-targeting peptides as recognizing sensors to NPs functionalized with
three different Raman reporters. This experiment has also showed that SERS-based
imaging of cells significantly improves spectral quality and temporal resolution of
Raman images [139]. Using SERS tags not only provides information about
organelles but also enables detecting specific molecules inside cells, e.g., telom-
erase. Telomerase ribonucleoprotein complex is an important biomarker in the early
clinical diagnostic, but its activity is difficult to recognize by conventional methods.
Xu et al. [140] have designed a DNA-driven NP self-assembling pyramid encoding
a Raman reporter (Cy5) which is specific for the recognition of telomerase at the
intracellular level [140]. These examples clearly show that SERS biosensing of
cellular environment requires the advanced development of nanotechnology com-
bined with new techniques of molecular conjugation.
6.3 Examination of Primary and Cultured Cells
Cell culture studies provide a valuable complement to in vivo experiments,

allowing for a more controlled manipulation of cellular functions and processes.
In contrast to cell lines, primary cells which are isolated directly from tissues have
a finite lifespan and limited expansion capacity. On the other hand, primary cells
have normal cell morphology and maintain many of the important markers and
functions seen in vivo [65, 141]. Therefore, in many cases the use of a cell line is
desirable, e.g., in cell biological studies on basic mechanisms or in drug devel-
opment; however, in other cases such as research on complex metabolic pro-
cesses, the use of primary cells becomes essential [142]. A particular example are
medical conditions involving alterations in cross talk between few cell types
constituting the tissue. Such situation takes place in, e.g., the liver during
development of fatty liver disease and its progression into more advanced stages
[143] or development of heart failure [144]. In both cases, pathogenic mechanism
involves disruption of homeostatic signaling between cells such as hepatocytes,
hepatic stellate cells, and liver sinusoidal endothelial cells (LSECs)—in the liver
or between cardiomyocytes—and cardiac microvascular endothelial cells
(CMECs)—in the heart. In both cases, the pathogenic mechanism can be induced
and studied with the use of in vitro models; however, it cannot fully replicate the
processes occurring in living organisms [145].
6.3.1 Liver Sinusoidal Endothelial Cells (LSECs)
LSECs are a highly specialized type of endothelial cells that constitute the barrier
between the blood and hepatocytes in liver sinusoids. They possess small pores—
fenestrations that determine the healthy phenotype of LSECs. Fenestrations with the
diameter ca. 150 nm enable the molecules’ trafficking between the blood and
hepatocytes across sinusoids [146, 147]. However, among other LSECs’ specific
features, the fenestrations are lost within 2–3 days after introducing cells into a
culture [148]. Therefore, it is necessary to take this fact into account while planning
the research involving LSECs. However, their isolation protocol has been well
established over past decades; it involves a few steps such as perfusion of the
selected organ, digestion of the connective tissue, purification of cell suspension,
and separation of chosen cells’ fractions [149].
The isolation of liver cells followed by Raman spectroscopy measurements
based on above-mentioned protocol was previously reported for hepatocytes [150],
hepatic stellate cells [151], and LSECs [152]. Morphologically, LSECs display a
typical fenestrated shape and have a size around 30 µm and round-shaped nucleus
(size around 6 µm) which seems to be densely packed in contrast to a very thin
cytoplasm. With use of Raman spectroscopy, LSECs can be characterized by rel-
atively intense DNA/RNA signals at 787 and 1585 cm−1 (Fig. 6.7) arising from the
nucleus. Raman spectroscopic studies reveal also the ability of these cells to form
LDs. Moreover, detailed analysis of LDs shows their chemical composition
heterogeneity; they can be composed mainly of triacylglyceroles of saturated or
monounsaturated fatty acids. LDs assigned as saturated are characterized by the
Raman profile of saturated fatty acids with specific marker bands at 1064, 1130,
1305, 1444, and 2880 cm−1, and on the other hand, bands at 1267, 1657, and
3015 cm−1 are characteristic for unsaturated lipids [90]. The comparison of Raman
spectra obtained from isolated LSECs and a cultured cell line (TSECs—Transfected
Sinusoidal Endothelial Cells [153]) reveals some differences. The characteristic,
strong nuclear DNA signals are no longer so intense as in primary LSECs, and the
cytoplasm seems to be much more thicker as the Raman spectrum is more pro-
nounced for TSECs. On the other hand, both cell types possess lipid droplets that
also, in some cases, are heterogenic in composition as it is observed for isolated
LSECs.
6.3.2 Cardiac Microvascular Endothelial Cells (CMECs)
The heart consists of different cell types including endothelial cells, cardiomy-
ocytes, and fibroblasts. The cross—talk between them has the greatest impact on
maintaining the cardiac homeostasis and autoregulation. The cell–cell communi-
cation is realized with numerous cardioactive factors secreted in both directions:
from microvascular endothelial cells to cardiomyocytes and vice versa [154, 155].
Fig. 6.7 Raman imaging of isolated primary LSEC. Representative Raman images of LSECs
(a) obtained by integration in the spectral region of 3050–2850 cm−1 (organic matter, cell body),
800–770 cm−1 (DNA and RNA, nucleus), 3030–3000 cm−1 (unsaturated lipids, lipid droplets),
2885–2875 cm−1 (saturated lipids, lipid droplets), and CA image with respective average Raman
spectra (b) of saturated LDs (purple), unsaturated LDs (green), nucleus (blue), and cytoplasm
(orange). Scale bars equal 5 lm
Changes in this communication may result in cardiac diseases. Complexity and

heterogeneity of the heart require, especially for studies of diseases development,
primary cells whose phenotype is more similar to cells in vivo than cell lines.
CMECs seem to be a satisfying in vitro model for studies on progression of heart
failure what was already confirmed [156]. Moreover, due to the reduction of NO
bioavailability and inflammation, endothelial dysfunction has the biggest impact on
myocardium, thus, endothelial cells’ condition may serve as early diagnostic bio-
marker and target for novel drug therapies.
Primary isolated CMECs are usually small with slightly elongated shape.
Compared to other cardiac endothelial cells, i.e., to the H5V cell line (a transformed
endothelial cell line from the murine heart), CMECs adhere to substrate in more
irregular shape, as well as they tend to create rather variable structures instead of a
homogeneous layer. Raman imaging unveils additional features within cells [157].
Due to high lateral resolution and confocality, Raman images of CMECs present
the subcellular environment (Fig. 6.8a). All organelles are clearly visible in the
Raman image obtained by integration of bands in the spectral range 3027–2800
cm−1 showing the distribution of organic compounds. The most distinct structure in
Raman images of CMECs is a relatively large nucleus. The nucleus, the densest
structure inside cell, is visible in Raman images as an ellipsoidal structure with
diameter of 15–20 lm and 2–3 small nucleoli within it. Nucleoli, due to lack of a
surrounding biomembrane, have irregular shape and do not exceed 6 lm.
Fig. 6.8 Raman imaging of isolated primary CMEC. Representative Raman images of a single
CMEC (a) presenting the distribution of all organic compounds, lipids, nucleic acids, and haem.
The CA image (b) and respective average Raman spectra (c) of main clusters nucleoli (purple),
nucleus (blue), and endoplasmic reticulum (green). Scale bars equal 4 lm
Application of CA enables extraction of average Raman spectra from the nucleus

and nucleoli (blue and purple classes in CA image, respectively; Fig. 6.8b). Due to
the increased density of nucleic chromatin in nucleoli, Raman bands at 786 and
1575 cm−1 (the ring breathing modes of the DNA/RNA bases) have higher
intensity in the Raman spectrum of the nucleoli class (Fig. 6.8c).
Lipids in isolated primary CMECs are distributed mainly in the endoplasmic
reticulum. These cells do not exhibit any LDs in contrast to HMEC-1 cells shown in
Sect. 6.2.3. As it was mentioned above, LDs have an important role in inflam-
mation, and their presence, development or changes in the lipid composition, may
be related to stressors and factors like TNF-a or interleukins [90]. The Raman
spectrum of the endoplasmic reticulum (a green class) shows a typical spectral
profile attributed to lipids. The most characteristic bands are 2855 cm−1 (the C–H
stretching vibrations) and 1300 cm−1 (the –CH2– deformations) are a proof for a
presence of fatty acids in a small amount inside the cell.
Beside lipids, the average Raman spectrum of the endoplasmic reticulum class
shows a distinguishable band originating from the heme group. Due to RR effect
under the wavelength of 532 nm, all hemeproteins have an enhanced intensity of
some bands, i.e., at 1587 cm−1 that enables to easily recognize them in the spectra.
The Raman image obtained by integration of this band shows that haem is irreg-
ularly distributed in the whole endoplasmic reticulum. Precise identification of the
source of heme in CMECs is not possible due to the low concentration of this group
which results in low-intensity bands in the Raman spectra. The only visible bands in
the average spectrum, which are located at 1587 cm−1 (m19 mode: the asymmetric
C–C vibrations) and 749 cm−1 (m15 mode: the porphyrin breathing modes), are not
enough to determinate if heme comes from cytochrome c located in cell’s mito-
chondria or if the source is exogenous (haem from blood cells as an impurity from
isolation) [158–160]. In comparison with primary CMECs, H5V cells show a
different phenotype. H5V is rich in lipids and LDs within cytosol and is much more
resistant to stimulation with TNF-a. The concentration of 10 nM of this cytokine is
enough to initiate apoptosis in primary CMECs, while the H5V cell line does not
react to stimulation.
6.4 Label-Free and Label Raman Spectroscopic Imaging

as a Potential Tool for Diagnosis of Diseases
of Affluence in Tissues
6.4.1 Label-Free Raman Spectroscopic Imaging
A conventional medical diagnosis performed by pathologists relies on visualization

techniques used to detect, define, and estimate the extent of a disease in biopsies.
The most common tool is the microscopic assessment of pathological changes in
tissues based on histological and immunohistochemical staining. The obtained
results, although in many cases extremely valuable for diagnosis of disease or
choice of treatment, are subject to some restrictions, for instance difficulties in
distinguishing the type of cancer or the need of the use of a few dyes/fluorophores.
A histological examination requires also long time of the analysis and is dependent
on the pathologist’s experience. Raman spectroscopy resolves most of these issues.
Label-free Raman characteristics showing overall information about all constituent
biomolecules can be obtained from tissue cross-sections prepared from the biopsy
material (ex vivo) as well as from in situ measurements directly from organs inside
the living body (in vivo). Finally, the lack or minimal invasiveness of this technique
leads to its successive application in clinical diagnostics [161].
Researchers working in the field of Raman tissue diagnostics have employed a
variety of methods for sample preparation, and unfortunately, there is no universal
standard offered for each application. In general, the tissue slices are placed directly
onto CaF2 or glass windows that possess a simple or well-defined Raman spectrum
beyond the studied spectral region. An optimal option is to study raw, not fixed
tissue samples, since most fixation agents contribute to a Raman spectrum. This is
achievable in vivo using Raman fiber optic probes using fast-data collecting
techniques such as coherent anti-Stokes Raman spectroscopy [162] or simply for
tissues not sensitive to external conditions, i.e., for aorta cross-sections [72]. Tissue
cross sections are usually preserved because of long time of spectral acquisition. To
extend storage of tissue samples, two common methods can be employed, i.e.,
chemical or non-chemical fixation using, for example, air-drying. Among chemical
methods, the most frequently used protocol is fixation in the formalin solution.
Hydrated formalin cross-links the primary and secondary amine groups of proteins
preventing their autolysis and stabilizing a tissue structure. Effects of formalin
fixation on Raman spectra of mammalian animal tissues and cells have been pre-
viously studied and summarized [163].
The choice of a laser excitation line is another crucial element of Raman imaging
of tissues. Most of the research is performed with using near-infrared lasers, mainly
with a 785-nm line, due to huge autofluorescence arising from connective tissues
when they are irradiated by visible light sources. In case of tissues containing blood,
a NIR laser excitation is required to omit the contribution of a resonance Raman
effect of heme moieties which strongly overlaps Raman signal of other biocom-
ponents. However, one must know that high lateral resolution is achieved in
microscopic Raman imaging when UV and Vis lasers are employed.
In recent decades, Raman spectroscopy has been successfully applied to show
tissue alterations occurring due to development of lifestyle diseases and various
types of cancers [71, 161, 164, 165]. A proper Raman analysis of these changes
associated with a disease requires to understand biochemistry of healthy tissue. In a
review article published by us [166], the chemical composition of tissue
cross-sections and homogenates of various organs, inter alia, kidney, brain, liver,
and aorta, were discussed in terms of their characteristic Raman signatures. As was
above mentioned, the vascular endothelium status is a determinant of the condition
of the cardiovascular system, and its dysfunction is at the root of diseases of
affluence. An examination of the endothelium in ex vivo cross-sections of the
murine aorta has been performed for animal models of diabetes type 2 [74],
hypertension [60], and cancer metastasis [72]. Common features of endothelial
dysfunction in the three pathologies are variations in an overall amount of proteins
and/or lipids in this layer of the vessel. In addition, confocal Raman and AFM
imaging clearly show the formation of lipid rafts’ clusters on the endothelium
surface in the diabetic aorta (Fig. 6.9i). Raman spectroscopy was also used to study
a blood vessel wall to characterize the chemical composition of atherosclerotic
plaque in mice fed with different diets [167]. For example, in ApoE/LDLR−/− mice
on low-carbonate–high-protein (LCHP) diet within brachiocephalic arteries, various
components in the plague deposit were identified without labelling, including
cholesterol esters and inorganic calcium salts (Fig. 6.9ii). A high lipid-to-protein
ratio and an increased content of cholesterol were also found in a murine brain
Fig. 6.9 Raman imaging of ex vivo murine artery in micro- and macroscale. (i) White light,
Raman, and AFM images of the en face vascular wall of diabetic (db/db) mouse (a).
A microphotograph of the studied tissue (100 magnification), Raman image showing distribution
of organic matter (the 2800–3100 cm−1 spectral region) (b), the corresponding AFM images of
topography (c), and phase (d). Adapted from [170], open access: Creative Commons Attribution
(CC BY) license. (ii) White light, Raman, and CA images of a cross section of brachiocephalic
artery of ApoE/LDLR−/− mouse fed a LCHP diet. A microphotograph (a) with area imaged by
Raman microscopy (marked with red color, 100 magnification), a CA image (b) showing
localization of calcification, remodelled media, heme, internal elastic lamina/fibrous cap and lipids,
Raman images showing distribution of organic matter, autofluorescence originated from elastin
and hemoglobin (Hb), respectively (c–e). Adapted from [167], open access: Creative Commons
Attribution (CC BY) license
affected by atherosclerosis [168]. Furthermore, Raman spectroscopy was applied in

classification of various types of brain tumors [169] as well as in detection of
tumors in breast [170, 171], skin [172], lung [173], liver [174], cervical [175, 176]
and prostate [177] tissues. These works showed the applicability of Raman spec-
troscopy to characterize chemically malignant tissues and to define cancer margins.
The above-mentioned (Sect. 6.4.1) research have some major limitations related
to, i.e., sample preparation and fixation methodology of ex vivo tissues and gene
knockout modifications in animal model systems. Thus, it is necessary to reduce
these obstacles and gain non-falsified information about the system. Nowadays, a
lot of effort is put to develop in vivo Raman diagnostics. Spontaneous Raman
spectroscopy can be combined with optical fibers for endoscopy of hollow organs.
This nondestructive label-free approach enables to gather specific information about
chemical structure of the tissue and its alternation from even very early stages of
disease expansion. However, work with fiber optic probes encounters some prob-
lems such as a strong background signal arising from silicon-based material of
probe and weak Raman scattering of biological species. To solve these issues,
various sophisticated schemes and optical elements are designed and integrated into
common fiber probes for medical applications [178]. So far, Raman spectroscopy
via fiber optic probes in in vivo biomolecular characterization and diagnosis was
performed on skin [179], brain [180], components of the digestive system including
esophagus [181], larynx [182], stomach [183] and colon [183], breast [184], arteries
[185], lung [186], liver [187], cervix [188], and bladder [189]. The main aim of
these studies was to identify spectroscopic markers associated with the early
detection of cancerous lesions. Recent advances in in vivo Raman spectroscopy are
also summarized in two reviews [161, 165].
6.4.2 Tissue Imaging With Immuno-SERS Microscopy
Label-free conventional Raman spectroscopy has been proven to be a powerful and

nondestructive tool for investigation even such complex chemical structures like
tissues. However, it requires long data collection time to obtain reliable and high
signal-to-noise spectral information. Advantages of SERS over conventional
Raman technique were introduced to the reader in the previous paragraphs of this
chapter (Sects. 6.1 and 6.2) and include recording a selective signal from a target
molecule. Furthermore, because of enhancement, a standard SERS data collection
allows reducing laser powers and acquisition times what is a key in the diagnostic
field. Selectivity of SERS nanoprobes is achieved by functionalization of metal
nanoparticles with antibodies or aptamers. An antibody-based method has been
developed since 1941 when Albert H. Coons had described a new way of local-
ization of antigens in tissues using a fluorescent-labelled antibody which was called
thereafter as immunohistochemical staining [190]. A similar approach was trans-
ferred to design SERS probes and develops immuno-SERS staining (iSERS).
Typical SERS tags employed in an immunoassay are composed of a metallic
nanoparticle, Raman reporter, protective shell which also facilities conjugation of a
targeting ligand, and an antibody as a biorecognition molecule (Fig. 6.10a) [48].
Immuno-SERS microscopy in recent time was proposed as an alternative
immunohistochemistry method. Several obstacles typical for immunohistochemical
fluorescence can be overcome by using iSERS nanotags, e.g., the lack of
Fig. 6.10 Fabrication of SERS nanotags decorated with antibody (a). An example of using iSERS
staining to detect SMCs in atherosclerotic brachiocephalic artery (b). White light image (left down)
of the cross section with ROI selected for microscopic Raman imaging and CA iSERS maps
showing the SMCs distribution in the aortic wall (right down)
photobleaching of SERS labels and autofluorescence of samples as well as narrow

SERS signals compared to fluorescence that permits multiplexing capabilities
[191].
iSERS imaging of prostate tissues to detect prostate-specific antigen (PSA) was
proposed firstly by Schlücker et al. in 2006 [192]. Next, they extended this
methodology for monitoring PSA expression [193] and determining co-localization
of p63 and PSA on healthy prostate tissue [194]. The latter was the first report
showing multiplexing capability of iSERS in tissues; for this purpose, two different
Raman reporters: 4-nitrothiobenzoic acid (4-NTB) and 4-mercaptobenzoic acid
(4-MBA) were employed to construct iSERS nanotags. The Schlücker group also
demonstrated a successful iSERS-based localization and expression of the HER2 in
normal and malignant breast tissues [195]. A similar approach was used by Wang
and co-workers for quantitative multiplexed imaging of freshly excised human
tissues for surgical guidance applications [196]. Ex vivo experiments carried out by
Quynh et al. [197] confirmed the utility of iSERS imaging to detect tumor regions.
Here, BerEP4 antibody conjugated to SERS NPs covered by 4-ATP
(4-aminothiophenol) as a Raman reporter specifically recognized basal cell carci-
noma (BCC) in skin during surgical procedures. SERS detection of antibody–
antigen interactions was also used for a rapid and sensitive recognition of LMP1
protein expression specific for nasopharyngeal cancer [198].
An example of iSERS imaging is shown in Fig. 6.10b. Here, gold nanoparticles

covered with a Raman reporter molecule and ethylene glycol spacer (mixture of
4-NTB-MEG-OH and 4-NTB-TEG-COOH according to protocol in [195]) were
conjugated with an anti-actin smooth muscle cells (SMCs) antibody labelled by a
cy3 fluorophore. This iSERS probes allowed detecting one of the most important
factors determining the formation of the atherosclerotic plaque in the murine bra-
chiocephalic artery.
References
1. Atkins PW, Barron LD (1969) Rayleigh scattering of polarized photons by molecules. Mol
Phys 16:453–466
2. Barron LD, Buckingham AD (1971) Rayleigh and Raman scattering from optically active
molecules. Mol Phys 20:1111–1119
3. Barron LD, Bogaard MP, Buckingham AD (1973) Raman scattering of circularly polarized
light by optically active molecules. J Am Chem Soc 95:603–605
4. Hecht L, Barron LD, Blanch EW, Bell AF, Day LA, Ziegler LD (1999) Raman optical
activity instrument for studies of biopolymer structure and dynamics. J Raman Spectrosc
30:815–825
5. Hug W (2003) Virtual enantiomers as the solution of optical activity’s deterministic offset
problem. Appl Spectrosc 57:1–13
6. Hug W (2010) Raman optical activity, spectrometers A2. In: Lindon JC (ed) BT—
encyclopedia of spectroscopy and spectrometry, 2nd edn. Academic Press, Oxford,
pp 2387–2396
7. Kapitán J, Barron LD, Hecht L (2015) A novel Raman optical activity instrument operating
in the deep ultraviolet spectral region. J Raman Spectrosc 46:392–399
8. Kubota K, Shingae T, Foster ND, Kumauchi M, Hoff WD, Unno M (2013) Active site
structure of photoactive yellow protein with a locked chromophore analogue revealed by
near-infrared Raman optical activity. J Phys Chem Lett 4:3031–3038
9. Profant V, Pazderková M, Pazderka T, Maloň P, Baumruk V (2014) Relative intensity
correction of Raman optical activity spectra facilitates extending the spectral region.
J Raman Spectrosc 45:603–609
10. Shingae T, Kubota K, Kumauchi M, Tokunaga F, Unno M (2013) Raman optical activity
probing structural deformations of the 4-hydroxycinnamyl chromophore in photoactive
yellow protein. J Phys Chem Lett 4:1322–1327
11. Unno M, Kikukawa T, Kumauchi M, Kamo N (2013) Exploring the active site structure of a
photoreceptor protein by Raman optical activity. J Phys Chem B 117:1321–1325
12. Yamamoto S, Watarai H (2010) Incident circularly polarized Raman optical activity
spectrometer based on circularity conversion method. J Raman Spectrosc 41:1664–1669
13. Zhang Y, Wang P, Jia G, Cheng F, Li C (2017) A short-wavelength Raman optical activity
spectrometer with laser source at 457 nm for the characterization of chiral molecules. Appl
Spectrosc 71:2211–2217
14. Barron LD, Hecht L, Blanch EW, Bell AF (2000) Solution structure and dynamics of
biomolecules from Raman optical activity. Prog Biophys Mol Biol 73:1–49
15. Blanch E (2003) Vibrational Raman optical activity of proteins, nucleic acids, and viruses.
Methods 29:196–209
16. Barron LD, Hecht L, McColl IH, Blanch EW (2004) Raman optical activity comes of age.
Mol Phys 102:731–744
17. Hecht L, Phillips AL, Barron LD (1995) Determination of enantiomeric excess using Raman
optical activity. J Raman Spectrosc 26:727–732
18. Blanch EW, Bell AF, Hecht L, Day LA, Barron LD (1999) Raman optical activity of
filamentous bacteriophages: hydration of a-helices. J Mol Biol 290:1–7
19. Wen ZQ, Barron LD, Hecht L (1993) Vibrational Raman optical activity of monosaccha-
rides. J Am Chem Soc 115:285–292
20. Bell AF, Hecht L, Barron LD (1994) Disaccharide solution stereochemistry from vibrational
Raman optical activity. J Am Chem Soc 116:5155–5161
21. Bell AF, Hecht L, Barron LD (1995) Vibrational Raman optical activity of ketose
monosaccharides. Spectrochim Acta A 51:1367–1378
22. Blanch EW, Hecht L, Barron LD (2003) Vibrational Raman optical activity of proteins,
nucleic acids, and viruses. Methods 29:196–209
23. Kessler J, Yamamoto S, Bouř P (2017) Establishing the link between fibril formation and
Raman optical activity spectra of insulin. Phys Chem Chem Phys 19:13614–13621
24. Gąsior-Głogowska M, Malek K, Zajac G, Baranska M (2016) A new insight into the
interaction of cisplatin with DNA: ROA spectroscopic studies on the therapeutic effect of the
drug. Analyst 141:291–296
25. Polavarapu PL (1990) Ab initio vibrational Raman and Raman optical activity spectra.
J Phys Chem 94:8106–8112
26. Bose PK, Polavarapu PL, Barron LD, Hecht L (1990) Ab initio and experimental Raman
optical activity in (+)-(R)-methyloxirane. J Phys Chem 94:1734–1740
27. Ruud K, Helgaker T, Bouř P (2002) Gauge-origin independent density-functional theory
calculations of vibrational Raman optical activity. J Phys Chem A 106:7448–7455
28. Jovan Jose KV, Raghavachari K (2016) Raman optical activity spectra for large molecules
through molecules-in-molecules fragment-based approach. J Chem Theory Comput 12:585–
594
29. Kessler J, Kapitán J, Bouř P (2015) First-principles predictions of vibrational Raman optical
activity of globular proteins. J Phys Chem Lett 6:3314–3319
30. Abdali S, Blanch EW (2008) Surface enhanced Raman optical activity (SEROA). Chem Soc
Rev 37:980–992
31. Hiramatsu K, Leproux P, Couderc V, Nagata T, Kano H (2015) Raman optical activity
spectroscopy by visible-excited coherent anti-Stokes Raman scattering. Opt Lett 40:4170–
4173
32. Tatarkovič M, Miškovičová M, Šťovíčková L, Synytsya A, Petruželka L, Setnička V (2015)
The potential of chiroptical and vibrational spectroscopy of blood plasma for the
discrimination between colon cancer patients and the control group. Analyst 140:2287–2293
33. Šebestík J, Teplý F, Císařová I, Vávra J, Koval D, Bouř P (2016) Intense chirality induction
in nitrile solvents by a helquat dye monitored by near resonance Raman scattering. Chem
Commun 52:6257–6260
34. Šebestík J, Bouř P (2011) Raman optical activity of methyloxirane gas and liquid. J Phys
Chem Lett 2:498–502
35. Šebestík J, Kapitán J, Pačes O, Bouř P (2016) Diamagnetic Raman optical activity of
chlorine, bromine, and iodine gases. Angew Chem—Int Ed 55:3504–3508
36. Šebestík J, Bouř P (2014) Observation of paramagnetic Raman optical activity of nitrogen
dioxide. Angew Chem—Int Ed 53:9236–9239
37. Wu T, Kapitán J, Mašek V, Bouř P (2015) Detection of circularly polarized luminescence of
a Cs–EuIII complex in Raman optical activity experiments. Angew Chem—Int Ed
54:14933–14936
38. Nafie LA (1996) Theory of resonance Raman optical activity: the single electronic state
limit. Chem Phys 205:309–322
39. Vargek M, Freedman TB, Lee E, Nafie L (1998) Experimental observation of resonance
Raman optical activity. Chem Phys Lett 287:359–364
40. Merten C, Li H, Nafie LA (2012) Simultaneous resonance Raman optical activity involving
two electronic states. J Phys Chem A 116:7329–7336
41. Haraguchi S, Hara M, Shingae T, Kumauchi M, Hoff WD, Unno M (2015) Experimental
detection of the intrinsic difference in Raman optical activity of a photoreceptor protein
under preresonance and resonance conditions. Angew Chem—Int Ed 54 11555–11558
42. Magg M, Kadria-Vili Y, Oulevey P, R. Weisman B, Bürgi T (2016) Resonance Raman
optical activity spectra of single-walled carbon nanotube enantiomers. J Phys Chem Lett
7:221–225
43. Zajac G, Kaczor A, Pallares Zazo A, Mlynarski J, Dudek M, Baranska M (2016)
Aggregation-induced resonance raman optical activity (AIRROA): a new mechanism for
chirality enhancement. J Phys Chem B 120:4028–4033
44. Dudek M, Zajac G, Kaczor A, Baranska M (2016) Aggregation-induced resonance Raman
optical activity (AIRROA) and time-dependent helicity switching of astaxanthin
supramolecular assemblies. J Phys Chem B 120:7807–7814
45. Zajac G, Lasota J, Dudek M, Kaczor A, Baranska M (2017) Pre-resonance enhancement of
exceptional intensity in aggregation-induced Raman optical activity (AIROA) spectra of
lutein derivatives. Spectrochim Acta A 173:356–360
46. Dudek M, Zajac G, Kaczor A, Baranska M (2017) Resonance Raman optical activity of
zeaxanthin aggregates. J Raman Spectrosc 48:673–679
47. Procházka M (2016) Surface-enhanced Raman spectroscopy. Bioanalytical, biomolecular
and medical applications. Springer International Publishing, Switzerland
48. Schlücker S (ed) (2011) Surface-enhanced raman spectroscopy. Analytical, biophysical and
life science applications. Wiley-VCH Verlag & Co. KGaA, Weinheim
49. Aroca R (2006) Surface-enhanced vibrational spectroscopy. Wiley, Chichester
50. Kaczor A, Malek K, Baranska M (2010) Pyridine on colloidal silver. Polarization of surface
studied by surface-enhanced Raman scattering and density functional theory methods. J Phys
Chem C 114:3909–3917
51. Jaworska A, Malek K (2014) A comparison between adsorption mechanism of tricyclic
antidepressants on silver nanoparticles and binding modes on receptors. Surface-enhanced
Raman spectroscopy studies. J Colloid Interface Sci 431:117–124
52. Jaworska A, Wietecha-Posłuszny R, Woźniakiewicz M, Koscielniak P, Malek K (2011)
Evaluation of the potential of surface enhancement Raman spectroscopy for detection of
tricyclic psychotropic drugs. Case studies on imipramine and its metabolite. Anal 136:4704–
4709
53. Živanović V, Madzharova F, Heiner Z, Arenz C, Kneipp J (2017) Specific interaction of
rricyclic antidepressants with gold and silver nanostructures as revealed by combined one-
and two-photon vibrational spectroscopy. J Phys Chem C 121:22958–22968
54. Cîntǎ-Pînzaru S, Peica N, Küstner B, Schlücker S, Schmitt M, Frosch T, Faber JH,
Bringmann G, Popp J (2006) FT-Raman and NIR-SERS characterization of the antimalarial
drugs chloroquine and mefloquine and their interaction with hematin. J Raman Spectrosc
37:326–334
55. Jaworska A, Malek K, Marzec KM, Baranska M (2012) Nicotinamide and trigonelline
studied with surface-enhanced FT-Raman spectroscopy. Vib Spectrosc 63:469–476
56. Jaworska A, Malek K, Marzec KM, Baranska M (2014) An impact of the ring substitution in
nicorandil on its adsorption on silver nanoparticles. Surface-enhanced Raman spectroscopy
studies. Spectrochim Acta A 129:624–631
57. Marzec KM, Jaworska A, Malek K, Kaczor A, Baranska M (2013) Substituent effect on
structure and surface activity of N-methylpyridinium salts studied by FT-IR, FT-RS, SERS
and DFT calculations. J Raman Spectrosc 44:155–165
58. Ostovar Pour S, Rocks L, Faulds K, Graham D, Parchaňský V, Bouř P, Blanch EW (2015)
Through-space transfer of chiral information mediated by a plasmonic nanomaterial. Nat
Chem 7:591–596
59. Rygula A, Majzner K, Marzec KM, Kaczor A, Pilarczyk M, Baranska M (2013) Raman
spectroscopy of proteins: a review. J Raman Spectrosc 44:1061–1076
60. Pacia MZ, Mateuszuk L, Chlopicki S, Baranska M, Kaczor A (2015) Biochemical changes
of the endothelium in the murine model of NO-deficient hypertension. Analyst 140:2178–
2184
61. Schulz H, Baranska M (2007) Identification and quantification of valuable plant substances
by IR and Raman spectroscopy. Vib Spectrosc 43:13–25
62. Czamara K, Majzner K, Pacia MZ, Kochan K, Kaczor A, Baranska M (2015) Raman
spectroscopy of lipids: a review. J Raman Spectrosc 46:4–20
63. Byrne HJ, Baranska M, Puppels GJ, Stone N, Wood B, Gough KM, Lasch P, Heraud P,
Sulé-Susoj J, Sockalingum GD (2015) Spectropathology for the next generation: quo vadis?
Analyst 140:2066–2073
64. Aird WC (2007) Phenotypic heterogeneity of the endothelium: I. Structure, function, and
mechanisms. Circ Res 100:158–173
65. Bouis D, Hospers GAP, Meijer C, Molema G, Mulder NH (2001) Endothelium in vitro: a
review of human vascular endothelial cell lines for blood vessel-related research.
Angiogenesis 4:91–102
66. Aird WC (2003) Endothelial cell heterogeneity. Crit Care Med 31:S221–S230
67. Wilson SE, Lloyd SA, He YG, McCash CS (1993) Extended life of human corneal
endothelial cells transfected with the SV40 large T antigen. Invest Ophthalmol Vis Sci
34:2112–2123
68. Edgell CJ, McDonald CC, Graham JB (1983) Permanent cell line expressing human factor
VIII-related antigen established by hybridization. Proc Natl Acad Sci 80:3734–3737
69. Ades E, Candal F, Swerlick R, George VG, Summers S, Bosse DC, Lawley TJ (1992)
HMEC-1: establishment of an immortalized human microvascular endothelial cell line.
J Invest Dermatol 99:683–690
70. Lidington EA, Moyes DL, McCormack AM, Rose ML (1999) A comparison of primary
endothelial cells and endothelial cell lines for studies of immune interactions. Transpl
Immunol 7:239–246
71. Baranska M, Kaczor A, Malek K, Jaworska A, Majzner K, Staniszewska-Slezak E,
Pacia MZ, Zajac G, Dybas J, Wiercigroch E (2015) Raman microscopy as a novel tool to
detect endothelial dysfunction. Pharmacol Reports 67:736–743
72. Pacia MZ, Buczek E, Blazejczyk A, Gregorius A, Wietrzyk J, Chlopicki Baranska M,
Kaczoret A (2016) 3D Raman imaging of systemic endothelial dysfunction in the murine
model of metastatic breast cancer. Anal Bioanal Chem 408:3381–3387
73. Pacia MZ, Mateuszuk L, Buczek E, Chlopicki S, Blazejczyk A, Wietrzyk A, Baranska M,
Kaczor A (2016) Rapid biochemical profiling of endothelial dysfunction in diabetes,
hypertension and cancer metastasis by hierarchical cluster analysis of Raman spectra.
74. Pilarczyk M, Mateuszuk L, Rygula A, Kepczynski M, Chlopicki S, Baranska M, Kaczor A
(2014) Endothelium in spots—high-content imaging of lipid rafts clusters in db/db mice.
PLoS ONE 9:e106065
75. Pilarczyk M, Rygula A, Mateuszuk L, Chlopicki S, Baranska M, Kaczor A (2013)
Multi-methodological insight into the vessel wall cross-section: Raman and AFM imaging
combined with immunohistochemical staining. Biomed Spectrosc Imaging 2:191–197
76. Rygula A, Pacia MZ, Mateuszuk L, Kaczor A, Kostogrys RB, Chlopicki S, Baranska M
(2015) Identification of a biochemical marker for endothelial dysfunction using Raman
spectroscopy. Analyst 140:2185–2189
77. Buschman HP, Marple ET, Wach ML, Bennett B, Bakker Schut TC, Bruining HA,
Bruschke AV, van der Laarse A, Puppels GJ (2000) In vivo determination of the molecular
composition of artery wall by intravascular Raman spectroscopy. Anal Chem 72:3771–3775
78. Wang HW, Le TT, Cheng JX (2008) Label-free imaging of arterial cells and extracellular
matrix using a multimodal CARS microscope. Opt Commun 281:1813–1822
79. Majzner K, Kaczor A, Kachamakova-Trojanowska N, Fedorowicz A, Chlopicki S,
Baranska M (2013) 3D confocal Raman imaging of endothelial cells and vascular wall:
perspectives in analytical spectroscopy of biomedical research. Analyst 138:603–610
80. Elsheikha HM, Alkurashi M, Kong K, Zhu X-Q (2014) Metabolic footprinting of
extracellular metabolites of brain endothelium infected with Neospora caninum in vitro.
BMC Res Notes 7:406
81. Große C, Bergner N, Dellith J, Heller R, Bauer M, Mellmann A, Popp J, Neugebauer U
(2015) Label-free imaging and spectroscopic analysis of intracellular bacterial infections.
Anal Chem 87:2137–2142
82. Kong K, Rowlands CJ, Elsheikha H, Notingher I (2012) Label-free molecular analysis of
live Neospora caninum tachyzoites in host cells by selective scanning Raman
micro-spectroscopy. Analyst 137:4119–4122
83. Majzner K, Kochan K, Kachamakova-Trojanowska N, Maslak E, Chlopicki S, Baranska M
(2014) Raman imaging provides insights into chemical composition of lipid droplets of
different size and origin: in hepatocytes and endothelium. Anal Chem 86:6666–6674
84. Majzner K, Chlopicki S, Baranska M (2016) Lipid droplets formation in human endothelial
cells in response to polyunsaturated fatty acids and 1-methyl-nicotinamide (MNA); confocal
Raman imaging and fluorescence microscopy studies. J Biophotonics 9:396–405
85. Astanina K, Koch M, Jüngst C, Zumbusch A, Kiemer AK (2015) Lipid droplets as a novel
cargo of tunnelling nanotubes in endothelial cells. Sci Rep 5:11453
86. Melo RC, D’Avila H, Wan HC, Bozza PT, Dvorak AM, Weller PF (2011) Lipid bodies in
inflammatory cells. J Histochem Cytochem 59:540–556
87. Pi J, Li T, Liu J, Su X, Wang R, Yang F, Bai H, Jin H, Cai J (2014) Detection of
lipopolysaccharide induced inflammatory responses in RAW264.7 macrophages using
atomic force microscope. Micron 65:1–9
88. Szymonski M, Targosz-Korecka M, Malek-Zietek KE (2015) Nano-mechanical model of
endothelial dysfunction for AFM-based diagnostics at the cellular level. Pharmacol Rep
67:728–735
89. Pober JS, Sessa WC (2007) Evolving functions of endothelial cells in inflammation. Nat Rev
Immunol 7:803–815
90. Czamara K, Majzner K, Selmi A, Baranska M, Ozaki Y, Kaczor A (2017) Unsaturated lipid
bodies as a hallmark of inflammation studied by Raman 2D and 3D microscopy. Sci Rep
7:40889
91. Zhang D, Feng Y, Zhang Q, Su X, Lu X, Liu S, Zhong L (2015) Raman spectrum reveals the
cell cycle arrest of Triptolide-induced leukemic T-lymphocytes apoptosis. Spectrochim
Acta A 141:216–222
92. Moritz TJ, Taylor DS, Krol DM, Fritch J, Chan JW (2010) Detection of doxorubicin-
induced apoptosis of leukemic T-lymphocytes by laser tweezers Raman spectroscopy.
Biomed Opt Express 1:1138–1147
93. Huang H, Shi H, Feng S, Chen W, Yu Y, Lina D, Chen R (2013) Confocal Raman
spectroscopic analysis of the cytotoxic response to cisplatin in nasopharyngeal carcinoma
cells. Anal Methods 5:260–266
94. Lipiec E, Bambery KR, Heraud P, Kwiatek WM, McNaughton D, Tobin MJ, Vogel C,
Wood BR (2014) Monitoring UVR induced damage in single cells and isolated nuclei using
SR-FTIR microspectroscopy and 3D confocal Raman imaging. Analyst 139:4200–4209
95. Li B, Lu MQ, Wang QZ, Shi G, Liao W, Huang S (2015) Raman spectra analysis for single
mitochondrias after apoptosis process of yeast cells stressed by acetic acid. Fenxi Huaxue
Chin J Anal Chem 43:643–650
96. Yao H, Tao Z, Ai M, Peng L, Wang G, He B, Li Y (2009) Raman spectroscopic analysis of
apoptosis of single human gastric cancer cells. Vib Spectrosc 50:193–197
97. Panza J, Maier J (2007) Raman spectroscopy and Raman chemical imaging of apoptotic
cells. Imag Manip Anal Biomol Cells Tissue V 6441:6441081–64410812
98. Ong YH, Lim M, Liu Q (2012) Comparison of principal component analysis and
biochemical component analysis in Raman spectroscopy for the discrimination of apoptosis
and necrosis in K562 leukemia cells. Opt Express 20:22158–22171
99. Brauchle E, Thude S, Brucker SY, Schenke-Layland K (2015) Cell death stages in single
apoptotic and necrotic cells monitored by Raman microspectroscopy. Sci Rep 4:4698
100. Jiang X, Jiang Z, Xu T, Su S, Zhong Y, Peng F, Su Y, He Y (2013) Surface-enhanced

Raman scattering-based sensing in vitro: facile and label-free detection of apoptotic cells at
the single-cell level. Anal Chem 85:2809–2816
101. Pliss A, Kuzmin AN, Kachynski AV, Prasad PN (2010) Biophotonic probing of
macromolecular transformations during apoptosis. Proc Natl Acad Sci U S A 107:12771–
12776
102. Verrier S, Notingher I, Polak JM, Hench LL (2004) In situ monitoring of cell death using
Raman microspectroscopy. Biopolymers 74:157–162
103. Zoladek A, Pascut FC, Patel P, Notingher I (2011) Non-invasive time-course imaging of
apoptotic cells by confocal Raman micro-spectroscopy. J Raman Spectrosc 42:251–258
104. Fazio E, Trusso S, Franco D, Nicolò MS, Allegra A, Neri F, Musolino C, Guglielmino SP
(2016) A micro-Raman spectroscopic investigation of leukemic U-937 cells in aged cultures.
105. Okada M, Smith NI, Palonpon AF, Endo H, Kawata S, Sodeoka M, Fujita K (2012)
Label-free Raman observation of cytochrome c dynamics during apoptosis. Proc Natl Acad
Sci 109:28–32
106. Czamara K, Petko F, Baranska M, Kaczor A (2016) Raman microscopy at the subcellular
level: a study on early apoptosis in endothelial cells induced by Fas ligand and
cycloheximide. Analyst 141:1390–1397
107. Hengartner MO (2000) The biochemistry of apoptosis. Nature 407:3–6
108. Schneider-Poetsch T, Ju J, Eyler DE, Dang Y, Bhat S, Merrick WC, Green R, Shen B,
Liu JO (2010) Inhibition of eukaryotic translation elongation by cycloheximide and
lactimidomycin. Nat Chem Biol 6:209–217
109. Farhane Z, Nawaz H, Bonnier F, Byrne HJ (2018) In vitro label-free screening of
chemotherapeutic drugs using Raman microspectroscopy: towards a new paradigm of
spectralomics. J Biophotonics 11. https://doi.org/10.1002/jbio.201700258
110. Siddique MR, Rutter AV, Wehbe K, Cinque G, Bellisolac G, Sulé-Suso J (2017) Effects of
nilotinib on leukaemia cells using vibrational microspectroscopy and cell cloning. Analyst
142:1299–1307
111. Nawaz H, Bonnier F, Meade AD, Lynga FM, Byrne HJ (2011) Comparison of subcellular
responses for the evaluation and prediction of the chemotherapeutic response to cisplatin in
lung adenocarcinoma using Raman spectroscopy. Analyst 136:2450–2463
112. Farhane Z, Bonnier F, Casey A, Byrne HJ (2015) Raman micro spectroscopy for in vitro
drug screening: subcellular localisation and interactions of doxorubicin. Analyst 140:4212–
4223
113. Gewirtz DA (1999) A critical evaluation of the mechanisms of action proposed for the
antitumor effects of the anthracycline antibiotics adriamycin and daunorubicin. Biochem
Pharmacol 57:727–741
114. Minotti G, Menna P, Salvatorelli E, Cairo G, Gianni L (2004) Anthracyclines: molecular
advances and pharmacologic developments in antitumor activity and cardiotoxicity.
Pharmacol Rev 56:185–229
115. Ciesielska E, Studzian K, Wąsowska M, Oszczapowicz I, Szmigiero L (2015) Cytotoxicity,
cellular uptake and DNA damage by daunorubicin and its new analogues with modified
daunosamine moiety. Cell Biol Toxicol 21:139–147
116. Szafraniec E, Majzner K, Farhane Z, Byrne HJ, Lukawska M, Oszczapowicz I, Chlopicki S,
Baranska M (2016) Spectroscopic studies of anthracyclines: structural characterization and
in vitro tracking. Spectrochim Acta A 169:152–160
117. Majzner K, Wojcik T, Szafraniec E, Lukawska M, Oszczapowicz I, Chlopicki S,
Baranska M (2015) Nuclear accumulation of anthracyclines in the endothelium studied by
bimodal imaging: fluorescence and Raman microscopy. Analyst 140:2302–2310
118. Zhang Q, Lu X, Tang P, Zhang D, Zhong L, Tian J (2016) Gold nanoparticle (AuNP)-based
surface-enhanced Raman scattering (SERS) probe of leukemic lymphocytes. Plasmonics
11:1361–1368
119. Fabris L (2015) Gold-based SERS tags for biomedical imaging. J Opt 17:114002
120. Huang JY, Zong C, Xu LJ, Cui Y, Ren B (2011) Clean and modified substrates for direct
detection of living cells by surface-enhanced Raman spectroscopy. Chem Commun
47:5738–5740
121. Taylor J, Huefner A, Li L, Wingfieldc J, Mahajan S (2016) Nanoparticles and intracellular
applications of surface-enhanced Raman spectroscopy. Analyst 141:5037–5055
122. Jaworska A, Wojcik T, Malek K, Kwolek U, Kepczynski M, Ansary AA, Chlopicki S,
Baranska M (2015) Rhodamine 6G conjugated to gold nanoparticles as labels for both SERS
and fluorescence studies on live endothelial cells. Microchim Acta 182:119–127
123. Kneipp J, Kneipp H, McLaughlin M, Brown D, Kneipp K (2006) In vivo molecular probing
of cellular compartments with gold nanoparticles and nanoaggregates. Nano Lett 6:
2225–2231
124. Panikkanvalappil SR, Hira SM, El-Sayed MA (2016) Elucidation of ultraviolet
radiation-induced cell responses and intracellular biomolecular dynamics in mammalian
cells using surface-enhanced Raman spectroscopy. Chem Sci 7:1133–1141
125. Huang J, Zong C, Shen H, Cao U, Ren B, Zhang Z (2013) Tracking the intracellular drug
release from graphene oxide using surface-enhanced Raman spectroscopy. Nanoscale
5:10591–10598
126. Kim JH, Kim JS, Choi H, Lee SM, Jun BH, Yu KN, Kuk E, Kim YK, Jeong DH, Cho MH,
Lee YS (2006) Nanoparticle probes with surface enhanced Raman spectroscopic tags for
cellular cancer targeting. Anal Chem 78:6967–6973
127. Cialla-May D, Zheng X-S, Weber K, Popp J (2017) Recent progress in surface-enhanced
Raman spectroscopy for biological and biomedical applications: from cells to clinics. Chem
Soc Rev 46:3945–3961
128. Jaworska A, Jamieson LE, Malek K, Campbell CJ, Choo J, Chlopicki S, Baranska M (2015)
SERS-based monitoring of the intracellular pH in endothelial cells: the influence of the
extracellular environment and tumour necrosis factor-a. Analyst 140:2321–2329
129. Talley CE, Jusinski L, Hollars CW, Lane SM, Huser T (2004) Intracellular pH sensors based
on surface-enhanced Raman scattering. Anal Chem 76:7064–7068
130. Wang F, Widejko RG, Yang Z, Nguyen KT, Chen H, Fernando LP, Christensen KA,
Anker JN (2012) Surface-enhanced Raman scattering detection of pH with
silica-encapsulated 4-mercaptobenzoic acid-functionalized silver nanoparticles. Anal Chem
84:8013–8019
131. Wei H, Willner MR, Marr LC, Vikesland PJ (2016) Highly stable SERS pH nanoprobes
produced by co-solvent controlled AuNP aggregation. Analyst 141:5159–5169
132. Lim JK, Joo S-W (2006) Gold nanoparticle-based pH sensor in highly alkaline region at
pH > 11: surface-enhanced Raman scattering study. Appl Spectrosc 60:847–852
133. Wang Z, Bonoiu A, Samoc M, Cui Y, Prasad PN (2008) Biological pH sensing based on
surface enhanced Raman scattering through a 2-aminothiophenol-silver probe. Biosens
Bioelectron 23:886–891
134. Chen P, Wang Z, Zong S, Zhu D, Chen H, Zhang Y, Wu L, Cui Y (2016) pH-sensitive
nanocarrier based on gold/silver core–shell nanoparticles decorated multi-walled carbon
manotubes for tracing drug release in living cells. Biosens Bioelectron 75:446–451
135. Song J, Zhou J, Duan H (2012) Self-assembled plasmonic vesicles of SERS-Encoded
amphiphilic gold nanoparticles for cancer cell targeting and traceable intracellular drug
delivery. J Am Chem Soc 134:13458–13469
136. Bobba KN, Saranya G, Alex SM, Velusamy N, Maiti KK, Bhuniya S (2018) SERS-active
multi-channel fluorescent probe for NO: guide to discriminate intracellular biothiols. Sensors
Actuators B Chem 260:165–173
137. Cao Y, Li DW, Zhao LJ, Liu XY, Cao XM, Long YT (2015) Highly selective detection of
carbon monoxide in living cells by palladacycle carbonylation-based surface enhanced
raman spectroscopy nanosensors. Anal Chem 87:9696–9701
138. Pissuwan D, Hobro AJ, Pavillon N, Smith NI (2014) Distribution of label free cationic
polymer-coated gold nanorods in live macrophage cells reveals formation of groups of
intracellular SERS signals of probe nanoparticles. RSC Adv 4:5536
139. Kang JW, So PTC, Dasari RR, Lim D-K (2015) High resolution live cell Raman imaging
using subcellular organelle-targeting sers-sensitive gold nanoparticles with highly narrow
intra-nanogap. Nano Lett 15:1766–1772
140. Xu L, Zhao S, Ma W, Wu X, Li S, Kuang H, Wang L, Xu C (2016) Multigaps embedded
nanoassemblies enhance in situ Raman spectroscopy for intracellular telomerase activity
sensing. Adv Funct Mater 26:1602–1608
141. Kaur G, Dufour JM (2012) Cell lines: valuable tools or useless artifacts. Spermatogenesis
2:1–5
142. Pan C, Kumar C, Bohl S, Klingmueller U, Mann M (2009) Comparative proteomic
phenotyping of cell lines and primary cells to assess preservation of cell type-specific
functions. Mol Cell Proteomics 8:443–450
143. Maslak E, Gregorius A, Chlopicki S (2015) Liver sinusoidal endothelial cells (LSECs)
function and NAFLD; NO-based therapy targeted to the liver. Pharmacol Rep 67:689–694
144. Edelberg JM, Aird WC, Wu W, Rayburn H, Mamuya WS, Mercola M, Rosenberg RD
(1998) PDGF mediates cardiac microvascular communication. J Clin Invest 102:837–843
145. Wayhart JP, Lawson HA (2017) Animal models of metabolic syndrome A2. In: Conn P,
Michael BT (eds) Animal models for the study of human disease. 2nd edn. Academic Press,
pp 221–243 (Chapter 9)
146. Sørensen KK, Simon-Santamaria J, McCuskey RS, Smedsrød B (2011) Liver sinusoidal
endothelial cells. Compr Physiol 5:1751–1774
147. Wisse E (1972) An ultrastructural characterization of the endothelial cell in the rat liver
sinusoid under normal and various experimental conditions, as a contribution to the
distinction between endothelial and Kupffer cells. J Ultrastruct Res 38:528–562
148. Elvevold K, Smedsrod B, Martinez I (2008) The liver sinusoidal endothelial cell: a cell type
of controversial and confusing identity. Am J Physiol Gastrointest Liver Physiol 294:
G391–G400
149. Meyer J, Lacotte S, Morel P, Gonelle-Gispert C, Bühler L (2016) An optimized method for
mouse liver sinusoidal endothelial cell isolation. Exp Cell Res 349:291–301
150. Schie IW, Wu J, Weeks T, Zern MA, Rutledge JC, Huser T (2011) Label-free
characterization of rapid lipid accumulation in living primary hepatocytes after exposure
to lipoprotein lipolysis products. J Biophotonics 4:425–434
151. Testerink N, Ajat M, Houweling M, Brouwers JF, Pully VV, van Manen H-J, Otto C,
Helms JB, Vaandrager AB (2012) Replacement of retinyl esters by polyunsaturated
triacylglycerol species in lipid droplets of hepatic stellate cells during activation. PLoS ONE
7:e34945
152. Kochan K, Kus E, Filipek A, Szafrańska K, Chlopicki S, Baranska M (2017) Label-free
spectroscopic characterization of live liver sinusoidal endothelial cells (LSECs) isolated from
the murine liver. Analyst 142:1308–1319
153. Huebert RC, Jagavelu K, Liebl AF, Huang BQ, Splinter PL, LaRusso NF, Urrutia RA (2010)
Immortalized liver endothelial cells: a cell culture model for studies of motility and
angiogenesis. Lab Invest 90:1770–1781
154. Brutsaert DL (2003) Cardiac endothelial-myocardial signaling: its role in cardiac growth,
contractile performance, and rhythmicity. Physiol Rev 83:59–115
155. Tirziu D, Giordano FJ, Simons M (2010) Cell communications in the heart. Circulation
122:928–937
156. Lee JF, Barrett-O’Keefe Z, Garten RS, Nelson AD, Ryan JJ, Nativi JN, Richardson RS,
Wray DW (2016) Evidence of microvascular dysfunction in heart failure with preserved
ejection fraction. Heart 102:278–284
157. Garlanda C, Parravicini C, Sironi M, DeRossi M, Wainstok de Calmanovici R, Carozzi F,
Bussolino F, Colotta F, Mantovani A, Vecchi A (1994) Progressive growth in immunod-
eficient mice and host cell recruitment by mouse endothelial cells transformed by polyoma
middle-sized T antigen: implications for the pathogenesis of opportunistic vascular tumors.
Proc Natl Acad Sci U S A 91:7291–7295
158. Yeo BS, Mädler S, Schmid T, Zhang W, Zenobi R (2008) Tip-enhanced Raman
spectroscopy can see more: the case of cytochrome c. J Phys Chem C 112:4867–4873
159. Ogawa M, Harada Y, Yamaoka Y, Fujita K, Yaku H, Takamatsu T (2009) Label-free
biochemical imaging of heart tissue with high-speed spontaneous Raman microscopy.
Biochem Biophys Res Commun 382:370–374
160. Wood BR, Asghari-Khiavi M, Bailo E, McNaughton D, Deckert V (2012) Detection of
nano-oxidation sites on the surface of hemoglobin crystals using tip-enhanced Raman
scattering. Nano Lett 12:1555–1560
161. Kallaway C, Almond LM, Barr H, Wood J, Hutchings J, Kendall C, Stone N (2013)
Advances in the clinical application of Raman spectroscopy for cancer diagnostics.
Photodiagnosis Photodyn Ther 10:207–219
162. Galli R, Uckermann O, Koch E, Schackert G, Kirsch M, Steiner G (2013) Effects of tissue
fixation on coherent anti-Stokes Raman scattering images of brain. J Biomed Opt 19:071402
163. Mariani MM, Lampen P, Popp J, Wood BR, Deckert V (2009) Impact of fixation on in vitro
cell culture lines monitored with Raman spectroscopy. Analyst 134:1154–1161
164. Kong K, Kendall C, Stone N, Notingher I (2015) Raman spectroscopy for medical
diagnostics—from in-vitro biofluid assays to in-vivo cancer detection. Adv Drug Deliv Rev
89:121–134
165. Kendall C, Isabelle M, Bazant-Hegemark F, Hutchings J, Orr L, Babrah J, Baker R, Stone N
(2009) Vibrational spectroscopy: a clinical tool for cancer diagnostics. Analyst 134:
1029–1045
166. Dybas J, Marzec KM, Pacia MZ, Kochan K, Czamara K, Chrabąszcz K, Staniszewska-
Ślęzak E, Małek K, Barańska M, Kaczor A (2016) Raman spectroscopy as a sensitive probe
of soft tissue composition—imaging of cross-sections of various organs vs. single spectra of
tissue homogenates. Trends Anal Chem 85:117–127
167. Wrobel TP, Marzec KM, Chlopicki S, Maślak E, Jasztal A, Franczyk-Żarów M,
Czyżyńska-Cichoń I, Moszkowski T, Kostogrys RB, Baranska M (2015) Effects of cow
carbohydrate high protein (LCHP) diet on atherosclerotic plaque phenotype in ApoE/
LDLR−/− mice: FT-IR and Raman imaging. Sci Rep 5:14002
168. Kochan K, Chrabaszcz K, Szczur B, Maslak E, Dybasa J, Marzec KM (2016) IR and Raman
imaging of murine brains from control and ApoE/LDLR−/− mice with advanced
atherosclerosis. Analyst 141:5329–5338
169. Zhang J, Fan Y, He M, Ma X, Song Y, Liu M, Xu J (2017) Accuracy of Raman
spectroscopy in differentiating brain tumor from normal brain tissue. Oncotarget 8:
36824–36831
170. Surmacki J, Musial J, Kordek R, Abramczyk H (2013) Raman imaging at biological
interfaces: applications in breast cancer diagnosis. Mol Cancer 12:48
171. Brozek-Pluska B, Musial J, Kordek R, Bailo E, Dieing T, Abramczyk H (2012) Raman
spectroscopy and imaging: applications in human breast cancer diagnosis. Analyst
137:3773–3780
172. Bratchenko IA, Artemyev DN, Myakinin OO, Khristoforova YA, Moryatov AA,
Kozlov SV, Zakharov VP (2017) Combined Raman and autofluorescence ex vivo
diagnostics of skin cancer in near-infrared and visible regions. J Biomed Opt 22:27005
173. Magee ND, Beattie JR, Carland C, Davis R, McManus K, Bradbury I, Fennell DA,
Hamilton PW, Ennis M, McGarvey JJ, Elborn JS (2010) Raman microscopy in the diagnosis
and prognosis of surgically resected nonsmall cell lung cancer. J Biomed Opt 15:026015
174. Andreou C, Neuschmelting V, Tschaharganeh DF, Huang CH, Oseledchyk A, Iacono P,
Karabeber H, Colen RR, Mannelli L, Lowe SW, Kircher MF (2016) Imaging of liver tumors
using surface-enhanced Raman scattering nanoparticles. ACS Nano 10:5015–5026
175. Cm Krishna, Rubina S (2015) Raman spectroscopy in cervical cancers: An update. J Cancer
Res Ther 11:10–17
176. Kamemoto LE, Misra AK, Sharma SK, Goodman MT, Luk H, Dykes AC, Acosta T (2010)
Near-infrared micro-Raman spectroscopy for in vitro detection of cervical cancer. Appl
177. Kast RE, Tucker SC, Killian K, Trexler M, Honn KV, Auner GW (2014) Emerging
technology: applications of Raman spectroscopy for prostate cancer. Cancer Metastasis Rev
33:673–693
178. Latka I, Dochow S, Krafft C, Dietzek B, Popp J (2013) Fiber optic probes for linear and
nonlinear Raman applications—Current trends and future development. Laser Photonics Rev
7:698–731
179. Lui H, Zhao J, McLean D, Zeng H (2012) Real-time Raman spectroscopy for in vivo skin
cancer diagnosis. Cancer Res 72:2491–2500
180. Desroches J, Jermyn M, Pinto M, Picot F, Tremblay MA, Obaid S, Marple E, Urmey K,
Trudel D, Soulez G, Guiot MC, Wilson BC, Petrecca K, Leblond F (2018) A new method
using Raman spectroscopy for in vivo targeted brain cancer tissue biopsy. Sci Rep 8:1792
181. Bergholt MS, Zheng W, Lin K, Ho KY, The M, Yeoh KG, So JBY, Huang Z (2011) In vivo
diagnosis of esophageal cancer using image-guided Raman endoscopy and biomolecular
modeling. Technol Cancer Res Treat 10:103–112
182. Bergholt MS, Lin K, Zheng W, Lau DP, Huang Z (2012) In vivo, real-time, transnasal,
image-guided Raman endoscopy: defining spectral properties in the nasopharynx and larynx.
J Biomed Opt 17:770021
183. Huang Z, The SK, Zheng W, Lin K, Ho KY, Teh M, Yeoh KG (2010) In vivo detection of
epithelial neoplasia in the stomach using image-guided Raman endoscopy. Biosens
Bioelectron 26:383–389
184. Haka AS, Volynskaya Z, Gardecki JA, Nazemi J, Shenk R, Wang N, Dasari RR,
Fitzmaurice M, Feld MS (2009) Diagnosing breast cancer using Raman spectroscopy:
prospective analysis. J Biomed Opt 14:54023
185. Matthäus C, Dochow S, Bergner G, Lattermann A, Romeike BFM, Marple ET, Krafft C,
Dietzek B, Brehm BR, Popp J (2012) In vivo characterization of atherosclerotic plaque
depositions by Raman-probe spectroscopy and in vitro coherent anti-stokes Raman
scattering microscopic imaging on a rabbit model. Anal Chem 84:7845–7851
186. Huang Z, McWilliams A, Lui H, McLean DI, Lam S, Zeng H (2003) Near-infrared Raman
spectroscopy for optical diagnosis of lung cancer. Int J Cancer 107:1047–1052
187. Krafft C, Sobottka SB, Schackert G, Salzer R (2005) Near infrared Raman spectroscopic
mapping of native brain tissue and intracranial tumors. Analyst 130:1070–1077
188. Duraipandian S, Zheng W, Ng J, Low JJ, Ilancheran A, Huang Z (2012) Simultaneous
fingerprint and high-wavenumber confocal Raman spectroscopy enhances early detection of
cervical precancer in vivo. Anal Chem 84:5913–5919
189. Draga ROP, Grimbergen MCM, Vijverberg PLM, van Swol CFP, Jonges TGN,
Kummer JA, Ruud Bosch JLH (2010) In vivo bladder cancer diagnosis by high-volume
Raman spectroscopy. Anal Chem 82:5993–5999
190. Kumar GL, Rucbeck L (eds) (2009) Immunohistochemical (IHC) Staining Methods. Dako
North America, Carpinteria
191. Vendrell M, Maiti KK, Dhaliwal K, Chang Y-T (2013) Surface-enhanced Raman scattering
in cancer detection and imaging. Trends Biotechnol 31:249–257
192. Schlücker S, Küstner B, Punge A, Bonfig R, Marx A, Ströbel P (2006) Immuno-Raman
microspectroscopy: in situ detection of antigens in tissue specimens by surface-enhanced
Raman scattering. J Raman Spectrosc 37:719–721
193. Jehn C, Küstner B, Adam P, Marx A, Ströbel P, Schmuckd C, Schlücker S (2009) Water
soluble SERS labels comprising a SAM with dual spacers for controlled bioconjugation.
194. Salehi M, Schneider L, Ströbel P, AMarx A, Packeisend J, Schlücker S (2014) Two-color
SERS microscopy for protein co-localization in prostate tissue with primary antibody–
protein A/G–gold nanocluster conjugates. Nanoscale 6:2361–2367
195. Wang X-P, Zhang Y, König M, Papadopoulou E, Walkenfort B, Kasimir-Bauer S,
Bankfalvic A, Schlücker S (2016) iSERS microscopy guided by wide field immunofluores-
cence: analysis of HER2 expression on normal and breast cancer FFPE tissue sections.
Analyst 141:5113–5119
196. Wang Y, Kang S, Khan A, Ruttner G, Leigh SY, Murray M, Abeytunge S, Peterson G,
Rajadhyaksha M, Dintzis S, Javid S, Liu JTC (2016) Quantitative molecular phenotyping
with topically applied SERS nanoparticles for intraoperative guidance of breast cancer
lumpectomy. Sci Rep 6:21242
197. Quynh LM, Nam NH, Kong K, Nhung NT, Notingher I, Henini M, Luong NH (2016)
Surface-enhanced Raman spectroscopy study of 4-atp on gold nanoparticles for basal cell
carcinoma fingerprint detection. J Electron Mater 45:2563–2568
198. Chen Y, Zheng X, Chen G, He C, Zhu W, Feng S, Xi G, Chen R, Lan F, Zeng H (2011)
Immunoassay for LMP1 in nasopharyngeal tissue based on surface-enhanced Raman
scattering. Int J Nanomedicine 7:73–82
Chapter 7
Hydantoins and Mercaptoimidazoles:
Vibrational Spectroscopy as a Probe
of Structure and Reactivity in Different
Environments, from the Isolated
Molecule to Polymorphs
Rui Fausto, Gulce O. Ildiz, Elisa M. Brás and Bernardo A. Nogueira
Abstract In this chapter, we highlight the power of vibrational spectroscopy as

central technique to investigate the structure and reactivity of two relevant families
of nitrogen-containing heterocyclic molecules: hydantoins and mercaptoimidazoles.
Infrared spectroscopy is used in connection with the matrix isolation technique to
investigate the structures of the isolated molecules and their photochemistry, while
both infrared and Raman spectroscopies, supplemented by thermodynamics,
microscopy, and diffraction techniques, are used to investigate neat condensed
phases of the compounds and transitions between these phases. The experimental
studies are supported by extensive computational studies, which include several
approaches for detailed analysis of the electron density.
7.1 Introduction
Vibrational spectroscopy has long been recognized as a powerful technique for

investigation of structure and reactivity of chemical systems in different conditions.
The technological progresses achieved in the last decades in instrumentation,
together with the development of new, powerful variants of the fundamental
vibrational spectroscopy methods and sampling techniques, allowed the modern
vibrational spectroscopies to reach an unprecedented level of sophistication.
Moreover, the availability of fast computers and efficient software for data col-
lection and analysis, as well as for prediction of vibrational spectra, has also
R. Fausto (&) G. O. Ildiz E. M. Brás B. A. Nogueira

CQC, Department of Chemistry, University of Coimbra, 3004-535 Coimbra, Portugal
e-mail: rfausto@ci.uc.pt
G. O. Ildiz
Department of Physics, Faculty of Sciences and Letters, Istanbul Kultur University,
Ataköy Campus, Bakirköy, 34156 Istanbul, Turkey

200 R. Fausto et al.
strongly contributed to the success of vibrational spectroscopy in the present days.

One has nowadays available experimental data of superior quality, and their
interpretation can be ground in solid theoretical foundations, facilitated by the
comparison of experimental and calculated spectra, and helped by use of modern,
effective data post-processing methods, including multivariate statistical-based
procedures.
If per se vibrational spectroscopy is already an authoritative technique for
structure and reactivity evaluation, when used simultaneously with other experi-
mental methods it strongly increases its predictive and interpretative power. Among
these, X-ray diffraction (XRD), allowing access to details of the structure of
crystalline materials and thermodynamic methods (such as differential scanning
calorimetry (DSC)), providing information on phase transition properties, deserve
here to be mentioned. Vibrational spectroscopy can use structural information
obtained from XRD and extract detailed information on several important proper-
ties of the crystalline materials not easy accessible (or not accessible at all) to X-ray
diffraction methods, in particular those that are related with hydrogen atoms, which
are weak X-ray scatters. Hydrogen bonding, above all, has been a relevant subject
for which vibrational spectroscopy has proved to be a superior investigation
method. On the other hand, vibrational spectroscopy, being a molecule-based
technique, magnifies the interpretative capabilities of the thermodynamic approa-
ches on differentiating and characterizing structurally different phases of a com-
pound and interpreting phase transition data at a molecular level.
For the studies on isolated molecules, matrix isolation vibrational spectroscopy
(in particular matrix isolation infrared spectroscopy) has emerged as a very powerful
and elegant method, overcoming some of the difficulties of gas phase spectroscopy
such as the complexity of the spectra resulting from rotational structure, extensive
anharmonic contributions, and the need for the compound under study to have a
substantial vapor pressure in order to acquire good quality spectra.
Matrix isolation is a technique where molecules are trapped from the gas phase
in an environment of solidified inert gases at temperatures close to absolute zero.
The method was first developed almost simultaneously by the groups of Pimentel at
Berkeley (USA) [1], and Porter at Cambridge (UK) [2] and had its name coined in
the historical paper by Whittle et al. [1] “Matrix isolation method for the experi-
mental study of unstable species”.
Though matrix isolation was invented as a tool for study reactive species, the
method has also many advantages over other techniques for the study of stable
molecules and has been used extensively for this purpose too. The main advantages
of the method can be attributed to the very low work temperature, virtual absence of
intermolecular interactions between the molecules under investigation, and rigidity
and chemical inertness of the matrix medium. Because inert gases are generally
used as matrix host, solute–solvent interactions can be neglected for most of the
practical applications. On the other hand, molecular diffusion and rotation are
inhibited (except in the case of very small molecules), allowing registering of pure
vibrational spectra, which can be easily compared with theoretically predicted
spectroscopic data (that are usually obtained for molecules in vacuo). Also, hot
7 Hydantoins and Mercaptoimidazoles: Vibrational … 201
vibrational transitions as well as subtractive combination tones are suppressed,

since excited vibrational states are depopulated at the low work temperatures typical
of the matrix isolation experiments, introducing an additional simplification in the
spectra and contributing to their increased resolution.
Another interesting feature of matrix isolation, this time particularly useful for
reactivity studies (in particular in photochemistry) , results from the fact that isolated
molecules in matrices are cage confined. Thus, if fragmentation of a matrix-isolated
species takes place, for example, in result of photoexcitation, most of the times the
obtained fragments stay in the matrix cage where they were formed. Thus, no sub-
sequent cross-reactions involving species resulting from fragmentation of different
reactant molecules can occur, strongly reducing the number of possible products in
comparison with gas phase or solution studies and open the gate for a detailed study of
the mechanisms of unimolecular photochemically induced processes.
In addition to the above-mentioned advantages, standard matrix isolation setups
can be also used, without any modification, for preparation of samples of a solid
pristine compound, by the sole condensation of its vapor. This method of sample
preparation in general leads to producing of amorphous solids that can then be
submitted to a controlled temperature program to induce in situ crystallization,
which can be spectroscopically followed. The temperature program can be chosen
to mimic a temperature program used, for example, in a DSC experiment, so that
spectroscopic data can be directly compared with thermodynamic information.
In the studies described in the present chapter, matrix isolation infrared spec-
troscopy has been used to investigate the structures of the isolated molecules of the
targeted compounds and their photochemistry. The neat condensed phases and
phase transitions have been investigated by infrared and Raman spectroscopies,
supplemented by DSC, polarized-light thermomicroscopy (PLTM), and XRD.
Computational studies, including the use of several approaches for detailed analysis
of the electron density of the investigated molecules, were used to help in inter-
preting the experimental data. On the whole, it is our aim to highlight here the
power of vibrational spectroscopy as central technique to investigate the structure
and reactivity of two relevant families of nitrogen-containing heterocyclic mole-
cules: hydantoins and mercaptoimidazoles.
7.2 Structures and Infrared Spectra of the Isolated

Molecules
7.2.1 Mercaptoimidazoles
Mercaptoimidazoles constitute an important family of nitrogen-containing hetero-

cyclic molecules that have been shown to present antioxidant properties toward
several oxygen-derived toxic species. The mercaptoimidazole moiety is present in
natural antioxidant compounds such as ergothioneine or ovothiol [3–6].
Mercaptoimidazoles are also similar compounds to thiouracil, a relevant antithyroid
drug that was used for the first time in 1943 as therapy of Graves’ disease [7], and
that is still in use nowadays.
An interesting property of this type of compounds is that they may undergo slow
oxidation to the disulfide dimeric form in acidic media [8]. In turn, structurally
appropriate disulfides may undergo homolytic S–S bond photodissociation into a
pair of radicals exhibiting suitable magnetic and/or optical properties to act as
photoswitches [9, 10].
The molecules studied in our laboratory were 2-thiobenzimidazole (TBI),
1-methyl-2-thioimidazole (MTI), and 1-methyl-2-thiobenzimidazole (MTBI). All
these molecules can exist in two tautomeric forms, thione and thiol, whose struc-
tures are represented in Fig. 7.1. DFT(B3LYP)/6-311++G(d,p) calculations
demonstrated that for all molecules, the thione tautomer is more stable than the thiol
form [11, 12]. The relative energy difference between the tautomers increases in the
order MTI < TBI < MTBI (38.8, 44.6 and 49.8 kJ mol−1, respectively), indicat-
ing that the presence of both benzo and methyl substituents stabilizes the thione
tautomer. In the thiol tautomer, the SH group points toward the imine ring nitrogen
atom (Fig. 7.1), both to minimize repulsions between the sulfhydryl hydrogen atom
and the ortho hydrogen atom or methyl substituent and to maximize the attractive
interaction between the nearly antiparallel bond dipoles associated with the SH and
C=N ring bonds.
Fig. 7.1 Structures of TBI,

MTI, and MTBI tautomeric
forms (thione: left; thiol:
right)
In practical terms, the heavy-atom skeleton of the two tautomers of all studied
molecules is planar. In the case of the thione tautomers, the heavy-atom skeleton
planar structures correspond to true minima on the investigated potential energy
surfaces. For thiol forms, the planar geometry is also a minimum energy structure
for TBI, while for the methyl-substituted molecules (MTI, MTBI), the calculations
predict two slightly nonplanar equivalent-by-symmetry minima, with N=C–S–H
dihedral angles of a few degrees. However, these forms are separated from each
other by a very low energy barrier (below 0.1 kJ mol−1), which stays below the
zero-point vibrational level of these structures. Under these circumstances, the
heavy-atom skeleton planar geometry corresponds also to the most probable
structure for the two molecules in the ground vibrational state, thus being the one
relevant in practical terms.
An interesting distinct structural feature in the two methyl-substituted molecules
is the orientation of the methyl group (Fig. 7.1). In MTI, the methyl group is
oriented in such a way that the methyl hydrogen atom placed in the plane of the
heavy-atom skeleton of the molecule points away from the sulfur substituent, in
order to minimize the methyl/S steric repulsions. This is true for both thione and
thiol tautomers. On the other hand, in MTBI, the in-plane methyl hydrogen atom
points toward the sulfur substituent in both tautomeric forms, because in this case,
the presence of the benzo substituent introduces a stronger steric/electrostatic
repulsion between the methyl group and the closest located phenyl hydrogen atom.
Such interaction can in fact be expected to attain its maximum importance for an
orientation of the methyl group identical to that found in MTI.
More important is the conclusion, extracted from the structural data obtained for
all molecules, that no intramolecular hydrogen bond exists in any of the tautomers.
In fact, the ∡N–H⋯S and ∡S–H⋯N angles in the thione and thiol tautomers,
respectively, are in the range 67–82°, i.e., they are considerably smaller than the
commonly accepted minimum value for a ∡D–H⋯A (D, donor; A, acceptor) angle
allowing the establishment of a hydrogen bond (*110° [13]). As it will be shown
in Sect. 7.3.1, this structural feature is of extreme importance in determining the
mechanism of unimolecular photo tautomerization in these molecules.
The infrared spectra of the matrix-isolated compounds are shown in Fig. 7.2,
where they can be compared with the theoretically predicted spectra for the two
tautomers. The results are clear regarding the sole presence in the matrices of the
thione tautomer. This result allows us also to conclude that no tautomerization takes
place upon sublimation of the compounds, and that the thione form, which was
initially present in the crystals [14–16], was kept upon the phase transition.
The excellent agreement between the calculated and experimental data facilitated
the assignment of the bands, which can be found elsewhere [11, 12]. The two
striking observations are:
• the extensive site splitting and broadening observed for the bands assigned to
vibrations with dominant contributions of the NH modes (in particular the
bending and rocking modes), which demonstrates the involvement of the NH
moieties in specific interactions with the host matrix atoms and reflects in the
Fig. 7.2 Experimental infrared spectrum of MTI (top), TBI (middle), and MTBI (bottom)
isolated in an argon matrix (I), compared with the B3LYP/6-311++G(d,p) calculated spectra of
thione and thiol tautomers (II and III, respectively). Reprinted with permission from [11, 12],
copyright (2018) Elsevier
observed high sensitivity of the vibrations of the NH groups to local

environment,
• for the methyl-substituted compounds, also the splitting and relative broadening
of the bands associated with the bending vibrations of the methyl group, which
points to a partially free rotation of this group even in the matrix media, thus
implying a large amplitude methyl torsional vibration that, in turn, leads to an
increased dispersion of the vibrational levels associated with the bending modes.
In consonance with the spectroscopically suggested appreciable conformational
freedom of the methyl groups, the B3LYP/6-311++G(d,p) calculated methyl
internal rotation barriers for MTI and MTBI amount only to 1.2 and 2.7 kJ mol−1,
respectively [11, 12].
7.2.2 Hydantoins
Hydantoins are heterocyclic compounds that are derived from imidazolidine. They
exhibit relevant physiological activities, for example, as anticonvulsive,
antiepileptic, anti-inflammatory, and anticancer drugs [17–22] and have also been
suggested as potential therapeutic agents for the treatment of HIV-1 [23, 24].
Like mercaptoimidazoles, hydantoins may also exhibit tautomerism, and tau-
tomeric forms of hydantoins bearing either one or two OH moieties may also be
conceivable to exist. Taking the di-keto most stable tautomer as reference, these
tautomers can be viewed as resulting from migration of the lactam hydrogen atoms
to the carbonyl oxygen atoms. However, these tautomers have been shown to be
higher in energy than the most stable tautomer by at least 70 kJ mol−1 [25] and
have never been observed experimentally for the simplest members of the family.
In our studies on hydantoins [26–30], we have considered both 1-methyl- and
5-methyl-substituted hydantoins (1MH, 5MH) and 5-acetic acid hydantoin (AAH),
besides the parent, unsubstituted hydantoin (H) (Fig. 7.3). Those studies allowed us
to establish the basic structural properties of the hydantoin moiety. It could be
concluded, for example, that the hydantoin ring tends to assume a planar (or
quasi-planar) geometry, sharing common geometrical and electronic features in all
investigated compounds. To mention just the most important ones:
Fig. 7.3 Representation of the studied hydantoin molecules

• The relative lengths of the two distinct C=O bonds in the ring are dictated by the
degree of p electronic delocalization from the nitrogen atoms, being longer
when the C=O fragment is connected to two nitrogen atoms than when con-
nected to a single nitrogen;
• the internal angles of the ring with a nitrogen atom in the apex are much larger
(around 113°) than those with carbon atoms in the apex (*101–106°), due to
different s–p compositions of the hybrid orbitals of the N and C atoms used to
make the ring bonds;
• the p charge of the oxygen atom of the carbonyl group connected to two
nitrogen atoms is more negative than that of the oxygen atom of the second
carbonyl group (connected to a single nitrogen atom), in agreement with its
longer bond length and involvement in a more extended p mesomerism;
• the r charges of the two oxygen atoms are considerably less negative than the
corresponding p charges and have opposite relative values, showing that a larger
p bond polarization toward the oxygen atom leads to reduce the trend for the
associated r bond to be polarized in the same direction. These structural char-
acteristics have relevant consequences for the reactivity of the hydantoin moiety,
as it will be pointed out in Sect. 7.3.2.
The parent hydantoin, as well as its studied methyl derivatives (1MH, 5MH)
have no conformational flexibility, being, by that reason, quite rigid structurally. On
the other hand, AAH bears a conformationally flexible substituent, with three
different conformationally relevant degrees of freedom, and has 13 different con-
formers: six forms bearing a cis carboxylic group (O=C–O–H dihedral of *0°) and
seven possessing this group in the trans configuration (*180°). The most stable
conformer of AAH is shown in Fig. 7.3, being a cis carboxylic acid form. The
remaining cis carboxylic acid conformers have relative energies within 17 kJ mol−1
[B3LYP/6-311++G(d,p) calculated values], while all trans carboxylic acid con-
formers have relative energies of at least 19 kJ mol−1, with the highest energy
conformer (Fig. 7.3) having a relative energy of about 40 kJ mol−1.
The electronic structure of the studied hydantoins was investigated by several
methods of electron density analysis, starting with the simple picture provided by
atomic charges p/r partition, and going through natural bond orbital (NBO) and
Atoms in Molecules (AIM) analyses. All these analyses were found to be consistent
with the usual representation of the structure of the hydantoin moiety in terms of
different resonance structures (Fig. 7.4), with the nitrogen atoms exhibiting a
positive p charge and the oxygen atoms bearing a negative p charge. The most
positively charged nitrogen atom was found to be that showing a positive charge in
a larger number of mesomeric structures (structures III and IV in Fig. 7.4), while
the relative p charges of the oxygen atoms are also in agreement with the relative
number of resonance structures where these atoms appear negatively charged. On
the other hand, the results also showed that the polarization of the r system follows
the inverse trends compared to those characterizing the p system. For the carbonyl
groups, a higher p bond polarization toward oxygen leads to reduce the trend for the
Fig. 7.4 I: dominant canonic form of the hydantoin ring. II, III, and IV: Mesomeric structures
assuming delocalization of p electrons in the NCO fragments. R,R′ = H,H (H); CH3,H (1MH); H,
CH3 (5MH); H,CH2COOH (AAH)
associated r bond to be polarized in the same direction, and the same applies to the
relative polarization of the p and r systems associated with the carbon–nitrogen
bonds.
Orbital interactions obtained from the NBO analyses are also in good agreement
with the picture obtained based on both the geometrical parameters and the atomic
charges [26, 28, 29], and the same applies when relevant parameters extracted from
the AIM analysis for electron distribution characterization are taken into account.
For example, a direct correlation between the value of the charge density at the
bond critical points of the five bonds of the hydantoin ring in 1MH and the
corresponding bond lengths was found to exist, the shorter the bond, the greater the
charge density at the associated bond critical point [28].
On the whole, the charge density analyses reveal that the main electronic effects
in the hydantoin ring are, by one side, the N-to-O p electron donation, and, by the
other side, the r system backdonation from the r lone electron pairs of the oxygen
atoms to the ring.
The infrared spectra of the investigated hydantoins isolated in argon matrices
were found to fit well the predicted infrared spectra for the calculated minimum
energy structures (see example in Fig. 7.5, concerning 1MH). For AAH, the
analysis of the experimental spectrum revealed that only the most stable conformer
of the molecule subsists in the matrix upon deposition. According to the calculated
relative conformational energies, the room temperature gas phase population of the
most stable AAH conformer prior to matrix deposition should be *93%. The
predicted room temperature gas phase populations for all other conformers but
Fig. 7.5 Experimental infrared spectrum of monomeric 1MH isolated in an argon matrix at 10 K
(I; top) and B3LYP/6-311++G(d,p) calculated spectrum (wavenumbers scaled by 0.978) (II;
bottom). Reprinted with permission from [28], copyright (2014) American Chemical Society
the second more stable one are smaller than 0.8%, i.e., below the detection limit of
the experimental technique; that of the second more stable conformer is *6%, and
in principle large enough to allow for its experimental detection. However, the
energy barrier separating this form from the most stable conformer is very small
(ca. 5 kJ mol−1), and under these conditions, it shall promptly convert to the lowest
energy conformer during matrix deposition. Indeed, conformational cooling leading
to conversion, at the time of matrix deposition, of higher energy conformers into
lower energy forms from which they are separated by small energy barriers (of a
few kJ mol−1) is a well-known common phenomenon [31, 32].
Detailed assignments for the matrix isolation spectra of the studied hydantoins
have been described by Ildiz et al. [26] and Nogueira et al. [28, 29].
7.3 Photochemistry for the Matrix-Isolated Molecules
7.3.1 Mercaptoimidazoles
The matrix-isolated thione tautomers of the studied mercaptoimidazoles were

irradiated with narrowband ultraviolet (UV) light provided by a tunable Laser/
MOPO system. The chosen UV wavelengths (261 nm, in the case of MTI, and
307 nm for both TBI and MTBI) were selected taking into account the absorption
spectra of the compounds in ethanol, the time-dependent DFT (TD-DFT) calculated
UV spectra for the two tautomeric forms of the compounds, as well as literature
data for similar compounds [11, 12, 33, 34].
In the result of the performed irradiations, the thione tautomers were converted
into the corresponding thiol forms (Fig. 7.6). Subsequent irradiation of the matrices
with shorter wavelength UV light (246 nm) successfully converted the previously
generated thiol tautomers of both benzo-substituted molecules (TBI, MTBI) back
to the corresponding thione forms (Fig. 7.6). On the other hand, the thiol ! thione
photochemical back conversion could not be induced for MTI, upon irradiation
with UV light of wavelengths 230 nm.
On the whole, the experimental results clearly demonstrate that the two tau-
tomers of the benzo-substituted mercaptoimidazoles can be photochemically
interconverted in a selective way, with appropriate choice of the excitation wave-
length, i.e., the populations of the two tautomers can be optically controlled in an
efficient manner through UV-driven reactions taking place under different excitation
conditions. Another important observation is that both tautomers were found to
show great stability regarding photodecomposition (no signals of photodecompo-
sition products could be observed in the spectra along all performed irradiation
experiments), what makes these chemical systems promising candidates for acting
as molecular switches.
The observed tautomeric reactions could be rationalized in terms of the
PhotoInduced Detachment Association mechanism (PIDA), first proposed in the
theoretical studies of Chmura et al. [35]. This mechanism involves the hydrogen
photodetachment and subsequent recombination of the radical species formed upon
excitation. According to the PIDA mechanism, the thiyl radical is the common
intermediate species in the thione ! thiol and thiol ! thione phototautomeriza-
tions. Absorption of the UV light photons takes the reactant species to an excited
state of (n/p)p* type, with subsequent internal conversion to a singlet state of (n/p)r
* type, leading to dissociation through cleavage of the NH or SH bonds, thus
generating a hydrogen atom and the corresponding thiyl radical. The recombination
of these might lead to the original species or attachment of the H atom to the S atom
or N atom (depending if one is considering the thione ! thiol or the inverse
process) yielding the corresponding photoproduct.
After formation of the intermediate radical, its subsequent recombination with
the H atom to form the final products can be expected to follow a product branch
that should depend essentially on the relative stability of the products, which favors
the thione tautomer and justifies the observed higher efficiency of the thiol !
thione process compared to the thione ! thiol isomerization. Note also that results
of DFT calculations support the proposed mechanism, since for both direct and
reverse tautomerizations, the energy of the radical pair stays below the excitation
energies. For example, in the case of TBI [12], for the thione ! thiol isomerization
E[(radical + H) − thione] = 381 kJ mol−1 and the used excitation wavelength was
Fig. 7.6 Results of UV irradiations giving rise to tautomerization reactions in MTI (top), TBI
(middle), and MTBI (bottom). Simulated difference spectra of the relevant species are shown for
comparison. Reprinted with permission from [11, 12], copyright (2018) Elsevier
307 nm, i.e., 390 kJ mol−1, whereas for the thiol ! thione isomerization E
[(radical + H) − thiol] = 325 kJ mol−1 and the used excitation wavelength was
246 nm, i.e., 486 kJ mol−1.
The fact that for MTI, the photoinduced conversion of the thiol tautomer into the
thione form could not be observed is also in agreement with the involvement of the
thiyl radical in the photochemical process and the PIDA mechanism. Indeed, while for
TBI and MTBI, the intermediate radicals were shown to have a significant spin
density associated with the unpaired electron in both N and S atoms, in the case of
MTI, the calculated spin density associated with the unpaired electron in the corre-
sponding thiyl radical is practically localized only in the sulfur atom, thus making
more difficult, for this molecule, recombination of the H atom at the N position. An
additional factor might also contribute to the inefficiency of the thiol ! thione tau-
tomerization in MTI: The required excitation wavelength to induce the hydrogen
atom release in MTI is shorter, so that larger energies must be applied to the molecule;
this leads to photodecomposition to take place, as testified by observation of bands
ascribable to photofragmentation products, in particular N-vinylidenemethanamine
[which shall be formed together with isothiocyanic acid (SCNH) or thiocyanic acid
(NCSH)] [12].
7.3.2 Hydantoins
The matrix-isolated hydantoins were subjected to UV irradiations in a similar way

to what has been described in the previous section for the mercaptoimidazoles. For
these compounds, no tautomerization reactions were observed. On the other hand,
the compounds were found to undergo fragmentation, following a common pattern
of reactivity. The common photoproducts are isocyanic acid (OCNH) and CO, the
first molecule being produced by cleavage of the weakest (longest) hydantoin C–C
and N–C ring bonds, as represented in Fig. 7.7. The process occurs, with all
probability, through a concerted mechanism, the third photoproduct being an imine
whose precise structure depends on the reactant molecule. Illustrative spectral data
showing the generation of the products of photofragmentation of 5MH is depicted
in Fig. 7.8.
Fig. 7.7 Schematic

representation of the
concerted mechanism for the
common photofragmentation
of the studied hydantoins. R,R
′ = H,H (H); CH3,H (1MH);
H,CH3 (5MH); H,CH2COOH
(AAH)
Fig. 7.8 B3LYP/6-311++G

(d,p) calculated (scaled)
infrared spectra (2320–
965 cm−1 region) for
isocyanic acid, CO and
methanimine (top; I), infrared
difference spectrum
(irradiated matrix) minus
(as-deposited) (middle, II),
and spectrum of the
as-deposited 5MH in argon
matrix (bottom; III).
Reprinted with permission
from [30], copyright (2017)
American Chemical Society
7.4 Neat Condensed Phases—Polymorphism and Phase

Transitions
7.4.1 Mercaptoimidazoles: The Case of the Thioimidazole

Dimer
Thioimidazole undergoes easy dimerization upon crystallization from tetrahydro-

furan (THF) solution, which results in production of the disulfide derivative {(2-
[(1H-imidazol-2-yl) disulfanyl]-1H-imidazole]; IDI}. Suitable crystals for single
crystal XRD experiments were collected, and their structure was solved. Very
interestingly, the material was found to constitute a new polymorph of the com-
pound, which, as described below, presents unique structural features.
The structure of a crystalline phase of IDI had been reported previously by
Bazargani et al. [36]. That polymorph (I) crystallizes in the monoclinic space group
C2/c, with a = 14.083(3) Å, b = 6.3928(13) Å, c = 9.922(2) Å and b = 122.29(3)°,
with four molecules per unit cell (Z = 4). The asymmetric unit of the crystals of
polymorph I contains one half-molecule (Z′ = ½). The S–S bond distance is 2.0713
(14) Å, the planar imidazole rings form an angle between them of 21.83(19)°, and
the C–S–S–C torsion angle is 83.62(17)°, with the two hydrogen atoms connected
to the ring nitrogen atoms pointing to opposite directions. Intermolecular N–H⋯S

hydrogen bonds result in the formation of linear chains along the c-direction;
further cg⋯cg interactions between the imidazole rings of adjacent chains in the a-
direction (cg⋯cg distance: 3.4466(19) Å) define the supramolecular structure in the
crystal.
The structure of the newly synthesized polymorph (II) was found to be con-
siderably more complex. The crystals of polymorph II belong to the noncen-
trosymmetric Ia monoclinic space group, with Z = 16 and cell parameters:
a = 7.45910(10) Å, b = 44.1680(8) Å, c = 11.3522(2) Å, and b = 103.0240(10)°.
The crystals belong to the rare category of crystals having an asymmetric unit with
Z′ > 1 (only about 9% of the ca. 800,000 structures contained in the Cambridge
Structural Database share this feature [37]). In the present case, the unit cell con-
tains four symmetry independent molecules. However, the most fascinating struc-
tural feature of this new polymorph is that the symmetry independent molecules
exhaust the whole set of conformers of the molecule: the abundant lowest energy
conformer and two high-energy forms not present in solution before crystallization.
The four IDI independent molecules present in the crystal of polymorph II
(extracted from the XRD data; A-D) together with the calculated conformers of the
molecule (1-3) are shown in Fig. 7.9. The calculated relative energies for the dif-
ferent conformers as isolated species, obtained at the B3LYP/6-311++G(d,p) level
of theory, are also given in the figure, as well as the C–S–S–C dihedral angles of all
represented structures. In the crystal, two molecules (C, D) are similar to conformer
1, and the remaining two, A and B, correspond to conformers 3 and 2, respectively.
Taking into account the relative energies of the IDI conformers (Fig. 7.9), the
isolated IDI molecule exists almost exclusively in conformer 1, both in gas phase
and solution (the energies shown in Fig. 7.9 correspond to gas phase data; in THF,
the calculated relative energies for conformers 2 and 3 are 9.7 and 9.3 kJ mol−1,
respectively). The most stable IDI conformer exhibits an intramolecular H-bond
interaction established between the unprotonated nitrogen atom of one of the
imidazole rings and the NH fragment of the second imidazole ring of the molecule.
Because of this interaction, in conformer 1 the number of positions readily available
for establishing intermolecular H-bonds is reduced to half, when compared with the
higher energy conformers. Indeed, to fulfill the four H-bond valences with inter-
molecular H-bonds in the crystal, the most stable IDI conformer has to break the
intramolecular H-bond. This fact makes the energy balance associated with the
establishment of the intermolecular H-bonds less favorable for conformer 1 than for
both conformers 2 and 3, which require only minor structural rearrangements to
establish intermolecular H-bonds.
It can then be concluded that the simultaneous observation of the three con-
formers of IDI in polymorph II results from a compromise between the greater
intrinsic stability of the isolated conformer 1, and the easier and more efficient in
energetic terms packing achieved by the higher energy conformers 2 and 3. In
polymorph I, the IDI molecules assume geometry similar to that of conformer 2.
The Raman spectrum (500–50 cm−1 region) of polymorph II is shown in
Fig. 7.10, where it is compared with the simulated spectrum built based on the
Fig. 7.9 Structures of the four nonequivalent molecules of IDI in the crystal of polymorph II
(top), and equilibrium geometries for the three conformers of the isolated IDI molecule (middle).
The C–S–S–C dihedral angles are indicated near each structure, and relative energies (in kJ mol−1)
are provided within parentheses for the calculated conformers. The bottom image corresponds to
the XRD determined unit cell (viewed along the a axis), with the four molecules constituting the
asymmetric unit labeled according to the names given here to each one of the nonsymmetrically
equivalent units in the crystal
calculated Raman spectra for the three conformers of IDI, summed according to the
relative abundance of the conformers in the crystalline state. As seen in Fig. 7.10,
these two spectra show a very good agreement, indicating that the intermolecular
potential in the crystalline phase does not affect too much the intramolecular
potentials, and also that the intermolecular hydrogen bonds involving conformer 1
in the crystal are comparable in strength to the intramolecular hydrogen bond
existing in the isolated molecule. The strong overlapping doublet of bands observed
at 331/326 cm−1 is characteristic of the m S–S stretching vibration, and its frequency
is well reproduced by the calculations (scaled value).
Figure 7.10 shows also the calculated Raman spectrum of conformer 2, which is
the one existing in polymorph I, for which the experimental spectrum has not yet
been reported. Assuming that the predicted spectrum of 2 would also fit properly
the experimental spectrum of polymorph I, one can expect that the spectrum of that
polymorph could be easily distinguishable from that of polymorph II by direct
comparison of the spectral region shown in the figure. For example, the doublets
observed in the Raman spectrum of polymorph II at 176/167, 226/222, 240/
236 cm−1 and the overlapped doublet at 331/326 cm−1 can be expected to be
Fig. 7.10 Room temperature Raman spectrum of polymorph II of IDI (top), simulated Raman
spectrum of polymorph II obtained by adding the calculated spectra of conformers 1, 2, and 3, in a
ratio 2:1:1 (middle), and calculated spectrum of conformer 2 (the one existing in the polymorph
I of the compound; bottom). Calculated spectra were scaled by 0.71
replaced by single bands at nearly the same frequencies in the Raman spectrum of
polymorph I. Ongoing studies will allow us to confirm these predictions.
7.4.2 Hydantoins: Polymorphism in 1MH and 5MH
The study of neat 1MH and 5MH, by combined use of DSC, PLTM, and infrared
and Raman spectroscopies, allowed to identify several new polymorphs of these
compounds, characterize them structurally, and shed light on the energetics and
other relevant features of the observed phase transitions [28–30].
For 1MH, two polymorphs were identified. Polymorph I was obtained by
crystallization of the low-temperature amorphous phase produced from fast con-
densation of the vapor of the compound, as well as from crystallization of most of
the solvents used during the polymorph screening work. Polymorph II was
obtained by sublimation, and by crystallization from methanol. This last polymorph
converts into polymorph I upon warming, in a low energy solid–solid transition
which takes place from about 110 to 135 °C. The so formed polymorph I then melts
at a temperature of 155.7 ± 0.7 °C with an enthalpy of fusion of
21.5 ± 0.3 kJ mol−1. The crystals of polymorph I are monoclinic, P21/c, with
Z = 4, a = 5.601(10) Å, b = 12.178(3) Å, c = 8.090(2) Å, a = c = 90°, and
b = 105.62(2)° [28]. The most relevant pattern of hydrogen bonds noticeable in the
crystal of polymorph I is the centrosymmetric ring formed by N–H⋯O hydrogen
bond interactions, which group 1MH molecules in dimeric units. In turn, poly-
morph II is orthorhombic, Pna21 space group, with Z = 4, a = 19.0258(4) Å,
b = 3.91210(10) Å, and c = 6.82880(10) Å [29]. The striking difference between
the H-bond network in the two polymorphs is that while in polymorph I, the N–
H⋯O interactions group the molecules in dimers (which are then associated by C–
H⋯O weak interactions); in polymorph II, the N–H⋯O interactions are used to
form zigzag chains, which are interspersed with other chains through the interac-
tions involving oxygen atoms and form an angle of ca. 36° between them [29].
In the case of 5MH, four polymorphs were identified. Polymorph I corresponds
to the commercial sample, which could be transformed into the remaining forms
(II–IV) by thermal treatment. The melting points of the polymorphs (following
their numbering) were found to be 147.2 ± 0.6, 143.0 ± 0.4, 120.1 ± 0.8, and ca.
94–96 °C [30]. Only the structure of polymorph III could be solved by XRD,
which resulted to be triclinic, P-1 space group, with Z = 2, a = 4.3618(2) Å,
b = 6.1535(2) Å, c = 10.3145(5) Å, a = 76.196(2)°, b = 80.860(3)°, and
c = 84.904(3)° [30]. In this crystal, the molecules are packed in layers, where the
molecules are joined in infinite chains through a head-to-tail pattern of hydrogen
bonding involving the C=O and NH groups.
The different polymorphs of the two compounds were investigated by infrared
and Raman spectroscopies.
In the case of 1MH, analysis of the H-bond sensitive regions of the infrared
spectra of the two polymorphs allowed to estimate the average energy per H-bond
in each polymorph. This was achieved by use of Rozenberg and coworkers’
empirical correlations [38, 39], which relate the frequency redshifts of the stretching
vibrations and/or the frequency blueshifts of the out-of-plane bending modes of the
H-bond donors (compared to the corresponding frequencies for the free groups;
matrix isolation data was used as reference) with H-bond properties, including
energies. The results indicate that slightly stronger N–H bonds exist in polymorph
I than in polymorph II (DH(H-bond) = −28 and −26 kJ mol−1, respectively). These
results are in agreement with the structural crystallographic data, which indicate that
the N–H bond lengths are longer, and both the N⋯O and H⋯O distances associated
with the N–H⋯O dominant intermolecular interactions are shorter in polymorph
I than in polymorph II.
A very interesting result was obtained by following the transformation of a
single crystal of polymorph II of 1MH into polymorph I, by temperature variation
Raman microspectroscopy. As shown in Fig. 7.11, upon heating, the crystal of
polymorph II converts into polymorph I in a transformation that kept crystal
integrity [29].
For 5MH, the Raman spectra of polymorphs I and II were found to be almost
identical (Fig. 7.12), except in relation to relative band intensities, indicating that
the orientation of the 5MH molecules in both crystals shall be very similar. On the
Fig. 7.11 (left): Raman spectra of a single crystal of 1MH (1850–1000 cm−1 spectral region)
showing the transformation of polymorph II into polymorph I, upon heating. Spectra are obtained
without polarization analyzer, i.e., for the reactant polymorph the spectrum corresponds to the X(Z
[YZ]X) spectrum. (right): single crystal of polymorph II used in the Raman measurements. The
Miller indices of the planes corresponding to the faces of the crystals are indicated in the figure; the
laser beam (X direction) was focused on the upper faces of the crystals, and the spectra collected at
180° scattering geometry ðXÞ; the crystal was sampled with the planes defining their side faces
approximated aligned along the Y and Z directions. Reprinted with permission from [29], copy-
right (2017) Elsevier
Fig. 7.12 Raman spectra of polymorphs I–III of 5MH. Reprinted with permission from [30],
copyright (2017) American Chemical Society
other hand, the Raman spectrum of polymorph III is clearly distinctive from those
of polymorphs I and II, suggesting a significant dissimilarity in the H-bond network
of polymorph III compared to those of polymorphs I and II (no Raman spectrum
could be obtained for polymorph IV, which was found to be very labile, promptly
converting to other polymorphs).
7.4.3 Hydantoins: The Unusual Conformational Selection

in AAH upon Crystallization
Several polymorphs were also identified for AAH, which could be characterized by
DSC, PLTM, Raman spectroscopy, and XRD. Polymorph I corresponds to the
commercial sample used, while the remaining polymorphs (II-V) were obtained by
recrystallizations from different solvents. The melting points of polymorphs I, II,
III, and V were found to be 214.3 ± 0.8, 184–185, 175–178, and 198–200 °C,
respectively. Polymorph IV transforms into V at 180–185 °C. Until now, only in
the case of polymorph III suitable crystals could be produced for single crystal
structure determination. The crystal was found to be orthorhombic, P212121 space
group, Z = 4, a = 7.6148(5) Å, b = 8.5592(6) Å, and c = 9.3406(6) Å. The
molecules form chains interconnected by H-bonds (N–H⋯OH), which define
parallel sheets of molecules.
Very interestingly, the molecules in the crystal assume the conformation of the
highest energy form predicted for the isolated molecule situation (Fig. 7.4), which
has a predicted relative energy of ca. 40 kJ mol−1. Considering that AAH has 13
conformers, it is impressive that the highest energy form is the one present in the
crystal. It is relatively frequent that crystallization takes place with selection of
conformers that are not the lowest energy form of the isolated molecule. However,
the selection of the most energetic among 13 conformers, with a relative energy as
high as 40 kJ mol−1 is, certainly, a very rare case, implying a strongly favorable
packing of the molecules in the crystal that can overcome the intrinsic unfavorable
energetic requirements of the selected constituting unit. The Raman spectroscopy
data obtained for the different AAH polymorphs is in agreement with this con-
clusion. In fact, the spectra of the polymorphs (Fig. 7.13) exhibit notorious dif-
ferences, which are compatible with substantially different crystal structures that, in
turn, can be correlated with existence of strong intermolecular interactions and
specific packing requisites. The accentuated dissimilarity between the Raman
spectra of the different polymorphs may also indicate that the conformation(s)
assumed by the individual molecules in each polymorph might be different.
Confirmation of this hypothesis requires the determination of the crystal structure of
the various polymorphs. Studies are on the way to achieve this goal.
Fig. 7.13 Raman spectra of the polymorphs I–IV of AAH
7.5 Conclusion
Modern vibrational spectroscopies play a central role in the investigation of the

structure and reactivity of chemical systems, from isolated molecules to complex
condensed phase structures. When used together with complementary techniques,
such as XRD, DSC, PLTM, and contemporary theoretical approaches, they allow
for elucidation of subtle structural details of the systems under investigation and
successfully extract fundamental information on the chemical transformations they
undergo, either induced by light or thermally.
This chapter intended to demonstrate some of the capabilities of vibrational
spectroscopy as major research tool to investigate the structure and reactivity of two
relevant closely-related families of nitrogen-containing heterocyclic molecules:
hydantoins and mercaptoimidazoles. Infrared spectroscopy, used in connection with
the matrix isolation technique and a set of complementary computational approa-
ches, allowed to characterize in details the structures of the isolated molecules of the
compounds selected for investigation and their photochemistry. Infrared and Raman
spectroscopies, supplemented by thermodynamics, microscopy and diffraction
techniques, were used to investigate neat condensed phases of the compounds and
the observed transitions between these phases.
The tautomerism in the studied mercaptoimidazoles was addressed in various
perspectives, from the determination of the preferred tautomeric species present in
the different phases investigated, till the photochemically induced (inter)conversion
between the thione and thiol tautomers of these compounds. The crystal structure of
a new polymorph of the dimer of thioimidazole was described in details by a

combined use of XRD, quantum chemical calculations, and Raman spectroscopy.
These studies reveal a fascinating crystal architecture, where four nonequivalent
molecules are present in the unit cell, which exhaust the existing conformers of the
isolated molecule of the compound.
The electronic and geometric major features of the hydantoin moiety were
described, leading to the very interesting picture of reciprocal influence of the r and
p electronic systems (through p electron charge donation from the nitrogen atoms to
the carbonyl oxygens and r backdonation from these latter to the ring), which
determine the details of the geometries of the studied molecules and, ultimately,
also determine their photofragmentation pattern, which takes place by cleavage of
the weakest bonds in the ring and keeping intact the fragment (OCNH) constituted
by the strongest (shorter) bonds of the hydantoin ring. Polymorphism shown by the
hydantoins has been studied, and several new polymorphs are identified for 1MH,
5MH, and AAH. For 1MH, an unusual transformation between the two poly-
morphs of the compound which takes place in keeping crystal integrity was
observed. For AAH, an intriguing case of conformational selection where the
highest energy conformer for the isolated molecule situation (higher in energy than
the most stable conformer by ca. 40 kJ mol−1) is chosen to form a crystal was
described.
Acknowledgements This investigation has been performed within the Project PTDC/QEQ-QFI/
3284/2014–POCI-01-0145-FEDER-016617, funded by the Portuguese “Fundação para a Ciência
e a Tecnologia” (FCT) and FEDER/COMPETE 2020-EU. The Coimbra Chemistry Centre
(CQC) is supported by FCT, through the project UI0313/QUI/2013, also cofunded by FEDER/
COMPETE 2020-EU. E.M.B. and B.A.N. thank FCT for the grant CCMAR/BI/0013/2017, within
Project PTDC/MAR-BIO/4132/2014, and the Ph.D. grant SFRH/BD/129852/2017, respectively.
R.F. and G.O.I. acknowledge the financial support through the project MATIS–Materiais e
Tecnologias Industriais Sustentáveis (FCT and CENTRO-01-0145-FEDER-000014).
References
1. Whittle E, Dows DA, Pimentel GC (1954) Matrix isolation method for the experimental study
of unstable species. J Chem Phys 22:1943–1944
2. Norman I, Porter G (1954) Trapped atoms and radicals in a glass ‘cage’. Nature 174:508–509
3. Holler TP, Ruan F, Spaltenstein A, Hopkins PB (1989) Total synthesis of marine
mercaptohistidines: ovothiols A, B, and C. J Org Chem 54:4570–4575
4. Hand CE, Honek JF (2005) Biological chemistry of naturally occurring thiols of microbial
and marine origin. J Nat Prod 68:293–308
5. Zoete V, Vezin H, Bailly F, Vergoten G, Catteau J-P, Bernier J-L (2000)
4-mercaptoimidazoles derived from the naturally occurring antioxidant ovothiols 2.
Computational and experimental approach of the radical scavenging mechanism. Free Rad
Res 32:525–533
6. Crépin A, Wattier N, Petit S, Bischoff L, Fruit C, Marsais F (2009) Aminoacid-derived
mercaptoimidazoles. Org Biomol Chem 7:128–134
7. Nagasaka A, Hidaka H (1976) Effect of antithyroid agents 6-propyl-2-thiouracil and

l-methyl-2-mercaptoimidazole on human thyroid iodide peroxidase. J Clin Endocrinol Metab
43:152–158
8. Heath H, Toennies G (1958) The preparation and properties of ergothioneine disulphide.
Biochem J 68:204–210
9. Phan H, Lekin K, Winter SM, Oakley RT, Shatruk M (2013) Photoinduced solid state
conversion of a radical r-dimer to a p-radical pair. J Am Chem Soc 135:15674–15677
10. Matsuzaki H, Fujita W, Awaga K, Okamoto H (2003) Photoinduced phase transition in an
organic radical crystal with room-temperature optical and magnetic bistability. Phys Rev Lett
91:017403
11. Brás EM, Fausto R (2018) An insight into methimazole phototautomerism: Central role of the
thiyl radical and effect of benzo substitution. J Mol Struct. https://doi.org/10.1016/j.molstruc.
2018.02.013 (in press)
12. Brás EM, Fausto R (2018) Controlled light-driven switching in 2-thiobenzimidazole.
J Photoch Photobio A 357:185–192
13. Arunan E, Desiraju GR, Klein RA, Sadlej J, Scheiner S, Alkorta I, Clary DC, Crabtree RH,
Dannenberg JJ, Hobza P, Kjaergaard HG, Legon AC, Mennucci B, Nesbitt DJ (2011)
Definition of the hydrogen bond. Pure Appl Chem 83:1637–1641
14. Lodochnikova OA, Bodrov AV, Saifina AF, Nikitina LE, Litvinov IA (2013) A new
polymorph of methimazole: Single crystal and powder X-ray diffraction study. J Struct Chem
54:140–147
15. Khan H, Badshah A, Shaheen F, Giek C, Qureshi RA (2008) 1-methyl-1H-benzimidazole-2
(3H)-thione. Acta Cryst E 64:o1141
16. Form GR, Raper ES, Downie TC (1976) The crystal and molecular structure of
2-mereaptobenzimidazole. Acta Cryst B 32:345–348
17. Block SS (2003) Disinfection, sterilization and preservation, 4th edn. Lea & Febiger Inc.,
Philadelphia, USA
18. Kumar CSA, Kavitha CV, Vinaya K, Prasad SBB, Thimmegowda NR, Chandrappa S,
Raghavan SC, Rangappa KS (2009) Synthesis and in vitro cytotoxic evaluation of novel
diazaspiro bicyclo hydantoin derivatives in human leukemia cells: a SAR study. Invest New
Drugs 27:327–337
19. Kavitha CV, Nambiar M, Kumar CSA, Choudhary B, Muniyappa K, Rangappa KS,
Raghavan SC (2009) Novel derivatives of spirohydantoin induce growth inhibition followed
by apoptosis in leukemia cells. Biochem Pharmacol 77:348–363
20. Sarges R, Schnur RC, Belletire JL, Peterson MJ (1988) Spiro hydantoin aldose reductase
inhibitors. J Med Chem 31:230–243
21. Yang K, Tang Y, Iczkowski KA (2010) Phenyl-methylene hydantoins alter CD44-specific
ligand binding of benign and malignant prostate cells and suppress CD44 isoform expression.
Am J Transl Res 2:88–94
22. Park HS, Choi HJ, Shin HS, Lee SK, Park MS (2007) Synthesis and characterization of novel
hydantoins as potential COX-2 inhibitors: 1,5-Diarylhydantoins. Bull Korean Chem Soc
28:751–757
23. Comber RN, Reynolds RC, Friedrich JD, Manguikian RA, Buckheit RW, Truss JW,
Shannon WM, Secrist JA (1992) 5,5-disubstituted hydantoins: syntheses and anti-HIV
activity. J Med Chem 35:3567–3572
24. Cruz-Cabeza AJ, Schwalbe CH (2012) Observed and predicted hydrogen bond motifs in
crystal structures of hydantoins, dihydrouracils and uracils. New J Chem 36:1347–1354
25. Faris WM, Safi ZS (2014) Theoretical investigation of tautomerism stability of hydantoin in
the gas phase and in the solution. Orient J Chem 30:1045–1054
26. Ildiz GO, Nunes CM, Fausto R (2013) Matrix isolation infrared spectra and photochemistry of
hydantoin. J Phys Chem A 117:726–734
27. Ildiz GO, Boz I, Unsalan O (2012) FTIR spectroscopic and quantum chemical studies on
hydantoin. Opt Spectrosc 112:665–670
28. Nogueira BA, Ildiz GO, Canotilho J, Eusébio MES, Fausto R (2014) Molecular structure,
infrared spectra, photochemistry, and thermal properties of 1-methylhydantoin. J Phys
Chem A 118:5994–6008
29. Nogueira BA, Ildiz GO, Henriques MSC, Paixão JA, Fausto R (2017) Structural and
spectroscopic characterization of the second polymorph of 1-methylhydantoin. J Mol Struct
1148:111–118
30. Nogueira BA, Ildiz GO, Canotilho J, Eusébio MES, Henriques MSC, Paixão JA, Fausto R
(2017) 5-methylhydantoin: from isolated molecules in a low-temperature argon matrix to
solid state polymorphs characterization. J Phys Chem A 121:5267–5279
31. Rosado MTS, Lopes Jesus AJ, Reva ID, Fausto R, Redinha JS (2009) Conformational cooling
dynamics in matrix-isolated 1,3-butanediol. J Phys Chem A 113:7499–7507
32. Reva ID, Lopes Jesus AJ, Rosado MTS, Fausto R, Eusébio ME, Redinha JS (2006) Stepwise
conformational cooling towards a single isomeric state in the four internal rotors system
1,2-butanediol. Phys Chem Chem Phys 8:5339–5349
33. Reva I, Nowak MJ, Lapinski L, Fausto R (2015) Hydrogen atom transfer reactions in
thiophenol: photogeneration of two new thione isomers. Phys Chem Chem Phys 17:4888–
4898
34. Rostkowska H, Lapinski L, Khvorostov A, Nowak MJ (2003) Photochemical
double-proton-transfer reactions in 2,6-dithiopurine. A matrix isolation study. J Phys
Chem A 107:804–809
35. Chmura B, Rode MF, Sobolewski AJ, Lapinski L, Nowak MJ (2008) A computational study
on the mechanism of intramolecular oxo–hydroxy phototautomerism driven by repulsive pr*
state. J Phys Chem A 112:13655–13661
36. Bazargani MF, Talavat L, Naderi S, Khavasi HR (2011) 2-[(1H-Imidazol-2-yl)disulfanyl]-
1H-imidazole. Acta Cryst E 67:o2585
37. Steed KM, Steed JW (2015) Packing problems: high Z′ -crystal structures and their
relationship to cocrystals, inclusion compounds, and polymorphism. Chem Rev 115:2895–
2933
38. Rozenberg MS (1996) IR spectra and hydrogen bond energies of crystalline acid salts of
carboxylic acids. Spectrochim Acta A 52:1559–1563
39. Rozenberg MS, Shoham G, Reva I, Fausto R (2004) Low temperature FTIR spectroscopy and
hydrogen bonding in cytosine polycrystals. Spectrochim Acta A 60:463–470
Chapter 8
Vibrational Spectroscopy in Analysis
of Stimuli-Responsive Polymer–Water
Systems
Marcin Kozanecki, Marcin Pastorczak and Krzysztof Halagan
Abstract Over the last years, a rapid development in the material science, which is
an answer to an increasing demand for functional, smart systems, has taken place.
The recent progress in design, synthesis and characterization of stimuli-responsive
polymer systems (SRPS) fits in this trend very well. However, extensive experi-
ments, simulations as well as theoretical works are still conducted to deepen the
knowledge about these systems, their complexity and diversity result in still
insufficient understanding of some crucial phenomena. One of them is inter-
molecular interactions which change during swelling/deswelling processes, phase
transitions (commonly leading to the phase separation) and loading or a release of
various additives. Since the vibrational spectroscopy is considered to be the most
powerful tool to study molecular interactions, this chapter presents various aspects
related to the usage of vibrational spectroscopy in the field of SRPS.
8.1 Stimuli-Responsive Polymer Systems
Stimuli-responsive polymer systems are able to change dramatically and abruptly

physicochemical properties under the influence of a single specific stimulus or a set
of stimuli. Stimuli are commonly classified into three groups: physical, chemical
The authors would like to dedicate this chapter to Prof. Jacek Ulanski from the Department of
Molecular Physics, Lodz University of Technology—our teacher and mentor, who laid
foundations of our knowledge, and believed (and still believes) in our progress and
self-development.
M. Kozanecki (&) K. Halagan

Faculty of Chemistry, Department of Molecular Physics, Lodz University
of Technology, 116 Zeromskiego Str., 90-924 Lodz, Poland
e-mail: marcin.kozanecki@p.lodz.pl
M. Pastorczak
Institute of Physical Chemistry, Polish Academy of Sciences,
44/52 Kasprzaka Str., 01-244 Warsaw, Poland

224 M. Kozanecki et al.
and biological ones. Physical stimuli (light, temperature, ultrasound, as well as

magnetic, mechanical and electrical fields) usually influence the chain dynamics
(they change the energy level of the polymer–solvent system), chemical stimuli
(co-solvent, ionic strength, pH, various non-ionic compounds), they impact inter-
molecular interactions while biological stimuli (enzymes, receptors) change the
molecule functioning (e.g., enzymatic or receptor properties) [1–4].
Two strategies for SRPS design can be used. One of them assumes a direct
application of stimuli-sensitive polymers while, in the case of the other one, the
polymer is combined with a stimuli-sensitive compound(s) and plays only a role of a
passive matrix or a carrier [1]. In this chapter, only the systems containing active
polymers will be discussed. Taking into account this limitation, the subject is still very
broad as the polymer sensitivity to a specific stimulus can be achieved in different
manners, including changes in the hydrophilic–hydrophobic balance, conformations
of macromolecules, solubility, degradation, and bond cleavage [1, 5]. One of the most
spectacular stimuli-responsive systems in nature is the protein of the egg white, which
turns into gel in high temperature, with the presence of salts or changes in pH. The sol–
gel transition resulting from protein denaturation is irreversible, while from applied
engineering systems it is expected to exhibit reversible transitions, ideally without
hysteresis. A variety of homo- and co-polymers and many different ways of their
synthesis open a gate to a vast array of different responsive materials and devices, such
as micro-valves and attenuators, lenses with the adjustable focal lengths, sensors, drug
carriers, smart textronics and many others. Furthermore, polymer properties can be
modified at different levels of the hierarchy, starting from a chemical structure (e.g., a
type and sequences of co-monomers) and isomerism (stereo-regularity, cis-trans
isomerism) to supramolecular structures. A particularly attractive possibility is to
form various polymer topologies (such as stars, combs and brushes, ladder-like,
dendrimers, hyperbranched and networks) and to locate stimuli-responsive groups in
different parts of a macromolecule [1, 5–9].
Depending on the chemical structure, including substitution by functional groups
and topology, polymer systems with sensitivity to different stimuli can be achieved.
Moreover, it is also possible to create dual and multi-stimuli-responsive polymers
that simultaneously respond to more than one stimulus. Depending on the polymer
architecture, different parameters are used to control the threshold value of a
stimulus. In the case of linear chains, these are molecular mass, length and a type of
side groups, a substitution degree, a co-monomer composition and end groups (for
telehelic macromolecules) [10–14]. As regards dendrimers, this is a generation of
dendrimer and functionality of their cores that determines the size of a polymer
molecule and the number of ending groups [15–28]. For star-like and brush-like
polymers, crucial parameters include the number of arms, their length and space
distribution. Space distribution is especially important in the case of arms made of
block co-polymers (including systems capable of sol–gel transition resulting from
aggregation of star-like macromolecules) and so-called mictoarms (stars containing
arms made of different polymers) [29–38]. It is also possible to design SRPS in a
form of permanent gel. The continuity of a polymer network (nano- and microgels
8 Vibrational Spectroscopy in Analysis of Stimuli-Responsive … 225
usually differ from macrogels in their properties), the network density, the presence
of additional dangling chains and free ends of polymer chains impact the stimuli
threshold in this class of stimuli-responsive materials. Most SRPS are designed in a
form of polymer solutions, mixtures and dispersions. Thus, the polymer concen-
tration and a type of a solvent also influence the response of the system to a
stimulus.
Besides single macromolecules and gels, polymer-modified surfaces, polymer
micelles and vesicles can also be used as SRPS—see Fig. 8.1. However, our further
considerations will be limited only to solutions of linear macromolecules and gels.
The first ones are important not only as stimuli-responsive materials, but also (or
even primarily) as building blocks for more complex systems. Gels should be
considered as the most popular and promising smart materials. In these systems, the
intermolecular interactions between components constitute the crucial issue.
The roots of polymer solutions theory date back to the 1940s and Huggins [39]
and Flory’s works [40–44] developed later by Stockmayer and Krigbaum [45–49].
A detailed description of the Flory’s theory exceeds the framework of this study,
but it can be found in many polymer handbooks [50–52]. A broad review of
development of Flory’s theory has also been done lately by Bhattacharjee et al.
[43]. However, it is important to notice that polymer solubility (especially in case of
Fig. 8.1 Selected mechanisms of responsivity to stimuli in water–polymer systems: a formation

of micelles and vesicles by co-polymers capable of self-assembling; b surfaces decorated by
stimuli-responsive polymer chains able to form ‘mushrooms’ and ‘pancakes’ conformations (this
can be realized on macro-, micro- as well as nanoscale, for example by nanoparticle decoration);
c hydrogels including micro- and nanogels
amphiphilic polymers) is often strongly limited compared to that of low molecular

weight compounds. To understand the nature of sensitivity to different stimuli in
polymer–solvent systems, the h conditions should be defined. In the simplest case
of a binary system containing a linear polymer and a low molecular weight solvent,
the h conditions relate to the temperature (Th) and a type of a solvent in which the
polymer coil acts in the solution like an ideal chain i.e. can be considered as a
random walk. Such a situation arises if the following conditions are fulfilled:
(i) Gaussian statistics of polymer chains, (ii) ignorance of the segment–segment
correlation and (iii) ignorance of topological constraints. In reality, these conditions
are satisfied if the repulsions between polymer segments leading to polymer chain
expansion and contributing to energy U are balanced by the entropic contracting
force determining entropic part S of Helmholtz free energy F = U − TS. The
reduced Helmholtz free energy of mixing may be expressed according to the fol-
lowing formula:
F Up
¼ ln Up þ Us ln Us þ vUs Up ð8:1Þ
kB T M
where kB is Boltzmann constant, M—average molar mass of polymer chain, while

Up and Us are molar fractions of polymer segments and solvent molecules,
respectively. The mixing parameter v represents the intermolecular interactions in
the system and determines the solvent quality:
Z
v¼ ð8:2Þ
ð2kB T Þ 2esp ess epp
where Z is a coordination number (the number of the nearest neighbouring mole-

cules), while e represents energies of intermolecular interactions between particular
elements (s and p indices correspond to solvent molecules and polymer segments,
respectively). If v = 0.5, a polymer–solvent system is in equilibrium in the given
temperature and pressure; for v > 0.5, the solvent is called a ‘poor solvent’ and the
polymer chain is susceptible to globular conformation (collapse); for v < 0.5, the
polymer chain is in the ‘good solvent’—well soluble and expanded—see Fig. 8.2.
The performed short analysis showed that the equilibrium in the polymer–sol-
vent system may be perturbed by changes in temperature, addition of a co-solvent
or other components which impact polymer–polymer, solvent–solvent and poly-
mer–solvent interactions. As a result, coil-to-globule transitions can be triggered. In
the ideal case of the infinite chain, the h temperature corresponds to the critical
solution temperature (the coil-to-globule transitions temperature)—see Fig. 8.3b.
In real systems, the length of a polymer chain is finite and the temperature–
concentration phase diagram looks like in Fig. 8.3a. Instead of h temperature, the
critical point occurs determining the upper critical solution temperature (UCST).
The systems exhibiting the lower critical solution temperature (LCST) are also
known, and from a practical point of view, they are even more important (they will
be presented in section Thermo-responsive systems—“non-ionisable” polymers).
Fig. 8.2 Polymer chain in the solution could be compared to a sunbather. In favourable
conditions (good solvent, T > Th), it is extended and its radius of gyration (R)—being a measure of
mass distribution around the centre (or more trivially: a measure of space covered by a single
polymer chain) is proportional to Nm, the scaling exponent m (called also a Flory exponent) is a
dimension-dependent parameter: m = 3/(d + 2), where d is a so-called mass fractal dimension (for
a random chain in the 3-D system m = 3/5). In ideal (h) conditions, the polymer chain fulfills
criteria of the statistical random chain and its radius of gyration is defined by the random walk law:
R * N1/2. If the conditions are unfavourable (bad solvent, T < Th), the polymer tries to occupy the
minimal volume and to reduce the contact with surrounding environment. In this state, the radius
of gyration is reciprocally proportional to d
Fig. 8.3 Phase diagram (temperature T vs. concentration c) for a chain of finite (a) and infinite
(b) length. The true h (tricritical) point occurs for the infinite (ideal) chain only if c reached 0.
Moreover, only for c = 0 swollen and collapsed states exist. For c 6¼ 0, semi-dilute solution exists.
The coexistence curve corresponds to states for which the osmotic pressure is zero (critical points
of the phase separation in polymer solution with given concentration). For the finite polymer chain,
a critical point for the polymer–solvent phase separation occurs. Furthermore, solution behaviour
is controlled by concentration fluctuations near the critical point. The h-like condition area is
extended to the finite range of temperature (blue horizontal dashed line marks the crossover to the
repulsive polymer solution) and concentration for high enough concentrations system crossovers
from diluted to semi-diluted (red-dotted line). [49] © IOP Publishing. Reproduced with
permission. All rights reserved
The discussed simple model of a linear chain in a solvent becomes more

complex in the case of macromolecules with branched architecture including net-
works, dendrimers, star-like and brush-like. They usually have an ‘inhomogeneous’
structure in the molecular scale containing solvophilic centres distributed between
solvophobic parts. As a result, various types of intermolecular interactions (sensi-
tive to various factors) should be taken into consideration to properly describe a
given system. These interactions are also a factor that determines the type of a
stimulus to which the polymer is sensitive. The case of polymer–water systems is
extremely difficult, as both static and dynamic properties are strongly dependent on
water behaviour, well-known as anomalous. Techniques based on vibrational
spectroscopy are invaluable tools in this field, as it is discussed in the next section.
8.2 Water and Polymer–Water Systems as Seen

by Vibrational Spectroscopy
In most aqueous polymer systems, water constitutes the major component (in hy-
drogels, it is usually over 90% of the weight fraction), which substantially con-
tributes to the final properties of polymer solutions, dispersions and gels. The
presence of a high fraction of water is responsible for high biocompatibility and
biodegradability of most of these systems, their ability to dissolve additional
components like salts, non-ionic or even apolar molecules. Water also facilitates
fast transport of both mass and energy through polymer solutions or gels.
What is crucial from the perspective of SRPS are water properties and water–
polymer interactions which are decisive for the stimuli-responsiveness itself.
Regardless of actual triggers and the nature of mixing/demixing transitions of these
systems, the observed macroscopic action is most commonly based on highly
cooperative polymer–solvent interactions. To be more precise, stimuli-
responsiveness lies in the delicate balance between hydrophilic and hydrophobic
interactions in the solution which, when perturbed, provokes micro- and macro-
scopic phase transitions. Breaking of hydrogen bonds between hydrophilic groups
of a polymer and water molecules leads to their phase separation in temperature-
responsive systems. The decrease in the hydration of ionic polymer groups after
their neutralisation (protonation of anionic or deprotonation of cationic groups)
provokes demixing in pH-responsive systems. Other methods for triggering
demixing transition, like addition of a co-solvent or a specific molecule, are based
on very similar mechanisms associated with weakening of water–polymer and
strengthening of polymer–polymer interactions. It is also important to stress that the
presence of a dominating fraction of water in the system is responsible for trans-
portation of a triggering signal through a stimuli-responsive material, and thus, it
limits the rate of transition. Water has a thermal conductivity (0.591 W/m •K) much
higher than other popular solvents (alcohols, oils around 0.1 W/m •K) and many
times higher than most polymers, which plays a key role in homogenous transition
of thermo-responsive polymer systems. The pH-responsive systems benefit from an

ultrafast proton transfer of water through the Grotthuss mechanism, where only a
charge, not mass is transported between water molecules [53, 54]. In fact, on
account of the above-described properties of water, rehydration of a polymer is
usually a much slower process than dehydration, as the transport of heat, proton or
mass in neat polymers is significantly slower than in aqueous solutions or gels [55].
In the next part of this section, we will first characterize in detail a vibrational
spectrum of liquid water and then describe specific ‘states’ of water present in
polymer solutions and gels, and finally, we will shortly review ultrafast vibrational
spectroscopy methods applied to aqueous systems.
In the Raman (as well as in infrared) spectrum of liquid water—as shown in
Fig. 8.4—three fundamental modes occur: symmetric and antisymmetric stretching
(m1 and m3—superimposed on the broad multi-mode band in the range 3000–
3700 cm−1), the bending mode (m2 at 1640 cm−1) and the broad librational band
located between 300 and 900 cm−1. The OH stretching band (called further ms) is,
undoubtedly, the most studied band in the water spectrum. Its nature is very
complex due to the strong overlapping of m1 and m3, inhomogeneous broadening due
to hydrogen bonds formation, Fermi resonance between m1 and overtone of m2.
There is an agreement between experimentalists and theoreticians that the shoulder
of the band at lower wavenumbers should be assigned to water molecules engaged
in the higher number of hydrogen bonds with neighbours (three to four H-bonds)
and the shoulder at higher wavenumbers is related to the less hydrogen-bonded
fraction of water (engaged in one H-bond and not H-bonded water molecules)
[56–59].
A very unique feature of the Raman spectrum of water is the presence of a
polarized component (depolarization ratio q 0.2; with possible q values spanning
0–0.75 range) centred at around 3250 cm−1 in the OH stretching multi-mode band.
Since the intensity of this mode decreases with the rising temperature and shifts
Fig. 8.4 Raman steady-state

spectrum of water (blue) and
hexagonal ice (black) with
marked fundamental and
intermolecular modes of
water. Inset: low-frequency
Raman spectrum of water
with intermolecular modes
marked with arrows.
Figure from the Ph.D.
dissertation of Marcin
Pastorczak (Lodz University
of Technology, Lodz 2010)
down towards the polarized band observed in ice (at 3150 cm−1, see: spectrum of
ice, in Fig. 8.4) when the temperature is decreased, it has been linked to the
‘structured’ fraction of water [60, 61]. The ‘collective’ nature of this unique mode
has been recently confirmed with computational studies [62]. Other water modes of
the collective nature present in the Raman spectrum are two broad low-frequency
bands (marked with arrows in the inset to Fig. 8.4). They are both assigned to
inter-monomer damped harmonic vibrations of water molecules [63]. The band at
ca. 60 cm−1 is interpreted as the inter-monomer bending of the whole H2O mole-
cule engaging at least three molecules (O–O–O bending) and the mode centred at
ca. 175 cm−1 is assigned to stretching of a hydrogen bond between two molecules
(O–O stretching). These two modes are located in the spectral range difficult to
access in IR, and thus, Raman spectra are an exclusive source of information.
The dynamic micro-structure of water is mostly manifested in intense stretch
vibrational band at 3100–3700 cm−1. The disturbance of bulk water structure
(decoupling of OH oscillators) is caused not only by the influence of temperature,
but also by the presence of ions, surfactants, polymer chains, etc., and it is reflected
in the shape of this broad band. It is worth noting here that the ratio of a low- and
high-frequency component of the OH stretching multi-mode band changes with the
temperature. This ratio may thus be used as a molecular thermometer capable of
measuring in situ temperature of a polymer–water system [60, 61, 64].
The analysis of the changes in water structure in polymer systems is often only
qualitative—the shapes of normalised spectra are compared [65, 66]. The attempts
of presenting such changes quantitatively usually comprise either determination of
the collective parameter IC (equal to difference between intensities of ms for parallel
and perpendicular polarizations normalized to the depolarization of the band), as
proposed by Green et al. [67] or by deconvolution of ms band into components and
subsequent calculation of a ratio of the integrated intensities of high I3400 and low
I3200 frequency components [68, 69]. The value of the I3400/I3200 ratio in distilled
water at the room temperature was determined as ca. 0.87, and it is often taken as a
reference [69]. What has to be stressed is the fact that the value is strongly
dependent on the Raman excitation wavelength. Pastorczak et al. [70] showed that
the polarized component around 3200 cm−1 is enhanced by the resonance with red
light. Weak absorption of water in the red part of the visible range comes from
vibrational overtones and combinational modes. It is supposed that strongly cou-
pled OH oscillators, which contribute the most to the component at 3200 cm−1, are
much more anharmonic than OH vibrations of loosely bonded water, and therefore,
they are resonantly enhanced by excitation light from the red range [71] (Fig. 8.5).
As it was mentioned in many technically useful SRPS, such as solutions, dis-
persions and gels, water is a dominant component. Therefore, a spectroscopist
frequently faces a problem of a strong water signal (either in the OH stretching or
H–O–H bending region) masking polymer-related bands (it is especially frequent in
IR spectroscopy due to a much higher activity of water in this technique in com-
parison to Raman spectroscopy). In such a case, a useful solution is substituting
H2O with heavy water (D2O), whose bands are downshifted to 2504 cm−1
(stretching) and 1209 cm−1 (bending) [72]. Such an approach may be applied
Fig. 8.5 Raman steady-state spectra of water acquired with the use of different excitation
wavelengths (marked with colourful circles) superimposed on the absorption spectrum of water in
the visible range of radiation—black lines. The figure reprinted from the Ph.D. dissertation of
Marcin Pastorczak (Lodz University of Technology, Lodz 2010)
providing that no isotopic exchange between polymer hydrogen and water deu-
terium atoms takes place. However, it is necessary to remember that isotopic
H ! D exchange in water takes place easily (due to the already mentioned
Grothuss mechanism of proton/deuteron transport) and leads to the formation of
some fraction of HDO. The presence of HDO contributes to the vibrational spec-
trum with OD and OH stretching bands and an additional bending mode at
1450 cm−1. Similar isotopic substitution may happen when D2O-swollen sample is
exposed to humid air for a long time.
In polymer solutions, dispersions and gels, the structure of water is mainly
determined by a number of hydrophilic centres (binding sites) in polymer chains,
which decides of a number of hydrogen bonds between both components and of an
amount of ‘hydration water’. A model of the supramolecular structure of hydrogels
proposed by Lee et al. [73] and developed later by Maeda [74, 75] and other
researchers may be generalised as follows: the ‘hydration water’ (called also pri-
mary bound water), which is strongly bound to a polymer chain, disturbs also the
structure of water molecules located at the second hydration shell. This water,
usually named secondary bound water, is, according to Hoffman [76], also the one
bound by hydrophobic sites of a polymer (the contemporary view on hydropho-
bically bound water seen by ultrafast spectroscopy will be presented later in this
chapter). The rest of water (called bulk or free water) fills cavities between hydrated
chains and does not differ from ‘normal’ water in its properties. Nevertheless,
Lafleur et al. [77] distinguished also interstitial water trapped in small interstices
Fig. 8.6 Model of polymer gel proposed by Lee [73] and developed by other researchers [74, 75].
A general view—the figure drawn based on Maeda’s work [74] and zoom in allowing for
distinguishing between strongly and weakly bound water; hydrophilic centres marked in orange,
hydrophobic parts of a polymer chain in light grey
(1–2 nm; between phospholipids bilayers or between myosin molecules in muscle

fibres) and revealed a greater orientational disorder than in case of normal bulk
water. Particular states of water in hydrogel are presented schematically in Fig. 8.6.
Maeda observed a high influence of polymer hydrophilicity on the average
number of defects in water. According to his studies, in hydrogels with a low
crosslinking degree, the value of Ic was similar to that in the solution of the linear
polymer (with the same concentration), but a higher crosslinking degree resulted in
the decrease in Ic [75]. Because the number of hydrophilic groups was similar in
both cases, Maeda interpreted the additional quantity of defected water as the water
caged in interstitial regions. Similar dependency was observed when water structure
changes against polymer molecular weight w; i.e., no relation was observed up to
some critical value w (>104), for which Ic suddenly dropped. The critical value was
probably related to the chain length for which entanglements’ concentration was
high enough to produce water caged effect similar to that in gel.
The most experimentally documented differences in properties of primary and
secondary bound water comparing to normal water are their thermodynamic
properties and molecular dynamics. The applied investigation techniques often
imply the nomenclature of particular states of water. The lack of unified nomen-
clature corresponding to water states in gels and other polymer systems makes
discussion and consensus difficult in the field of investigation of water–polymer
systems by means of various methods.
Calorimetric methods make use of the fact that interactions with a polymer
prevent some quantity of water from crystallisation or result in melting some of it at
the temperature below 0 °C. In DSC thermograms, the peak of ice melting during
the heating scan should be proportional to the amount of water which crystallises.
Therefore, the comparison of the enthalpy of water melting in the hydrogels sample
(taking into account water concentration) to the enthalpy of the fusion of ice
(ΔH0 = 333 J/g) is thought to show an amount of freezable (normal) and

non-freezable water content.
While free water or bulk water is regarded as qualitatively similar to normal
water with respect to rotational and translational movements, the secondary bound
water is translationally restricted, but retains some rotational movements (thus, it
may freeze, but at temperatures below 0 °C). The primary bound water is con-
strained for both translational and rotational movements, and therefore, it cannot
freeze [78]. Traditional methods for studies of molecular relaxation processes are
the broadband dielectric spectroscopy (BDS) combined with the time-domain
reflectometry (TDR). They allow covering practically a full range of frequencies
desired to observe different states of water in hydrogels. Three main regions relating
to various water relaxations may be broken into:
• the kHz domain where the slowest processes of ice are visible [79];
• the MHz domain (up to 1 GHz) where the relaxations of bound water are
manifested [80];
• the GHz domain where the main process of free water can be distinguished at
the frequency of 17 GHz (at 20 °C), corresponding to the relaxation time
s1 = 8–9 ps [81, 82].
The origin of the highest frequency process of water (range 0.1–2 THz), which
corresponds to the relaxation time s2 1 ps, is still a matter of many speculations
[82, 83]. Broad gaps in the dielectric spectrum between loss peaks related to solid
ice and to liquid—bound and free water, allow for easy distinguishing of all the
states of water in gels. What is more, according to Kirkwood equation, a dielectric
strength Δe of a process is proportional to the number of dipoles rotating in the
electric field, it is possible to estimate roughly from this value an amount of water
molecules immobilised by polymer hydrophilic sites [84]. Qualitative comparison
of the polymer bound water in poly(vinyl methyl ether) hydrogels perceived by
several experimental techniques (BDS, DSC and Raman spectroscopy) may be
found in [85].
The water structure, dynamics and its interactions with solutes have been
extensively studied with the use of nonlinear spectroscopy methods like pump–
probe femtosecond spectroscopy, sum-frequency generation spectroscopy,
two-dimensional infrared spectroscopy and variations of these methods. A number
of interesting results collected with the use of these methods could be a subject of
several separate books [86–89]. In this chapter, we would only like to give a reader
a flavour of possibilities that these technologically advanced methods offer in
studies of liquid water and aqueous solutions.
A real breakthrough in studies of the structure and dynamics of liquid water
came with the development of laser setups capable of generating intense fem-
tosecond pulses in the mid-infrared range. The availability of tuneable sources of IR
pulses made it possible to study sub-ensembles of water molecules on the timescale
comparable to their lifetimes. In the 1990s, Bakker and co-workers applied the
ultrafast pump–probe IR spectroscopy to study energy transfers in organic solvents
[90] and then in liquid water [91, 92]. In one-colour pump–probe experiments, the
IR beam is split into an intense pump pulse, which excites a significant fraction of
oscillators from the ground to the first vibrationally excited state (0 ! 1) and a
weak probe pulse which monitors the relative absorbance change as a function of
the delay time between the pump and the probe (Fig. 8.7a). Depletion of the ground
state results in ‘bleaching’ at x01 frequency (a negative absorption change in the IR
transient spectrum). The population of excited molecules contributes to the excited
state absorption (1 ! 2) signal at x12 frequency (a positive absorption change in
the IR transient spectrum) (Fig. 8.7b). During the relaxation of vibrationally excited
states (2 ! 1 and 1 ! 0), both the ‘bleaching’ and the increased absorption signals
decay and one can determine the respective vibrational relaxation times
T (Fig. 8.7c). The probe pulse in the two-colour experiments is generated separately
from the pump pulse and may have a different wavelength. That makes it possible
to monitor the pump-induced transmission changes at arbitrary frequencies, e.g. to
pump the vibrational excitation 0 ! 1 of the OH stretching vibration and then to
monitor the vibrational relaxation channel through other modes. The time-resolved
polarization experiment offers an interesting opportunity to follow a decay of
rotational anisotropy of liquid water. The linearly polarized pump pulse vibra-
tionally ‘labels’ some fraction of the oscillator in HDO molecules and the probe
pulse, polarized either parallelly or perpendicularly to the pump pulse, monitors the
fraction of molecules that rotated out of their original orientation. The decay of this
rotational anisotropy in time contains information of water rotational movements,
determined by Rezus and Bakker as 2.6 ps [93, 94]. Interestingly, they later
observed with the same technique that water molecules are significantly slowed
down in the vicinity of hydrophobic groups; i.e., each methyl group rotationally
slows down four water hydroxyl groups [95]. The roots of this effect lie in the
mechanism of water rotation. As proposed by Laage and Hynes, presently com-
monly accepted rotational reorientation of water is composed of large angular
jumps rather than small diffusive steps. Each jump involves concerted cleavage and
formation of hydrogen bonds with two neighbouring water molecules [96].
Consequently, the presence of a ‘defect’—a C–H group instead of O–H group in
Fig. 8.7 Principle of vibrational pump–probe spectroscopy: a potential energy diagram of typical
vibrational excitation, b transient spectrum is the difference between the spectra in the presence
and absence of a pump pulse, c change in transient absorption at the frequency x01 vs. delay time
t reflects the kinetics of vibrational relaxation
the vicinity of water molecule reduces its probability to perform angular jumps and
slows down the rotational motion.
In the following years, the femtosecond pump–probe IR spectroscopy (fs-IR) has
been successfully applied to study liquid water and aqueous systems most intensly by
groups of Bakker, Nibbering [97, 98] and Tokmakoff [99]. Those researches resulted
in determination of the vibrational relaxation time 1 ! 0 of stretching OH
(T1 = 0.26 ps) [100] and bending H–O–H mode (T2 = 0.17 ps) [101] in neat H2O,
but also (with the use of two-colour experiments) the relaxation channel of OH
stretching excitation through H–O–H bending down to librational mode was unrav-
elled [101–103]. It should be noted that extremely fast kinetics of these processes
result from the resonant energy transfer between H-bonded OH groups [92]—in
isotopically diluted water, this relaxation proceeds much more slowly; T1 = 0.74 ps
for OH in HDO/D2O [104], while T1 = 1.8 ps for OD in HDO/H2O [105] (all values
of vibrational relaxation times are given for the temperature of 298 K).
Nonlinear vibrational spectroscopy methods have been scarcely applied to aque-
ous polymer systems so far. Hunger, Mazur and co-workers applied fs-IR together
with dielectric spectroscopy and DSC to study water dynamics around polysaccha-
rides (hyaluronan, dextran [106] and alginates [107]). Remarkably, they observed by
both fs-IR and DSC methods around 15 water molecules bound by hyaluronan
monomer unit. Moreover, their results suggest that the effect of studied polysaccha-
ride on the dynamics of water is restricted solely to the first hydration shell.
Applications of surface-specific vibrational sum-frequency generation spectro-
scopies (VSFG) to studies of the water structure and dynamics at interfaces,
although of paramount significance, are not covered in this review due to the
breadth of the topic. Some VSFG results on aqueous pH-responsive polymer sys-
tems will be discussed in the next section.
8.3 pH-Sensitive Systems—Polyelectrolytes
The group of pH-responsive polymers is together with thermo-responsive polymers

(described in the next section) the most commonly synthesized and investigated
types of SRPS. They undergo some sort of a structural transition (polymer chain
collapse/extension, swelling/deswelling, micellization, gelation, etc.) in response to
the change in pH of their environment. The pH-responsiveness of these systems is
linked with ionizable, basic or acidic groups incorporated into a polymer main
chain. Henceforth, the pH-responsive polymers are classified as polyacids and
polybases. Some authors single out an additional category—natural pH-responsive
polymers like chitosan or some ionizable polypeptides (polyglutamic acid,
polyaspartic acid, etc.) [108]. The most often applied polyacids are poly (carboxylic
acids), poly (phosphoric acids), poly (sulphonic acids) and poly (aminoacids).
Popular polybases are polypyridines, polyimidiazoles and polymers containing
tertiary amino groups. It should be also mentioned that polyacids and polybases
may form a polymer complex stable at the given pH range [109].
Given the fact that a structural transition in the pH-responsive polymer system is
practically always related to the degree of ionization of a polymer chain, and the
role of vibrational spectroscopy in studies of these systems is usually limited to
monitoring the fraction of ionized groups on a polymer chain. This is perfectly seen
in studies of the popular pH-responsive systems based on polyacrylic or poly-
methacrylic acids. A fully ionized (at high pH) polyacid chain is strongly hydrated
and expanded due to repulsive anion–anion interactions. After protonation of a
critical amount of carboxylate groups (at chain neutralization degree a = 0.3 [110]
at acidic pH, hydrophobic chain–chain interactions lead to the collapse of a polymer
chain and eventually to the polymer/water phase separation. Since the carbonyl
band (stretching of C=O, at 1705 cm−1) is well separated from the antisymmetric
stretching of –COO− anion (around 1566 cm−1), the analysis of intensity of these
modes allows for precise tracking of chain protonation [111]. Santonicola et al.
applied the FTIR titration method to study the pH-dependent collapse of poly-
methacrylic acid (PMAA) brushes grafted from a silicone surface [112]. They
showed evolutions of COOH and COO− related bands during pH changes and
based on them a final titration curve was constructed, useful to determine effective
pKa of polyacid brushes—see Fig. 8.8. It is also important to mention that vibra-
tional spectroscopy is very useful to characterise the chain–chain interaction in
detail. According to Dong et al. [113], the broadband of C=O stretching vibrations
is composed of four lines assigned to different forms of carboxylic groups in
polyacrylic acids, such as free monomers, cyclic dimers and oligomeric chain
forms: laterally hydrogen-bonded groups (intramolecular hydrogen bonds formed
between neighbouring COOH groups of the same polymer chain) and face-on
bonded structures (intermolecular hydrogen bonds formed between –COOH groups
of two neighbouring polymer chains).
Fig. 8.8 a Carbonyl absorption region in FTIR spectra of PMAA brushes grafted from silicon
surfaces by atom transfer radical polymerisation (ATRP) in water medium after incubation in
phosphate-buffered saline (PBS) solutions with different pH values, b a degree of dissociation of
carboxylic acid groups (I1566/I1705) as a function of solution pH. The solid line is a fit to sigmoidal
function. Reprinted with permission from [112], copyright (2010) American Chemical Society
In Raman spectroscopy studies on crosslinked polyacrylic acid hydrogel,

Tsukida et al. [110] studied changes in the structure of water in the course of
pH-induced transition of polyacid chains. Walczak et al. [114], with the use of the
polarized Raman spectroscopy, analysed the dependence of the polyacrylic acid
chain conformation on the degree of polymer neutralization and the results corre-
lated well with earlier SAXS and NMR studies of the subject.
The important issue related to pH-responsive polymers is an influence of the
counterion on the polymer chain conformation, hydration and stability of ionomers
in water. Interestingly, the effect of ‘salting out’ and ‘salting in’ of proteins in water
and related Hofmeister series has been known for many years. Franz Hofmeister
published between 1888 and 1898 [115–117] the results of his experiments aimed
to quantify the effect of particular cations and anions on the egg white protein
stability in a solution. Hofmeister observed that for salts having the same cation,
anions could be organized in the following series from a high to a low ability to
precipitate protein (protein ‘salting out’):
SO4 2 [ HPO4 2 [ F [ CH3 COO [ Cl [ Br [ NO3 [ I [ ClO4 [ SCN

ð8:3Þ
Also cations (having the same anion) could be ordered in similar series:
ðCH3 Þ4 N þ [ Rb þ [ K þ [ Na þ [ Li þ [ Mg2 þ [ Ca2 þ ð8:4Þ
Later, it was observed that the presence of some ions actually increases solubility of
proteins in water, which refers to the ‘salting in’ effect. Swann et al. [118] investigated
influence of anions (acetate, chloride, bromide, iodide, nitrate and thiocyanate) on the
expansion of block cationic copolymer (poly(methyl methacrylate)-block- poly(2-
(diethylamino)ethyl methacrylate)-block-poly(methyl methacrylate)) in water. They
found that the magnitude of the influence of particular anions follows reverse order to
the classical Hofmeister series. Similar studies on the influence on ions on the stability
of several non-ionic polymers were performed recently by Sadeghi and Jahani [119].
A very interesting study by Fu and Schlenoff was focused on the mechanism of
associations between oppositely charged polyions (poly(diallyldimethylammonium)
with poly-(styrenesulfonate)) [111]. They investigated the influence of Hofmeister
series of counterions on polyions complexation and with the use of Raman spec-
troscopy related to changes with water perturbation seen in the OH stretching region.
Most often, the effect of small (low molecular weight) ions on the collapse of
macromolecules is linked with two phenomena related each other: ion pairing and
changes in hydration of the ionic group of macromolecule by the presence of a small
counterion. This is confirmed by direct observation (with the use femtosecond
infrared spectroscopy and dielectric spectroscopy) of the so-called phenomenon of
‘cooperative hydration’ when an anion and cation cooperatively lock the motion of
several water molecules located between them [120]. Further works performed in
Bakker’s group showed cooperative hydration between a carboxylate anion and alkali
cations [121]. It was observed that the carboxylate group cooperatively binds 2–3
water molecules when combined with Na+ and Li+ cations, and no more than one
water molecule when combined with weakly hydrated cations like Cs+, K+ and
NH4 þ . It points at the formation of the solvent separated ion pair between carboxylate
anions and Na+ and Li+ cations. What is more, the observed propensity of cations
forming solvent separated ion pairs with formate anion ðCs þ \K þ ; NH4 þ
\Li þ \Na þ Þ followed well Hofmeister series for monovalent cations. The speci-
ficity of ions interactions with macromolecules in water was reviewed by Liu and Liu
in [122], and specific ion interactions with polycationic brushes were described in
detail by Willott et al. [123].
Nonlinear spectroscopy studies on aqueous polymer pH-responsive systems are
rather limited. A particular potential of such studies was certainly shown by VSFG
spectroscopy, which is a surface-specific method providing spectra of the layer a
few molecules thick. The surface specificity is inherent to this method, as VSFG
measures the second-order nonlinear susceptibility v(2) which can be generated only
in non-centrosymmetric media. The net signal from the bulk of the material cancels
out and only a few molecular layers present at an interface contribute to a VSFG
spectrum [124]. Initial studies on the surface propensity of the pH-responsive poly
(acrylic acid) [125, 126] and poly (methacrylic acid) [126] were performed by the
Richmond’s group. They observed a strongly pH-dependent multi-step process of
surface absorption of both polyacids. Further studies on the surface propensity of
polyacrylic acid were performed by Balzerowski et al. [127] with the use of
heterodyne-detected VSFG. The heterodyne scheme of detection of this method
allows for disentangling real and imaginary parts of v(2) as a function of the fre-
quency and the sign of imaginary part of the spectrum is directly connected to the
absolute orientation of molecules at the interface. Thus, it may be easily deduced
from spectra which chemical groups point towards and which outward an interface.
Balzerowski et al. observed that the polyacrylic acid has the increased surface
propensity below the degree of ionization of 0.203, and above this value, the
polymer is well dissolved in the bulk. Moreover, it was shown that the ionization of
the carboxylic acid groups is accompanied by an increase in the VSFG signal of
interfacial water molecules as a result of enhanced orientation of water.
The high surface specificity of VSFG makes it an excellent tool for studying
systems such as stimuli-responsive brushes, carpets. Kondo et al. [128] investigated
zwitterionic poly(carboxymethylbetaine) brushes prepared with the use of atom
transfer radical polymerisation (ATRP) method on the fused quartz substrate. With
the use of VSFG, they studied orientation of brushes on a surface by polarization-
resolved analysis of the polymer C–H stretching bands. They also noticed only tiny
orientation of water molecules located in the vicinity of a zwitterionic brush. One can
easily imagine numerous applications of VSFG to studies of stimuli-responsive layers
and interfaces. It should be kept in mind, however, that VSFG signal is generated
through focusing two femtosecond beams of high pulse peak power on the very
interfacial layer. A strong optical field often leads here to thermal decomposition of
organic molecules present at the interface. Hence, an VSFG experiment on the layer
with high concentration of a polymer would be difficult to perform.
8.4 Thermo-Responsive Systems—“Non-Ionisable”

Polymers
Thermo-responsive polymer systems are designed mainly with the use of

non-ionisable macromolecules. Therefore, a mechanism responsible for triggering
the coil-to-globule transition or VPT is different in comparison to pH-sensitive
systems and bases mainly on the disturbance of hydrogen bonds and/or dipole–
dipole and/or hydrophobic interactions . For gels, the most popular SRPS, a broad
review of dynamic and kinetic behaviours, was performed by Shibayama and
Tanaka [129] and lately by Koetting et al. [130]. The network formation strongly
reduces the number of freedom degrees in the polymer–solvent system. Thus, the
free transition between collapsed and expanded states becomes limited. As a result,
the volume phase transition (VPT) is observed (an interesting comparison between
coil-to-globule and volume phase transitions can be found in [131]. Similarly to
liquid–gas transition, it could be continuous or discontinuous and its kinetics
depends on the network elasticity, polymer–solvent friction and the presence of
additional substances [132, 133]. For practical applications, systems exhibiting
discontinuous VPT are more attractive.
For the first time, the volume phase transition was observed experimentally by Tanaka
for partially ionized acrylamide gel [134]. Interestingly, it was previously predicted by
Dusek and Patterson [135], who developed theoretical considerations for linear polymers
done by Pristin and Eizner, de Gennes and Lifshitz et al. [136–138] and assumed the
possibility of conformational transitions in gels analogously to the coil-to-globule
transition in the solution of linear polymers. Currently, the family of thermo-responsive
polymers is very broad and includes: poly (N-alkyl substituted acrylamides) [10], poly
(N-alkoxyalkylacrylamides) [11, 12], poly (N-vinylalkylamides) [12], polyvinyl-
methylether (PVME) [140, 141], poly (oligoether methacrylates and acrylates) [13, 142]
and their various co-polymers. Also, some natural polymers such as cellulose and its
derivatives are sensitive to temperature [143, 144].
As it was mentioned, one of the most studied issues related to thermo-responsive
polymers is the polymer chain hydration and its changes during the coil-to-globule
transition or VPT. The presence of polymer chains in water influences its
supramolecular structure leading to hydrogen bond redistribution. Especially, hy-
drophilic centres located along the polymer chain (or network) can strongly impact
the organization and dynamics of water molecules. However, also hydrophobic
parts can be stabilized by water molecules by so-called hydrophobic interactions
[76, 77, 95, 145–147].
As it was mentioned in the previous section, various states of water differ in their
thermal, diffusive and other properties [148–154]. They may be determined by various
techniques; however, they are mainly studied by vibrational spectroscopy—a fast,
cheap and non-invasive method giving a broad range of information, simultaneously,
about the state of both a polymer and water. Moreover, information on particular
components of the complex systems can be achieved and analysed separately.
The series of IR spectroscopic studies done for various polymers by Maeda and
co-workers deserves special attention [65, 74, 75, 110, 139, 146, 155–160]. They
proposed a very useful procedure to study the hydration degree of particular

polymer groups by IR spectroscopy. This procedure is based on the analysis of
differential spectra and was originally proposed to monitor VPT in
thermo-responsive systems by IR spectroscopy. However, the method is universal
and may be applied to study VPT triggered by other stimuli as well as other types of
the phase transition. Firstly, a series of time-resolved vibrational spectra should be
collected. Next, the differential spectra between the spectra at various temperatures
(T) and the spectrum collected at the reference temperature (T0) should be obtained
DA(T–T0) = AT–AT0, where A relates to the signal intensity (absorbance for
absorption IR spectroscopy). After that, the difference between the intensity of
positive and negative peaks of a selected vibrational mode in each difference
spectrum should be calculated [146, 157]:
DDATT0 ðm1 m2 Þ ¼ DATT0 ðm1 Þ DATT0 ðm2 Þ ð8:5Þ
where m1 and m2 relate to the wavenumber at the positive and negative peaks in
difference spectrum, respectively. It was assumed that m1 < m2, so positive values of
DDAðTT0 Þ ðm1 m2 Þ correspond to the redshift of a given band, while the negative
ones are characteristic of the blue shift. The procedure proposed by Maeda et al.
allows investigator to monitor easily the shift of all bands in the spectrum and also
to cancel the baseline drift, which often accompanies thermal experiments. The
DDAðTT0 Þ ðm1 m2 Þ parameter can be very useful to study hydration of particular
groups in the system and to monitor VPT as it is shown in Fig. 8.9.
In the case of polymers with relatively simple chemical structure and only one
type of hydrophilic centres (PVME or poly(N-isopropylacrylamide)—PNIPAM),
the analysis of hydration degree is relatively easy. More complex systems with
several hydrophilic centres require a more complex spectroscopic analysis and
should be often supported by computer simulations. As an example of such a
system, poly(2-(2-methoxyethoxy) ethyl methacrylate PMEO2MA with four dif-
ferent oxygen atoms may be pointed out. As it was shown by Maeda et al. [146] for
the linear polymer and by Olejniczak et al. [161] for PMEO2MA hydrogels, the
energy of water–oxygen interactions is different and decreases in series: carbonyl
group > ether bridge in the middle of side chain > terminal methoxy
group. Moreover, it was shown that both Raman and IR spectroscopy may be used
to estimate the percentage of hydrophilic centres accessible for water. The analysis
is based on the shift towards lower wavenumbers of the mode characteristic for
C=O stretching (from ca. 1728 to 1709 cm−1 for PMEO2MA) induced by the
formation of C=O water complex. Thus, the fraction of the hydrated (H-bonded)
carbonyl groups (fH-bond) may be calculated according to the general formula:
1
fHbond ¼ e1709 A1728
ð8:6Þ
1þ e1728 A1709
Fig. 8.9 a IR absorption, b Fourier self-deconvoluted, c second-derivative, and d difference

spectra of PMEO2MA-D2O system (polymer concentration was equal to 10% wt.). The main
bands are assigned at the top of chart (a). Green arrows in (b) indicate the peak shifts upon heating,
while violet triangles in (d) indicate the peaks used to calculate DDA(T–T0) (m1 – m2) presented in the
chart (e): violet up triangles mas(C–H) (peak marked A on the chart (a)), blue circles mas(C–O) (peak
marked C on the chart (a)) of C–O–C bridges in oligoether side group, red down triangles m(C=O).
The green line on the chart (e) presents the transmittance of visible light (520 nm) through the
sample—the deflection point corresponds to the so-called cloud point and is regarded at VPT
temperature. Reprinted with permission from [146], copyright (2010) American Chemical Society
where A is an integral intensity and e—absorption coefficient (indices 1709 and

1728 correspond to wavenumbers of absorption maxima characteristic for
H-bonded and non-bonded C=O groups).
It was shown that for linear PMEO2MA, only 50–60% of C=O groups are
hydrated even below LCST [146], while for gels this value is lower and equals ca.
40% independently of the network architecture [161]. This surprisingly low degree
of hydration found for the group, which interacts with water the strongest, results
from the low accessibility of carbonyl groups hindered by the oligoether part. Even
a small, hydrophobic, a-methyl group may influence C=O hydration, as Maeda
showed by comparing determined by the IR analysis a hydration degree of anal-
ogous acrylates and methacrylates [159]. It is also worth noticing that in the case of
diluted solutions of linear PMEO2MA (the system with the water content inac-
cessible for PMEO2MA gels), the C=O groups hydrated by two water molecules
were also found instead of the simultaneous presence of 40% non-hydrated car-
bonyl groups. Their presence was manifested by the stronger shift of C=O line
towards a lower wavenumber (line 1698 cm−1) [146].
Another interesting issue is the relation between the water structure and the
hydration degree in polymer gels. In the previous section, it was mentioned that the
ratio between the intensities of 3400 and 3200 cm−1 Raman lines may be used to
monitor the water structure in a system. As Pastorczak et al. [162] showed, Raman
spectroscopy allowed them to correlate changes in the water structure with the
hydration degree of hydrophilic centres. Investigations were performed for PVME
gels, and hydration of O–CH3 groups was monitored by the peak position of the
mode relating to CH3 stretching—see Fig. 8.10 (applicability of the vibrational
modes relating to alkyl groups for monitoring the oxygen hydrophilic centre will be
discussed later). These results unequivocally proved that the bulk water dominating
in a fully swollen state is removed from the network firstly, and the polymer chain
dehydration occurs when the bulk water is removed completely. These results also
showed that the point where the shift of CH3 vibration reached the maximal value
may be used to estimate the amount of strongly bound water in the system. The
comparison of the literature data showed that for many polymers, independently of
their chemical structure, the amount of bound water is similar and varies between
10 and 13 water molecules per monomer unit [146, 148, 161, 162].
Fig. 8.10 Correlation

between water structure
(monitored by I3400/I3200
ratio) and hydration of PVME
network (monitored by
position of vs(CH3) mode).
Experimental data were taken
from Ph.D. dissertation of
Marcin Pastorczak (Lodz
University of Technology,
Lodz 2010). Data were
published in [162]
Since the polymer hydrogels may be potentially applied as carriers for biolog-
ically important molecules (drugs, fertilizers and others), interactions between those
molecules and other components (water and polymer chains) seem to be especially
interesting. Moreover, the presence of additives usually changes the system stability
and impacts the phase transition temperature. Until now, it was found that many
various ionic substances, for example simple inorganic salts such as halogens or
surfactants [158, 163, 164], influence the VPT temperature and dynamics, shifting
the VPT to lower or to higher temperatures. It is postulated that an explanation of
this phenomenon relates to structure-making and structure-breaking properties of
various salts ordered in Hofmeister series—see the previous section [164, 165].
However, also non-ionic substances (for example hydroquinone, phenol, glycerol
or urea [166, 167]) strongly impact VPT.
Based on spectroscopic results, Maeda postulated that additives do not interact
with the polymer skeleton directly, but disturb the water structure simultaneously
influencing the polymer hydration. The direct proof of this thesis was delivered by
Olejniczak et al. [133]. Studies performed on PMEO2MA gels loaded by solutions
of various non-steroidal anti-inflammatory drugs (namely ibuprofen and naproxen
sodium salts) showed a significant increase in the VPT temperature, linearly
dependent on drug concentration. The comparison of Raman spectrum recorded for
the drug solution with the differential spectrum obtained by subtraction of Raman
spectrum characteristics for gels filled with deionized water from the spectrum
collected for gel loaded with an appropriate drug solution was performed. As a
result, both spectra should be representative of the hydrated polymer network. The
positions of all bands characteristics for polymer were found unchanged in the drug
presence as any shifts of Raman lines were not observed [133]. Interesting work
about the hydration of ionizable, amphiphilic amino acids in aqueous solutions was
published by Ide et al. [156]. The relative intensity of water collective band (nor-
malized to the signal for pure water) was measured for aqueous solutions of amino
acids differing as regards hydrophobicity. It was found that at neutral pH, the water
structure is not changed by the presence of amino acids independently of their
chemical structure. In acidic conditions, the intensity of collective band decreases
with the increase in hydrophobicity, which revealed that the degree of H-bonding
between water molecules increases in the neighbourhood of hydrophobic species,
while the opposite tendency is observed next to hydrophilic centres.
The limits of application of classical methods of vibrational spectroscopy to
study changes in the hydration degree during VPT or the coil-to-globule transition
results from a relatively low time resolution restricted by the acquisition time of a
single spectrum with the quality sufficiently high for further mathematical treatment
(commonly it is a timescale of single minute for Raman and tens of seconds for
FTIR spectroscopy). Thus, many investigations of VPT are performed in
quasi-dynamic conditions (a sample is heated to a given temperature, shortly sta-
bilized to reach the uniform temperature distribution, a spectrum is collected, next
the temperature is changed, and the procedure is repeated). It could be expected that
such a procedure does not mimic properly the dynamics of VPT. The comparison of
the thermo-optical analysis and Raman results shows a high agreement—see
Fig. 8.11 a Changes of polymer network hydration during VPT for PVME gels differing in
density of crosslinks (PVME-18 low and PVME-65 high density of crosslinks, respectively)
monitored by the position of vs(CH3) mode. The Raman results are presented on the background of
thermo-optical analysis results (TOA) measured the white light transmittance through a sample.
Experimental data were taken from Ph.D. dissertation of Marcin Pastorczak (Lodz University of
Technology, Lodz 2010) and MSc diploma thesis of Agnieszka Lazuchiewicz (Lodz University of
Technology, Lodz 2008). b Arrhenius-type plot for changes of PVME conformation during VPT
monitored by the ratio of the intensities of symmetric and asymmetric CH2 stretching bands. Part
of data (Raman experiment results) presented in Fig. 8.11a was partially published in [162].
Figure 8.11b is reprinted with permission from [162], copyright (2017) Elsevier
Fig. 8.11a. Moreover, the difference in dynamics of VPT between the PVME of the
different network density is clearly seen. However, it is necessary to remember that
hydrogels are often heterogeneous systems and their properties can be
‘space-dependent’. The average of some experiments is obligatory to eliminate
artefacts and minimize errors. Data reproducibility is sometimes very poor, and
thus, only estimated values are often accessible. It is also important to notice that
the VPT transition may create a gradient of the water content. Water from the outer
zone of the gel can be removed readily, while the water molecules from the inner
parts need more time to flow out from the polymer network. The loss of water in the
outer part of the gel leads to the partial collapse of the polymer chain and reduction
of gel porosity in this zone. This undesirable process is called ‘skin formation’, and
as a result, water from the interior is trapped significantly, slowing down the VPT
[168]. To study this process, both high spatial and time resolutions are required. In
spite of the fast development in advanced microscopic techniques offering a pos-
sibility of depth profiling (for example confocal micro-Raman spectroscopy), this
problem has still remained unresolved and challenged.
Both VPT and the coil-to-globule transition result also in the change of polymer
conformation relating to transition from the highly stressed and expanded chain in a
swollen state to the strongly jammed globular structure in a collapsed state. The
conformational changes induce changes in vibrations of groups forming the main
chain (there are mainly CH2 groups as the vinyl or acrylic monomers are usually
used to synthesize artificial polymer hydrogels) . It is commonly accepted that the
isosbestic point found for series of spectra recorded as a function of any factor
(temperature, pressure, pH, time, etc.) proves that under the given factor, the system
transits from one state to another and it should be also useful to study conforma-
tional changes in organic compounds. If the bands related to isosbestic point are
separated and do not overlap with other bands, it can be easily found by an analysis
of differential spectra as a signal amplitude at the isosbestic point that they are equal
to 0 and located between two extremes, which should be proportional to the part of
molecules at the given state. However, in the case of polymer systems, overlapping
of many lines often occurs. Therefore, the deconvolution process is the only way to
analyse the integral intensities of bands related to various states. In the case of
PVME gels, it was shown [162] that the ratio of symmetric and asymmetric
stretching vibrations of CH2 groups located in the polymer main chain may be used
to monitor conformational changes during VPT and to estimate the activation
energy of a VPT process. The spectra were recorded as a function of the temper-
ature. The natural logarithm of the ratio of integral intensities of symmetric and
asymmetric stretching vibrations of CH2 groups as a function of inverse tempera-
ture according to Arrhenius-type equation was shown in Fig. 8.11b. In the range of
temperatures related to VPT, the relationship was found linear and the slope cor-
responding to activation energy was estimated at ca. 88 kJ/mol. A similar value
(80 kJ/mol) was found for the PNIPAM coil-to-globule transition [169].
As it was shown in this section, the vibrational spectroscopy can be very useful
to study various issues related to thermo-responsive polymers, such as a degree of
polymer hydration and its changes induced by the temperature, the water structure,
polymer chain conformations related to the coil-to-globule transition or VPT, etc.
However, a lot of information on polymer systems that can be extracted from IR
and Raman spectra is usually averaged over billions of molecules. To overcome
this, computer simulations seem to be most useful as they give access to specific
local molecular states.
8.5 Computer Simulations of Thermo-Responsive

Polymer–Water Systems
A short overview of computational methods applicable to study ‘smart’ polymer–

water systems is presented in this section. In spite of water importance in aqueous
SRPS, computational studies performed for pure water are not included in the
discussion as a separate and very broad topic. A wide variety of different methods
applied for liquid water modelling makes them impossible to be presented here
even briefly. However, for readers interested in this matter some references can be
recommended [170, 171].
Modelling and simulation have constituted an important part of the research
process in material sciences for a few decades. This field plays an important role in
predicting and explaining experimentally observed phenomena. Newly developed
smart materials, including stimuli-responsive systems with a hierarchical structure,

have to be designed on the molecular level in the length scale of 0.1–10 nm and
need a deep understanding of synthesis–structure–property–processing–perfor-
mance relations. A predictive description of synthesis processes leading to the
creation of sophisticated macromolecules requires consideration of complex
molecular models. This frequently includes such types of architecture as multi-arm
stars, hyperbranched polymers, dendrimers, comb-like or dendronized polymers,
rings, gels, etc. Furthermore, simulations of polymer blends, melts, solutions and
composites (and soft matter, in general) are somewhat challenging, because their
properties usually involve a variety of timescales and a broad range of sizes, as
Fig. 8.12 shows. The timescales cover the domain from 10−13 to 10−14 s (for bond
vibrations) up to, at least, a few seconds for phase separation processes. The size
scales range from single angstroms (for the bond length) up to centimetres (for
domains or crystals). Also, various intermolecular interactions should be taken into
consideration, often between amphiphilic polymers and solvents with complex
supramolecular structure like water. Therefore, a successful solution to a problem in
the material design usually requires hierarchical, multiscale simulation involving a
combination of atomistic (below *10−7 m), mesoscopic (up to *10−4 m) and
macroscopic methods. How to link them together remains a challenging problem
and there is still no single universal method covering the whole above-mentioned
range of size and time. We are still and probably we will always be doomed only to
‘… relatively good approximation to truth - which is much too complicated to allow
anything but approximations’, as John von Neumann said. This is the reason why
various methods are used to simulate different levels of molecular complexity. The
more complex and detailed atomistic model, the shorter times are available in the
simulation. This is mainly caused by hardware limitations; however, the limits shift
systematically towards a more precise description of complex systems in longer
timescales.
A detailed description of various simulation methods is available in books [172],
and thus in next paragraphs, only the most useful techniques in relation to polymer
science (excluding macroscopic methods) will be introduced shortly.
Quantum mechanical calculations use generally two approaches, ab initio and
empirical or semi-empirical methods, to solve a set of simplified Schrödinger
equations. At present, the most popular first principles calculation method is Kohn–
Sham density functional theory (DFT) [173] with B3LYP hybrid functional [174,
175]. In DFT, the energy of a system is expressed in terms of electron density rather
than the wave function as in ab initio. Empirical methods use some parameters
determined experimentally. Quantum mechanics (QM) is particularly useful for the
structure optimization, calculation of energy (including the energy of interactions),
IR, Raman and NMR spectra prediction. The structure optimization is performed
without the temperature; moreover, solvent, e.g. water, can be present explicit or
with a reaction field using the integral equation formalism model (IEFPCM [176]).
From a technical point of view, QM calculations are practically limited to only a
few hundreds of atoms, which mean a few monomer units in the field of polymer
science. This limitation arises from strong dependence of computing time on the
Fig. 8.12 A range of applicability of various computing methods compared with typical length
and timescales in polymer materials
number of atoms—computational complexity is at least an order of magnitude

higher than for a molecular dynamics method described below. Much larger sys-
tems are possible to be investigated only using high-performance computing
(HPC) hardware. Moreover, an initial ‘guess’ (initial configuration) must be very
close to ‘real’ molecular structure obtained from diffraction experiments.
Popular QM software packages are e.g. GAUSSIAN, GAMESS, ABINIT.
Molecular dynamics (MD) is a classical method that applies molecular
mechanics (MM), i.e. classical Newton mechanics to molecular systems. This
simulation method enables to study systems containing many thousands up to
millions of atoms. Usually each atom is simulated as one particle; however, in the
united atom approach, also called coarse-graining, some groups, e.g. methylene unit
or even the whole monomer unit, can be represented by one particle according to
original Huggies’ proposal [39]. An exemplary united atom approach to PVME–
water system is presented in Fig. 8.13 [148]. Each particle possesses a radius
(typically the van der Waals radius), polarizability and eventually a charge. Bonds
are usually described by the harmonic potential with an equilibrium distance equal
to the experimental or calculated bond length. Accurate reproduction of vibrational
spectra requires a usage of more computationally expensive Morse potential.
Non-bonded interactions, such as dispersion and repulsion, are described by
Lennard-Jones 12-6 potential. To speed up the calculations, their range is usually
limited by the use of the cut-off radius (typically 1 nm order of magnitude).
Fig. 8.13 Exemplary of the

coarse-graining procedure
applied for PVME–water
system (the chemical structure
of PVME is included). The
equilibrium distances between
given elements were
estimated from QM
calculations. Open Access
[148] by The Royal Society of
Chemistry
Long-range electrostatic interactions are problematic (due to the slow potential

fall-off) and a number of methods can be used to address this problem, like particle
mesh Ewald (PME) [177] or particle–particle particle–mesh (PPPM) techniques
[178]. An ab initio QM is often used to evaluate molecular charges for MD sim-
ulations using a restrained electrostatic potential protocol (RESP) [179]. The exact
form of the potential function, called overall the force field, depends on the par-
ticular simulation and is essential to obtain results comparable with the experiment.
The popular force fields used for polymers are AMBER (94 [180], 99SB [181], ff03
[182]), GROMOS (87 [183], 96 53A6 [184]), OPLS-AA [185], CHARMM [186],
PCFF [187]. Water molecules, if used as solvents, are usually described by SPC/E
[188], TIP3P [189] or TIP4P/2005 [190] models. Typical time step (time resolu-
tion) is of the single femtoseconds order of magnitude and the accessible timescale
reaches hundreds of nanoseconds for the research-level NPT-ensemble simulation
(only some special dedicated supercomputers offer simulations up to hundreds of
microseconds of the real time). Periodic boundary conditions are commonly used to
minimize the effect of the finite size of a simulation box and emulate the infinite
sample. Popular MD software packages are e.g. GROMACS, LAMMPS, AMBER,
CHARMM, NAMD. The MM can be combined with QM to increase the theory
level in a desired region of the simulation space [191].
In dissipative particle dynamics (DPD) [192] (a broad review can be found in
[193]), the simulated objects are composed of molecules described as a set of
soft-like beads that move along the Newton momentum equation and interact in the
continuous space. DPD uses coarse-grain mapping to transform molecular system
into a mesoscopic model composed of beads. The total force acting on a given bead
is a sum of various forces: a conservative repulsive force (determining the ther-
modynamic behaviour of the system), a dissipative force (including the friction
forces), a distance-dependent random force (representing omitted by
coarse-graining atomistic details) and a bonding force (representing bonds and
described by harmonic functions). In the DPD method, a conservative force
parameter (maximum repulsion between two beads) can be determined from the
relationship with the Flory–Huggins parameter. The soft nature of interactions in
DPD allows to use larger time steps (this speeds up the calculations), as compared
with MD that deals with steep repulsive potentials. However, one should avoid a
too large time step that leads to numerical errors (strongly dependent on the
numerical algorithm used). Popular software packages including DPD dynamics
include e.g. GROMACS-DPD, LAMMPS, ESPResSo.
Monte Carlo (MC) methods represent a variety of models and algorithms based
on random events. MC methods often use a coarse-graining approach and a lattice
environment, instead of the continuous space, which increases the simulation speed
due to fast recognition of neighbours of a given molecule. Various methods have
been proposed in the field of simulation of conformation and dynamics of polymer
systems. The bond fluctuation model (BFM), first introduced by Carmesin and
Kremer [194] and later modified by Shaffer [195], is a good example of the MC
approach. It works on the regular cubic lattice originally proposed by Meyer for
calculation of entropy in polymer solutions [196] and next used also in the early
stage of the Flory–Huggins theory [39–44]. In this model, the excluded volume is
preserved (one element can occupy one lattice site), but the algorithm requires
usually, at least, *30% of empty lattice sites to work. The monomers can be joined
by a bond vector, which is taken from a set of 108 allowed vectors in Carmesin and
Kremer version of the algorithm. In the athermal case, a movement attempt is
realized in a set of steps: (i) a random selection of a monomer and move direction,
(ii) check if the excluded volume condition would be fulfilled and (iii) if the bond
can be created in a new position. If checks are positive, the movement of the
monomer is performed. Taking into account the temperature, additional energetic
test (using e.g. Metropolis algorithm [197]) is required, which leads to the signif-
icant extension of the calculation time.
Other classical MC methods applied for polymers are the general reptation
algorithm [198] (also empty sites are needed, but the algorithm is not ergodic), the
pivot algorithm [199] and the bond breaking algorithm [200] (working on the fully
occupied lattice, but do not preserve chains integrity). Interesting approaches called
the Cooperative Motion Algorithm (CMA) [201] and the Dynamic Lattice Liquid
(DLL) model [202] were proposed by Pakula. Both methods are based on the
cooperative movement of system elements and work on the completely occupied
lattice (no vacancies are needed). In the CMA method, system elements (solvent,
chain segments) are moved as cooperative loops starting from the temporal
vacancy, which randomly moves through the system, moving encountered elements
on the vacancy actual place until the starting point is reached. This way leads to
preservation of the excluded volume of beads. CMA is regarded as a good tool to
predict static properties and is very efficient in equilibration of polymer melts [203].
The DLL model is a dynamic method and is also based on a lattice structure; i.e.,
positions of beads (representing solvent or chain segments) are consistent with
lattice sites. It is assumed that molecular systems have some excess volume.
Therefore, molecules have enough space to vibrate around their positions (this is
considered as displacement attempts). However, moves cannot be easily performed
over a larger distance because all the neighbouring lattice sites are occupied (DLL
can also work with vacancies included). Consequently, most of the attempted
displacements finish with the return to the initial position. Only joint attempts
undertaken in groups of beads (at least two-member groups) can be successful, as it
is illustrated in Fig. 8.14. In the considered system, only paths in the form of closed
loops (a sum of displacement vectors for a given loop equals zero) can satisfy this
condition. The algorithm defined by the DLL model starts with a random selection
of unit vectors which are assigned to all beads and represents movement attempts.
An exemplary assignment for simple liquid on the triangular lattice is shown in
Fig. 8.14. All beads which do not contribute to correlated sequences (loops) are
immobilized; the rest is moved by one lattice constant. Many rearrangements can
appear simultaneously in the system as the algorithm is fully parallel [202]. The
introduction of bonds between beads, representing polymer chains, leads to two
additional constraints—movement attempts leading to breaking a bond or violating
bond excluded volume are immobilized. DLL can be implemented on sequential
computers (even PC class). However, full capabilities of DLL model could be
reached only on parallel machines like Analyzer of Real Complex Systems—
ARUZ (built in Lodz in 2016). DLL has been successfully used in simulations of a
variety of problems related to stimuli-responsive polymers, including gelation
processes [204], polymer stars formation [205], polymer network crosslinking
[206], diffusion in complex environments [207].
The complexity of stimuli-responsive systems and difficulties in simulations in a
broad range of time and size scales lead to the limited number of works devoted to
these materials. However, some crucial points were discussed in the literature in the
light of results of computer simulations. These are hydration of polymer chains,
diffusion properties of particular components (polymer chains, water and eventual
additives), changes in polymer conformation. Further discussion is restricted only to
synthetic SRPS in water with the exception of biological systems, such as proteins,
lipids, polysaccharide and others.
The most well-studied thermo-responsive polymer is poly(N-iso-
propylacrylamide) (PNIPAM) and its co-polymers. It forms hydrogel that exhibits
the volume phase transition in water at 32 °C (305 K) [208]. One of the first MD
simulations of PNIPAM–water system was performed in 2004 by Longhi et al.
[209], who investigated the hydration of PNIPAM monomer and 50-unit oligomer
in dilute aqueous solution at 300 and 310 K. The study of water–polymer hydrogen
bonding showed that the water molecules constituting bound water (called by the
authors’ first hydration shell) were mainly localized in the hydrophilic region of the
monomer. This observation was confirmed also for other polymers such as PVME
[210]. It was also shown that the number of water–water hydrogen bonds in the
region near to the hydrophilic groups of PVME is smaller by 1–1.9 than those in
pure water. Furthermore, this effect was not compensated by polymer–water
hydrogen bonds in case of PVME. This result well corresponds to the correlation
between the water structure and the hydration degree of PVME chain found
experimentally by Pastorczak et al. [162]—see Fig. 8.10. In the region near the
hydrophobic groups, the average number of water–water hydrogen bonds was
Fig. 8.14 (a) A schematic illustration of particle movement in cooperation with neighbours.
Numbers indicate molecule location in next time steps, A and B mark loops moving the molecule
between position 1–2 and 2–3, respectively. (b) Schematic representation of molecular simple
liquid in the DLL algorithm. The vector field represents displacement attempts in the direction of
the neighbouring lattice sites. An example of rejected attempts of movement: (1) collision,
(2) creation of a vacant site, (3) molecule not participating in closed loop; accepted move:
(4) cooperative loop
found to be only slightly smaller in comparison to pure water, while bulk water
does not differ from pure water [210]. The above-mentioned results were obtained
for linear polymers using united atom approximation.
Beside the non-uniform distribution of water molecules along the PVME chain,
it was also found that a significant amount (10%) of ether oxygen atoms does not
interact with water via a hydrogen bond, even if the concentration of the solution is
very low (3.3 wt%) and the temperature is below LCST (300 K) [211]. This
number increased with the increasing polymer content, however, the increase rate
changes if the concentration reaches *27 wt%. This point was recognized as
relating to demixing transition found experimentally by colorimetric method at 30
wt% [211]. Partial hydration of hydrophilic centres of different polymer chains
(PMEO2MA) in water was postulated not only basing on simulations [212], but also
on vibrational spectroscopy studies (see section Thermo-responsive systems
—“non-ionisable” polymers and [161]). It was explained in the light of steric effects
[146, 161, 210], which was well supported by comparative MD studies on various
model hydrogels, namely PVME, PNIPAM and poly(vinyl alcohol) (PVA), which
showed that the number of water–polymer hydrogen bonds is the largest for PVA
and the smallest for PVME [210]. This relates to accessibility of a polymer hy-
drophilic centre for water–oxygen in the PVME structure is partially hidden by both
a polymer chain and a methyl group, while in PVA the chain is the only steric
hindrance for water molecules. Additionally, some of hydrophilic (oxygen) centres
form hydrogen bonds with two water molecules despite the presence of
non-hydrated hydrophilic groups, as it was shown for PMEO2MA [213]. MD
simulations performed for PVME confirmed such a possibility. As it was mentioned
previously, the total bound water for many polymers and gels oscillated in the range
of 10–13 molecules per monomer unit in the swollen state. Some light was shed on
these, rather unexpected, behaviours by MD simulations, which showed that no
more than two hydration shells differ in their properties from bulk water [213]. It
was also found that the amount of not structured water molecules (one-bonded or
non-bonded to other water molecules) increased with polymer concentration at the
expense of double water–water hydrogen-bonded molecules. At the sufficiently
high polymer concentration, a significant amount of water was hydrogen bonded to
two polymer units [149].
Even for simple polymers, the knowledge about the strength of different types of
interactions (water–water, water–polymer, polymer–polymer) is important.
Saramak et al. [149] used ab initio QM to calculate the interaction energy as a
function of a distance of water molecules from the PVME trimer. Water–O–CH3
group and water–water interactions were found to have similar strength with the
energy minimum at *3 Å. Other interactions were significantly weaker. In case of
polymers with more than one hydrophilic centre (e.g. PMEO2MA) and many
co-polymers, it is important to know the thermal stability of hydrogen bonds
between water and a particular polar group, as a crucial issue influencing the
kinetics of the coil-to-globule or volume phase transitions. Maeda et al. [146]
showed for PMEO2MA that the strongest interactions are formed between water
and a carbonyl group, while the energy of water–ether oxygen interactions
decreases with increasing the distance from the main chain. They also showed that
the H-bond between the ether oxygen and water influences vibrations relating to the
stretching of adjacent alkyl groups. They can thus be experimentally monitored by
both the shift of C–O stretching modes as well as by the shift of C–H stretching
modes. However, in the second case two mechanisms for observed changes can be
proposed: (i) a direct interaction between water and the C–H groups (including
formation of the C–HO–H systems) and (ii) hyperconjugation between the ether
oxygen and the adjacent alkyl groups leading to the shortening of C–H bond. This
fact is important in the light of the results discussed in the previous section, where it
was shown that the C–H stretching vibrations are very useful to study the hydration
degree in polymer gels and solutions.
The dehydration during the coil-to-globule transition was also investigated using
computer methods. It was found that for polymers exhibiting LCST hydration
decreased with increasing the temperature primarily due to the fact that the polymer
hydrophobic groups tend to hide from the solvent. For PNIPAM, the number of
water molecules residing in the first coordination shell is ca. 6% lower at 310 K
than at 300 K [209]. A little bit higher value (ca. 14%) was given by Tavagnacco
et al. [214], who performed MD simulations of oligomer containing 30 monomer
units. However, in this case the range of studied temperatures was significantly
broader (from 243 to 323 K). Below the coil-to-globule transition temperature,
PNIPAM was surrounded by a network of hydrogen-bonded water. The change of
hydrogen bond pattern between water molecules as well as between water and
amide groups was observed above the transition temperature. Moreover, the hy-
dration shell was only partially lost during the coil-to-globule transition and even
above the LCST polymers still remained largely hydrated. It is fully compatible
Fig. 8.15 Snapshots of PMEO2MA polymers at T < LCST (290 K) and T > LCST (330 K).
Reprinted with permission form [212], copyright (2018) John Wiley and Sons Inc.
with the experimental data [155, 159, 160]. It was also proved that dehydration of
PNIPAM chains induced by the increasing temperature and leading to the
coil-to-globule transition is mainly an entropy-driven process [215].
Although results concerning water interactions with strongly polar or ionised
groups are rather well understood, the hydrophobic interactions are still discussed in
the literature. Dalgakiran and Tatlipinar [212] showed that water solubility of
PMEO2MA below LCST was provided by hydrophobic hydration around the side
chain carbon atoms involving cage-like water structures. Those cages were dis-
rupted in the temperature above LCST and accompanied by a reduction in local
water-ordering—see snapshots in Fig. 8.15. Based on the analysis of a number of
hydrogen bonds and hydrogen bond lifetimes, the authors postulated that a direct
intermolecular hydrogen bonding between a polymer and water molecules has only
a minor impact on the phase transition. The decrease in water structuration on the
polymer–water interface across the transition temperature was shown also by
Bhandary et al. [216], who investigated the wetting behaviour of PNIPAM polymer
brushes grafted onto the silicon surface using MD simulations, and by Tavagnacco
et al. [214], for free oligomers in aqueous solution.
Studies of multi-stimuli-sensitive systems, like poly(2-(dimethylamino)ethyl
methacrylate) (PDMAEMA) [217] , are much more difficult. This polymer holding
ability of protonation of amine groups responsible for pH sensitivity is also a
thermo-responsive material. The answer to thermal stimuli strongly depends on pH
in such a system. A protonated dimethylamino group maintains its partial hydration
even in the collapsed state, while the unprotonated one loses most of the bound
water in the globule state. This behaviour influences also water interactions with a
carbonyl group and impacts thermo-responsiveness. Accessibility of charges on the
PDMAEMA molecule was found to be close to *50% [218].
Diffusive properties of water in polymer hydrogels and solutions are one of the
most critical issues in VPT and coil-to-globule transitions. Translational
self-diffusion coefficient D for water molecules in the first hydration shell (bound
water) formed around 50-unit PNIPAM chain was investigated with MD simula-
tions [209]. At 300 K, it was found ca. 19% smaller than that of bulk water.
Moreover, D for water molecules surrounding the PNIPAM oligomers doped with
amino acid-based co-monomers was found as characteristically dependent on the
distance from the solute molecule [219]. The D values were small in the range of
short distances and asymptotically increased far apart the polymer. Water domains
sufficiently far from macromolecule (>2 nm) revealed D values close to those found
for bulk water at the same temperature. Such behaviours were related to the water–
polymer interaction, including, in particular, the water bridges between distant
oligomer units. Moreover, all investigated doped oligomers revealed the same
influence on the solvent. MD simulations performed by Tavagnacco et al. [214]
showed a correlation between polymer segmental dynamics and a diffusion motion
of bound water occurring with the same activation energy. Below the transition
temperature, the ratio of bonded water diffusion coefficient to bulk water diffusion
coefficient ranged from 0.2 to 0.3 and it increased up to 0.4 for temperature above
the coil-to-globule transition. In case of PVA, PVME and PNIPAM, the mobility of
water molecules was highly reduced around the polymer chains for both transla-
tional and rotational motions [220]. There are two reasons for this reduction: (i) the
existence of hydrogen bonds between water and polymers around the hydrophilic
groups and (ii) structuration of water around the hydrophobic groups. The mean
square displacement for water, used to calculate D, was approximately linear only
in time of 0.5–5 ps due to statistics in the selected region. For PVA and PNIPAM,
the water motions were suppressed more significantly in the hydrophilic region than
in the hydrophobic region with diffusion coefficients approximately 10–40% of
those in pure water. For the PVME hydrogel in the hydrophilic region, the motions
were hindered because of tight hydrogen bonds between water and polymers and
stabilization of water–water hydrogen bonds. In the hydrophobic region, the
motions were hindered because of the structuration of water. For the PVME,
hydrogels with the water content higher than 75 wt%, D values in the bulk region
were almost the same as for pure water, while for the systems with a lower amount
of water D decreased in the bulk region, influenced by the presence of polymers.
For the high PVME concentration (> 54 wt%), a water sub-diffusive behaviour was
observed for time below * 0.1 ns (the mean square displacement slope in the log–
log scale was lower than unity) and a major decrease of water diffusion coefficients
(connected with the slope in the normal scale) [213]. It is important to notice that
water does not crystallise any more in the PVME–water system with the polymer
content higher than 61% [154].
Studies of PDMAEMA revealed that water dynamics in the vicinity of the
carbonyl group was more restricted than that of other functional groups (it is in
agreement with the strongest energy of this interaction, as it was mentioned above)
with the increasing temperature. The extended residence time for water molecules
surrounding the polymer chain resulted from both, a relatively slow decrease of
relaxation time of water–C=O group H-bonds, and the extended relaxation time of
water–water H-bonds in the vicinity of the carbonyl group [217].
To investigate bigger systems, less molecularly detailed methods (using the

coarse-graining procedures) are useful. As an example, studies performed for the
model PVME–water system on the 3D lattice containing 503 beads with the use of
DLL algorithm can be referred [149]. PVME chains were polymerized up to
DP 190, which for 250 chains in the calculation box allowed to achieve the
maximal water–polymer molar content of 3:1 (this value corresponds approxi-
mately to 500,000 atoms constituting polymer chains in the system). The diffusive
behaviour of water close to the chains (bound water) and more far away (bulk
water) was analysed as a function of the polymer content. The chains influence
strongly the water mobility only in the first coordination shell. However, only the
athermal casus was considered (such a system corresponds to h conditions). These
studies were continued to check the influence of both polymer chain length and
concentration on the mobility of solvent molecules in the PVME–water system by
Kozanecki et al. [148]. The obtained results for self-diffusion coefficients showed a
good agreement with a free volume theory. In diluted solutions, diffusion of the
solvent near the chains was affected by the polymer chain length. In dense solu-
tions, the effect of polymer concentration was dominating. This behaviour was well
correlated with the spatial polymer homogeneity in the system.
The last part of this chapter is devoted to conformational changes induced in
water–polymer systems by the temperature. The conformational changes of
PNIPAM hydrogel as a function of the temperature was first studied by Walter et al.
[215] using MD simulations. The authors reported that these changes are caused by
a conformational transition of the single polymer chains in the hydrogel backbone.
The studies were performed for chains containing 15, 30, 50 and 75 monomer units
starting from a stretched initial conformation. It was found that sufficiently long
chains (with more than 15 monomer units) collapse and remain collapsed in the
equilibrium. Moreover, longer chains need more time for the conformation tran-
sition (from 3 to 7.5 ns for 30 and 75 monomers, respectively). Different force
fields for describing the hydrogel backbone and water molecules were compared
showing surprisingly that not all combinations are able to reveal the conformational
transition. Only two combinations (GROMOS-96 53A6 + TIP4P and
OPLS-AA + SPC/E) yielded the conformational change upon the temperature
variation. It was also shown that the results concerning h-temperature for
OPLS-AA + SPC/E (320 K) were in fair agreement with the experimental results
(305 K) [221].
Tucker and Stevens [222] investigated the polymer length dependence (over the
range of 3–30 mers) of the transition temperature for a single syndiotactic PNIPAM
oligomers and confirmed higher LCST for shorter chains [223]. The interactions
were simulated in a longer than 100-ns MD simulation. The differences in the
vibrational spectra in PNIPAM–water system at various temperatures (during
heating and cooling cycles) were investigated with MD simulation [224] to
understand the coil-to-globule and globule-to-coil transitions and identify the
changes in the relative interactions between various polymer groups and water.
Oligomer–water interactions were found to be dominant below the LCST, whereas
oligomer–oligomer interactions predominated above the LCST. Moreover,
interactions between oligomer–oligomer and between the oligomer and water were
less perturbed during the cooling cycle. Results suggested that a hysteresis, if it
occurs, is a consequence of ultrafast heating–cooling kinetics, which results in the
insufficient relaxation for the solvated oligomer. The coil-to-globule transition was
also confirmed for PMEO2MA by MD simulations [212]. The analysis of the pair
interaction between PNIPAM polymer chains, in a concentration regime close to
the chain overlap concentration, by comparing atactic and isotactic-rich PNIPAM
solutions [225] provided information on the features of interchain junctions as a
function of tacticity. MD simulations carried out above the LCST showed the
PNIPAM transition to the insoluble state and do not reveal much influence of
stereochemistry on the structure of the polymer ensemble. The results for the
temperature below LCST gave the estimate of the swelling ratio of
non-stereocontrolled PNIPAM microgels as equal to 12, which is a typical value—
see section Thermo-responsive systems—“non-ionisable” polymers.
Additives can influence the coil-to-globule transition; e.g., a comparison of pure
PNIPAM with PNIPAM doped with amino acid-based co-monomers showed that
the presence of co-monomers in the chain prevents against the chain collapse upon
heating above 315 K [219]. MD simulations were successfully used to model
PNIPAM oligomers with chain lengths of 8, 18 and 30 units. The results compared
with the PNIPAM–alkyl copolymer surfactant revealed that in case of pure
PNIPAM oligomers, the backbone folded onto itself above the LCST to shield the
hydrophobic groups from water. The surfactant, however, did not show confor-
mational change with the temperature [226]. The effect of salts addition (NaCl,
NaBr, NaI, KCl) on lowering the LCST temperature for a single 50-unit PNIPAM
chain was simulated by Du et al. [227]. The authors found that cations had strong
affinity with the amide oxygen in the polymer. However, the affinity can be
weakened by a stronger cation–anion interaction. Coarse-grained MD simulations
were used to study thermo-responsive and pH-responsive behaviours for
PAA-PDMAEMA co-polymers in water [228]. In these studies, water was treated
as a continuous medium, and for polymer short-range, long range Coulombic
interactions and hydrogen bonding were taken into account as a bead-spring model.
It was found that a maximum pH value exists for which PAA-PDMAEMA solu-
tions exhibit pronounced thermo-responsive behaviours. The effect of the polymer
molecular mass on the aggregation behaviour had been more profound than that of
the polymer composition.
PNIPAM was also investigated in a form of brush grafted onto the silica surface
[229]. It was showed that above the LCST, the PNIPAM chains (50-unit oligomers)
aggregate on the silica support forming a hydrophobic surface. At temperatures
below the LCST, the PNIPAM chains were stretched and broadcast the surface
properties to the hydrophilic one. A more sophisticated grafting was investigated by
Lin et al. [230], i.e. the effect of grafting PNIPAM on the edge of nanoscale silicate
platelets. In this case, the DPD method was used to study the effects of the
polymer-grafted architecture and grafting density on the micellar behaviour.
The use of a finite-difference simulation (FDM) to calculate the electric field
intensity and its distributions in smart thermo-stimulated PNIPAM/silver
nanoparticles hybrid composites for surface-enhanced Raman scattering (SERS)

studies is an interesting application of macroscopic methods to solve application
problems related to stimuli-responsive systems [231]. Finally, it is worth citing
Moghadam and Larson [232], who studied aqueous solutions of hydrophobic drug
molecules (phenytoin) with PNIPAM and its co-polymers to assess its efficacy for
drug delivery applications (effects of the chain length, tacticity, and the
co-monomer content were investigated). The authors noted that: ‘(..) the length and
time scales reachable by atomistic simulations are not sufficient to observe rare
nucleation events and crystal growth nor the effect of polymer on these processes.
Furthermore, even though our results clearly show the collapse of single chains in
aqueous solutions as the temperature increases above the LCST, to understand the
dynamics of gels, which are composed of a complex network of these chains, we
would need to simulate chains containing hundreds of monomers, which is far
beyond the capability of atomistic simulations and would require the development
of coarse-grained models of both polymer and drug, perhaps also using implicit
solvent to increase simulation time/length scale’.
8.6 Outlooks and Perspectives
Currently, stimuli-responsive polymer materials, especially hydrogels, gain a great

attention as potential biomaterials for tissue engineering, drug carriers and
bio-sensors. In this field, the bio-inert systems containing water as a solvent are
mostly demanded. Oligo- and poly (ethylene glycol) and its derivatives (including
POEGMAs), and co-polymers seem to be most attractive. Also, water treatment
technologies important for the environment protection have started to use
stimuli-responsive polymer hydrogels for ion recovery, organic compounds
adsorption and bio-separation. Lenses with a adjustable focal length, attenuators,
micro-valves are examples of technical applications of SRPS. However, some
polymer smart systems have been already commercialized and they are produced at
the industrial scale. Many interesting fields have been still undiscovered and require
further systematic investigation. The so-called reconfigurable materials have risen
to the forefront here [6, 233]. This new class of materials can change not only the
volume, but also the shape in a controlled manner. This new functionality can be
used to design smart connectors and adhesives, grippers, self-folding (origami)
systems, artificial muscles and attenuators and potentially many other devices.
The creation of smart materials still requires new ways of the synthesis of
polymer systems with a strictly demanded size, shape, functionality and properties.
Currently, a broad range of controlled radical polymerisations, such as many
variants of the atom transfer radical polymerisation (ATRP) or the reversible/
addition fragmentation chain transfer (RAFT) polymerization should be mentioned
here [234–241]. The relatively new approach is the polymerization-induced
self-assembly (PISA) (as an alternative to physical methods of controlled
micro-phase separation of di- and multiblock co-polymers), allowing to create
Fig. 8.16 3D PNIPAM printed micro-structures and their changes in size and shape controlled by
temperature. The PNIPAM-based photocurable resin was obtained by mixing its monomer with N,
N′-methylene-bis(acrylamide) (crosslinker) and phenylbis (2,4,6-tri-methylbenzoyl) phosphine
oxide (photo-initiator) in ethanol. To control the depth of light penetration, Sudan I (photoab-
sorber) was also added, and Rhodamine B was used to visualize the final structure. a Gripper arm
bendability was achieved by the use of various light intensities to cure the resin in a ‘white’ and
‘grey’ part of each arm; these parts differ in network density and in the degree of swelling. b A
dumb-bell-shaped structure is able to change its shape as it consists of two parts differing in the
VPT temperature. The shift of VPT temperature was achieved by addition of anionic co-monomer
to the resin (methacryl-amidopropyl-trimethyl-ammonium chloride). Open Access [244] by
Springer Nature
polymer systems with various morphologies and a controlled nanostructure

[242, 243].
The chemical methods allow to control the structure at the nano- or, at most,
micrometer level. Thus, the development of polymer processing is also necessary to
find new, innovative, ways of designing stimuli-responsive devices on the higher
size scale. One of the last very spectacular examples in this field is the usage of 3D
printing to create PNIPAM thermo-sensitive items with complex shapes—as
Fig. 8.16 shows [244].
The big challenge still ahead for the time-resolved vibrational spectroscopy
methods is to fully determine kinetics and identify molecular origins of all the
stages of various types of stimuli-responsive phase transitions revealed by SRPSs.
This task is difficult for several reasons. First of all, there is a problem of delivering
an instantaneous trigger of a phase transition in the whole studied volume of a
sample. In the case of thermo-responsive polymers, such a trigger could be a
high-energy near-infrared pulse being in resonance with a water absorption band.
The vibration excitation of water would be transferred into heat within tens of
picosecond in the irradiated volume of a sample (an excellent review of the
so-called temperature-jump method can be found in [245]). In the case of
pH-responsive systems, a fast decrease in the pH of a system may be provided, e.g.
by addition of the so-called photoacids into a sample. A unique feature of these
compounds is a large change in their acidity constant pKa upon electronic exci-
tation to S1 state. In the aqueous solution or gel, it may lead to the dissociation of
photoacid into a photobase and a proton. The proton stays in a solution as long as
the photoacid is in its S1 state. An example of similar ‘pH-jump method’ can be
found in [246].
Another problem is that most likely particular stage of stimuli-responsive phase
transition spans from picoseconds (e.g. proton transfer [54]) to milliseconds (col-
lapsing of a polymer chain into a globule [247] or even to minutes (water diffusion
[248])—see Fig. 8.17. In a typical ultrafast time-resolved spectroscopy system, a
probe pulse is delayed with respect to a pump due to a different length of paths they
travel (provided by motorized delay stages). Such a method provides a time win-
dow of a few nanoseconds. Longer delays (which provide an observation time
window of many microseconds) may be achieved by using the system of two
electronically synchronized lasers, which is a very expensive and technically dif-
ficult solution.
Different computational methods are very useful to predict the structure and
properties of stimuli-responsive materials in different ranges of time and size.
Furthermore, they can be a good support for experimental techniques, as the results
delivered by simulations correspond well to experimental ones. However, up to
now molecular simulations have been limited to a few hundred thousands of atoms
and times of a few hundreds of nanoseconds. To be able to fully compare simu-
lations with experiments of both time and length scales simulated in ‘all-atom’
approach would require, at least, 2–3 orders of magnitude more. Currently, the
spatial and temporal resolutions easily accessible for computer simulations and
experiments are practically disjunctive in the field of SRPS. The local phenomena
occurring at short times could be precisely modelled, but require very sophisticated
and expensive equipment to be studied experimentally. Contrarily, slow macro-
scopic processes may be monitored by various simple techniques, while their
investigations by computer methods have been still suppressed by hardware limi-
tations. Therefore, further progress in the field of both computational and experi-
mental methods is needed. In the case of simulations, it is directly connected with
the development of faster hardware and new algorithms involving parallel com-
puting on clusters, support of GPU-based calculations, etc., enabling the
Fig. 8.17 Schematic presentation on a current view on a coil-to-globule transition of a polymer in

water after an external stimulus (e.g. increase in temperature above LCST): left below—a relaxed
chain in a solution below LCST, left upper—rearrangement of chains conformations after
receiving an external stimulus with desorption of water, upper right—aggregation of globules,
right below—slow diffusion of water from polymer aggregates. Timescales of particular stages
were estimated from the following studies on conformational changes in aqueous solutions of
polymer or peptides [54, 247, 248]
investigation of larger systems in longer timescales. Still, new original and inno-
vative alternatives appear in the form of dedicated machines, like Anton 2 for MD
simulations or ARUZ for MC methods. From the experimental point of view, the
high-resolution time (pump-and-probe) and space (such as nanoIR)-resolved
methods seem to be the most promising to open new doors towards a better
understanding of complex phenomena in polymer–water systems responsible for
their stimuli responsivity.
Acknowledgements The authors acknowledge the financial support within the projects granted
by National Science Centre, Poland: No. 2013/09/B/ST4/03010 (MK), No. 2014/14/A/ST5/00204
(KH, MK), No. DEC-2013/08/S/ST4/00556 (MP) and UMO-2015/17/B/ST4/04035 (MP). Special
thanks to Prof. Jacek Ulanski and Prof. Piotr Polanowski from the Department of Molecular
Physics for long and fruitful discussions. Also, the authors would like to acknowledge Prof. Piotr
Ulanski and Prof. Slawomir Kadlubowski from the Institute of Applied Radiation Chemistry at
Lodz University of Technology for everyday support and cooperation.
References
1. Cabane E, Zhang X, Langowska K, Palivan CG, Meier W (2012) Stimuli-responsive

polymers and their applications in nanomedicine. Biointerphases 7:9
2. Gil ES, Hudson SM (2004) Stimuli-reponsive polymers and their bioconjugates. Prog Polym
Sci 29:1173–1222
3. Delcea M, Mohwald H, Skirtach AG (2011) Stimuli-responsive LbL capsules and nanoshells
for drug delivery. Adv Drug Deliv Rev 63:730–747
4. Liechty WB, Kryscio DR, Slaughter BV, Peppas NA (2010) Polymers for drug delivery
systems. Annu Rev Chem Biomol Eng 1:149–173
5. Schmaljohann D (2006) Thermo- and pH-responsive polymers in drug delivery. Adv Drug
Deliv Rev 58:1655–1670
6. Wei M, Gao Y, Li X, Serpe MJ (2017) Stimuli-responsive polymers and their applications.
Polym Chem 8:127–143
7. Hu J, Meng H, Li G, Ibekwe SI (2012) A review of stimuli-responsive polymers for smart
textile applications. Smart Mater Struct 21:053001
8. Gao Y, Wei M, Li X, Xu W, Ahiabu A, Perdiz J, Liu Z, Serpe MJ (2017) Stimuli-responsive
polymers: fundamental considerations and applications. Macromol Res 25:513–527
9. Štular D, Simončič B, Tomšič B (2017) Stimuli-responsive hydrogels for textile
functionalisation: a review. Tekstilec 60(2):76–96
10. Ito S (1989) Phase transition of aqueous solution of poly (N-alkylacrylamide)
derivatives-effects of side chain structure. Kobunshi Ronbunshu 46(7):437–443
11. Ito S (1990) Phase transition of aqueous solutions of poly (N-alkoxyalkylacrylamide)
derivatives effects of side chain structure. Kobunshi Ronbunshu 47(6):467–474
12. Suwa K, Morishita K, Kishida A, Akashi M (1997) Synthesis and functionalities of poly(N-
vinylalkylamide). V. Control of a lower critical solution temperature of poly(N-
vinylalkylamide). J Polym Sci A 35:3087–3094
13. Lutz JF (2008) Polymerization of oligo(ethylene glycol) (meth)acrylates: toward new
generations of smart biocompatible materials. J Polym Sci A 46:3459–3470
14. Roth PJ, Jochum FD, Forst FR, Zentel R, Theato P (2010) Influence of end groups on the
stimulus-responsive behavior of poly[oligo(ethylene glycol) methacrylate] in water.
Macromolecules 43:4638–4645
15. Kojima C (2010) Design of stimuli-responsive dendrimers. Expert Opin Drug Deliv 7:
307–319
16. Haba Y, Harada A, Takagishi T, Kono K (2004) Rendering poly(amidoamine) or poly
(propylenimine) dendrimers temperature sensitive. JACS 126:12760–12761
17. Haba Y, Kojima C, Harada A, Kono K (2007) Comparison of thermosensitive properties of
poly(amidoamine) dendrimers with peripheral N-isopropylamide groups and linear polymers
with the same groups. Angew Chem Int Ed 46:234–237
18. Li W, Zhang A, Chen Y, Feldman K, Wud H, Schlüter D (2008) Low toxic,
thermoresponsive dendrimers based on oligoethylene glycols with sharp and fully reversible
phase transitions. Chem Commun 2008:5948–5950
19. Parrott MC, Marchington EB, Valliant JF, Adronov A (2005) Synthesis and properties of
carborane-functionalized aliphatic polyester dendrimers. JACS 127:12081–12089
20. Pistolis G, Malliaris A, Tsiourvas D, Paleos CM (1999) Poly(propyleneimine) dendrimers as
pH-sensitive controlled-release systems. Chem Eur J 5:1440–1444
21. Sideratou Z, Tsiourvas D, Paleos CM (2000) Quaternized poly(propylene imine) dendrimers
as novel pH-sensitive controlled-release systems. Langmuir 16:1766–1769
22. Sideratou Z, Tsiourvas D, Paleos CM (2001) Solubilization and release properties of
PEGylated diaminobutane poly(propylene imine) dendrimers. J Colloid Interface Sci
242:272–276
23. Paleos CM, Tsiourvas D, Sideratou Z, Tziveleka L (2004) Acid- and salt-triggered
multifunctional poly(propylene imine) dendrimer as a prospective drug delivery system.
Biomacromol 5:524–529
24. Kimura M, Kato M, Muto T, Hanabusa K, Shirai H (2000) Temperature-sensitive dendritic
hosts: synthesis, characterization, and control of catalytic activity. Macromolecules
33:1117–1119
25. You YZ, Hong CY, Pan CY, Wang PH (2004) Synthesis of a dendritic core–shell
nanostructure with a temperature-sensitive. Shell Adv Mater 16:1953–1957
26. Xu J, Luo S, Shi W, Liu S (2006) Two-stage collapse of unimolecular micelles with double
thermoresponsive coronas. Langmuir 22:989–997
27. Yang Z, Zhang W, Zou J, Shi W (2007) Synthesis and thermally responsive characteristics
of dendritic poly(ether-amide) grafting with PNIPAAm and PEG. Polymer 48:931–938
28. Lee HI, Lee JA, Poon Z, Hammond PT (2008) Temperature-triggered reversible micellar
self-assembly of linear–dendritic block copolymers. Chem Commun 2008:3726–3728
29. Plummer R, Hill DTJ, Whittaker AK (2006) Solution properties of star and linear poly(N-
isopropylacrylamide). Macromolecules 39:8379–8388
30. Xu J, Liu S (2009) Synthesis of well-defined 7-arm and 21-arm poly(N-isopropylacrylamide)
star polymers with b-cyclodextrin cores via click chemistry and their thermal phase
transition behavior in aqueous solution. J Polym Sci A 47:404–419
31. Liu YY, Zhong YB, Nan JK, Tian W (2010) Star polymers with both temperature sensitivity
and inclusion functionalities. Macromolecules 43:10221–10230
32. Lambeth RH, Ramakrishnan S, Mueller R, Poziemski JP, Miguel GS, Markoski LJ,
Zukoski CF, Moore JS (2006) Synthesis and aggregation behavior of thermally responsive
star polymers. Langmuir 22:6352–6360
33. Ni C, Wu G, Zhu C, Yao B (2010) The preparation and characterization of amphiphilic star
block copolymer nano micelles using silsesquioxane as the core. J Phys Chem C 114:
13471–13476
34. Bai Y, Wei J, Yang L, He C, Lu X (2012) Temperature and pH dual-responsive behavior of
polyhedral oligomeric silsesquioxane-based star-block copolymer with poly(acrylic
acid-block-N-isopropylacrylamide) as arms. Colloid Polym Sci 290:507–515
35. Zhu W, Nese A, Matyjaszewski K (2011) Thermoresponsive star triblock copolymers by
combination of ROP and ATRP: from micelles to hydrogels. J Polym Sci A 49:1942–1952
36. Guo Y, Li M, Li X, Shang Y, Liu H (2017) Stimuli-responsive and micellar behaviors of
star-shaped poly[2-(dimethylamino)ethyl methacrylate]-b-poly[2-(2-methoxyethoxy)ethyl
methacrylate] with a b-cyclodextrin core. React Funct Polym 116:77–86
37. Das S, Chatterjee DP, Ghosh R, Das P, Nandi AK (2016) Water soluble stimuli-responsive
star copolymers with multiple encapsulation and release properties. RSC Adv 6:8773–8785
38. Kuckling D, Wycisk A (2013) Stimuli-responsive star polymers. Inc J Polym Sci A
51:2980–2994
39. Huggins ML (1941) Solutions of long chain compounds. J Chem Phys 9:440
40. Flory PJ (1941) Thermodynamics of high polymer solutions. J Chem Phys 9:660–661
41. Flory PJ (1941) Molecular size distribution in three dimensional polymers. I. Gelation. JACS
63:3083–3090
42. Flory PJ (1941) Molecular size distribution in three dimensional polymers. II. Trifunctional
branching units. JACS 63:3091–3096
43. Flory PJ (1941) Molecular size distribution in three dimensional polymers. III.
Tetrafunctional branching units. JACS 63:3096–3100
44. Flory PJ (1942) Thermodynamics of high polymer solutions. J Chem Phys 10:51–61
45. Stockmayer WH (1944) Theory of molecular size distribution and gel formation in Branched
polymers II. General cross linking. J Chem Phys 12:125–131
46. Flory PJ, Krigbaum WR (1950) Statistical mechanics of dilute polymer solutions II. J Chem
Phys 18:1086–1094
47. Krigbaum WR, Flory PJ (1953) Statistical mechanics of dilute polymer solutions. IV.
Variation of the osmotic second coefficient with molecular weight. JACS 75:1775–1784
48. Stockmayer WH (1950) Light scattering in multi-component systems. J Chem Phys 18:
58–61
49. Stockmayer WH (1960) Problems of the statistical thermodynamics of dilute polymer
solutions. Macromol Chem Phys 35:54–74
50. Flory JP (1953) Principles in polymer chemistry. Cornell University Press, Ithaca
51. Yamakawa H (1971) Modern theory of polymer solutions 1971. Harper & Row Publishers,
New York
52. Teraoka I (2002) Polymer solutions: an introduction to physical properties. Wiley, New
York
53. de Grotthuss CJT (1806) Memoir on the decomposition of water and of the bodies that it
holds in solution by means of galvanic electricity. Ann Chim (Paris) 58:54–73 (Reprint in
Eng de Grotthuss CJT Biochim Biophys Acta (2006) 1757:871–875)
54. Siwick BJ, Bakker HJ (2007) On the role of water in intermolecular proton-transfer
reactions. JACS 129:13412–13420
55. Imran AB, Seki T, Takeoka Y (2010) Recent advances in hydrogels in terms of fast stimuli
responsiveness and superior mechanical performance. Polym J 42:839–851
56. Sadlej J (2007) On the calculations of the vibrational Raman spectra of small water clusters.
Chem Phys 342:163–172
57. Walrafen GE, Hokmabadi MS, Yang WH (1986) Raman isosbestic points from liquid water.
J Chem Phys 85:6964–6965
58. Scatena LF, Brown MG, Richmond GL (2001) Water at hydrophobic surfaces: weak
hydrogen bonding and strong orientation effects. Science 292:908–912
59. Bakker HJ, Skinner JL (2009) Vibrational spectroscopy as a probe of structure and dynamics
in liquid water. Chem Rev 110:1498–1517
60. Walrafen GE (1967) Raman spectroscopy studies of the effect of temperature on water
structure. J Chem Phys 47:114–126
61. Green JL, Lacey AR, Sceats MG (1986) Spectroscopic evidence for spatial correlation of
hydrogen bonds in liquid water. J Phys Chem 90:3958–3964
62. Auer BM, Skinner JL (2008) IR and Raman spectra of liquid water: theory and
interpretation. J Chem Phys 128(22):224511
63. Tominaga Y, Fujiwara A, Amo Y (1998) Dynamical structure of water by Raman
spectroscopy. Fluid Phase Equilibira 144:323–330
64. Walrafen GE, Hokmabadi MS, Yang WH (1988) Raman investigation of the temperature
dependence of the bending m2 and combination m 2 + m L bands from liquid water. J Phys
Chem 92:2433–2438
65. Maeda Y, Kakinoki K, Kitano H (1996) Raman spectroscopic study of water in lipid
dispersions: changes in structure of hydrating water caused by gel-liquid crystal phase
transition. J Raman Spec 27:425–427
66. Luu DV, Cambon L (1990) Perturbation of liquid-water structure by ionic substances. J Mol
Struc 237:411–419
67. Green JL, Lacey AR, Sceats MG (1987) Collective proton motions in H2O/H2O2 mixture:
evidence for defects and network reconstruction. J Chem Phys 86:1841–1847
68. Marinov VS, Matsuura H (2002) Raman spectroscopic study of temperature dependence of
water structure in aqueous solutions of a poly(oxyethylene) surfuctant. J Mol Struc 610:
105–112
69. Joachimiak A, Halamus T, Wojciechowski P, Ulanski J (2005) Structure of hydrogels based
on lyotropic phases of cellulose derivative as studied by Raman spectroscopy. Macromol
Chem Phys 206:59–65
70. Pastorczak M, Kozanecki M, Ulanski J (2008) Raman resonance effect in liquid water.
J Phys Chem A 112:10705–10707
71. Sándorfy C (2006) Hydrogen bonding: how much anharmonicity? J Mol Struc 790:50–54
72. Waldron RD (1957) Infrared spectra of HDO in water and ionic solutions. J Chem Phys
26:809–814
73. Lee HB, Jhon MS, Andrade JD (1975) Nature of water in synthetic hydrogels.
I. Dilatometry, specific conductivity, and differential scanning calorimetry of polyhydrox-
yethyl methacrylate. J Coll Interface Sci 51:225–231
74. Maeda Y, Kitano H (1995) The structure of water in polymer systems as revealed by Raman
spectroscopy. Spectrochim Acta A 51:2433–2446
75. Maeda Y, lde M, Kitano H (1999) Vibrational spectroscopic Study on the structure of water
in polymer systems. J Mol Liq 80:149–163
76. Hoffman AS (2002) Hydrogels for biomedical applications. Adv Drug Deliv Rev 43:3–12
77. Lafleur M, Pigeon M, Pezolet M, Caille JP (1989) Raman spectrum of interstitial water in
biological systems. J Phys Chem 93:1522–1526
78. Cerveny S, Colmenero J, Alegria A (2005) Dielectric investigation of the low-temperature
water dynamics in the poly(vinyl methyl ether)/H2O system. Macromolecules 38:7056–7063
79. Johari GP (1981) The dipolar correlation factor, the electrostatic field, the dipole moment,
and the Coulombic interaction energy of water molecules in clathrate hydrates. J Chem Phys
74:1326–1336
80. Henry F, Gaudilla M, Costa LC, Lakkis F (2003) Free and/or bound water by dielectric
measurements. Food Chem 82:29–34
81. Craig D (1995) Dielectric analysis of pharmaceutical systems. CRC Press, Boca Raton
82. Beneduci A (2008) Which is the effective time scale of the fast Debye relaxation process in
water? J Mol Liq 138:55–60
83. Ellison WJ, Moreau JM (1996) Water: a dielectric reference. J Mol Liq 68:171–279
84. Shinyashiki N, Shimomura M, Ushiyama T, Miyagawa T, Yagihara S (2007) Dynamics of
water in partially crystallized polymer/water mixtures studied by dielectric spectroscopy.
J Phys Chem B 111:10079–10087
85. Pastorczak M, Dominguez-Espinosa G, Okrasa L, Pyda M, Kozanecki M, Kadlubowski S,
Rosiak JM, Ulanski J (2014) Poly(vinyl methyl ether) hydrogels at temperatures below the
freezing point of water—molecular interactions and states of water. Colloid Polym Sci
292:1775–1784
86. Shen YR (2016) Fundamentals of sum-frequency spectroscopy. Cambridge molecular
science. Cambridge University Press, Cambridge
87. Fayer M (2013) Ultrafast infrared vibrational spectroscopy. CRC Press, Boca Raton
88. Hamm P, Zanni M (2011) Concepts and methods of 2D infrared spectroscopy. Cambridge
University Press, Cambridge
89. Cheng J, Xie X. (2012) Coherent Raman scattering microscopy. CRC Press, Boca Raton
90. Bakker HJ, Planken PCM (1990) Role of solvent on vibrational energy transfer in solution.
Nature 347:745–747
91. Woutersen S, Emmerichs U, Bakker HJ (1997) Femtosecond mid-IR pump-probe
spectroscopy of liquid water: evidence for a two-component structure. Science 278:658–660
92. Woutersen S, Bakker HJ (1999) Resonant intermolecular transfer of vibrational energy in
liquid water. Nature 402:507–509
93. Rezus YLA, Bakker HJ (2005) On the orientational relaxation of HDO in liquid water.
94. Rezus YLA, Bakker HJ (2006) Orientational dynamics of isotopically diluted H2O and D2O.
95. Rezus YLA, Bakker HJ (2007) Observation of immobilized water molecules around
hydrophobic groups. Phys Rev Lett 99:148301
96. Laage D, Hynes JT (2006) A molecular jump mechanism of water reorientation. Science
311:832–835
97. Mohammed OF, Pines D, Dreyer J, Pines E, Nibbering ETJ (2005) Sequential proton
transfer through water bridges in acid-base reactions. Science 310:83–86
98. Cowan ML, Bruner BD, Huse N, Dwyer JR, Chugh B, Nibbering ETJ, Elsaesser T,
Miller RJD (2005) Ultrafast memory loss and energy redistribution in the hydrogen bond
network of liquid H2O. Nature 434:199–202
99. Fecko CJ, Eaves JD, Loparo JJ, Tokmakoff A, Geissler PL (2003) Ultrafast hydrogen-bond
dynamics in the infrared spectroscopy of water. Science 301:1698–1702
100. Lock AJ, Bakker HJ (2002) Temperature dependence of vibrational relaxation in liquid H2O.
J Chem Phys 117:1708–1713
101. Ashihara S, Huse N, Espagne A, Nibbering ETJ, Elsaesser T (2006) Vibrational couplings
and ultrafast relaxation of the O-H bending mode in liquid H2O. Chem Phys Lett 424:66–70
102. Ashihara S, Huse N, Espagne A, Nibbering ETJ, Elsaesser T (2007) Ultrafast structural
dynamics of water induced by dissipation of vibrational energy. J Phys Chem A 111:
743–746
103. Lindner J, Vöhringer P, Pshenichnikov MS, Cringus D, Wiersma DA, Mostovoy M (2006)
Vibrational relaxation of pure liquid water. Chem Phys Lett 421:329–333
104. Nienhuys H-K, Woutersen S, van Santen RA, Bakker HJ (1999) Mechanism for vibrational
relaxation in water investigated by femtosecond infrared spectroscopy. J Chem Phys
111:1494–1500
105. Kropman MF, Nienhuys H-K, Woutersen S, Bakker HJ (2001) Vibrational relaxation and
hydrogen-bond dynamics of HDO:H2O. J Phys Chem A 105:4622–4626
106. Hunger J, Bernecker A, Bakker HJ, Bonn M, Richter RP (2012) Hydration dynamics of
hyaluronan and dextran. Biophys J 103:L10–L12
107. Mazur K, Buchner R, Bonn M, Hunger J (2014) Hydration of sodium alginate in aqueous
solution. Macromolecules 47:771–776
108. Kocak G, Tuncer C, Butun V (2017) pH-responsive polymers. Polym Chem 8:144–176
109. Chun MK, Cho CS, Choi HK (2002) Mucoadhesive drug carrier based on interpolymer
complex of poly(vinyl pyrrolidone) and poly(acrylic acid) prepared by template polymer-
ization. J Control Release 81:327–334
110. Tsukida N, Muranaka H, Ide M, Maeda Y, Kitano H (1997) Effect of neutralization of poly
(acrylic acid) on the structure of water examined by Raman spectroscopy. J Phys Chem B
101:6676–6679
111. Tamura T, Kawauchi S, Satoh M, Komiyama J (1997) Infrared spectroscopic study and
ab initio calculation for dissociation of poly(a-hydroxy acrylic acid) in aqueous solutions.
Polymer 38:2093–2098
112. Santonicola MG, de Groot GW, Memesa M, Meszyńska A, Vancso GJ (2010) Reversible
pH-controlled switching of poly(methacrylic acid) grafts for functional biointerfaces.
Langmuir 26:17513–17519
113. Dong J, Ozaki Y, Nakashima K (1997) Infrared, Raman, and near-infrared spectroscopic
evidence for the coexistence of various hydrogen-bond forms in poly(acrylic acid).
114. Walczak WJ, Hoagland DA, Hsu SL (1992) Analysis of polyelectrolyte chain conformation
of polarized Raman-spectroscopy. Macromolecules 25:7317–7323
115. Hofmeister F (1888) Zur Lehre Von Der Wirkung Der Salze. Naunyn-Schmiedeberg’s. Arch
Pharmacol 24:247–260
116. Hofmeister F (1888) Zur Lehre Von Der Wirkung Der Salze. Naunyn-Schmiedeberg’s. Arch
Pharmacol 25:1–30
117. Kunz W, Henle J, Ninham BW (2004) ‘Zur Lehre von der Wirkung der Salze’ (about the
science of the effect of salts): Franz Hofmeister’s historical papers. Curr Opin Coll Interface
Sci 9:19–37
118. Swann JMG, Bras W, Topham PD, Howse JR, Ryan AJ (2010) Effect of the hofmeister
anions upon the swelling of a self-assembled pH-responsive hydrogel. Langmuir 26:10191–
10197
119. Sadeghi R, Jahani F (2012) Salting-in and salting-out of water-soluble polymers in aqueous
salt solutions. J Phys Chem B 116:5234–5241
120. Tielrooij KJ, Garcia-Araez N, Bonn M, Bakker HJ (2010) Cooperativity in ion hydration.
Science 328:1006–1009
121. Pastorczak M, van der Post ST, Bakker HJ (2013) Cooperative hydration of carboxylate
groups with alkali cations. Phys Chem Chem Phys 15:17767–17770
122. Liu L, Kou R, Liu G (2017) Ion specificities of artificial macromolecules. Soft Matter 13:
68–80
123. Willott JD, Murdoch TJ, Webber GB, Wanless EJ (2017) Physicochemical behaviour of
cationic polyelectrolyte brushes. Prog Polym Sci 64:52–75
124. Shen YR (1989) Surface properties probed by second-harmonic and sum-frequency
generation. Nature 337:519–525
125. Beaman DK, Robertson EJ, Richmond GL (2011) Unique assembly of charged polymers at
the oil − water interface. Langmuir 27:2104–2106
126. Beaman DK, Robertson EJ, Richmond GL (2012) Ordered polyelectrolyte assembly at the
oil–water interface. Proc Natl Acad Sci 109:3226–3231
127. Balzerowski P, Meister K, Versluis J, Bakker HJ (2016) Heterodyne-detected sum frequency
generation spectroscopy of polyacrylic acid at the air/water-interface. Phys Chem Chem
Phys 18:2481–2487
128. Kondo T, Nomura K, Murou M, Gemmei-Ide M, Kitano H, Noguchi H, Uosaki K, Ohno K,
Saruwatari Y (2012) Structure of water in the vicinity of a zwitterionic polymer brush as
examined by sum frequency generation method. Colloid Surf B 100:126–132
129. Shibayama M, Tanaka T (1993) Volume phase transition and related phenomena of polymer
gels. Adv Polym Sci 109:1–62
130. Koetting MC, Peters JT, Steichen SD, Peppas NA (2015) Stimulus-responsive hydrogels:
theory, modern advances, and applications. Mater Sci Eng R Rep. 93:1–49
131. Wu C (1998) A comparison between the ‘coil-to-globule’ transition of linear chains and the
‘‘volume phase transition’’ of spherical microgels. Polymer 39:4609–4619
132. Li Y, Tanaka T (1992) Phase transitions of gels. Annu Rev Mater Sci 22:243–277
133. Olejniczak MN, Piechocki K, Kozanecki M, Koynov K, Adamus A, Wach RA (2016) The
influence of selected NSAIDs on volume phase transition in poly(2-(2-methoxyethoxy)ethyl
methacrylate) hydrogels. J Mater Chem B 4:1528–1534
134. Tanaka T (1978) Collapse of gels and the critical endpoint. Phys Rev Lett 40(12):820–823
135. Dusek K, Patterson D (1968) Transition in swollen polymer networks induced by
intramolecular condensation. J Polym Sci A 6:1209–1216
136. Pititsen OB, Eizner YE (1965) The theory of helix-coil transitions in macromolecules.
Biofizika 10:3–6
137. DeGennes PG (1972) Exponents for the excluded volume problem as derived by the Wilson
method. Phys Lett A 38:339–340
138. Lifshitz IM, Grosberg AY, Khokhlov AR (1978) Some problems of the statistical physics of
polymer chains with volume interaction. Rev Modern Phys 50:683–713
139. Maeda Y, Nakamura T, Ikeda I (2002) Hydration and phase behavior of poly(N-
vinylcaprolactam) and poly(N-vinylpyrrolidone) in water. Macromolecules 35:217–222
140. Schafer-Soenen H, Moerkerke R, Berghmans H, Koningsveld R, Dusek K, Solc K (1997)
Zero and off-zero critical concentrations in systems containing polydisperse polymers with
very high molar masses. 2. The system water − poly(vinyl methyl ether). Macromolecules
30:410–416
141. Pyda M, Van Durme K, Wunderlich B, Van Mele B (2005) Heat capacity of poly(vinyl
methyl ether). J Polym Sci B 43:2141–2153
142. Vancoillie G, Frank D, Hoogenboom R (2014) Thermoresponsive poly(oligo ethylene glycol
acrylates). Prog Polym Sci 39:1074–1095
143. Wach RA, Mitomo H, Yoshii F, Kum T (2002) Hydrogel of radiation-induced cross-linked
hydroxypropylcellulose. Macromol Mater Eng 287:285–295
144. Xia X, Tang S, Lu X, Hu Z (2003) Formation and volume phase transition of hydroxypropyl
cellulose microgels in salt solution. Macromolecules 36:3695–3698
145. Cho EC, Lee J, Cho K (2003) Role of bound water and hydrophobic interaction in phase
transition of poly(N-isopropylacrylamide) aqueous solution. Macromolecules 36:9929–9934
146. Maeda Y, Kubota T, Yamauchi H, Nakaji T, Kitano H (2007) Hydration changes of poly(2-
(2-methoxyethoxy)ethyl methacrylate) during thermosensitive phase separation in water.
Langmuir 23:11259–11265
147. Deshmukht SA, Sankaranarayanan SKRS, Suthar K, Mancini DC (2012) Role of solvation
dynamics and local ordering of water in inducing conformational transitions in poly(N-
isopropylacrylamide) oligomers through the LCST. J Phys Chem B 116:2651–2663
148. Kozanecki M, Halagan K, Saramak J, Matyjaszewski K (2016) Diffusive properties of water
molecules in neighborhood of polymer chain as seen by Monte-Carlo simulations. Soft
Matter 12:5519–5528
149. Saramak J, Halagan K, Kozanecki M, Polanowski P (2014) Computational studies of
intermolecular interactions in aqueous solutions of poly (vinylmethylether). J Mol Model
20:2529
150. Loozen E, Van Durme K, Nies E, Van Mele B, Berghmans H (2006) The anomalous melting
behavior of water in aqueous PVME solutions. Polymer 47:7034–7042
151. Van Durme K, Loozen E, Nies E, Van Mele B (2005) Phase behavior of poly(vinyl methyl
ether) in deuterium oxide. Macromolecules 38:10234–10243
152. Meeussen F, Bauwens Y, Moerkerke R, Nies E, Berghmans H (2000) Molecular complex
formation in the system poly(vinyl methyl ether)/water. Polymer 41:3737–3743
153. Zhang J, Berge B, Meeussen F, Nies E, Berghmans H, Shen D (2003) influence of the
interactions in aqueous mixtures of poly(vinyl methyl ether) on the crystallization behavior
of water. Macromolecules 36:9145–9153
154. Nies E, Li T, Berghmans H, Heenan RK, King SM (2006) Supper critical solution
temperature phase behavior, composition fluctuations, and complex formation in poly (vinyl
methyl ether)/D2O solutions: small-angle neutron-scattering experiments and wertheim
lattice thermodynamic perturbation theory predictions. J Phys Chem B 110:5321–5329
155. Hidaka T, Sugihara S, Maeda Y (2013) Infrared spectroscopic study on LCST behavior of
poly(N,N-bis(2-methoxyethyl)acrylamide). Eur Polym J 49:675–681
156. Ide M, Maeda Y, Kitano H (1997) Effect of hydrophobicity of amino acids on the structure
of water. J Phys Chem B 101:7022–7026
157. Maeda Y, Higuchi T, Ikeda I (2000) Change in hydration state during the coil-globule
transition of aqueous solutions of poly(n-isopropylacrylamide) as evidenced by FTIR
spectroscopy. Langmuir 16:7503–7509
158. Maeda Y, Yamamoto H, Ikeda I (2004) Micro-raman spectroscopic investigation on the
phase separation of poly(vinyl methyl ether)/alcohol/water ternary mixtures. Langmuir
20:7339–7341
159. Maeda Y, Yamauchi H, Kubota T (2009) Confocal micro-Raman and infrared spectroscopic
study on the phase separation of aqueous poly(2-(2-methoxyethoxy)ethyl (meth)acrylate)
solutions. Langmuir 25:479–482
160. Maeda Y (2011) hydration of temperature-responsive polymers observed by ir spectroscopy.
Macromol Symp 303:63–70
161. Olejniczak MN, Kozanecki M, Saramak J, Matusiak M, Kadlubowski S, Matyjaszewski K
(2017) Raman spectroscopy study on influence of network architecture on hydration of poly
(2-(2-methoxyethoxy)ethyl methacrylate) hydrogels. J Raman Spectr 48:465–473
162. Pastorczak M, Kozanecki M, Ulanski J (2009) Water-polymer interactions in PVME
hydrogels—Raman spectroscopy studies. Polymer 50(19):4535–4542
163. Inomata H, Goto S, Otaka K, Saito S (1992) Effect of additives on phase transition of N-
isopropylacrylamide gels. Langmuir 8:687–690
164. Kokufuta E, Zhang YQ, Tanaka Y, Mamada A (1993) Effects of surfactants on the phase
transition of poly (N-isopropylacrylamide) gel. Macromolecules 26:1053–1059
165. Van Durme K, Rahier H, Van Mele B (2005) Influence of additives on the thermoresponsive
behavior of polymers in aqueous solution. Macromolecules 38:10155–10163
166. Suzuki Y, Suzuki N, Takasu Y, Nishio I (2007) A study on the structure of water in an
aqueous solution by solvent effect on a volume phase transition of N-isopropylacrylamide
gel and low-frequency Raman spectroscopy. J Chem Phys 107:5890–5897
167. Otake K, Inomata H, Konno M, Saito S (1990) Thermal-analysis of the volume phase
transition with N-isopropylacrylamide gels. Macromolecules 23:283–289
168. Yoon JA, Gayathri C, Gil RR, Kowalewski T, Matyjaszewski K (2010) Comparison of the
thermoresponsive deswelling kinetics of poly(2-(2-methoxyethoxy)ethyl methacrylate)
hydrogels prepared by ATRP and FRP. Macromolecules 43:4791–4797
169. Zhu PW, Napper DH (1996) Volume phase transitions of poly(N-isopropylacrylamide) latex
particles in mixed water-N, N-dimethylformamide solutions. Chem Phys Lett 256:51–56
170. Ouyang JF, Bettens RPA (2015) Modelling water: a lifetime enigma. Chimia 69:104–111
171. Chen M, Ko HY, Remsing RC, Calegari Andrade MF, Santra B, Sun Z, Selloni A, Car R,
Klein ML, Perdew JP, Wu X (2017) Ab initio theory and modeling of water. PNAS
114:10846–10851
172. Kotelyanskii M, Theodorou DN (2004) Simulation methods for polymers. Marcel Dekker,
New York
173. Kohn W, Sham LJ (1965) Self-consistent equations including exchange and correlation
effects. Phys Rev 140(4A):A1133–A1138
174. Becke AD (1993) Density-functional thermochemistry. III. The role of exact exchange.
J Chem Phys 98:5648–5652
175. Stephens PJ, Devlin FJ, Chabalowski CF, Frisch MJ (1994) Ab initio calculation of
vibrational absorption and circular dichroism spectra using density functional force fields.
J Phys Chem 98:11623–11627
176. Miertuš S, Scrocco E, Tomasi J (1981) Alectrostatic interaction of a solute with a continuum.
A direct utilization of ab initio molecular potentials for the prevision of solvent effects. Chem
Phys 55:117–129
177. Darden T, York D, Pedersen L (1993) Particle mesh Ewald: an Nlog(N) method for Ewald
sums in large systems. J Chem Phys 98:10089–10092
178. Hockney R, Eastwood J (1981) Computer simulation using particles. McGraw-Hill, New
York
179. Bayly CI, Cieplak P, Cornell W, Kollman PA (1993) A well-behaved electrostatic potential
based method using charge restraints for deriving atomic charges: the RESP model. J Phys
Chem 97:10269–10280
180. Cornell WD, Cieplak P, Bayly CI, Gould IR, Merz KM, Ferguson DM, Spellmeyer DC,
Fox T, Caldwell JW, Kollman PA (1995) A second generation force field for the simulation
of proteins, nucleic acids, and organic molecules. JACS 117:5179–5197
181. Hornak V, Abel R, Okur A, Strockbine B, Roitberg A, Simmerling C (2006) Comparison of
multiple Amber force fields and development of improved protein backbone parameters.
Proteins Struct Funct Genet 65:712–725
182. Duan Y, Wu C, Chowdhury S, Lee MC, Xiong GM, Zhang W, Yang R, Cieplak P, Luo R,
Lee T, Caldwell J, Wang JM, Kollman P (2003) A point-charge force field for molecular
mechanics simulations of proteins based on condensed-phase quantum mechanical
calculations. J Comput Chem 24:1999–2012
183. van Gunsteren WF, Berendsen HJC (1987) Groningen molecular simulation (GROMOS)
library manual. Biomos, Groningen, The Netherlands
184. van Gunsteren WF, Billeter S, Eising A, Hünenberger P, Krüger P, Mark A, Scott W,
Tironi I (1996) Biomolecular simulation: the Gromos 96 manual and user guide. vdf
Hochschulverlag AG an der ETH Zürich, Zürich, Switzerland
185. Jorgensen WL, Tirado-Rives J (1988) The OPLS [optimized potentials for liquid
simulations] potential functions for proteins, energy minimizations for crystals of cyclic
peptides and crambin. JACS 110:1657–1666
186. Vanommeslaeghe K, Hatcher E, Acharya C, Kundu S, Zhong S, Shim J, Darian E,
Guvench O, Lopes P, Vorobyov I, Mackerell AD (2010) CHARMM general force field: a
force field for drug-like molecules compatible with the CHARMM all-atom additive
biological force fields. J Comput Chem 31:671–690
187. Sun H (1994) Force field for computation of conformational energies, structures, and
vibrational frequencies of aromatic polyesters. J Comput Chem 15:752–768
188. Berendsen HJC, Grigera JR, Straatsma TP (1987) The missing term in effective pair
potentials. J Phys Chem 91:6269–6271
189. Jorgensen WL, Chandrasekhar J, Madura JD, Impey RW, Klein ML (1983) Comparison of
simple potential functions for simulating liquid water. J Chem Phys 79:926–935
190. Abascal JLF, Vega C (2005) A general purpose model for the condensed phases of water:
TIP4P/2005. J Chem Phys 123:234505
191. Warshel A, Levitt M (1976) Theoretical studies of enzymic reactions: dielectric, electrostatic
and steric stabilization of the carbonium ion in the reaction of lysozyme. J Mol Biol
103:227–249
192. Hoogerbrugge P, Koelman J (1992) Simulating microscopic hydrodynamic phenomena with
dissipative particle dynamics. Europhys Lett 19:155–160
193. Espanol P, Warren PB (2017) Perspective: dissipative particle dynamics. J Chem Phys
146:150901
194. Carmesin I, Kremer K (1988) The bond fluctuation method: a new effective algorithm for the
dynamics of polymers in all spatial dimensions. Macromolecules 21:2819–2823
195. Shaffer SJ (1994) Effects of chain topology on polymer dynamics: bulk melts. J Chem Phys
101:4205
196. Meyer KH (1940) Proprietes de polymeres en solution XVI. Interpretation statistique des
proprietes thermodynamiques de systemes binaires liquides. Helv Chirn Acta 23:1063–1070
197. Metropolis N, Rosenbluth AW, Rosenbluth MN, Teller AH, Teller E (1953) Equation of
state calculations by fast computing machines. J Chem Phys 21:1087–1092
198. Murat M, Witten TA (1990) Relaxation in bead-jump polymer simulations. Macromolecules
23:520–527
199. Lal M (1969) ‘Monte Carlo’ computer simulation of chain molecules. Mol Phys 1:
1757–1764
200. Olaj OF, Lantschbauer W (1982) Simulation of chain arrangement in bulk polymer, 1a,b).
Chain dimensions and distribution of the end‐to‐end distance. Macromol Chem Rapid
Commun
201. Pakula T (1987) Cooperative relaxations in condensed macromolecular systems. 1. A model
for computer simulation. Macromolecules 20:679–682
202. Pakula T (2000) Collective dynamics in simple supercooled and polymer liquids. J Mol Liq
86:109–121
203. Polanowski P, Jeszka JK (2007) Microphase separation in two-dimensional athermal
polymer solutions on a triangular lattice. Langmuir 23:8678–8680
204. Polanowski P, Jeszka JK, Li W, Matyjaszewski K (2011) Effect of dilution on branching and
gelation in living copolymerization of monomer and divinyl cross-linker: modeling using
dynamic lattice liquid model (DLL) and Flory-Stockmayer (FS) model. Polymer 52:
5092–5101
205. Polanowski P, Jeszka JK, Matyjaszewski K (2014) Synthesis of star polymers by “core-first”
one–pot method via ATRP: Monte Carlo simulations. Polymer 55:2552–2561
206. Polanowski P, Jeszka JK, Krysiak K, Matyjaszewski K (2015) Influence of intramolecular
crosslinking on gelation in living copolymerization of monomer and divinyl cross-linker.
Monte Carlo simulation studies. Polymer 79:171–178
207. Polanowski P, Sikorski A (2017) Comparison of different models of motion in a crowded
environment: a Monte Carlo study. Soft Matter 13:1693–1701
208. Aseyev V, Tenhu H, Winnik FM (2010) Non-ionic thermoresponsive polymers in water.
Adv Polym Sci 242:29–89
209. Longhi G, Lebon F, Abbate S, Fornili SL (2004) Molecular dynamics simulation of a model
oligomer for poly(N-isopropylamide) in water. Chem Phys Lett 386:123–127
210. Tamai Y, Tanaka H, Nakanishi K (1996) Molecular dynamics study of polymer–water
interaction in hydrogels. 1. Hydrogen-bond structure. Macromolecules 29:6750–6760
211. Wu R, Ji Q, Kong B, Yang X (2008) Molecular dynamics simulations of the hydration of
poly(vinyl methyl ether): hydrogen bonds and quasi-hydrogen bonds. Sci China Ser B Chem
51:736–742
212. Dalgakiran E, Tatlipinar H (2018) Atomistic insights on the LCST behavior of PMEO2MA
in water by molecular dynamics simulations. J Polym Sci B 56:429–441
213. Wu R, Qiu X, Yang X (2016) Molecular dynamics simulations of atomistic hydration

structures of poly(vinyl methyl ether). Chin J Polym Sci 34:1396–1410
214. Tavagnacco L, Zaccarelli E, Chiessi E (2018) On the molecular origin of the cooperative
coil-to-globule transition of poly(N-isopropylacrylamide) in water. Phys Chem Chem Phys
20:9997–10010
215. Walter J, Ermatchkov V, Vrabec J, Hasse H (2010) Molecular dynamics and experimental
study of conformation change of poly(N-isopropylacrylamide) hydrogels in water. Fluid
Phase Equilib 296:164–172
216. Bhandary M, Benkova Cordeiro MN, Singh JK (2016) Molecular dynamics study of wetting
behavior of grafted thermo-responsive PNIPAAm brushes. Soft Matter 12:3093–3102
217. Min SH, Kwak SK Kim BS (2015) Atomistic simulation for coil-to-globule transition of
poly(2-dimethylaminoethyl methacrylate). Soft Matter 11:2423–2433
218. Samsonova O, Glinca S, Biela A, Pfeiffer C, Dayyoub E, Sahin D, Klebe G, Kissel T (2013)
The use of isothermal titration calorimetry and molecular dynamics to show variability in
DNA transfection performance. Acta Biomater 9:4994–5002
219. Gangemi F, Longhi G, Abbate S, Lebon F, Cordone R, Ghilardi GP, Fornili SL (2008)
Molecular dynamics simulation of aqueous solutions of 26-unit segments of p(NIPAAm)
and of p(NIPAAm) “doped” with amino acid based comonomers. J Phys Chem B
112:11896–11906
220. Tamai Longhi Y, Tanaka H, Nakanishi K (1996) Molecular dynamics study of polymer–
water interaction in hydrogels. 2. Hydrogen-bond dynamics. Macromolecules 29:6761–6769
221. Caykara T, Kiper S, Demirel G (2006) Network parameters and volume phase transition
behavior of poly(N-isopropylacrylamide) hydrogels. J Appl Polym Sci 101:1756–1762
222. Tucker AK, Stevens MJ (2012) Study of the polymer length dependence of the single chain
transition temperature in syndiotactic poly(N-isopropylacrylamide) oligomers in water.
223. Shan J, Zhao Y, Granqvist N, Tenhu H (2009) Thermoresponsive properties of N-
isopropylacrylamide oligomer brushes grafted to gold nanoparticles: effects of molar mass
and gold core size. Macromolecules 42:2696–2701
224. Deshmukh SA, Kamath G, Suthar KJ, Mancini DC, Sankaranarayanan SKRS (2014)
Non-equilibrium effects evidenced by vibrational spectra during the coil-to-globule
transition in poly(N-isopropylacrylamide) subjected to an ultrafast heating–cooling cycle.
Soft Matter 10:1462–1480
225. Paradossi G, Chiessi E (2017) Tacticity-dependent interchain interactions of poly(N-
isopropylacrylamide) in water: toward the molecular dynamics simulation of a thermore-
sponsive microgel. Gels 3:13
226. Abbott LJ, Tucker AK, Stevens MJ (2015) Single chain structure of a poly(N-
isopropylacrylamide) surfactant in water. J Phys Chem B 119:3837–3845
227. Du H, Wickramasinghe R, Qian X (2010) Effects of salt on the lower critical solution
temperature of poly (N-isopropylacrylamide). J Phys Chem B 114:16594–16604
228. Han X, Feng J, Dong F, Zhang X, Liu H, Hu Y (2014) Thermo-/pH-responsive behaviours
of base-rich diblock polyampholytes in aqueous solution: experiment and simulation. Mol
Phys 112:2046–2057
229. Kai K, Diannan L, Zheng L (2012) Temperature-triggered protein adsorption and desorption
on temperature-responsive PNIPAAm-grafted-silica: molecular dynamics simulation and
experimental validation. Chin J Chem Eng 20:284–293
230. Lin H-C, Hsieh B-Z, Lin Y-L, Sheng Y-J, Lin J-J (2012) Effect of grafting architecture on
the surfactant-like behavior of clay-poly(NiPAAm) nanohybrids. J Coll Interface Sci
387:106–114
231. Wang L, Zhao X, Zhang Y, Zhang W, Ren T, Chen Z, Wang F, Yang H (2015) Fabrication
of intelligent poly(N-isopropylacrylamide)/silver nanoparticle composite films with dynamic
surface-enhanced Raman scattering effect. RSC Adv 5:40437–40443
232. Moghadam S, Larson RG (2017) Assessing the efficacy of poly(N-isopropylacrylamide) for

drug delivery applications using molecular dynamics simulations. Mol Pharmaceutics
14:478–491
233. Ko H, Javey A (2017) Smart actuators and adhesives for reconfigurable matter. Acc Chem
Res 50:691–702
234. Braunecker WA, Matyjaszewski K (2007) Controlled/living radical polymerization: features,
developments, and perspectives. Prog Polym Sci 32:93–146
235. Matyjaszewski K, Tsarevsky NV (2009) Nanostructured functional materials prepared by
atom transfer radical polymerization. Nat Chem 1:276–288
236. Matyjaszewski K (2012) Atom transfer radical polymerization (ATRP): current status and
future perspectives. Macromolecules 45:4015–4039
237. Matyjaszewski K, Tsarevsky NV (2014) Macromolecular engineering by atom transfer
radical polymerization. JACS 136:6513–6533
238. Pan X, Tasdelen MA, Laun J, Junkers T, Yagci Y, Matyjaszewski K (2016) Photoinduced
atom transfer radical polymerization with ppm-level Cu catalyst by visible light in aqueous
media. JACS 137(49):15430–15433
239. Pan X, Tasdelen MA, Laun J, Junkers T, Yagci Y, Matyjaszewski K (2016) Photomediated
controlled radical polymerization. Prog Polym Sci 62:73–125
240. Moad G, Rizzardo E, Thang SH (2005) Living radical polymerization by the RAFT process.
Aust J Chem 58:379–410
241. Zoppe JO, Ataman NC, Mocny P, Wang J, Moraes J, Klok H-A (2017) Surface-initiated
controlled radical polymerization: state-of-the-art, opportunities, and challenges in surface
and interface engineering with polymer brushes. Chem Rev 117:1105–1318
242. Warren NJ, Armes SP (2014) Polymerization-induced self-assembly of block copolymer
nano-objects via RAFT aqueous dispersion polymerization. JACS 136:10174–10185
243. Gröschel AH, Müller AHE (2015) Self-assembly concepts for multicompartment nanos-
tructures. Nanoscale 7:11841–11876
244. Han D, Lu Z, Chester SA, Lee H (2018) Micro 3D printing of a temperature-responsive
hydrogel using projection micro-stereolithography. Scientific Reports 8:1963
245. Kubelka J (2009) Time-resolved methods in biophysics. 9. Laser temperature-jump methods
for investigating biomolecular dynamics. Photochem Photobiol Sci 8:499–512
246. Causgrove TP, Dyer RB (2006) Nonequilibrium protein folding dynamics: laser-induced
pH-jump studies of the helix–coil transition. Chem Phys 323:2–10
247. Xu J, Zhu Z, Luo S, Wu C, Liu S (2006) First observation of two-stage collapsing kinetics of
a single synthetic polymer chain. Phys Rev Lett 96:027802
248. Pastorczak M, Okrasa L, Yoon JA, Kowalewski T, Matyjaszewski K (2017) Kinetics of the
temperature-induced volume phase transition in poly(2-(2-methoxyethoxy)ethyl methacry-
late) hydrogels of various topologies. Polymer 110:25–35
249. Bhattacharjee SM, Giacometti A, Maritan A (2013) Flory theory for polymers. J Phys
Condens Matter 25:503101
250. Fu J, Schlenoff JB (2016) Driving forces for oppositely charged polyion association in
aqueous solutions: enthalpic, entropic, but not electrostatic. JACS 138:980–990
251. von Neumann J (1947) The Mathematician. In: Heywood RB (ed) The works of the mind.
University of Chicago Press, Chicago
Chapter 9
Mössbauer Spectroscopy
of Magnetoelectric Perovskite Oxides
Paweł Stoch and Agata Stoch
Abstract Magnetoelectrics have been one of the most widely studied materials in
recent years. They are very interesting from the fundamental point of view due to
joining magnetic and electric orderings in the same phase, which tends to exclude
each other. On the other hand, the orderings are coupled each other what makes
them promising material to be applied in many electronic devices. This phe-
nomenon is strongly related to the structure and its changes which can be tested by
Mössbauer spectroscopy. In the chapter, we look closer to this technique in
application to magnetoelectric perovskite solid solutions of BiFeO3–Pb(Fe0.5Nb0.5)
O3. The possibility of confirmation of random cation distribution, magnetic
ordering temperature, and iron magnetic properties will be presented and discussed.
The presented experimental hyperfine interaction parameters will be compared to
those theoretically calculated using ab initio methods.
9.1 Introduction
In this chapter, we would like to present the possibility of application of Mössbauer

spectrometry in structural and magnetic studies of magnetoelectric compounds.
Magnetoelectric compounds are the materials which try to join in the same phase
two mutually excluding orders like magnetic and electric. This is a very interesting
problem from the scientific point of view. Additionally, these two orders can be
coupled each other what leads to the observation of a magnetoelectric effect. This is
the effect of the occurrence of electrical polarization due to the external magnetic
field or vice versa. This gives the possibility of a wide application of the materials
in modern electronic devices, e.g., magnetic field sensors, four-state electronics,
P. Stoch (&)
Faculty of Materials Science and Ceramics, AGH University of Science and Technology, 30
Mickiewicza Av., 30-059 Krakow, Poland
e-mail: pstoch@agh.edu.pl
A. Stoch
Institute of Electron Technology Krakow Division, Zablocie 39, 30-701 Krakow, Poland

274 P. Stoch and A. Stoch
magnetic storage of information, energy harvesters, antennas, transformers [1].

Unfortunately, the coupling and thus the effect is small what strongly reduce the
possibility of application of the materials. The magnetoelectricity is related to the
structure and its changes what additionally makes that they are really rare in nature.
In this chapter, we focus on an exemplary group of magnetoelectrics which are
oxides with a perovskite structure and especially on two examples like BiFeO3 and
Pb(Fe0.5Nb0.5)O3 in which coupling between magnetic and electric systems is
realized in a different way. However, in the beginning, we want to bring the reader
closer to 57Fe Mössbauer spectrometry and the hyperfine interactions.
9.2 Mössbauer Spectroscopy
Mössbauer spectroscopy is one of the methods which gives the possibility to look at
the structure of the tested material from inside, from the point of view of the tested
nuclei. The fundamental of the method is an observation of recoilless emission and
resonant absorption of c-radiation by atomic nuclei. The resonant nuclear absorp-
tion is possible to be observed only if the given system is provided the amount of
energy equal the difference between lower and higher energy nuclear levels. This
can be possible between two the same atomic nuclei of which the first one is in the
excited state and the second in the ground state. The nucleus during the transition
from the excited to the ground state emits c quantum of energy precisely equal to
the difference in energy between those two states. The emitted quantum of energy
can be absorbed by the second nucleus which is in the ground state. The nucleus
absorbs the energy and is transferred to the excited state. This is the nuclear res-
onance transition between excited and ground states in the two nuclei. The process
is possible only if the energy difference between the excited and ground state is the
same for emitting and absorbing nuclei, and in this way a resonance condition is
achieved. Unfortunately, during the c quantum emission, the nucleus undergoes a
recoil process. Based on simple energy conservation rule, the emitting energy is less
by the recoil energy. The similar but opposite process is observed during the
absorption. The absorbing nucleus undergoes the recoil process, and the transferred
energy must be increased by the recoil energy. Thus, taking into account these two
processes, the energy of the c quantum should be higher by exactly twice the recoil
energy. The most well-known and frequently used Mössbauer isotope is 57Fe for
which the c emission energy is 14.4 keV and the recoil energy is below
2 10−2 eV. Thus, to fulfill the resonance conditions in this case, the c quantum
energy should be increased by 4 10−2 eV. Although comparing these energies,
the difference is several orders, and the resonant condition cannot be observed. The
recoil energy is in inverse proportion to the mass of the system. If we were able to
somehow increase this mass, the recoil energy would be much smaller. Fortunately,
if we place the emitting and absorbing nuclei in a crystal lattice, there is a finite
probability that the momentum of the recoil accompanying the emission or
absorption of a photon is taken over by the crystal lattice. Thus, the mass in recoil
9 Mössbauer Spectroscopy of Magnetoelectric Perovskite Oxides 275
equation will be much higher, and the recoil energy will be transferred to the entire
crystal which is built of thousands of atoms and the recoil problem should disap-
pear. This process is called as Mössbauer effect. To observe resonant emission and
absorption of c-radiation, we should have the emitting and absorbing nuclei placed
in a solid material. The emitting energy should not be too high. The difference in
energy between the ground and excited state should be lower than the energy of the
crystal lattice phonon creation. Too high energy can make the nucleus vibrations
which also can affect the transition energy.
The Mössbauer effect is the fundamental of Mössbauer spectroscopy. In this
method, the radioactive source emits c radiation which lights the absorber and
behind which is placed a detector (in transmission variant of the method). If the
difference between ground and excited energy levels of the absorber is equal, the
emitted c energy then is absorbed by the measured nucleus which is transferred to
the excited state. The nucleus returns to the ground state with emission of the c
quantum but this time in the full solid angle. The source is placed in a vibrator
which by the Doppler effect tune slightly the emitted quantum of energy. If the
energy is higher or lower than the resonance condition, the radiation is not absorbed
and is detected by the counter. In the resonance, the strong absorption is evidenced.
The vibration of the source gives the possibility to tune up the emitting energy to
small changes in the nuclear levels positions between the source and the measured
absorber. In this way, the differences in the nuclear levels structure between the
source and the absorber can be detected.
The structure of the nuclear levels is affected by the bonding properties of the
neighbors in the crystal. The effect gives the unique possibility to draw conclusions
concerning the movement of cations in a crystal lattice, the influence of additive
atoms, chemical bond, and coordination properties, internal magnetic and electric
fields, and so on from the point of view of the tested nuclei. Thus, it gives the
possibility to observe structural properties from inside the selected nucleus.
The vast majority of the Mössbauer measurements are focused on iron. In this
case, the Mössbauer source is 57Co which decay to 57Fe by electron capture. The
half-life is approximately 271 days. After the decay, the 57Fe nuclei are in excited
state and shift to the ground state with the emission of a c quantum of energy
14.4 keV. The source of c radiation is one of the iron isotopes 57Fe which abun-
dance in natural iron is about 2%. It should be noted that in the method from all the
iron nuclei in the absorber, only 57Fe atoms are Mössbauer active and can be
measured.
The measured Mössbauer effect spectrum is characterized by a set of hyperfine
interaction parameters which describe the structure of the nuclear levels and its
changes. The spectrum is then deconvoluted and calculated the hyperfine interac-
tion parameters can be used to obtain information concerning local structure,
magnetic, electric properties, etc., of the studied material.
More details concerning Mössbauer spectroscopy one can find in some excellent
books like [2–6].
9.2.1 Hyperfine Interactions
The Mössbauer effect gives the possibility to resolve the hyperfine interactions
which are the interactions between a nuclear (moment) property and an appropriate
electronic or atomic property. The information of the interactions gives information
regarding electron- and spin-density distributions.
The total interaction Hamiltonian for the atom contains terms relating to inter-
actions between the nucleus and the electrons. The Hamiltonian can be written as:
^ ¼H
H ^ 0 þ E0 þ M1 þ E2 þ ð9:1Þ
where Ĥ0 represents all terms in the Hamiltonian for atom except the hyperfine
interactions; E0 is an electric monopole (Coulombic) interaction between the
nucleus and the electrons; M1 is a magnetic dipole hyperfine interaction; E2 is
electric quadrupole interaction. Higher terms are negligible [2].
There are three main hyperfine interactions [5]:
• electronic monopole interaction—isomer shift;
• magnetic dipole interaction—nuclear Zeeman effect;
• electric quadrupole interaction—quadrupole splitting.
9.2.1.1 Isomer Shift
The isomer shift (IS) also known as the chemical shift is an effect of differences in
electron densities at the nuclei site between an absorber and source. The differences
are the result of differences in the crystalochemical surrounding of the test nuclei.
The isomer shift value gives the information concerning the character of atomic
bonds, valence, and coordination of the studied atom. It is observed as a shift of the
Mössbauer absorption line relative to zero velocity of the relative motion.
The IS is the electronic monopole interaction and is the electrostatic interaction
of atom nuclei with electrons. It depends on the size of nucleus and the electron
density at the nuclei site. The larger the size of the atomic nucleus, the greater the
contribution of the interaction to the nuclear energy level position.
In the case of a spherical nucleus with a radius R, the change of the energy level
caused by the interaction with the electron charge is given by dependence [2, 7]:
2p 2
DE ¼ Ze jWð0Þj2 R2 ð9:2Þ
5
where Z is an atomic number and W(0) is wave function of the electrons at the
nucleus.
A nonzero value of the electron density at the nucleus which the size is
negligibly small compared to the size of the atom has only s-electrons. Thus,
distribution of only these electrons has an influence on the energy shift. Other
electrons (p, d, f, …) only play an indirect role changing by the so-called screening
effect the s-electrons density [2].
The energy shift of nuclear levels will be different for the ground state than for
the excited state due to the different radii of the nucleus in these states. The shift can
be expressed (in the non-relativistic) approximation as [7]:
dR
IS ¼ C jWA ð0Þj2 jWS ð0Þj2 ð9:3Þ
R
where C is a constant for a given isotope containing nuclear parameters, dR/R is the
relative change of nuclear radius between excited and ground states, and the term in
parenthesis represents the difference in the total electron density at the nucleus
between absorber and source. For the 57Fe, the nuclear radius of the excited state is
smaller than the radius of the ground state, so dR has a negative sign and with
increasing s electron density the resonance line shift is decreasing. Electrons in 1s,
2s, 3s, … shells all contribute to jWð0Þj2 but in decreasing amounts as the principal
quantum number rises. The inner shells are almost not affected by chemical bonding
and the principal influence on the isomer shift will by the outermost occupied
s-orbital [8].
A value of the isomer shift for a given material is given in relation to a substance
taken as the standard. In case of the 57Fe, the most frequently used standard is
metallic iron (a-Fe) at room temperature.
Measurement of the isomer shift provides information about the effective s-
electrons density in the nucleus. It allows you to draw conclusions about the nature
of chemical bonds. It can be correlated with valence, spin state, coordination
number, and so on of the tested nucleus.
A good example of pure isomer shifted spectrum can be iron in MgO matrix
(Fig. 9.1). The sample was prepared by melting pure MgO powder with small
addition of iron [9].
The obtained spectrum is composed of only one single line (singlet) which
indicates only one iron site in the crystal lattice. The shape of the line is given be
Lorentz function and its center is shifted toward higher energies (higher velocities)
due to isomer shift effect. The IS value is approximately 1 mm/s. The IS values in
the range of 0.1–0.5 mm/s are characteristic for ferric iron and in the range of 0.6–
1.7 mm/s for ferrous iron in the high spin state [2, 7]. So, the value around 1 mm/s
can be ascribed to iron in the high spin ferrous state.
Isomer shift is also sensitive to iron coordination number but in this case, the
interpretation of the isomer shift should be more careful. Generally, in simple
oxides and silicates Fe3+ in tetrahedral coordination adopt lower IS values (0.1–
0.35 mm/s) than in octahedral (0.35–0.5 mm/s) [2]. Similar behavior is character-
istic for ferrous iron ions, and this division is below 1 mm/s.
Fig. 9.1 Transmission 57Fe

Mössbauer effect spectrum of
Fe2+ in MgO matrix
9.2.1.2 Magnetic Hyperfine Field (Nuclear Zeeman Splitting)
The second important hyperfine interaction is the nuclear Zeeman splitting which
will occur if there is a magnetic field at the nucleus. A nucleus in a state with the
spin quantum number I > 0 will interact with a magnetic field by means of its
magnetic dipole moment. The nuclear Zeeman effect splits the nuclear state with the
spin quantum number I into 2I + 1 equally spaced nondegenerate states. For 57Fe,
the excited state with I = 3/2 is split into four magnetic states and the ground state
with I = 1/2 into two states. Taking into account the selection rules for magnetic
dipole transitions, there are six allowed transitions. The effect is observed as a
resonance sextet. The relative line intensities exhibit a 3:2:1:1:2:3 pattern.
The magnetic hyperfine splitting can be used to determine the effective magnetic
field at the nucleus which is a superposition of an applied external field and an
internal field. The internal field (Bint) can be given as [10]:
Bint ¼ BF þ BL þ Bd ð9:4Þ
where BF is the Fermi contact field and is a result of the interaction of the nucleus
with an imbalance in the s-electron spin density at the nucleus; BL is the contri-
bution from the orbital motion of valence electrons with the quantum number L for
the total orbital moment and Bd is spin-dipolar contribution arising from
non-spherical distribution of the electronic spin density.
The most important term is the Fermi contact field which is approximately 90%
of the total internal field. The term is a result of polarization effects on filled s-
orbitals. These can occur if the atom has unpaired electrons in d- or f-orbitals or if it
is chemically bonded to such an atom. The interaction of an unpaired d-electron
with the s-electrons of parallel spin will be different to that with the s-electrons of
opposed spin. The result is unbalanced of spin density at the nucleus [11].
Fig. 9.2 Transmission 57Fe

Mössbauer effect spectrum of
a-Fe
A good example of Zeeman splitted spectrum is metallic iron (a-Fe) which is the
most frequently used in Mössbauer spectroscopy as a spectrometer calibration
standard, and most of the IS values are given relative to this standard (Fig. 9.2).
The typical Zeeman splitted spectrum is a composition of six Lorentzian lines
called as a sextet. Based on the position of the lines, the magnetic hyperfine field
can be calculated, which in this case is 33 T.
The magnetic hyperfine field gives information concerning the magnetic moment
of the iron atoms in the material. In case of an isolating material, the magnetic
hyperfine field can be approximated as proportional to the iron magnetic moment
with a proportionality constant about 13 T/µB [12]. Thus, by measuring of the field,
the magnetic moments of iron cations at different crystallographic sites in the
non-conducting oxides can be estimated. In case of conducting materials which
possess metallic bonds, the magnetic hyperfine field is proportional not to the iron
magnetic moment but rather to the mean magnetic moment of the crystal site [2].
9.2.1.3 Quadrupole Splitting
Any nucleus with a spin quantum number greater than I > 1/2 has a non-spherical
charge distribution and the nucleus may have an electric quadrupole moment (eQ).
If at the nucleus is high enough electric field gradient (EFG), then due to the
interaction of eQ with EFG the nuclear state will be split into sublevels.
In case of 57Fe, the nucleus in the ground state has I = 1/2 and has no quad-
rupole moment and the nuclear state is not splitted. In the excited state, the nucleus
has I = 3/2 and has a quadrupole moment. Thus, the level is splitted into two
sublevels with the eigenvalues [2]:
1=2
1 1
EQ ¼ eQVzz 1 þ g2 ð9:5Þ
4 3
where ƞ is the asymmetry parameter given by:
Vxx Vyy
g¼ ð9:6Þ
Vzz
where Vxx, Vyy, Vzz are components of the EFG tensor.

There are two main sources of the EFG at the nucleus which are an asymmetry
of the atom electrons charge distribution or an asymmetry of the electric charges
around the atom induced by an asymmetry of charge distribution of neighboring
ions in the crystal lattice. The second contribution is usually much higher and
important. It gives the possibility to determine the geometrical or electrical sym-
metry of the crystal site occupied by the Mössbauer nucleus. If all neighboring ions
have the same charge distribution, then only disorder of geometric symmetry of the
ions arrangement (coordination polyhedral symmetry disorder) will produce the
EFG. If surrounding ions’ charges are different, then the asymmetry and the EFG
will be due to this effect rather than due to the coordination polyhedral asymmetry.
In the case when iron ions are in octahedral coordination and are surrounded by
the same ions of regular symmetry, then the EFG will be equal zero. If the only
small deviation of the geometric or charge symmetry will be in the material, then
the EFG will exist and the quadrupole splitted spectrum will be registered.
A ferric iron has an electron configuration 3d5 whereas ferrous has the config-
uration 3d6. Half-filled 3d shell of Fe3+ has spherical symmetry, and it does not
produce additional electric field gradient. Thus, quadrupole split for Fe3+ ions is
rather small up to about 1 mm/s. In case of the ferrous iron, additional 3d electron
produces additional electric field gradient and for Fe2+ ions, the quadrupole split
parameter is much higher than for Fe3+ (1.5–3.5 mm/s). The ferrous iron ions are a
very sensitive probe of the EFG and local symmetry of the site.
The unique example is the spectrum presented in the Fig. 9.1 where Fe2+ ions
substitute Mg2+ ions. The obtained quadrupole split parameter for this site is equal
to zero. In the MgO crystal lattice, Fe2+ ions are placed in the middle of the perfect
oxygen octahedra so no geometrical asymmetry is evidenced [9]. Taking into
account the high sensitivity of the Fe2+ as a local EFG probe, this is the special case.
9.2.1.4 Calculation of the Hyperfine Interaction Parameters
The Mössbauer spectroscopy is a very powerful tool but the interpretation of the
Mössbauer spectra is not straightforward. Modern ab initio methods can give the
possibility to new look at the effect and can be very helpful in the prediction of the
hyperfine interaction parameters. Theoretical support can gain deeper knowledge
about the relationship between the materials properties and its atomic and electronic
structure. State-of-the-art quantum chemical methods have been used to calculate
the parameters for last over 20 years. The parameters have been calculated using
coupled-cluster (CC) theory, density functional theory (DFT), full linearized plane
waves (FLAPW) method, Hartree–Fock (HF) and with the application of different
basis functions, exchange-correlation potentials, etc. [13]. While most of the
methods work well for the vast majority of solids, they may fail for so-called
correlated electrons in 3d transition metal oxides or 4f and 5f compounds.
Unfortunately, iron belongs often to this class of materials and prediction of its
electronic structure may not be so reliable. Except for the problem of treating the
electron–electron interaction, there is a problem which magnetic order should be
considered (paramagnetic, ferromagnetic, different antiferromagnetic or ferrimag-
netic configurations) or whether the relativistic effects should be included, e.g.,
orbital contribution in the magnetic hyperfine field [14–16].
One of the most frequently calculated hyperfine interaction parameter is the
isomer shift. For this, Eq. (9.3) can be used in the simplified following form [16]:

IS ¼ a qsample
0 qreference
0 ¼ aqsample
0 þb ð9:7Þ
where q0 is calculated total electron density and a is a proportionality constant that

can be estimated through calibration.
The q0 values are directly available from the calculations. The value of a
depends on many factors like calculation theory, basis functions,
exchange-correlation potentials, and the calculation details. Therefore, to predict the
isomer shift value of a calculated compound, the a should be determined using
exactly the same theory and calculation details. To do this, one should at first
choose the set of different iron-containing compounds of relatively simple crystal
structure parameters and well-defined experimental isomer shift values. For the
compounds, the q0 values need to be calculated and plotted as the function of the
experimental isomer shift values. To this point, a straight line should be fitted and
thus the proportionality constant can be obtained.
The exemplary set of compounds which are frequently used to obtain the cali-
bration constant is summarized in Table 9.1 [17]. The compounds are chosen to
cover a wide range of well-defined experimental values of the isomer shift
parameters. The crystal structure parameters for these compounds are also well
known and can be found in the most of the crystallographic databases.
The calculation of q0 values was conducted using FLAPW method implemented
in WIEN2k [18] package with default parameters for three different
exchange-correlation potentials: CA-LDA [19], PBE [20], and PBEsol [21]. The
last one is one of the most frequently used in calculations of the electronic band
structure of magnetic oxides. Comparing the results, one can see the dependence of
the obtained results on the used potential. For the PBEsol, the experimental shift
plotted versus the difference in electron density between given compound and a-Fe
for the calibration compounds is shown in the Fig. 9.3.
The calculated calibration constant is equal aPBEsol = −0.316(21) a.u.3 mm/s and
for the other data presented in Table 9.1 the values are aLDA = −0.305(24) a.u.3
Table 9.1 Isomer shift values and the calculated q0 values for different exchange-correlation
potentials for the calibration compounds
Compound ISexp (mm/s) qLDA
0 (a.u.−3) qLDA
0 (a.u.−3) qPBEsol
0 (a.u.−3)
TiFe –0.145 15,223.03976 15,310.50415 15,276.26512
C-Fe –0.100 15,222.73036 15,310.20858 15,275.97564
a-Fe 0.000 15,222.33304 15,309.80079 15,275.56425
Fe2O3 0.370 15,220.52611 15,308.07181 15,273.83556
FeF3 0.489 15,220.08541 15,307.60416 15,273.34467
FeCl3 0.530 15,220.18412 15,307.68599 15,273.44398
FeO 0.810 15,218.85291 15,306.29004 15,272.09560
FeI2 1.044 15,218.88889 15,306.39925 15,272.11893
FeTiO3 1.070 15,219.01978 15,306.43586 15,272.23207
FeCl2 1.092 15,218.76946 15,306.25103 15,272.00249
FeBr2 1.120 15,218.80125 15,306.30642 15,272.03606
FeF2 1.467 15,218.21831 15,305.73292 15,271.46350
Fig. 9.3 Experimental

isomer shift plotted versus the
difference in electron density
for various compounds
mm/s, aPBE = −0.310(25) a.u.3 mm/s. The similar values were also obtained by
Wdowik [17] (−0.291 a.u.3 mm/s) and Kurian [22] (–0.31 a.u.3 mm/s). However, the
isomer shift calibration constant values varied from −0.11 to −0.62 a.u.3 mm/s [17].
In a similar way, the quadrupole split based on ab initio calculations can be
estimated. The experimental quadrupole splitting parameter (QS) can be expressed
as [2]:

eQ 1 2 1=2
QS ¼ Vzz 1 þ g ð9:8Þ
2 3
The values of Vzz and ƞ (Eq. 9.6) parameters are the output of the calculation
1=2
software. Thus, one can easily calculate the effective field gradient Vzz 1 þ 13 g2
which versus the experimental quadrupole splitting creates a straight line. In this
way, value of the unknown nuclear quadrupole moment Q can be estimated. The
Q value varied from +0.082 barn to +0.016 barn depending on the calculation
method. For the similar as above set of compounds, the FLAPW methods predict
the Q = +0.17(1) barn [17].
It should be notated that errors for both of the above parameters are larger than
simple linear regression errors and they are about 10% due to various approxi-
mations used and uncertainty of the experimental results [16].
In magnetic materials, the magnetic hyperfine field can be also calculated.
According to the Eq. (9.4), the field is the sum of different terms. Usually domi-
nating term is the Fermi contact field BF. It can be obtained from the values of the
spin density at the nucleus and is given as [16]:
8p
BF ¼ q" ð0Þ q# ð0Þ ð9:9Þ
3
where q" ð0Þ and q# ð0Þ are spin-up and spin-down densities at the nucleus,
respectively, and can be calculated using ab initio methods. Unfortunately, in many
cases the contact term is underestimated by about 10% in DFT [16].
The BF term for metallic iron calculated above with application of the PBEsol
potential is 33.7 T which is in a very good agreement with the experimental field of
33.0 T at room temperature. Additionally, the calculations are conducted without
the influence of temperature so we can predict that the experimental hyperfine field
at 0 K should be higher. On the other hand, the field calculated for hematite (Fe2O3)
is approximately 40.0 T, whereas the experimental field at room temperature is
about 48.0 T. This gives about 20% error which could be related to underestimation
of the parameter by DFT and neglecting other than the Fermi contact terms.
9.3 Magnetoelectrics
Magnetoelectrics belong to a wider group of the multiferroic materials.

Multiferroics are the materials that exhibit more than one ferroic order parameter
simultaneously like ferromagnetism, ferroelectricity, ferroelasticity, or ferro-
toroidicity. This definition is now expanded to include also non-primary order
parameters such as antiferro- or ferri- [23, 24].
Magnetoelectrics are materials that exhibit in the same phase magnetic and
electric orderings and are characterized by the existence of a coupling between
magnetic and electric systems which is pronounced by a magnetoelectric effect. The
effect is an observation of electric polarization induced by the external magnetic
field or vice versa. The magnetoelectric effect was theoretically predicted by Curie
in 1894. The first successful observation was realized in Cr2O3 in 1961 [25].
Unfortunately, the effect in single-phase compounds is very small what makes it

difficult to register and limits its practical application. One of the candidates to
exhibit such phenomena are materials which are ferroelectrics with ferromagnetism.
Ferroelectricity is associated with a static charge distribution in a unit cell. Thus,
it needs the broken spatial inverse symmetry while the time reverse symmetry can
be invariant. It means that a spontaneous polarization would not appear unless a
structural distortion of the high-symmetry phase breaks the inversion symmetry.
Ferroelectric transition needs a change of symmetry in the system from a cen-
trosymmetric (paraelectric) phase to a non-centrosymmetric (ferroelectric) one.
Most conventional ferroelectrics which are technologically important are perovskite
structure (ABO3) oxides with the transition metal at B-site. These transition metal
ions should have an empty d-shell and have a formal configuration d0 like Ti4+,
Nb5+, Ta5+, W6+ [26–28].
Magnetism is a dynamic effect which is related to spin order and needs the
broken time-reversal symmetry while spatial inverse symmetry may be invariant.
Indeed, most of conventional magnetic materials have a centrosymmetric crystal
structure. Additionally, magnetism requires transition metal ions with partially filled
d- or f-shells. It is because the spins of electrons on the completely filled shells add
to zero and no magnetic moment is observed and in this way they do not participate
in magnetic ordering [27].
Based on symmetry consideration, magnetoelectrics should have broken spatial
and time-reversal symmetry. Thus, among all of the 233 Shubnikov magnetic point
groups, only 13 point groups fulfill the requirements and allow the possibility to
observe spontaneous magnetic and electric polarization. This is one of explanation
why multiferroics are rare in nature and there is practically no overlap between
magnetic and ferroelectric oxides. Even there are hundreds of magnetic and fer-
roelectric oxides, separately [27].
One of the possibilities to overcome those problems is to synthesize materials
which have two different sublattices like in BiFeO3 which is perovskite (ABO3)-
type oxide. The A-sites (Bi3+) favor the stability of the ferroelectric distorted
structure, and the sublattice is responsible for electric polarization. At the same
time, the B-sites contain magnetic (Fe3+) ions which provide magnetism. Thus,
ferroelectricity is induced by the ions at the A-sites, and the B-sites ions induce
magnetic order. In this way, the above exclusion rule of ferroelectricity and mag-
netism can be obeyed [26–28].
Another option is to mix both magnetic and ferroelectric active ions in the same
sublattice, e.g., Pb(Fe0.5Nb0.5)O3 where at the B-sites Nb5+ ions, which have empty
d-shell and are ferroelectric active, are partially substituted by Fe3+ ions, which
have partially filled d-shell and are magnetically active. In this system, Nb5+ sta-
bilizes off-centrosymmetric distortion of the regular perovskite phase while Fe3+
cations work in opposite way. Thus, the magnetic doping cannot be
over-concentrated. These systems are called independent. The dilution of the
electrical active system leads to a decrease of electric polarization and Curie tem-
perature. In this system, rather low electric and magnetic ordering temperatures are
observed. The coupling between magnetic and ferroelectric order in this kind of
magnetoelectrics is rather low because these two orders originate from different
kinds of ions [27, 29].
9.3.1 ABO3 Perovskites
Most of the conventional ferroelectric oxides have perovskite structure, and also
there are hundreds of magnetic oxide perovskite. Therefore, these materials were
one of the first choices to look for the unique ones which would join both of the
groups in the same crystal phase and temperature.
Perovskites are a class of compounds with a structure related to the mineral
Perovskite (CaTiO3) and can be characterized by general formula ABO3 where A is
usually a large cation and B is usually a medium-sized cation. The idealized per-
ovskite structure is cubic of Pm-3 m space group (No. 221). In the structure, large
A2+ cation occupies Wyckoff position 1a (0, 0, 0) at the corners of the cubic unit
cell. The B4+ ion lies at position 1b (0.5, 0.5, 0.5) at the center of the unit cell and is
surrounded by a regular octahedron of O2− ions at 3c (0.5, 0.5, 0) positions. The
large A2+ cations are coordinated to 12 O2− anions. Thus, they are surrounded by a
cuboctahedral cage of O2−. The smaller B4+ cations are placed in the middle of
[BO6] octahedra which create a regular framework. All of the B4+–O2− bond
lengths are equal, and the six O2−–B4+–O2− bonds are linear. Thus, each B4+ cation
has six neighbors in the B-sublattice. The ideal perovskite structure has no adjus-
table atomic position parameters, so any compositional change must be accom-
modated by a change in lattice parameter [30].
The ideal cubic perovskite structure is centrosymmetric what based on our
previous consideration about symmetry conditions in ferroelectrics and breaking the
spatial inverse symmetry leads to the conclusion that charge polarization in this
kind of materials cannot be observed. On the other hand, there are hundreds of
ferroelectrics in perovskite structure. There should exist a mechanism of cation
displacement which leads to structural distortion in order to obtain
non-centrosymmetric crystal lattice. This mechanism is the pseudo or second order
Jahn–Teller effect which is related to B4+ ions with formal empty d-shell (d0). The
effect is spontaneous symmetry breaking which is occurring in nondegenerate states
in the systems with sufficiently low-lying excited states. The driving force of this
distortion is an improvement of covalency. In result of this distortion, there is the
energy gain due to better covalence bonding between the atoms in the distorted
structure. To obtain a maximum degree of symmetry, cation displacement needs to
be realized along one of the symmetry axes of the octahedron. In case of ferro-
electric materials, it is realized very often by transformation from ideal perovskite
cubic (Pm-3 m) structure at temperatures above Curie (paraelectric state) to
tetragonal which is stable below Curie temperature (ferroelectric phase). The cubic
cell expands slightly along c-axis and is compressed along a- and b-axes and adopts
P4 mm (No. 99) space group. The change leads to off-center movement of the
octahedrally coordinated B4+ cation along c-axis. This elongation can be realized
not only in one axis direction but also in two axes direction which gives an
orthorhombic phase (space group Amm2, No. 38) or in three axes direction with a
rhombohedral cell of R3 m space group (No. 160) [30].
The most important magnetic species which can be incorporated into perovskite
structure are ions which have incompletely filled d- and f-shells. In case of tran-
sition metal cations, the magnetic moment is related to the electron spin quantum
number. At high temperatures the magnetic moments on these cations are disor-
dered and the material is in paramagnetic state. Below a certain temperature, the
moments start to interact what leads to parallel (ferromagnetic) or antiparallel
(antiferromagnetic, ferrimagnetic) alignment [31]. In magnetic oxides, the antifer-
romagnetic ordering is realized by a process called superexchange [32]. The su-
perexchange is the strong antiferromagnetic coupling between two cations through
non-magnetic anion. If the two next nearest magnetic cations are connected at 90°
to the non-magnetic anion, the interaction can lead to the parallel moment’s
alignment. If the cation–anion–cation bond angle is 180°, then the coupling is the
most frequently antiferromagnetic. Additionally, the Pauli Exclusion Principle
decides that between two magnetic ions with half-filled orbitals the coupling should
be antiferromagnetic, between half-filled and filled orbitals ions will be ferromag-
netic and between one ion which has half-occupied or filled orbital and the second
with an empty orbital the coupling may be antiferro- or ferromagnetic but the
second one is preferable. The theory of the superexchange interaction was devel-
oped by Anderson in 1950 [33] and then was reviewed by him in 1959 [34].
Briefly, a cation with a spin-up configuration interacts with a spin-down electron in
filled p-orbital of oxygen. The other spin-up electron in the orbital must then induce
spin-down configuration in the other cation. In case of the perovskite transition
metal ions are placed at B-sites, in the middle of oxygen octahedra and cation–
anion–cation geometry is linear, so the superexchange interaction favors the anti-
ferromagnetic alignment of the magnetic moments [35].
9.3.1.1 BiFeO3
One of the most important and widely studied magnetoelectrics is bismuth ferrite
BiFeO3 (BFO). This is one of only a few compounds which have electric polar-
ization and magnetic order at room temperature. The ferroelectric Curie temperature
is considerably high and is approximately TC = 1100 K. At room temperature,
bismuth ferrite is antiferromagnetic and is in this magnetically order state up to Néel
temperature TN = 643 K. This makes the material very interesting from an appli-
cation point of view [1, 36, 37].
Bismuth ferrite at room temperature adopts an R3c rhombohedral (no. 161) [38,
39] structure. The unit cell in the hexagonal coordinate system (Fig. 9.4a) is indeed
distorted perovskite-like crystal structure and belongs to trigonal perovskites, typ-
ified by the (3, 3) phase LaAlO3 at room temperature. The unit cell is the result of a
rotation of [FeO6] octahedra about a trigonal axis normal to a triangular octahedron
face compared to the cubic parent structure (Fig. 9.4d). This is trigonal crystal class,
while the axes are termed rhombohedral. The most frequently, for the convenience,
the unit cell is presented in a hexagonal coordinate. The ideal trigonal perovskites
belong to space group R-3c (167) with a center of symmetry. In case of space group
R3c, there is displacement of the B cations from the center of the octahedra which
leads to the loss of the symmetry center [30].
The main driving force of the structural distortions is a stereochemical activity of
the Bi3+ 6s2 lone pair, which hybridizes with 6p oxygen electrons. As the result of
the hybridization, the asymmetric wave function is achieved what leads to signif-
icant shortening of three of twelve Bi–O bonds and elongation of the opposite three
bonds. The displacement is realized along [1 1 1] direction of the ideal cubic
structure and is called as ferroelectric. On the other hand, the [FeO6] octahedron is
rigid and to accommodate the displacement it needs to be tilted. The tilt counteracts
a movement of the Fe3+ from the center of the octahedra and is called as antifer-
rodistortive. The combination of these two distortions lowers the total energy by
only 1 eV with respect to the cubic structure. As the consequence, BiFeO3 adopts
the R3c structure in which Bi3+ cation is shifted towered c-axis from the middle of
Fig. 9.4 a BiFeO3 unit cell (red arrows indicate direction the of electric polarization, black arrows
indicate the direction of Fe3+ magnetic moment) ; b G-type antiferromagnetic moments alignment
in BiFeO3; c Bi3+ and Fe3+ ions displacement in coordination polyhedral; d FeO6 octahedra tilting
and rotation. The figures were drawn by VESTA [40]
its coordination polyhedra (Fig. 9.4c). The [FeO6] tilt is not able to fully com-
pensate the ferroelectric distortion and Fe3+ cations are also slightly shifted along
the c-axis. The displacements of Bi3+ and Fe3+ cations are 0.613(1) and 0.212(1) Å,
respectively [39]. This clearly shows that the displacement of Bi3+ is about three
times larger than Fe3+ and because of the formal charges of both cations are the
same, the Bi3+ contributes to the total polarization about three times more. The total
unit cell polarization is along the c-axis.
Below Néel temperature, BFO has G-type antiferromagnetic moments alignment
(Fig. 9.4d). In this kind of alignment, each iron cation has as the nearest B-site iron
which magnetic moment is aligned opposite. Such behavior is a direct consequence
of the superexchange interaction. The Fe3+ cations of formal configuration 3d5 have
half-filled d-orbitals and they are separated by non-magnetic oxygen anions. In the
ideal perovskite structure, the Fe–O–Fe bond angle is 180°. Thus, a strongly anti-
ferromagnetic superexchange interaction between nearest neighbor Fe spins occurs.
In the ideal regular case, this interaction would lead to full compensation of magnetic
moments and total magnetic moment of the net should be zero. In fact, BFO has
spatially modulated magnetic structure of a cycloid type with a period of modulation
of about 62 nm and exhibits at room temperature a residual moment arising from a
canted spin structure [36, 41, 42]. Therefore, it can be concluded that there should be
another interaction which has a tendency for parallel alignment of the magnetic
moments which is much weaker than Fe–O–Fe superexchange interaction [43, 44].
This interaction is anisotropic exchange interaction due to the relativistic spin-orbit
coupling called Dzyaloshinskii–Moriya (DM) interaction [45–47]. When the Fe–O–
Fe angle is linear, the DM interaction is zero but in BFO the FeO6 octahedra are tilted
and it gives the possibility to turn on the interaction which is an order of magnitude
smaller than the superexchange. The DM interaction stabilizes spin canting and
induces an incommensurate magnetic structure with a spin cycloid propagating
along [1 1 0] with periodicity 62–78 nm. Additionally, the DM interaction can
stabilize the observed oxygen displacement which causes net electric polarization
what can be called as the inverse DM interaction [48]. Because of this, the interaction
can cause symmetry breaking rather than be caused by non-centrosymmetric bonds.
On the other hand, due to the inverse DM interaction, the opposite effect can take
place. Thus, electric polarization can induce magnetic polarization and vice versa
and magnetoelectric coupling can be achieved [28, 47, 49].
Bismuth ferrite can be the subject of Mössbauer effect measurements. One can
think that because of high iron concentration, the material is very good to measure
by this technique. Bismuth has a very high absorption coefficient of 14.4 keV
radiation and only very little resonant absorption could be obtained. The effect is a
very little what results in elongation time of measurement. Mössbauer spectroscopy
at first is used in checking the purity of the sample. This technique is very sensitive
to even small amounts of the secondary phase like unreacted Fe2O3 or Bi2Fe4O9
which is easily formed during the synthesis due to a strong evaporation of bismuth.
It was found that the magnetic hyperfine field extrapolated to 0 K is 54.6 T and
with increasing temperature is decreased. The temperature dependence of the field
which is proportional to iron magnetic moment fulfills the molecular field model
using the Brillouin function for S = 5/2. The isomer shift value at room temperature
is 0.38 mm/s [50] which is a value characteristic for high spin Fe3+. More precise
measurements have shown that although in BiFeO3 structure there is only one iron
position, the spectrum can be fitted with the application of two Zeeman sextets. This
means that there are two different iron positions which give two sets of hyperfine
interaction parameters. These two positions are characterized by the same values of
magnetic hyperfine field and isomer shift parameter but different values of the
quadruple split. The origin of this behavior is not clear up to now but it could be due
to a distribution in the direction of the magnetic moments relative to the crystal axis
caused by the canting spin structure of BFO [50].
9.3.1.2 Pb(Fe0.5Nb0.5)O3
Lead iron niobate Pb(Fe0.5Nb0.5)O3 (PFN) is another type of magnetoelectric

material which belongs to a group of the independent multiferroics system, in
which magnetoelectric properties are realized by mixing magnetically and ferro-
electric active cations in the same sublattice.
PFN belongs to relaxor ferroelectrics which are a large group of ferroelectric
perovskites with a general formula Pb(B1B2)O3 in which B1 can be Mg2+, Ni2+,
Zn2+, Fe3+, Sc3+, and B2 is Nb5+, Ta5+, W6+. All of them have mixed valence
cations at the B-sites. The relaxor ferroelectrics differ from the conventional fer-
roelectric in that the relative permittivity shows a wide diffuse peak which position
is frequency dependent on temperature [51], whereas non-relaxor ferroelectric are
characterized by a sharp transition of the relative permittivity at Curie temperature.
Contrary to the normal ferroelectrics, the crystal structure does not change signif-
icantly in the ferroelectric–paraelectric transition region. In case of PFN, this region
is in the temperature range of 370–380 K [52].
At room temperature, PFN adopts ferroelectric trigonal R3 m structure which
has a rhombohedral unit cell in which Fe3+ and Nb5+ atoms occupy 3a sites and are
distributed randomly. In hexagonal representation of the unit cell, the cell param-
eters are a = b = 5.6729 Å and c = 6.9493 Å [53, 54]. The spontaneous polar-
ization lies along the pseudocubic [1 1 1] direction. In this phase, there is no
octahedral tilt. On the other hand, it should be noted that according to some authors
the PFN unit cell at room temperature is rather monoclinic [29, 55, 56].
Magnetic properties of PFN are characterized by two diffuse transitions at 150 K
and around 10 K [54, 57, 58]. Below 150 K PFN is similarly like BFO a G-type
antiferromagnetic in which the magnetic moments of Fe3+ have aligned normally
the [1 1 1] direction [53, 57, 59]. Less clear is the origin of the low-temperature
magnetic anomaly which is believed to be a magnetic glass phase transition [52,
54]. In the unit cell, the Fe–O–Fe angle is 180° which leads to a strong superex-
change interaction. The magnetic properties are dependent on the number of Fe–O–
Fe linkages in a crystal lattice [32]. The number of the linkages can be governed by
changing the Nb5+/Fe3+ ratio or/and degree of Fe3+ and Nb5+ cations ordering.
However, no such ordering have been observed and Pb(Fe0.5Nb0.5)O3 could be
considered as a fully disordered in B-site sublattice of perovskite [41, 60]. On the

other hand, large scattering of magnetic ordering temperature could be explained by
local compositional ordering and formation of Nb-poor-Fe-rich and
Nb-rich-Fe-poor regions as postulated in [59, 61, 62].
We should also mention that in PFN system Pb2+ cations have also stereo-
chemical active 6 s2+ lone pairs which similarly like in BFO can stabilize the
non-centrosymmetric distortion of the lattice [12].
Room temperature Mössbauer spectrum is a doublet, which is a consequence of
the disorder in PFN which produces strongly inhomogeneous electric field. This is
related to substitution of Fe3+ by Nb5+ which differs considerably the formal
charges. The mean IS value is 0.41 mm/s, and mean QS is about 0.42 mm/s [63].
This implies that all iron is present as +3 in octahedral symmetry. The distribution
of quadrupole splitting at 300 K forms Gauss-like curve, indicating that arrange-
ment of Fe3+ cations is randomly distributed. At temperatures below Néel point,
Mössbauer effect spectra are composed of a number of Zeeman splitted sextets
which correspond to several iron environments. In PFN crystal structure, there is
only one inequivalent iron position. Therefore, the distribution of the magnetic
hyperfine field is due to differences in iron coordination related to disorder in
B-sites. Substitution of Fe3+ by Nb5+ in the nearest B-site surrounding influences
the superexchange interaction what in turn influences the iron magnetic moment.
Thus, the magnetic hyperfine field is disturbed. In this indirect way, the Mössbauer
spectroscopy can confirm the random distribution of cations in B-sites [12, 63]. On
the other hand, ab initio calculations of different configurations of cations in B-sites
of PFN showed the possibility of iron ions clustering in PFN [61].
9.3.1.3 Bi0.5Pb0.5(Fe0.75Nb0.25)O3
Bismuth ferrite easily forms solid solutions with different magnetoelectric and
ferroelectric perovskite oxides [64–71]. This is mostly done to tune electrical or
magnetic properties of the final material and also from the fundamental point of
view. One of the examples of such solid solutions can be 0.5BiFeO3–0.5Pb
(Fe0.5Nb0.5)O3. This is a mixture of two magnetoelectrics in which the multifer-
roicity is realized in two different ways. From one side is BiFeO3 where structural
distortion and the properties are governed by Bi3+ 6s2 lone pairs, ferroelectricity and
magnetism are related to different sublattices. On the other side is PFN were the
ferroelectricity and structural distortion is related to Nb5+ empty d-shell which
influence is weakened by substitution of magnetic Fe3+. In this case, the magnetism
and ferroelectricity are related to the same sublattice. In solid solution, there are
substitutions of Bi3+ by Pb2+ cations in A-sites. Both of the cations have 6 s2 lone
pairs which are stereochemically active and are able to stabilize the rhombohedral
distortion. The second substitution is realized in B-sites where magnetism related to
Fe3+ is diluted by Nb5+ cations which should increase the electric properties.
Bi0.5Pb0.5(Fe0.75Nb0.25)O3 crystal structure is rhombohedral R3c, the same as
BiFeO3 but the level of the distortion is smaller. It can be easily compared by a
rhombohedral angle which is 59.971° whereas in BiFeO3 the angle is 59.348° and
for cubic, in the rhombohedral coordinate system the angle is 60°. If we assume that
the distortion is only caused by the lone pairs, it means that the A-site substitutions
lead to a weakening of the lone pair mechanism. Similarly like in BiFeO3 in the
material, there are three Bi–O and Pb–O bonds which are considerably shorter than
the rest. Ab initio calculations confirm the highly covalent character of this shorter
bonds which is a driving force of the rhombohedral distortion. The Pb–O bond
covalency is lower than the Bi–O and in this way, the distortive force is reduced and
as a result, the distortion is smaller [12].
B-Site Disorder
Mössbauer spectroscopy can be used to check whether the arrangement of ions at

the B-site is random or ordered. In PFN, the answer to this problem is not so clear,
and there are measurements which confirmed both of the possibilities [41, 59, 60,
62]. In PFN, the B-sites are equally occupied by Nb5+ and Fe3+ cations. This system
is different; here the magnetic sublattice is only partially diluted by non-magnetic
ions and most (75%) of the ions are magnetic.
Mössbauer spectroscopy is sensitive to fine changes of the local environment
caused by the first and even second neighbors of tested iron atoms. In crystalline
compounds, every crystallographic iron site could be observed as one component of
the spectrum. In case of disordered systems, there are many different local iron
environments and every one of them is described by a different set of hyperfine
interaction parameters. As a consequence, the collected spectrum is a composition
of many individual components. In the studied solid solution, although iron could
be present only in one inequivalent site, the Fe3+/Nb5+ substitution lead to disorder
in the B-sublattice. Every change in the local B-site order should reflect the
observed effect and the hyperfine interaction parameters. In the case of disordered
systems, one method for processing the spectra is an Extended Voigtian-Based
analysis (xVBF) [72, 73]. In this method, every iron atom site is given by the
Gaussian distribution of independently isomer shift (IS) , quadrupole splitting (QS),
and hyperfine field (Bhf) due to some randomness in the local environment of iron.
In the considered system, there are at least two sources of the randomness, first
are local environment changes, but the second can be fluctuations of the iron
magnetic moments. So to check the local environmental disorder, one needs to do
the measurements in conditions in which magnetic moments will not change or the
changes will be negligible. This can be done by measuring the sample in a magnetic
saturation at a low temperature far away from magnetic ordering point. The Néel
temperature of Bi0.5Pb0.5(Fe0.75Nb0.25)O3 is about 460 K. So one can assume that at
liquid nitrogen (77 K), the sample is in a magnetic saturation. The 57Fe Mössbauer
effect spectrum (Fig. 9.5a) is Zeeman splitted sextet and at this temperature, all iron
ions are in the magnetically ordered state. Due to one iron site in the crystal
structure, only one component was used in which IS, QS, and Bhf parameters were
Fig. 9.5 a 57Fe Mössbauer effect spectra of the Bi0.5Pb0.5(Fe0.75Nb0.25)O3 at 77 K; b–d contour
maps of the probability distributions as a function of IS, QS, and Bhf
given by the Gaussian distributions, as an effect of randomness induced by Fe3+/

Nb5+ substitutions.
After application of the xVBF fitting procedure, one can obtain maps of the
different probability distributions as a function of the hyperfine interaction
parameters (Fig. 9.4b–d) and correlation parameters between the IS, QS, and Bhf. If
the correlation parameters are zero and the maps are circles, then it can be supposed
that in the studied material there is no special cation ordering. In the considered
material, the correlation parameters are close to zero and the maps are almost
circled, what supports the idea of the random cations arrangement in the B-sites of
the perovskite [12].
The xVBF fitting method gives only mean values of the hyperfine interaction
parameters. To get more information and additionally check the random cations
distribution in the B-sites, fitting model based on binomial distribution can be used.
The model is well developed and is known to work well in case of random alloys.
In the beginning, one can need to assume that area of each spectral component
(each Zeeman sextet) is proportional to the probability of given configuration of
cations in the nearest neighborhood.
The binomial distribution is given by:

n k
Pp ðk; n; pÞ ¼ p ð1 pÞnk ð9:8Þ
k
where Pp is probability to getting k successes in n trials and p is the probability of

success.
In ABO3 crystal structure, every iron cation has six other cations in the
B-sublattice in his coordination. The probability of finding iron in the coordination
is 0.75 and niobium 0.25 based on the stoichiometry of the considered material.
Thus, based on the binomial distribution, (Eq. 9.9) can be the calculated probability
of different Fe and Nb configurations in the iron coordination. From all the possible
configurations, we selected only those which probability exceeds about 5%. The
most probable configurations are (5, 1), (4, 2), (6, 0), (3, 3) where the first number
in the bracket represents the number of Fe and the second Nb ions in the coordi-
nation. The sum of the selected probabilities is 0.963, and this value is used to
renormalize them. It is assumed that the particular neighborhoods of iron atoms
produce the individual subspectra that contribute to the overall Mössbauer effect
pattern and areas of the subspectra should follow the probabilities. Therefore, the
obtained spectrum can be fitted using four Zeeman sextets of the Lorentzian line
shape. The calculated probabilities and the Bi0.5Pb0.5(Fe0.75Nb0.25)O3 hyperfine
interaction parameters are presented in Table 9.2.
We can easily find out that the probabilities of different configurations of the
measured iron neighbors in B-site follow very well area of the subspectral com-
ponents. This clearly confirms the random distribution of Fe3+ and Nb5+ cations in
B-sites. This kind of analysis gives the possibility to assign the specific configu-
ration set of the hyperfine interaction parameters. Thus, deeper conclusions con-
cerning local iron environment and properties can be drawn, and this will be
discussed further.
Magnetic Ordering Temperature
Mössbauer spectroscopy can be used to obtain magnetic ordering temperature. To

do this, at first, one should somehow estimate the range of searching the temper-
ature. In case of the Bi0.5Pb0.5(Fe0.75Nb0.25)O3, we can assume that the substitution
of Fe3+ by Nb5+ dilutes the magnetic sublattice and the magnetic ordering
Table 9.2 Hyperfine interaction parameters of the Bi0.5Pb0.5(Fe0.75Nb0.25)O3 at 77 K [12]

No (Fe, Nb) Pp Area (%) IS (mm/s) Bhf (T) QS (mm/s)
1. 5, 1 0.370 36.1(5) 0.453(5) 53.30(4) –0.013(6)
2. 4, 2 0.308 30.2(5) 0.517(8) 51.59(5) –0.010(7)
3. 6, 0 0.185 19.4(5) 0.631(10) 53.34(7) –0.029(10)
4. 3, 3 0.137 14.3(5) 0.532(10) 49.79(10) 0.030(14)
temperature will be probably lower than BiFeO3. An interesting observation was

done by Zachariasz et al. [71] that magnetic ordering temperature in multiferroic
solid solutions is linearly correlated with the magnetic hyperfine field in magnetic
saturation. This gives the possibility of rough estimation of the temperatures. Next,
the Mössbauer effect should be measured at temperatures higher and lower than the
magnetic ordering temperature. At temperatures above the magnetic ordering point,
the magnetic hyperfine field is zero and the characteristic Zeeman sextets transform
to simple quadrupole doublets (Fig. 9.6a).
The quadrupole doublets were fitted using only one quadrupole splitted com-
ponent. At the temperatures above the magnetic ordering, there are no magnetic
hyperfine fields and the doublets are narrow. Just below the ordering temperature,
the nonzero magnetic field appears. The field is so small that formed Zeeman sextet
is not well resolved. In fact, it looks like a wide doublet. By comparison, in
Fig. 9.6a are showed two quadrupole doublets above (the right panel) and below
magnetic ordering temperature (left panel). Dependence of the Mössbauer line
width as a function of temperature forms two straight lines (Fig. 9.6b). The
high-temperature and the low-temperature lines intersect in the magnetic ordering
temperature.
Mössbauer spectroscopy gives the possibility to determine the magnetic ordering
temperature very precisely even if the magnetic–paramagnetic transition is not
sharp. On the other hand, this is a local method which can give you information
rather about the beginning of the paramagnetic–magnetic transition. Therefore,
magnetic ordering temperatures determined using this method can be sometimes
higher in comparison with other less local methods.
Iron Magnetic Moments
Mössbauer spectroscopy can be used to get information concerning magnetic

properties of investigated materials. In this case, one of the most important
parameters is the magnetic hyperfine field. This field, among others, is mostly a
result of internal magnetic moments acting on the measured iron nuclei.
57
Fig. 9.6 a An exemplary high-temperature Fe Mössbauer effect spectra; b dependence of
Mössbauer line width on temperature
Looking at the Mössbauer spectrum at 77 K of the discussed magnetoelectric

(Fig. 9.5a), one can see that the spectrum is Zeeman splitted and all iron cations are
in a magnetically ordered state. The component line widths are narrow, and the
magnetic hyperfine fields are high (Table 9.2). This means that at the liquid
nitrogen, the sample is a magnetic saturation. The value of the fields about 50 T is
quite characteristic for oxides and ferric iron in octahedral coordination. Magnetic
hyperfine field depends on a number of Fe–O–Fe linkages due to strong exchange
interaction. It is known that the number of the linkages determines magnetic
properties of oxides [32, 58]. The substitution of Fe/Nb leads to decrease the
number of the linkages and increase the number Fe–O–Nb, and therefore, one can
expect that a magnetic moment at iron should also be reduced. In oxides, it can be
assumed that the magnetic moment at iron nucleus is proportional to the magnetic
hyperfine field [2]. According to the data summarized in Table 9.2, it can be seen
that increasing the number of niobium atoms in iron surrounding decreases the
magnetic hyperfine field, due to increasing the number of Fe–O–Nb linkages. Based
on the ab initio results, the calculated mean iron magnetic moment is approx. 4.0 µB
[12] and mean measured Bhf = 52.4 T. Assuming proportionality between these
two parameters, the proportionality constant is approx. 13.1 T/µB. Thus, depen-
dence of iron magnetic moment can be estimated from the number of Nb cations
surrounding iron in the B-sites. As one can see (Fig. 9.7b), substitution of one Fe
atom by Nb decreases very slightly the iron magnetic moment. The higher sub-
stitution decreases this moment almost linearly, and it can be predicted that when all
six nearest B-sites were substituted by Nb, some iron magnetic moment would
exist. Therefore, exchange interaction through Fe–O–Nb–O–Fe could not be
excluded [12]. At room temperature, the spectrum is quite different (Fig. 9.7a).
The highest magnetic hyperfine field is observed for a component for which six
Fe3+ cations are present in the B surrounding (blue curve). Assuming the value
13.1 T/µB of the proportionality constant, the magnetic moment of the site can be
estimated as 3.8 µB. Introducing one Nb5+ cation leads to a decrease of the Bhf to
42.3 T (red curve), what translates into a lower value of the iron magnetic moment
which is now 3.2 µB. Further substitution of iron in the B-sublattice results in the
Fig. 9.7 a 57Fe Mössbauer spectrum of the Bi0.5Pb0.5(Fe0.75Nb0.25)O3 at 300 K; b dependence of

the magnetic hyperfine field on the number of Nb5+ cations in the iron B-site neighborhood
continuation of reduction of exchange interaction and iron magnetic moment to

2.9 µB (magenta curve). For higher concentration of niobium in the iron sur-
rounding, iron ions become paramagnetic in room temperature (green curve) [74]. It
should also be noted that the reduction of magnetic moment is associated with the
reduction of superexchange interaction and magnetic ordering temperature. The
influence of temperature is visible in the half width of the spectral line. The
parameter is the highest for the Zeeman splitted component, for which two iron ions
in B-sites are substituted by niobium and predicted magnetic ordering temperature
is lower than in case of one Nb5+. The broadening of the line is most likely related
to the additional disorder caused by thermal fluctuations of iron magnetic moments.
This suggests that these two components are not in a magnetic saturation and are
approaching magnetic ordering point. The value of the parameter for the last
component is considerably lower, what suggests that it is still in the saturation and
the magnetic ordering temperature for this composition is higher.
It should also be mentioned here that the paramagnetic component is attributed
to the iron in which surrounding is more than two niobium cations. In such case,
magnetic ordering temperature may be below room temperature and can be close to
pure Pb(Fe0.5Nb0.5)O3, for which Néel temperature is about 150 K. The similar
doublet of the similar hyperfine interaction parameter values was observed previ-
ously in different solid solutions with BFO [75–77].
9.4 Conclusions
In this chapter, application and usefulness of Mössbauer spectroscopy in structural

studies of the magnetoelectric oxide solid solutions are presented. In the beginning,
a general introduction to the Mössbauer spectrometry has been given. The spec-
troscopy gives the possibility to determine hyperfine interaction parameters like
isomer shift, quadrupole splitting, and magnetic hyperfine field which are briefly
described. A proper description of the parameters is a very often non-trivial
problem. Application of different calculation methods can be very useful to solve it.
Short description of the possibility of the hyperfine interaction parameters deter-
mination based on ab initio calculations is also given. In the next part of the chapter,
problem of coexistence of magnetism and ferroelectricity is presented and structural
conditions of such coexistence are shown. On the selected example of BiFeO3–Pb
(Fe0.5Nb0.5)O3, step-by-step analysis of Mössbauer spectrum is shown. Basic
concepts of spectral analysis with application to the structure and magnetic prop-
erties are presented. The selected examples allowed demonstrating the usefulness of
57
Fe Mössbauer spectrometry in the understanding of chemical, structural, and
magnetic properties of multiferroic compounds.
References
1. Scott JF (2007) Data storage: multiferroic memories. Nat Mater 6:256–257

2. Greenwood NN, Gibb TC (1971) Mössbauer spectroscopy. Chapman and Hall Ltd., London
3. Dickson DPE, Berry FJ (1986) Mössbauer spectroscopy. Cambridge University Press,
Cambridge
4. Gütlich P, Bill E, Trautwein AX (2011) Basic physical concepts. Mössbauer spectroscopy and
transition metal chemistry: fundamentals and applications. Springer, Heidelberg, pp 7–24
5. Gonser U (1975) From a strange effect to Mössbauer spectroscopy. In: Gonser U
(ed) Mössbauer spectroscopy. Springer, Heidelberg, pp 1–51
6. Long GJ (1984) Basic concepts of Mössbauer spectroscopy. In: Long GJ (ed) Mössbauer
spectroscopy applied to inorganic chemistry. Springer, US, Boston, MA, pp 7–26
7. Gütlich P, Bill E, Trautwein AX (2011) Hyperfine interactions. Mössbauer spectroscopy and
transition metal chemistry: fundamentals and applications. Springer, Heidelberg, pp 73–135
8. Parish RV (1986) Mössbauer spectroscopy and the chemical bond. In: Dickson DPE, Berry FJ
(eds) Mössbauer spectroscopy. Cambridge University Press, Cambridge, pp 17–69
9. Szczerba J, Prorok R, Stoch P, Śniezek E, Jastrzebska I (2015) Position of Fe ions in MgO
crystalline structure. Nukleonika 60:143–145
10. Thomas MF, Johnson CE (1986) Mössbauer spectroscopy of magnetic solids. In:
Dickson DPE, Berry FJ (eds) Mössbauer spectroscopy. Cambridge University Press,
Cambridge, pp 143–197
11. Grant RW (1975) Mössbauer spectroscopy in magnetism characterization of
magnetically-ordered compounds. Springer, Heidelberg, pp 97–137
12. Stoch A, Maurin J, Kulawik J, Stoch P (2017) Structural properties of multiferroic
0.5BiFeO3–0.5Pb(Fe0.5Nb0.5)O3 solid solution. J Eur Ceram Soc 37:1467–1476
13. Pápai M, Vankó G (2013) On predicting Mössbauer parameters of iron-containing molecules
with density-functional theory. J Chem Theory Comput 9:5004–5020
14. Neese F, Petrenko T (2011) Quantum chemistry and Mössbauer spectroscopy. Mössbauer
spectroscopy and transition metal chemistry. Springer, Heidelberg, pp 137–199
15. Filatov M (2009) First principles calculation of Mössbauer isomer shift. Coord Chem Rev
253:594–605
16. Blaha P (2010) Calculations of Mössbauer parameters in solids by DFT bandstructure
calculations. J Phys Conf Ser 217:12009
17. Wdowik UD, Ruebenbauer K (2007) Calibration of the isomer shift for the 14.4-keV
transition in Fe57 using the full-potential linearized augmented plane-wave method. Phys
Rev B 76:155118
18. Blaha P, Schwarz K, Madsen GKH, Kvasnicka D, Luitz J (2001) WIEN2k, an augmented
plane wave + local orbitals program for calculating crystal properties. Karlheinz Schwarz,
Technische Universität Wien, Wien
19. Ceperley DM, Alder BJ (1980) Ground state of the electron gas by a stochastic method. Phys
Rev Lett 45:566–569
20. Perdew JP, Ernzerhof M, Burke K (1998) Rationale for mixing exact exchange with density
functional approximations. J Chem Phys 105:9982–9985
21. Perdew JP, Ruzsinszky A, Csonka GI, Vydrov OA, Scuseria GE, Constantin LA, Zhou X,
Burke K (2008) Restoring the density-gradient expansion for exchange in solids and surfaces.
Phys Rev Lett 100:136406
22. Kurian R, Filatov M (2010) Calibration of 57Fe isomer shift from ab initio calculations: can
theory and experiment reach an agreement? Phys Chem Chem Phys 12:2758–2762
23. Fiebig M (2005) Revival of the magnetoelectric effect. J Phys D Appl Phys 38:R123–R152
24. Spaldin NA, Fiebig M (2005) Materials science. The renaissance of magnetoelectric
multiferroics. Science 309:391–392
25. Astrov DN (1961) Magnetoelectric effect in chromium oxide. Sov Phys JETP 13:729–733
26. Spaldin NA, Pickett WE (2003) Computational design of multifunctional materials. J Solid
State Chem 176:615–632
27. Wang KF, Liu J-M, Ren ZF (2009) Multiferroicity: the coupling between magnetic and
polarization orders. Adv Phys 58:321–448
28. Park J-G, Le MD, Jeong J, Lee S (2014) Structure and spin dynamics of multiferroric BiFeO3.
J Phys Condens Matter 26:433202
29. Sim H, Peets DC, Lee S, Lee S, Kamiyama T, Ikeda K, Otomo T, Cheong S-W, Park J-G
(2014) High-resolution structure studies and magnetoelectric coupling of relaxor multiferroic
Pb(Fe0.5Nb0.5)O3. Phys Rev B 90:214438
30. Tilley RJD (2016) Perovskites: structure-property relationships. Wiley, Chichester
31. Coey JMD (2010) Magnetism and Magnetic Materials. Cambridge University Press,
Cambridge
32. Gilleo MA (1960) Superexchange interaction in ferrimagnetic garnets and spinels which
contain randomly incomplete linkages. J Phys Chem Solids 13:33–39
33. Anderson PW (1950) Antiferromagnetism theory of superexchange interaction. Phys Rev
79:350–356
34. Anderson PW (1963) Theory of magnetic exchange interactions: exchange in insulators and
semiconductors. Solid State Phys 14:99–214
35. Dionne GF (2009) Magnetic oxides. Springer, US, Boston
36. Sosnowska I, Loewenhaupt M, David WIF, Ibberson RM (1992) Investigation of the unusual
magnetic spiral arrangement in BiFeO3. Phys Rev B 180–181:117–118
37. Teague JR, Gerson R, James WJ (1970) Dielectric hysteresis in single crystal BiFeO3. Solid
State Comm 8:1073–1074
38. Kubel F, Schmid H (1990) Structure of a ferroelectric and ferroelastic monodomain crystal of
the perovskite BiFeO3. Acta Crystallogr B 46:698–702
39. Fujii K, Kato H, Omoto K, Yashima M, Chen J, Xing X (2013) Experimental visualization of
the Bi–O covalency in ferroelectric bismuth ferrite (BiFeO3) by synchrotron X-ray powder
diffraction analysis. Phys Chem Chem Phys 15:6779–6782
40. Momma K, Izumi F (2011) VESTA 3 for three-dimensional visualization of crystal,
volumetric and morphology data. J Appl Crystallogr 44:1272–1276
41. Zhang S-T, Zhang Y, Lu M-H, Du C-L, Chen Y-F, Liu Z-G, Zhu Y-Y, Ming N-B, Pan XQ
(2006) Substitution-induced phase transition and enhanced multiferroic properties of Bi1
−xLaxFeO3 ceramics. Appl Phys Lett 88:162901
42. Wang J, Neaton JB, Zheng H, Nagarajan V, Ogale SB, Liu B, Viehland D, Vaithyanathan V,
Schlom DG, Waghmare UV, Spaldin NA, Rabe KM, Wuttig M, Ramesh R (2003) Epitaxial
BiFeO3 multiferroic thin film heterostructures. Science 299:1719–1722
43. Ravindran P, Vidya R, Kjekshus A, Fjellvag H, Eriksson O (2006) Theoretical investigation
of magnetoelectric behavior in BiFeO3. Phys Rev B 74:224412
44. Sosnowska I, Schäfer W, Kockelmann W, Andersen KH, Troyanchuk IO (2002) Crystal
structure and spiral magnetic ordering of BiFeO3 doped with manganese. Appl Phys A
74:1040–1042
45. Dzyaloshinsky I (1958) A thermodynamic theory of “weak” ferromagnetism of antiferro-
magnetics. J Phys Chem Solids 4:241–255
46. Moriya T (1960) Anisotropic superexchange interaction and weak ferromagnetism. Phys Rev
120:91–98
47. Sergienko IA, Dagotto E (2006) Role of the Dzyaloshinskii-Moriya interaction in multiferroic
perovskites. Phys Rev B 73:94434
48. Cheong S-W, Mostovoy M (2007) Multiferroics: a magnetic twist for ferroelectricity. Nat
Mater 6:13–20
49. Arima T, Tokunaga A, Goto T, Kimura H, Noda Y, Tokura Y (2006) Collinear to spiral spin
transformation without changing the modulation wavelength upon ferroelectric transition in
Tb1−xDyxMnO3. Phys Rev Lett 96:97202
50. Blaauw C, van der Woude F (1973) Magnetic and structural properties of BiFeO3. J Phys C
Solid State Phys 6:1422–1431
51. Kulawik J, Szwagierczak D (2007) Dielectric properties of manganese and cobalt doped lead
iron tantalate ceramics. J Eur Ceram Soc 27:2281–2286
52. Levstik A, Filipic C, Holc J (2008) The magnetoelectric coefficients of Pb(Fe1/2Nb1/2)O3 and
0.8Pb(Fe1/2Nb1/2)O3−0.2Pb(Mg1/2W1/2)O3. J Appl Phys 103:66106
53. Raevski IP, Kubrin SP, Raevskaya SI, Stashenko VV, Sarychev DA, Malitskaya MA,
Zakharchenko IN, Smotrakov VG, Eremkin VV (2008) Dielectric and Mossbauer studies of
perovskite multiferroics. Ferroelectrics 373:121–126
54. Watanabe T, Kohn K (1989) Magnetoelectric effect and low temperature transition of Pb(Fe1/
2Nb1/2)O3 single crystal. Phase Transit 15:57–68
55. Bonny V, Bonin M, Sciau P, Schenk KJ, Chapuis G (1997) Phase transitions in disordered
lead iron niobate: X-ray and synchrotron radiation diffraction experiments. Solid State Comm
102:347–352
56. Singh SP, Yusuf SM, Yoon S, Baik S, Shin N, Pandey D (2010) Ferroic transitions in the
multiferroic (1 − x)Pb(Fe1/2Nb1/2)O3−xPbTiO3 system and its phase diagram. Acta Mater
58:5381–5392
57. Raevski IP, Kubrin SP, Raevskaya SI, Titov VV, Sarychev DA, Malitskaya MA,
Zakharchenko IN, Prosandeev SA (2009) Experimental evidence of the crucial role of
nonmagnetic Pb cations in the enhancement of the Néel temperature in perovskite Pb1
−xBaxFe1/2Nb1/2O3. Phys Rev B 80:24108
58. Bhat VV, Ramanujachary KV, Lofland SE, Umarji AM (2004) Tuning the multiferroic
properties of Pb(Fe1/2Nb1/2)O3 by cationic substitution. J Magn Magn Mater 280:221–226
59. Kleemann W, Shvartsman VV, Borisov P, Kania A (2010) Coexistence of antiferromagnetic
and spin cluster glass order in the magnetoelectric relaxor multiferroic PbFe0.5Nb0.5O3. Phys
Rev Lett 105:257202
60. Ivanov SA, Tellgren R, Rundlof H, Thomas NW, Ananta S (2000) Investigation of the
structure of the relaxor ferroelectric Pb(Fe1/2Nb1/2)O3 by neutron powder diffraction.
J Phys-Condens Matter 12:2393
61. Raevski IP, Kubrin SP, Raevskaya SI, Sarychev DA, Prosandeev SA, Malitskaya MA (2012)
Magnetic properties of PbFe1/2Nb1/2O3: Mössbauer spectroscopy and first-principles calcu-
lations. Phys Rev B 85:224412
62. Laguta VV, Rosa J, Jastrabik L, Blinc R, Cevc P, Zalar B, Remskar M, Raevskaya SI,
Raevski IP (2010) 93Nb NMR and Fe3+ EPR study of local magnetic properties of
magnetoelectric Pb(Fe1/2Nb1/2)O3. Mater Res Bull 45:1720–1727
63. Yang Y, Liu J-M, Huang HB, Zou WQ, Bao P, Liu ZG (2004) Magnetoelectric coupling in
ferroelectromagnet PbFe1=2Nb1=2O3 single crystals. Phys Rev B 70:132101
64. Sunder VS, Halliyal A, Umarji AM (1995) Investigation of tetragonal distortion in the
PbTiO3–BiFeO3 system by high-temperature X-ray diffraction. J Mater Res 10:1301–1306
65. Mahesh Kumar M, Srinivas A, Suryanarayanan SV, Bhimasankaram T (1998) Dielectric and
impedance studies on BiFeO3–BaTiO3 solid solutions. Phys Status Solidi A 165:317–326
66. Patel JP, Senyshyn A, Fuess H, Pandey D (2013) Evidence for weak ferromagnetism,
isostructural phase transition, and linear magnetoelectric coupling in the multiferroic
Bi0.8Pb0.2Fe0.9Nb0.1O3 solid solution. Phys Rev B 88:104108
67. Dadami ST, Matteppanavar S, Shivaraja I, Rayaprol S, Angadi B, Sahoo B (2016)
Investigation on structural, Mössbauer and ferroelectric properties of (1–x)PbFe0.5Nb0.5O3–(x)
BiFeO3 solid solution. J Magn Magn Mater 418:122–127
68. Stoch A, Zachariasz P, Stoch P, Kulawik J, Maurin J (2011) Structural and Mössbauer effect
studies of 0.5Bi0.95Dy0.05FeO3–0.5Pb(Fe2/3W1/3)O3 multiferroic. Acta Phys Pol A 119:59–61
69. Stoch A, Stoch P, Kulawik J, Zieliński P, Maurin J (2012) Magnetoelectric properties of
0.3Bi0.95Dy0.05FeO3–0.7Pb(Fe2/3W1/3)O3 multiferroic. Acta Phys Pol A 121:128–130
70. Stoch A, Maurin J, Stoch P, Kulawik J, Szwagierczak D (2017) Crystal structure and
Mössbauer effect in multiferroic 0.5BiFeO3–0.5Pb(Fe0.5Ta0.5)O3 solid solution. Nukleonika
62:177–181
71. Zachariasz P, Stoch A, Stoch P, Maurin J (2013) Hyperfine interactions in xBi0.95Dy0.05FeO3–
(1−x)Pb(Fe2/3W1/3)O3 multiferroics. Nukleonika 58:53–56
72. Lagarec K, Rancourt DG (1997) Extended Voigt-based analytic lineshape method for
determining N-dimensional correlated hyperfine parameter distributions in Mossbauer
spectroscopy. Nucl Instrum Meth B 129:266–280
73. Rancourt DG (1989) Accurate site populations from Mossbauer spectroscopy. Nucl Instrum
Meth B 44:199–210
74. Stoch A, Stoch P (2018) Magnetoelectric properties of 0.5BiFeO3–0.5Pb(Fe0.5Nb0.5)O3 solid
solution. Ceram Int 44:14136–14144
75. Kowal K, Jartych E, Guzdek P, Stoch P, Wodecka-Duś B, Lisińska-Czekaj A, Czekaj D
(2013) X-ray diffraction, Mössbauer spectroscopy, and magnetoelectric effect studies of
(BiFeO3)x–(BaTiO3)1−x solid solutions. Nukleonika 58:57–61
76. Park T-J, Papaefthymiou GC, Viescas AJ, Lee Y, Zhou H, Wong SS (2010)
Composition-dependent magnetic properties of BiFeO3–BaTiO3 solid solution nanostruc-
tures. Phys Rev B 82:24431
77. MacKenzie KJD, Dougherty T, Barrel J (2008) The electronic properties of complex oxides
of bismuth with the mullite structure. J Eur Ceram Soc 28:499–504
Chapter 10
Vibrational Spectroscopy of Zeolites
Theory Versus Experiment
Magdalena Król, Andrzej Koleżyński, Andrzej Mikuła

and Włodzimierz Mozgawa
Abstract In this chapter, the ab initio calculations have been used to analyze the
structural properties and vibrational spectra of selected zeolites. The spectra obtained
as a result of theoretical calculations along with their interpretation were used to
describe the experimental spectra of real zeolite structures. Presented results show
that in the experimental spectra of zeolites one can distinguish the bands associated
with characteristic vibrations of a bigger element of the structure, composed of
tetrahedra, the primary building blocks. It was also shown that the composite
envelopes of particular bands are significantly affected by component bands asso-
ciated with characteristic vibrations of building units that form zeolite structures.
10.1 Introduction
According to the classical definition of Smith [1], zeolites are 3D aluminosilicates

frameworks (tectosilicates) with free spaces and channels of strictly defined
molecular sizes in the structure, which can be occupied by cations and small
molecules with a high freedom of motion [2]. The above definition refers mainly to
natural zeolites, but it is worth to mention that among synthetic zeolites one can
also distinguish structures in which tetrahedral positions are occupied by other
cations like germanium or phosphorus (e.g., [3]).
Zeolites can be also defined as crystalline, inorganic spatial polymers consisting
of corner-shared tetrahedra [TO4]. Since in a part of tetrahedra silicon atoms are
substituted heterovalently by aluminum, the whole structure shows a positive charge
deficit and contains charge compensating extra-framework cations, mainly from
groups I and II of the periodic table (Na+, K+, Mg2+, Ca2+, Sr2+, Ba2+, etc.) [4, 5].
Both the Si/Al ratio and the content of extra-framework cations affect the properties
of zeolites [2], and so continuously conducted research on these valuable and
M. Król (&) A. Koleżyński A. Mikuła W. Mozgawa

e-mail: mkrol@agh.edu.pl

302 M. Król et al.
diverse physicochemical properties shows versatile possibilities of using these

minerals [4].
Among many experimental methods used to characterize the structure of zeo-
lites, spectroscopic methods such as vibrational spectroscopy (IR and Raman) and
nuclear magnetic resonance (NMR) spectroscopy (providing primary information
about the near and middle range ordering) occupy a special place [6, 7]. Presented
work is mainly devoted to the analysis of experimental vibrational spectra of
zeolites based on the results of quantum mechanical calculations. The detailed
interpretation of experimental spectra was based on the identification of bands
related to vibrations of characteristic structural elements of zeolites. The identifi-
cation of these bands provides information related to both, the framework structure
itself and the influence of the above-mentioned extra-framework cations introduced
into the structure on spectra envelope.
10.2 Systematics of Zeolite-Type Crystals
The tetrahedral structure of zeolites can be considered as a set of spatially connected

structural units—the so-called SBU (secondary building unit)—built of silicon and
aluminum tetrahedral ([SiO4] and [AlO4], respectively), i.e., PBU (primary building
units) [8]. SBUs can be a single or double rings of different sizes (e.g., S4R—single
4-ring, D6R—double 6-ring) or more complex tetrahedral units (e.g., 4=1 or 5–1
units). Figure 10.1 shows schematically several selected SBUs observed in zeolite
structures [9]. By linking together, these units form distinctive types of chambers
and channels, e.g., periodically reproducible CBUs (composite building units), such
as a-cage, b-cage (sodalite cage), or e-cage (cancrinite cage) [9]. Only taking into
account all three mentioned levels of hierarchy of construction of zeolite structure
gives a spatial framework that determines the properties of these materials [10].
Fig. 10.1 Examples of secondary building units (SBUs) occurring in zeolite structures: S4R (a),
S6R (b), S8R (c), D4R (d); and more complex fragment of LTA-type structure (e)
10 Vibrational Spectroscopy of Zeolites … 303
SBUs are the most characteristic part of the zeolite structure, and it is most
convenient to describe their framework in terms of such units [10]. SBUs are also
most commonly used as a criterion for designing zeolitic structure systematic and
assignment of particular zeolites into various structural groups. In commonly used
and recommended by IUPAC systematics based precisely on the SBU, seven basic
structural groups are distinguished: S4R, S6R, D4R, D6R, 5–1, 4–1, and 4=1–1. Of
course, such a division is a conventional one (and obviously not perfect), due to the
fact that in the vast majority of zeolite structures more than one SBU can be
distinguished (Fig. 10.1e). Nevertheless, the classification of zeolites based on
SBUs seems to be the most appropriate and is widely used in the literature devoted
to this subject [10, 11].
It is also worth noting that within each SBU group there may be different types
of frameworks, differing only in chemical composition, i.e., Si/Al ratio in the
framework or the content of extra-framework ions. Each such framework type,
included in given SBU group, is labeled with a three-letter code, e.g., LTA (zeolite
A, zeolite ZK-4), or FAU (zeolite X, zeolite Y).
10.3 Application of QM Methods in Interpretation

of Zeolite Spectra
As already mentioned, vibrational spectroscopy occupies an important place among

many research methods used to describe the zeolite frameworks [11–13]. However,
obtaining structural information based on such spectra is usually associated with a
rather cumbersome process of their interpretation. As indicated by Ermoshin et al.
[14] and Pechar and Rykl [15], difficulties arise, inter alia, when considering the
whole unit cell in vibrational analysis. The analysis of normal modes of vibration
for such large systems predicts several dozens or even several hundreds of IR or
Raman active modes in the vibrational spectra [16], while in the experimental IR
and Raman spectra the number of bands is much smaller. Among other things, for
this particular reason, the use of computational methods is now a very valuable tool
supplementing the interpretation process in vibrational spectroscopy.
There have been, and are, attempts to describe the vibrational spectra of zeolites
based on the assignment of bands to appropriate normal modes of vibration that use
interpretation methods developed for solids [17–19]. In such attempts [20], using
the normal mode analysis, vibration classification is made for the known structure
symmetry. For this purpose, for example, analytical or correlational methods are
used [21–23]. However, due to the usually very few information provided by the
spectrum, these methods are quite difficult to use experimental spectra interpreta-
tion. There is, therefore, a need to use more straightforward methods that would
allow the band assignments to be as precise as possible and bypass the presented
difficulties. Such methods include theoretical ones based on the formalism of
quantum mechanics, providing detailed information on the frequency and intensity
304 M. Król et al.
of normal modes of vibration in a given system and thanks to the ability to easily
visualize these vibrations, allowing for a simpler and much more accurate analysis
and interpretation of the vibrational spectrum.
The fundamental problem in theoretical calculations carried out for zeolite
structures is the necessity to apply far-reaching simplifications when creating the-
oretical models of such structures, resulting from the structural complexity of real
systems. Since experimental studies show the existence of well-defined structurally
hierarchical fragments of the zeolite framework, the first approaches using quantum
mechanics methods focused on the analysis of the smallest (and therefore simplest)
structure fragments, i.e., tetrahedra [TO4] and simple rings, for which geometry
optimization and vibrational spectra calculations were possible in an acceptable
time using available computer resources. Such calculations were usually performed
for simple cluster models terminated with protons, mono- or divalent cations, or
simple functional groups—such termination was necessary to stabilize the geometry
of the clusters, naturally unstable due to the excess negative charge of such frag-
ment “cut out” from the larger structure of the zeolitic framework. Such calculations
were made using available programs, initially employing Hartree–Fock or post-HF,
and later on more and more often also the density functional theory formalism, with
molecular orbitals defined in terms of localized atomic orbitals which in turn were
defined in terms of basis set functions using analytical Gaussian or Slater functions.
This allowed for the first, important results, especially for the mid-infrared range,
but nonetheless caused a number of problems related to the omission of significant
interactions with the rest of zeolite framework (and hence the lack of the possibility
to analyze far-infrared region with characteristic lattice vibrations) and artificial
impact of terminal groups on the properties of vibrational spectrum of given
structural units.
Only in recent years, thanks to both, the dynamic development of numerical
methods and an increase of available computing power, it has been possible to
perform calculations for more complex structural models, including translational
symmetry of real zeolite structures. As a result, a detailed analysis of the far-infrared
range, and thus the analysis of the influence of the type, amount and position of
non-tetrahedral cations on the structure and vibrational spectra of zeolites became
also possible. Moreover, periodic models allow the analysis of the impact on the
vibrational spectra of zeolites of the increasing complexity of structural units
(starting from the isolated tetrahedron, through double tetrahedra [T2O7] clusters,
SBU and CBU, up to the full frameworks), allowing verification of the degree of
correctness of the earlier analyses based on simple cluster models.
In the case of periodic models, classical quantum mechanics programs for solids
routinely used by physicists are normally used for vibrational properties’ calcula-
tions. Available programs allow performing calculations using Hartree–Fock,
post-HF, or DFT formalism (the latter for different exchange–correlation poten-
tials). Different programs use different types of crystal orbitals, based on classical,
localized atomic orbitals defined by means of analytic Gaussian (e.g., Crystal14) or
Slater (ADF) functions or in the form of numerical atomic orbitals (SIESTA,
DMol3), completely delocalized crystalline orbitals defined by means of plane
waves (VASP, CASTEP, Quantum ESPRESSO), or wavelets (DFT ++) or hybrid

orbitals (L)APW (WIEN2k, elk, FLEUR, EXCITING) and (L)MTO (TB-LMTO,
RSPt), and the calculations are carried out in full crystalline potential or for various
approximate pseudopotentials. The results obtained by various methods are to a
large extent comparable; hence, their selection depends mainly on individual
preferences or the availability of a given program. It is worth noting, however, that
when attempting to analyze the impact of the level of complexity of a structural
fragment (cluster) on the spectrum, it is necessary to perform calculations for all
model structures using the same formalism and identical parameters and in this case
the choice of crystal orbital representation is limited in practice only to methods
using localized atomic orbitals (for a more detailed description of the aforemen-
tioned problems related to practical theoretical calculations, see Chap. 1 of this
monograph).
Regardless of the computational methods used, the basic problem in studies of
vibrational properties of zeolites is the selection of the framework fragment that will
be as representative as possible for the whole structure. Such representative models
should be complex enough to carry information not only about the short-range,
local order, but also generally about the whole framework, and on the other hand,
simple enough so that the estimated calculation time, with a given computing
power, was acceptable. Using the hierarchical structure of the discussed group of
aluminosilicates, it is usual to model and interpret their spectra on several levels of
approximation, providing complementary interpretive data.
10.4 Vibrational Spectra of Silicates
Single tetrahedra [TO4] as well as proton-terminated oxygen bridges (which

undoubtedly constitute the simplest structural element of the tectosilicate frame-
works [24, 25] belong to the oldest models used to interpret experimental vibrational
spectra of silicates (Fig. 10.2). However, the use of this type of models for calcu-
lations is a very poor approximation, due to the size of such systems being too small
and devoid of the interactions with the structural environment of real structures. For
this reason, there are no attempts in the literature to interpret the spectra of zeolites
based on the models of PBU or oxygen bridges. On the other hand, it should be noted
that such models were used to determine the position of bands and interpret the
associated vibrations in other groups of silicates [26]. For example, Tossell [27]
carried out calculations for a single silicon tetrahedron (unit H4SiO4) using Hartree–
Fock method, while Kubicki and Sykes [28] for two connected tetrahedra (H6Si2O7)
system using DFT method, to determine the position of the vibration band ms
Si–O (being in the range of 730–710 cm−1 and 755–709 cm−1, respectively)
depending on the adopted angle Si–O–Si. Guided by this type of information, it is
possible to indirectly analyze more complex systems (bearing in mind the discrep-
ancy between the obtained results).
306 M. Król et al.
Fig. 10.2 Theoretical IR spectra of H4[SiO4], and H6[Si2O7] molecules
Figure 10.2 illustrates exemplary spectra of an isolated tetrahedron and a silicon

bridge (units H4[SiO4] and H7[Si2O7]). The sequence and arrangement of individual
bands in the presented theoretical spectra are consistent, but the intensity ratios of
individual bands are definitely disturbed, probably due to strong oscillations of
terminal –OH groups. It is worth noting that the two connected tetrahedra, i.e.,
systems composed of 15 atoms, have much more degrees of freedom than a single
tetrahedron and hence the spectrum of the silicon bridge is characterized by a much
larger number of bands, with nonzero integral intensity.
Returning, however, to the assessment of the suitability of PBUs for the inter-
pretation of real systems, one should agree with [29] that such models are definitely
not suitable for a detailed interpretation of zeolite spectra. The only practical
application of the silicon tetrahedron model is the definition and visualization of its
internal vibrations (Fig. 10.3). In the spectra of monosilicates, but also all com-
pounds containing silicon tetrahedron in their structure, the following bands with
associated vibrations are present:
Fig. 10.3 Internal vibrations

of [SiO4] tetrahedron: ms Si–
O− (a) and d O–Si–O (b)
Fig. 10.4 Visualization of atomic displacements during characteristic vibrations: RO vibrations in

S4R (a, b); maximum atomic displacements from the equilibrium positions during the PO-type
oscillation of the D4R unit (c)
• 1060–860 cm−1—antisymmetric stretching mas Si–O−;

• 850–820 cm−1—symmetric stretching ms Si–O−;
• 650–500 cm−1 and 440–400 cm−1—bending d O–Si–O.
Much more useful model is the presented silicon bridge (Fig. 10.2b). Comparing
to the spectra of monosilicates and the above-mentioned types of normal vibrations,
in disilicate spectra there are additional bands in the range 700–650 cm−1 coming
from bridge stretching vibrations [30]. A detailed analysis of the vibrations of this
system showed that it is possible to clearly distinguish the vibrations of the Si–O–Si
bridge from the vibrations of the O–Si–O bridge within the tetrahedron [31]. This is
confirmed by the difference in strength and charge distribution between the bridge
and terminal bonds [24, 29].
In cyclosilicates, another group of bands called “ring” bands appears in a similar
range of wave numbers (800–600 cm−1) [30, 32, 33]. Characteristic ring vibrations
—the so-called RO-type (ring opening) vibrations (Fig. 10.4a, b)—according to
definition of Bornhauser and Calzaferri [34] cause simultaneous in-phase stretching
of all T–O bridge bonds and/or bending of O–T–O bonds, symmetrically with
respect to the ring main axis. In the case of double rings (D4R and D6R units in
zeolite structures), PO (pore opening) spatial oscillations can be also identified
(Fig. 10.4c), which cause simultaneous, in-phase displacement of all atoms that
build both rings, with respect to center of symmetry of the unit [11]. Identification
of ring bands is particularly important in the case of zeolites, since the correct
determination of the position of a given band and assigning it to the vibrations of
the ring of a given multiplicity allows for preliminary identification of the type of
zeolite structure.
10.5 Spectrum Analysis Based on the SBUs
Since, on the one hand, the structure of zeolites cannot be described using only
PBUs, and on the other hand, it is difficult to use large framework fragments
constituting whole asymmetric unit cells (in zeolites, they are most often composed
of hundreds of atoms [35]), it begs the question whether in such a case SBUs cannot
308 M. Król et al.
be used as units made up of fewer atoms than unit cells and, at the same time,
representing the whole zeolite structure in the spectroscopic image. Such an
approach appeared in the literature [36–40] and allowed quite accurate interpreta-
tion of particular fragments of zeolite spectra.
10.5.1 SBU Terminated by Protons (Dependence

of Framework Type on Characteristic Vibration
Modes)
The literature data show that units made of tetrahedra terminated by protons can be
a good model for real crystalline structures [41, 42]. Attempts to interpret the
spectra of zeolites using model units were made, inter alia, by Mozgawa et al. [43],
and these were calculations carried out for isolated pseudomolecules composed of
single and double rings (SBU analogs) , terminated with protons for charge neu-
tralization. While this approach is relatively simple and does not disturb the initial
symmetry (bond angles) of the molecules (due to the very low mass of the proton
and the lack of unpaired electrons) [44, 45], it generates, however, quite a large
number of additional bands associated with vibrations of terminal oxygen with
“mobile” hydrogen (OH groups). As a result of the calculations, it was possible to
obtain, apart from the vibrational spectra of the units, also the animation of indi-
vidual normal modes of vibrations and determine the changes of the corresponding
internal coordinates (angles and/or bonds lengths) and in consequence precise
assignment of bands to a particular type of unit vibration.
On the basis of the proposed method of interpreting the vibrational spectra, it is
possible to describe the specific structure of the zeolite framework. By modeling
pseudomolecules that are equivalents of the SBU, one can recognize the charac-
teristic normal vibrations of these units and the corresponding bands in the spectra.
In Table 10.1, positions of PO- and RO-type bands identified in different zeolite
structures were collected.
Positions of bands associated with vibrations of individual rings are not fixed; in
different types of frameworks, individual rings are made of oxygen bridges that
differ in bonds angles, and thus the positions of individual bands in the spectra
change. It has been assumed that bands associated with vibrations of
low-multiplicity rings, such as S4R, occur at higher wave numbers, comparing to
the bands associated with vibrations of rings with higher multiplicities. It is also
worth noting that despite the occurrence of different ring systems in the structures of
zeolites, including 10- and 12-membered rings, in the literature, band assignments
to rings with multiplicity higher than 8 are not likely to be seen.
The identification of characteristic vibrations for individual structural groups has
allowed, in consequence, the use of vibrational spectra to determine the type of
zeolite framework. The verification of the use of the proposed models for the
interpretation of the vibrational spectra is the comparison of the obtained theoretical
Table 10.1 Characteristic bands in different types of zeolite structures with examples
SBU group IUPAC code Characteristic band position (cm−1) References
S4R GIS 712 (RO 4R), 620 (RO 6R) [46, 47]
S6R OFF, LEV 480–420 (RO S6R) [46, 47]
D4R LTA 557 (PO D4R); 466 and 337 (RO S4R) [36]
D6R FAU, CHA 630 (RO 6R) [36]
5–1 FER 520 (RO 5R); [31]
MOR 420 and 490 (RO 5R) [31]
4=1 NAT, THO 529 and 427 (RO 4R) Own materials
4–4=1 HEU, STI 600 (RO 4R); 520 (RO 5R) [48]
results with the measured experimental spectra of zeolites [36, 37, 43, 49].
Characteristic bands, for example zeolites with different structural groups, are
summarized in Table 10.1. Particularly important from the point of view of
structural analysis is the pseudolattice region of spectrum (800–500 cm−1), in
which characteristic vibrations of SBUs occur. Based on the sequence of positions
and the ratio of the intensities of integral bands present in this range, it is possible to
identify a given type of structure (provided that the tested material is a single
phase).
It is worth noting that in addition to RO and PO vibrations of single and double
4- and 6-membered rings, shown in Fig. 10.4, a “pseudoring” RO-type
5-membered rings in 5–1 (Fig. 10.5a) and 4–4=1 (Fig. 10.5b) [48] units were
also recognized.
10.5.2 Influence of Tetrahedral Substitution on the Spectra

Envelope
Based on model calculations for SBUs, in addition to the identification of char-

acteristic ring vibrations, it was also possible to determine the effect of the degree of
substitution of tetrahedral atoms on the spectra [43]. The analyzed SBUs differed
with silicon with aluminum substitution degree; however, according to the
Loewenstein rule [50], the Al/Si ratio did not exceed one. According to the
Fig. 10.5 An example of

characteristic ring-opening
(RO) vibrations of
5-membered ring in 5–1
(a) and 4–4=1 unit (b)
310 M. Król et al.
aforementioned rule, the Al–O–Al bridge bonds are not present in real alumi-
nosilicate structures; hence, the distribution of [AlO4] tetrahedra was assumed so
that they were not adjacent to each other.
Pseudomolecules constituting single rings have a symmetry markedly reduced in
relation to the one usually found for rings in real structures (they are clearly
deformed) [43]. The degree of deformation is greater in the case of 6-membered
rings and increases additionally with the increase in the number of [AlO4] tetra-
hedra introduced into the model. The analysis of vibrations in the case of units
Fig. 10.6 Calculated spectra of D4R units—dependence on the degree of Si with Al substitution
constructed of two connected 4-membered rings (unit D4R) is much more unam-
biguous, which is mainly due to the greater stiffness of the system. Compared to
single rings, they are much less deformed, and the connection of the rings with each
other stiffens the unit, which greatly facilitates the unequivocal identification of
characteristic vibrations. The interpretation scheme for this system is shown in
Fig. 10.6.
The effect of shifting most of the bands associated with particular vibrations
toward the lower wave numbers together with the increase in the substitution of Si
with Al is clearly visible. In the spectrum of aluminum-free pseudomolecules
(H8[Si8O20]), the band with the highest integral intensity (at 1114 cm−1) is derived
from the asymmetric stretching vibrations in Si–O–Si bridges. Substitution of one
[SiO4] tetrahedron by [AlO4] in a unit (H8[Si7AlO20]– spectrum) causes the band
associated with these vibrations to shift to lower wave numbers (1098 cm−1). At the
same time, the bands from asymmetric stretching vibrations of Si–O–Al at about
1120 cm−1 become clearly visible (in H8[Si7AlO20]– spectrum, it is visible only as a
slight inflection). The introduction of further aluminum tetrahedra (H8[Si5Al3O20]3–
pseudomolecule spectrum) causes the band associated with Si–O–Si vibrations to
change position to 1047 cm−1 and band related to mas Si–O–Al vibrations to appear
as a separate one at 1086 cm−1. Finally, in the spectrum of the H8[Si4Al4O20]4–
unit, in which only Si–O–Al bridges exist, the bands at 1050 and 958 cm−1 are
presented. Analysis of this type leads to the conclusion that the introduction of
aluminum in subsequent units modifies the bands associated with Si–O(T) vibra-
tions and simultaneously causes the appearance of additional bands not present in
the spectrum of pure silicon units. This is a significant difference to the experi-
mental spectra of aluminosilicates, in which the presence of [AlO4] tetrahedral does
not cause the appearance of additional bands originating from the asymmetric
stretching vibrations mas Si–O–Al, but only modifies the position and the envelope
of bands associated with Si–O–Si bridge vibrations [51]. The shifting of the mas Si–
O–(Si,Al) band toward the lower wave numbers together with the increasing alu-
minum content in the material framework is attributed to the longer Al–O bond
length compared to Si–O, and hence its lower force is constant [52].
In a similar way, the influence of aluminum substitutions on other types of
vibrations can be analyzed [46]. The tendency to change the position of the bands
with an increase in aluminum content in SBU also applies to the characteristic RO
vibrations, the positions of which are marked in Fig. 10.6. Similar observations
were made by Mozgawa [46] for calculations carried out for S4R, S6R, and D6R
units.
In summary, it can be concluded that isolated molecules are fairly good models
for describing vibrational spectra of 3D aluminosilicates, including zeolites, and the
calculated spectra can be used for the interpretation of experimental ones. It was
found that it is possible to determine the number and positions of bands originating
from RO and PO characteristic vibrations and that double rings are more rigid than
single ones and to a lesser extent deformable and therefore are a better model for
identifying characteristic vibrations, and the substitution with aluminum in tetra-
hedral positions reduces the frequency of SBU characteristic vibrations.
312 M. Król et al.
10.5.3 SBU Terminated by Cations
Infrared spectra are affected not only by the substitutions in tetrahedral positions but
also by the exchange of the so-called extra-framework cations. In a mid-IR region,
this influence is quite well known, thanks—among others—to the support of
experimental techniques of vibrational spectroscopy with computational methods
[36, 38] as well as other measurement methods [53, 54]. For example, Ma et al.
found [53] that the Na+ ! Ca2+ ion exchange in stilbite structure results in shifting
the bands associated with 4- and 5-membered rings’ characteristic vibrations by
about 20 cm−1 toward higher wave numbers. These changes are explained by the
position of individual cations relative to the STI framework (changes in unit cell
parameters) and the degree of hydration of these cations, which was confirmed on
the basis of XRD studies. Similar observations were made by Mozgawa et al. [48,
55, 56], who analyzed the process of ion exchange on natural clinoptilolite. In this
case, changes in the intensity of the band at about 675 cm−1, attributable to
4-membered ring vibrations in the 4–4=1 unit, were considered as an indicator of
the occurrence of ion exchange process.
In order to check how the presence of non-tetrahedral cations affects the spectra
envelope, including the change in the position of the bands associated with SBUs
characteristic vibrations, the theoretical computational methods have proved again
to be useful. As a model for the calculation, pseudomolecules being equivalents of
the SBUs were once again used, this time, however, terminated with metal cations
(K+, Na+, Ca2+, and Zn2+). Some controversy in this case may cause the level of
interaction of these cations with isolated molecules, especially in the case of
non-alkaline cations, since there is a possibility that small d block cations will
interact stronger with the isolated SBUs than with the same units embedded in
stiffening crystal lattice framework, which should translate into their significant
deformation, and thus change in spectrum envelope.
In contrast to earlier described SBUs terminated with protons, metal cations
change the distribution of electron density and thus change the nature of chemical
bonds within a given unit. This applies primarily to Si–O− terminal, but also to
bridge bonds. In contrast, the bonds between cations and terminal oxygen ions in
tetrahedra have already ionic character (much stronger than in the case of protons),
so the bonds become undirected; i.e., the individual cations do not form bonds with
a particular tetrahedron, but interact electrostatically with several surrounding
tetrahedra at the same time. This means that the proposed models will be the most
suitable for interpreting the vibrational spectra of cyclosilicates [57]; however,
attempts are also made to transfer the interpretation to tectosilicates.
For example, Fig. 10.7 shows the calculated spectra of D4R unit terminated with
sodium and potassium cations. Król et al. made an attempt [36] to use the obtained
results to interpret experimental spectra. It was found that vibrational spectroscopy
can be used in the study of sorption properties of zeolites and based on the model
spectra it is possible to confirm and analyze the changes in the experimental spectra
of zeolites due to the immobilization process of metal cations within their structure.
Fig. 10.7 IR spectra calculated for D4R model units terminated by Na+ and K+ cations [36]
Fig. 10.8 IR spectra calculated for 5–1 model units terminated by Na+ and K+ cations [58]
The Na+ and K+ cations have similar electronegativity (0.93 and 0.82, respectively,
on the Pauling scale); however, they differ in terms of ionic radius and atomic mass
and the observed differences in spectra can be associated with the mentioned
parameters. Similar observations were made by analyzing the D6R [36], 4–4=1
[48], and 5–1 (Fig. 10.8) [58] units.
314 M. Król et al.
10.5.4 Factors Affecting the Position of RO Vibration Band
In conclusion, it should be noted that the vibrational spectra of zeolites may indi-
rectly serve their phase identification, although sometimes it is a complicated
process. The position of the bands associated with the characteristic pseudolattice
vibrations is influenced by:
• Multiplicity of rings (its increase causes the presence of bands with ever lower
wavelengths);
• Al:Si ratio (its increase also reduces the wave number of bands);
• Type of vibration associated with the band;
• Type of extra-framework cations;
• The degree of deformation and the multitude of ring conformations;
• Structure ordering.
These factors cause that the ranges in which the characteristic bands for par-
ticular types of units occur may change and overlap, and determining the exact
limits of such ranges in the spectra for some groups is difficult.
10.6 SBU Versus Periodic Structure
Recent advances in computational techniques and increased availability of com-

puters with ever-increasing computing power allow constructing increasingly
complex and computationally demanding theoretical models that better describe the
actual structures and thus provide much more precise information on the properties
of zeolites. Despite the undoubted advantage of a very short calculation time, the
approach in which small fragments of the structure are used to interpret the whole
vibrational spectrum of a solid state materials makes it impossible to determine the
effect of long-range ordering (due to translational symmetry of crystal lattice) on the
spectrum [59], especially in far-infrared region, where bands associated with crystal
lattice and cationic vibrations should be expected.
Despite the aforementioned dynamic increase in the available computing power,
computational methods used routinely in ab initio calculations for periodic systems
are so far rarely used in the description of zeolitic structures, inter alia due to the
high complexity of these systems: a large number of symmetrically non-equivalent
atoms in the unit cell, common partial occupation of some of Wyckoff’s positions
in asymmetric unit cell by extra-framework cations, and water present in real
systems, e.g., in the form of hydration shells of non-tetrahedral cations. The few
reported results are based on significant approximations; e.g., Creighton et al. [60],
using the Wilson method modified for crystals calculated the theoretical vibrational
spectrum of pure silica form of sodalite (Si12O24), while Iyer and Singer [61, 62] in
their works devoted to crystalline lattices of zeolite A and sodalite stated that there
is no direct correlation between the individual structure elements and the properties
of the whole spectrum, and a significant portion of the described vibrations can be
presented as a linear combination of normal vibrations of 4- and 6-membered rings.
The literature on the subject lacks a more systematic description and comparison
of the theoretical vibrational spectra of all structural elements of zeolites in
accordance with their hierarchy. Therefore, the question arises—is the envelope of
the vibrational spectra of the individual elements of the zeolite structure an integral
part of it?
Figure 10.9 presents a list of unscaled spectra, calculated using the same theory
level for all structural units which can be specified in FAU-type periodic frame-
work: a single tetrahedron (H4[SiO4] molecule), two connected tetrahedra
(H7[Si2O7] molecule), S4R, S6R, and D6R units, and the spectrum calculated for
the FAU framework. Presented spectra are clearly different from each other, and
only the sequence of band occurrence is similar—from the highest wave numbers
successively, bands associated with: asymmetric stretching vibrations mas Si–O,
symmetric stretching vibrations ms Si–O, and bending vibrations d Si–O–Si. It is
worth noting that due to the presence of translational symmetry, the primitive unit
cell of the periodic lattice contains fewer atoms and hence less vibrational degrees
of freedom than clusters terminated with functional groups; hence in the spectra of
model units, there are much more bands coming from active IR vibrations com-
pared to spectra of the whole structure (here the spectrum described as FAU).
Taking the above into account, although the calculations carried out for SBUs
had an invaluable contribution to the interpretation of experimental spectra of
Fig. 10.9 Theoretical spectra of structural units of the FAU structure

316 M. Król et al.
Fig. 10.10 Experimental and theoretical IR spectra of LTA, FAU, CHA, MOR, FER, and
HEU-type structures
zeolites, an attempt was made to reinterpret the experimental spectra of zeolites

based on calculations carried out for periodic model structures. The theoretical
spectra in this case allow for a much more unambiguous (although for obvious
reasons, non-ideal) assignment of individual bands, among which one can specify
also RO bands [31]. The selected theoretical and experimental spectra with inter-
pretation are shown in Fig. 10.10.
Undoubtedly, the LTA structure, which is represented by zeolite A, belongs to
the best-known and described zeolite structures. Mozgawa et al. [36, 37] made
vibrational spectrum description based on calculations carried out for a 4-membered
double ring model (D4R unit)—all assignments agree with those obtained on the
basis of periodic model [63, 64]. Both in the spectrum of the D4R unit (Fig. 10.6)
and in the LTA structure (Fig. 10.10), a system of three bands in the range 700–
430 cm−1 is clearly present. One of them should be attributed to symmetrical
stretching vibrations ms Si–O–Si, and the other two at about 550 and 430 cm−1
(being in fact a superposition of several component bands) were assigned to
4-membered ring vibrations, including RO-type oscillations. The vibration descri-
bed as PO D4R, on the other hand, produces a band with a low integral intensity
included in the complex envelope of the spectrum with a maximum at about
550 cm−1.
A different structural group consists of zeolites built mainly from the D6R unit.
This group includes FAU structures (X and Y zeolites used in the catalysis and
natural faujasite) and CHA (one of the more common zeolites—chabazite). It is
worth noting that while the structure of the CHA type is built only from D6R units,
the S4R units can also be distinguished in the structure of the FAU type. The
visualization of D6R unit vibrations [37] allowed to distinguish in theoretical
spectrum the bands related to vibrations of the whole unit, i.e., PO D6R at
584 cm−1, as well as RO vibrations of 4- and 6-membered rings. Of particular
importance should be the band at 584 cm−1, which is reported [65] as a band related
to PO D6R vibrations. Meanwhile, in the calculated IR spectra of FAU- and CHA-
type periodic structures (Fig. 10.10), only the bands related to the vibration of
single 4- and 6-membered rings are observed. A small amount of infrared active
vibrations can be explained by the absence of aluminum and/or extra-framework
ions in the proposed periodic model.
However, when comparing the experimental and theoretical spectra, it can be
concluded that the spectra of the proposed models correspond very well to the
experimental spectra (although these models do not contain aluminum and
extra-framework ions). When analyzing the pseudolattice region, it should be noted
that in the range of 800–700 cm−1 there are bands, which until now were attributed
to ms Si–O–Si vibrations within 4-membered rings (RO S4R) [37]. Visualizations of
respective vibrations of FAU framework show that the bands present in this range
should be assigned to bridge vibrations (but these are not RO-type). Next, in both
experimental spectra (zeolite Y and chabazite) a band with a complex envelope at
about 630 cm−1 appears, which is assigned to RO vibrations of double 6-membered
rings. At lower wavelengths, bands appear at about 580 and 510 cm−1 in faujasite
spectrum and about 510 and 470 cm−1 in the chabazite spectrum, which by many
authors are attributed to PO vibrations of the D6R unit [52, 66, 67]. However, as
already mentioned, no PO-type D6R vibrations were identified on the basis of
aluminum-free periodic models, although the bands in the presented ranges should
actually be associated with vibrations of this structural element and more specifi-
cally with vibrations of 4- and 6-membered rings in D6R (RO S4R vibration).
The zeolite structures built of 5–1 units (i.e. the so-called pentasils) belong to the
most important ones due to their application in catalytic processes. In the literature,
however, only a few band assignments can be found for 5-membered ring vibra-
tions, e.g., at about 450 cm−1 in the experimental spectrum of stilbite [53] or about
675 cm−1 in the experimental spectrum of clinoptilolite (belonging to another group
4–4=1) [48]. However, these are not the results based on theoretical calculations.
The exception is the work of Król et al. [58], in which the authors identified bands
associated with vibrations of 5-membered rings at 560 and 524 cm−1; however, the
authors of the work generated the theoretical IR spectrum of the 5–1 unit which was
simply cut from the ferrierite structure and terminated with sodium and potassium
cations (but the geometry of the unit was not optimized so reported results should
be taken with a pinch of salt). Figure 10.10 shows spectra of pure silica structures
of MOR and FER types juxtaposed with experimental spectra of mordenite and
ferrierite, respectively. These theoretical spectra are characterized by a very good
318 M. Król et al.
agreement with the experimental counterparts, which can be explained by the fact
that both structures belong in fact to high-silica zeolites. In the experimental spectra
of pentasils, bands in 1400–1000 cm−1 range are associated with vibrations mas Si–
O–Si, bands in the range 900–450 cm−1 with vibrations ms Si–O, and bands
occurring below 450 cm−1 with vibrations d Si–O–Si.
Visualizations of vibrations related to bands in the pseudolattice range of fer-
rierite spectrum have shown that despite the presence of rings with higher multi-
plicities (6- and 10-membered) in ferrierite, bands in this range are associated with a
5-membered ring, except for bands at around 600 cm−1 that can be associated with
vibration of deformed (flattened) 6-membered rings. The characteristic RO S5R
vibration should be assigned to a maximum at approximately 470 cm−1, consisting
of several component bands.
The sequence of individual band arrangement in MOR spectrum is analogous to
the sequence of bands in FER spectrum, which confirms the belonging of both
structures to the same group. Differences appear only in the number and position of
characteristic RO-type bands. In mordenite structure, besides 5-membered rings
there are also 4- and 12-membered rings. As with ferrierite, MOR vibration visu-
alizations indicate that all pseudolattice bands are associated with 5-membered ring
vibrations, with the exception of a few bands that should be associated with
4-membered rings. The problem, however, is to identify the RO S5R band due to
the presence of 4-membered rings. The rigid arrangement of this ring causes that
during the RO S5R vibration one of the walls is almost “rigid,” which means that
one can only talk about “pseudoring” vibration. According to [31], the band
associated with this vibration is one of the components of the complex envelope of
the experimental spectrum with a maximum at about 450 cm−1.
The next and last structure discussed in this study is a HEU-type framework,
which represents the most common natural zeolite–clinoptilolite. Its structure can
be successfully reproduced based on 4–4=1 unit. In the theoretical spectrum of such
unit, band related to characteristic vibrations of S4R and S5R rings has been
identified [48]. These calculations were verified experimentally, assigning those
vibrations’ positions in the ranges 790–690 and 675–590 cm−1, respectively. These
results were confirmed in the latest calculations carried out for periodic model [31].
The analysis of the presented results indicates (Fig. 10.10) that the modeling of
periodic structures provides theoretical spectra reflecting significantly more accu-
rately the experimental ones than respective spectra of individual SBUs (Fig. 10.6).
In the case of FER-, MOR-, or HEU-type structures, the proposed aluminum-free-
framework models render the experimental spectrum to a very large extent, which
suggests that the presence of aluminum and/or extra-framework ions does not
introduce significant structural deformation. The same applies to structures built
with D6R, i.e., faujasite and chabazite. In the case of the LTA structure, the the-
oretical spectra are more consistent with experimental ones after considering the
aluminum atoms in tetrahedral positions and the presence of extra-framework ions
(results presented below), which suggests that in high alumina zeolites ion
exchange will have a significant impact on the position and shape of bands in
pseudolattice range.
Finally, it is worth noting that when analyzing different zeolite structures, an

attempt was also made to identify characteristic RO vibrations (fully symmetric) of
larger structural elements than SBUs—i.e., chambers (CBUs), confirming previous
evidence [36], that this type of vibrations was not identified for either CAN
structure (cancrinite chamber), LTA (a and b chambers) or FAU (b chamber). The
only identified vibration of this type concerns a hypothetical structure of the SOD
type, for which discussed vibration is active in Raman spectrum. The probable
cause here is the relatively high symmetry and atomic density of the discussed
structural fragment.
10.7 Influence of Extra-Framework Ions on Vibrational

Spectra
As already mentioned, introducing into zeolite framework extra ions (cations or

anions) causes specific structural changes that can be observed in spectrum
envelope changes, both in the mid-infrared (MIR) and in the far-infrared
(FIR) ranges. In the case of the mid-infrared, these changes mainly concern the
pseudolattice range of the spectrum and changes in the position or intensity of the
bands associated with characteristic vibrations. In the far-infrared range, changes in
the spectrum envelope are related to the interaction of individual ions with the
framework and therefore most often with the change in the nature of Me–O bonds
(vibrations of the cation sublattice), as well as the presence of bands associated with
lattice vibrations.
10.7.1 Ion Exchange Versus Spectrum
Analyzing the pseudolattice range of zeolite spectra, changes were found due to the
exchange of extra-framework cations [36, 56, 68]. This observation can help in
understanding one of the most important properties of zeolites, namely the ability to
exchange and retain foreign ions in their structure. The essence of this type of
research was to determine whether and what changes will be observed in the spectra
of selected structures under the influence of the exchange of extra-framework
elements. In the case of theoretical analysis of the influence of the presence of
extra-framework ions on the zeolite spectra, the calculations for which the SBU
models were used are inadequate. In order to thoroughly investigate the effect of
such ions on the structural properties and the shape of vibrational spectra, one
should take into account the influence of the presence of long-range order and the
position of particular cations/anions relative to coordinating them oxygen atoms.
The presence of extra-framework ions has a significant impact on charge distri-
bution between Si, Al, and O atoms, Si–O and Al–O force constants, as well as the
320 M. Król et al.
geometry of the entire zeolite framework. These subtle changes have a relatively
large impact on the position and shape of individual bands, both in the mid- and in
the far-infrared regions.
Accordingly, an attempt was made to use periodic models with a different
content of selected extra-framework ions (both cations and anions), to analyze and
interpret changes in the structure of zeolites due to the ion exchange process. This
approach also required appropriate modification of silicate framework through
substitution of part of the silicon atoms with aluminum in order to maintain charge
neutrality of model structure. The calculations were based on the SOD [69, 70] and
the LTA model structure [63, 64]. The initial positions of non-tetrahedral cations
were selected based on the literature data. The optimized structures were analyzed
in terms of structure (electron density topology, changes in lattice parameters,
changes of positions of extra-framework ions relative to the aluminosilicate
framework and framework deformation), and theoretical vibrational spectra were
simulated. Obtained results and calculated theoretical spectra were used to interpret
structural changes undergoing during sorption processes.
Figure 10.11 presents theoretical spectra of the optimized LTA structure with
various heavy metal ions [63]. Selected structural parameters of the single cationic
forms of LTA are collected in Table 10.2. The obtained results clearly indicate that
the tendency of cations to be coordinated by particular oxygen atoms and thus the
deformation of individual SBUs is strictly dependent on the type of position and
size of cations. The topological analysis of total electron density indicates that
cations can be coordinated by either all 6 or only by 3 bridge oxygen atoms
belonging to a 6-membered ring, as indicated by the number and position of
oxygen–cation bond critical points. The values of electron density qBCP and elec-
tron density Laplacian r2 qðrÞ (Table 10.2) for these bond critical points
unequivocally suggest that they are relatively weak chemical bonds with predom-
inantly ionic, closed-shell character. For comparison, all Si–O bonds are charac-
terized by significantly higher values of qBCP ¼ 0:140:15 and r2 q ¼ 1:041:09,
as well as a strongly negative value of the total electronic energy density He ½qðrÞ,
which confirms that they are strong, covalent–ionic bonds with predominantly ionic
character. This type of Si–O bond character agrees with the concept of Schwartz
and Berry [71]. The results of electron density topology analysis also confirm that
Al–O bonds are even more ionic in nature than Si–O (Table 10.2).
When assessing the effect of the cation type on the structure of zeolites, it should
be noted that cations with a small ionic radius such as Li+ occupy positions close to
the plane of the ring, whereby the oxygen atoms coordinating them move toward
the center of the ring. Thus, both 6- and 4-membered rings undergo the greatest
deformation, which translates into an increase in the number of bands in the IR
spectra (Fig. 10.11), best seen in mas Si–O–(Si,Al) vibration range. The calculated
spectra of the LTA structure with large ionic cations, in the range related to mas Si–O
(Si,Al) vibrations, are characterized by much smaller band splitting. Greater dis-
tance of cations with a large ionic radius from framework oxide atoms results in
much weaker interaction with the framework (Table 10.2), and thus the degree of
Fig. 10.11 IR spectra of LTA model structure calculated for various extra-framework cations [63]
deformation of individual elements of the structure is smaller. This affects the

apparent similarity of the K-, Rb-, and Cs-LTA spectra (Fig. 10.11). Me–O inter-
actions have little effect on the mid-infrared spectra [63]. The bands associated with
ring vibrations in this range agree on the position; however, due to the decrease in
local symmetry, their integral intensity was increased, especially when compared to
the spectra of models that did not include extra-framework ions (Fig. 10.10).
Significantly, greater differences between the different spectra are visible in the
far-infrared range. The band positions in this range are assigned depending on the
position of the ions in the structure [72, 73]. Recent observations [31] using
computational techniques indicate that absorption bands in the FIR range come
from several different types of vibrations:
• Tetrahedra rotations, without changing O–Si–O bond angle, but with a sec-
ondary change of the Si–O–Si bridge angle—these vibrations are called network
vibrations and occur in the range 200–0 cm−1 (Fig. 10.12a);
322
Table 10.2 Analysis of the electron density topology of various single-cation LTA forms [63]
Cation BCP k1 [e a.u.−5] k2 [e a.u.−5] k3 [e a.−5] qBCP ðrÞ [e a.−3] r2 qðrÞ [e a.−5] V[q(r)] [a.u.] He[q(r)] [a.u.] G[q(r)] [a.u.]
Li Si–O −0.292 −0.288 1.652 0.142 1.072 −0.369 −0.050 0.319
Al–O −0.180 −0.177 1.171 0.092 0.813 −0.193 0.005 0.198
LiI–O −0.036 −0.035 0.241 0.022 0.170 −0.026 0.008 0.034
LiII–O −0.039 −0.038 0.263 0.024 0.186 −0.029 0.009 0.038
Na Si–O −0.334 −0.331 1.712 0.154 1.046 −0.406 −0.072 0.334
Al–O −0.177 −0.176 1.167 0.092 0.814 −0.192 0.006 0.198
NaI–O −0.021 −0.021 0.146 0.017 0.104 −0.019 0.004 0.023
NaII–O −0.025 −0.024 0.166 0.019 0.118 −0.022 0.004 0.026
K Si–O −0.294 −0.291 1.677 0.144 1.092 −0.375 −0.051 0.324
Al–O −0.176 −0.174 1.154 0.092 0.803 −0.190 0.005 0.195
KI–O −0.011 −0.010 0.073 0.011 0.053 −0.010 0.002 0.012
KII–O −0.003 −0.003 0.027 0.004 0.020 −0.003 0.001 0.004
Rb Si–O −0.295 −0.292 1.680 0.144 1.093 −0.375 0.051 0.324
Al–O −0.177 −0.175 1.159 0.092 0.806 −0.191 0.005 0.196
RbI–O −0.006 −0.005 0.046 0.007 0.035 −0.006 0.001 0.007
RbII–O −0.009 −0.009 0.065 0.011 0.047 −0.009 0.001 0.010
Cs Si–O −0.336 −0.333 1.722 0.155 1.053 −0.410 −0.073 0.337
Al–O −0.175 −0.173 1.153 0.091 0.805 −0.190 0.005 0.195
CsI–O −0.006 −0.006 0.043 0.008 0.031 −0.006 0.001 0.007
CsII–O −0.012 −0.009 0.066 0.012 0.047 −0.010 0.001 0.011
k—Hessian eigenvalues, qBCP ðrÞ—electron density, r2 qðrÞ—Laplacian of electron density, V[q(r)]—potential energy density, G[q(r)]—kinetic energy
density, He[q(r)]—total electronic energy density
M. Król et al.
Fig. 10.12 Extreme atomic displacements during restrained tetrahedra rotations (a); cation–
oxygen vibrations (b); and vibrations of the cationic sublattice (c)
• Cation–oxygen vibrations, which usually accompany network vibrations in the

200–0 cm−1 range, but may also occur at higher wave numbers and take the
breathing character (Fig. 10.12b);
• Vibrations of the cationic and/or anionic sublattice that cannot be identified with
cation–oxygen oscillations (Fig. 10.12c);
• Bending vibrations d O–Si–O above 200 cm−1.
In the discussed range, the zeolite spectra differ significantly, which should be
explained by the influence of the chemical nature of extra-framework cations on
changes in geometry and thus topological properties of Si–O and Al–O bonds. It is
also worth noting that with increasing atomic mass, cations tend to oscillate with
ever lower wavelengths. Moreover, the higher the mass of the cation, the less
“mobile” it is; i.e., the number of lattice vibrations accompanied by oscillations of a
given cation is smaller (Fig. 10.11).
Figure 10.13 presents the experimental spectra of lithium, sodium, and potas-
sium form of zeolite A (Li-, Na-, and K-zeolite A, respectively), which were used
for the preliminary verification of the above-mentioned relationships. Significant
extension of the component bands in the mas Si–O(Si,Al) vibration range in
Li-zeolite A spectrum shows very high agreement with the theoretical spectrum and
indirectly confirms that small cations have the greatest influence on the deformation
of the LTA structure. This is a direct consequence of the position of the individual
cations relative to the 6-membered ring, since smaller cations tend to occupy a
position close to or even in the plane of the ring, often significantly deforming this
part of the structure. A similar effect was observed analyzing the LTA structure
model with introduced magnesium or zinc ions [64]. It can therefore be concluded
that the deformation of structural elements translates into changes in the
mid-infrared spectra (mainly the change in the position of mas Si–O(Si,Al) bands), as
well as changes in the number and position of bands originating from ring
vibrations.
According to the conclusions drawn on the basis of theoretical analyses
(Fig. 10.11), the dependence of the position of bands in far-infrared region
(FIR) originating from Me–O vibrations on cations’ atomic masses is preserved,
i.e., the bigger mass, the lower values of the wave numbers (223, 174, and
147 cm−1 for Li-, Na-, and K-form, respectively; Fig. 10.13).
324 M. Król et al.
Fig. 10.13 Experimental FT-IR spectra of Li-, Na-, and K-form of zeolite A [63]
As already mentioned, alkali metal cations as well as alkaline earth metal cations
have little effect on the spectrum envelope in MIR range [63]. The situation is
slightly different with d block elements. While large alkali metal or alkaline earth
metal cations (K+, Sr2+, or Ba2+) have a tendency to locate in the LTA structure far
from the ring plane (position A/E according to Mortier [74]) and thus to evenly
interact with each oxygen in neighboring ring, which results in the appearance of
bands from mas Si–O–(Si,Al) vibrations in similar positions, cations with a small
radius and d electron cations (Mg2+, Ca2+, Zn2+, Cd2+, or Hg2+) tend to locate
within or near S6R ring (position B according to [74]), which causes a strong
deformation of this SBU and a reduction of the unit cell parameter. This results in a
splitting and strong extension of the mentioned bands in MIR spectrum. In addition,
d block elements, due to the presence of a non-spherical symmetrical potential with
the value increasing with the decrease of the ionic radius and the tendency to form
strong directional bonds, more strongly attract some of the oxygen ions in neigh-
boring ring, which promotes further structural deformation.
The different natures of interactions between the zeolite framework and the
transition metal cations as compared to the interaction of this framework with alkali
metal cations are evidenced by the reversibility of ion exchange processes. While
the mechanism of binding of the latter to the framework can be described as
reversible physical adsorption, the results presented, for example, in the works [56,
58, 68] show that some heavy metal ions, such as Ag+ or Cr3+, connect with the
zeolite framework permanently and irreversibly (Fig. 10.14).
Fig. 10.14 Contribution of ion exchange in total sorption process on zeolites
The exchange of cations in the structure of zeolites, in contrast to chemisorption,

causes an increase in the intensity of the band, for example, at about 675 cm−1 in
the spectrum of clinoptilolite [47, 48], bands at 560 cm−1 in mordenite spectrum
[58], or the bands at 704 cm−1 in zeolite Y spectrum [36]. The greatest similarity
with the spectrum of the sodium form is observed in Cd- and Pb-zeolite spectra
(Fig. 10.15). In both cases, the dominating mechanism in the sorption process is ion
exchange (Fig. 10.14). On the other hand, the introduction of Cr3+ ions in the
zeolite structure causes relatively small changes in band intensity comparing to the
hydrogen form [36, 58]. The mentioned bands should therefore be considered as a
kind of indicator, based on which one can infer about the occurrence of an ion
exchange process, which leads to the change in surroundings of rings resulting in
their deformation, which in turn affects the change of the intensity of respective ring
bands. Regardless of whether SBUs or periodic structure models are used for
calculations, the positions of respective ring bands in the experimental spectra of
zeolites may differ from the positions in theoretical spectra. However, it can be
safely assumed that the exchange of extra-framework cations will cause changes in
the intensity and position of these bands in the IR spectra.
Fig. 10.15 Fragments of MIR spectra of clinoptilolite (a) zeolite Y (b) and mordenite (c) after the
sorption of heavy metal cations
326 M. Król et al.
In far-IR range of the spectrum, the changes caused by the presence of a given
extra-framework ion are more pronounced and the position and intensity of the
bands in this range are strongly dependent on the nature of interaction with the
framework, the mass of the ion, and the position relative to the 6-membered ring. It
should be noted that in the case of a detailed theoretical analysis of the far-infrared
range, it is of great importance that the proposed model reflects the state of the real
system as much as possible.
10.7.2 Anions in the Structure of Zeolites
In zeolite structures (especially in some natural zeolites), in addition to cations also

water molecules and selected inorganic anions are naturally present. For example,
in sodalite structure sulfide S2– or chloride Cl– ions [75] are present, while CO3 2
ions stabilize the ring system in the cancrinite [76, 77]. The nature of interactions of
anions with aluminosilicate framework has not been the subject of intensive
research so far; however, it is known that due to relatively large dimensions and
negative net charge they are located in centers of sufficiently large chambers in
which they are coordinated by positively charged extra-framework ions.
In Fig. 10.16, infrared spectra of sodalite synthesized in the presence of various
anions were collected. One can easily see that they differ distinctly from each other,
especially in far-infrared region. In calculated spectra, the bands associated with
Fig. 10.16 Experimental FT-IR spectra of different anionic forms of sodalite

vibrations of the anion sublattice are usually accompanied by Na–O vibrations

bands and the isolation of a band related the vibration associated with the anion
sublattice only seems impossible on the basis of theoretical spectra. This result is
also confirmed by experimental research. Differences in FIR range confirm that the
band associated with vibrations of the anionic sublattice is a part of the complex
band at around 200 cm−1; i.e., it coincides with the band attributed to the Na–O
vibrations. This effect is particularly evident when comparing the spectra of the
chloride and bromide form (Fig. 10.16), where it is possible to observe band
splitting with a maximum at 199 cm−1 into two component bands at 199 and
162 cm−1. Introduction of S2– ions to the system translates into larger changes also
in the pseudolattice range of vibrational spectrum, suggesting some modification of
this structure, or a change in the oxidation state of sulfur during synthesis.
Summing up, it can be concluded that based on the results of theoretical and
experimental data, it is possible to analyze in detail the structural changes taking
place as a result of the exchange of particular extra-framework ions. It is possible to
assess the deformation of the structure due to ion exchange as well as the unam-
biguous assignment of bands located both in mid- and in far-infrared region of
vibrational spectrum. The latter can be associated with both vibrations of individual
cationic and anionic sublattices.
10.8 Conclusions
Modern spectroscopic methods, in particular IR spectroscopy, have become an

extremely useful and valuable source of knowledge about the properties and
structure of both amorphous and crystalline materials, especially from the point of
view of their local ordering and chemical bond properties. The direct relations
between the shape of the vibrational spectrum and the structure and chemical
composition, as well as the relative ease of measuring such spectra in different
conditions, allow for not only the analysis of the structure of a given material but
also the observation and analysis of changes occurring in various processes. This is
particularly important in the case of materials used in such processes as sorption or
catalysis, where a key role plays local effects, influencing local structural changes
(creation or breaking of chemical bonds) or more subtle local changes in the
properties of chemical bonds in host framework near guest atoms or molecules
deposited on the surface or located inside channels and structural gaps in meso-
porous materials.
On the other hand, the development of both more efficient numerical algorithms
and the dynamic increase in the available computing power of modern computers
allows for the growing number of more and more complex simulations, the
so-called in silico experiments, verified experimentally, providing a lot of useful
information allowing for a better understanding of changes taking place in the
material during various processes and more and more effective design of new
(and modifications to existing) materials with specific functional properties.
328 M. Król et al.
The mentioned continuous increase in computing power of modern computers

allows for wider and wider applications of theoretical methods supporting experi-
mental research; however, more structurally complex materials such as zeolites still
remain—to large extent—behind theoretical methods, hence the constant need for
simplifications and approximations in theoretical models of real zeolite structures.
The aim of this chapter was to show the applicability of modern theoretical
methods based on the formalism of quantum mechanics for the calculation of
vibrational spectra and structural analysis of aluminosilicate mesoporous materials
from the group of zeolites. The obtained results showed the efficiency of theoretical
methods and their practical usefulness as a tool supporting description and inter-
pretation of vibrational spectra of zeolites and allowed to show that due to the
hierarchical structure of zeolitic materials, it is possible to study their properties
using theoretical models corresponding to structural units from different levels,
which allows obtaining valuable results providing insight into the structure and
properties of such structural units. These results also confirmed the far-reaching
identity of the SBU structural units, manifested by, inter alia, well-defined bands on
vibrational spectra in the mid-infrared range and hence the rationality of the clas-
sification of zeolites based on such units.
An important result is also the demonstration that the correct reconstruction of
the spectrum in the far-infrared range, where there are bands associated with lattice
vibrations with positions and intensities strongly dependent on the amount, type,
and nature of the interactions of foreign ions with the aluminosilicate framework, is
possible only using models that take into account translational symmetry of crystal
lattice, which is still a major difficulty due to the enormous structural complexity of
zeolites. Nevertheless, the first results of calculations made for simplified periodic
models corresponding to the structure of zeolite A showed that this approach allows
obtaining vibrational spectra largely consistent with the experiment, creating the
possibility of a more detailed analysis of the impact of extra-framework ions on the
aluminosilicate framework and changes in the structure of zeolite A occurring
during, for example, sorption of heavy metals.
The presented results show the high usability of theoretical methods in the
analysis of properties of zeolites, but at the same time indicate the still large
challenges facing these methods, related to the complexity of real materials.
Looking at the pace at which the computing power of modern computers is
growing, one can hope, however, that in the coming years it will be possible to
conduct in silico experiments on a much wider scale, based on models devoid of the
disadvantages resulting from the necessary far-reaching simplifications mentioned
earlier in this chapter, thanks to which it will be possible to quickly and effectively
study the properties of existing materials and design new ones with required
properties as well as modeling of dynamic processes (such as sorption or catalysis),
and therefore making computational methods an irreplaceable source of information
supplementing the knowledge obtained in experimental research.
Acknowledgements This work was partially supported by the National Science Centre, Poland,
under grant No. 2015/17/B/ST8/01200.
References
1. Smith JV (1984) Definition of a zeolite. Zeolites 4:309–310

2. Ciciszwili GW, Andronikaszwili TG, Kirov GN, Filizowa ŁD (1990) Zeolity naturalne.
Wydawnictwo Naukowo-Techniczne, Warszawa (in Polish)
3. Xu R, Pang W, Yu J, Huo Q, Chen J (2009) Chemistry of zeolites and related porous
materials: synthesis and structure. Wiley, New York
4. Auerbach SM, Carrado KA, Dutta PK (2003) Handbook of zeolite science and technology.
Marcel Dekker, New York
5. Gottardi G, Galli E (eds) (1985) Natural zeolites, mineral and rocks 18. Springer, Berlin
6. Förster H (1992) Infrared studies of zeolite complex. In: Davies JED (ed) Spectroscopic and
computational systems. Kluwer Academic Publishers, Amsterdam
7. Klinowski J (1997) Solid-state NMR studies of molecular sieve catalysts. Chem Rev
91:1459–1479
8. Breck DW (1974) Zeolite molecular sieves. Wiley, New York
9. Baerlocher C, McCusker L (2002) Database of zeolite structures
10. Meier WM (1968) Molecular sieves. Society of Chemical Industry, London
11. Flaningen EM, Khatami H, Szymanski HA (1971) Infrared structural studies of zeolites
frameworks. Adv Chem Ser 101:201–229
12. de Man AJM, van Santen RA (1992) The relation between zeolite framework structure and
vibrational spectra. Zeolites 12:269–279
13. Bordiga S, Lamberti C, Bonino F, Travert A, Frédéric T-S, Thibault-Starzyk F (2015) Probing
zeolites by vibrational spectroscopies. Chem Soc Rev 44:7262–7341
14. Ermoshin VA, Smirnov KS, Bougerard D (1996) Molecular dynamics calculation of the
vibrational spectra of OH groups in zeolites and on silica surfaces. Surf Sci 368:147–151
15. Pechar F, Rykl D (1985) Infrared Spectra of Natural Zeolites. Rozpr. Ceskosl. Akad. Ved.,
Praha
16. Uzunova EL, Niklov GS (2000) Vibrational modes of double four-member rings of
oxygen-bridged silicon and aluminum atoms: a DFT study. J Phys Chem B 104:7299–7305
17. Geidel E, Boehling H, Peuker Ch, Pliz W (1991) Approximate assignment of vibrational
frequencies of the NaX framework. Stud Surf Sci Catal 65:511–519
18. Geidel E, Lechert H, Döbler J, Jobic H, Calzaferri G, Bauer F (2003) Characterization of
mesoporous materials by vibrational techniques. Microporous Mesoporous Mater 65:31–42
19. Karge HG, Geidel E (2004) Vibrational spectroscopy. Vib Spectrosc 2:1–200
20. Sherwood PMA (1972) Vibrational spectroscopy of solids. Cambridge University Press,
Cambridge
21. Adams DM, Newton DC (1970) Tables for factor group and point group analysis. Beckmann,
Croydon
22. Bhagavantam S, Venkatarayudu T (1969) Theory of groups and its application to physical
problem. Academic Press, New York
23. Ferraro JR, Ziomek JS (1975) Introduction group theory and its application to molecular
structure. Plenum Press, New York
24. Handke M, Mozgawa W (1993) Vibrational spectroscopy of the amorphous silicates. Vib
Spectrosc 5:75–84
25. Handke M, Mozgawa W (1995) Model quasi-molecule Si2O as an approach in the IR spectra
description glassy and crystalline framework silicates. J Mol Struct 348:341–344
26. Sefcik J, Goddard WA (2001) Thermochemistry of silicic acid deprotonation: comparison of
gas-phase and solvated DFT calculations to experiment. Geochim Cosmochim Acta 65:4435–
4443
330 M. Król et al.
27. Tossell JA (2005) Theoretical study on the dimerization of Si(OH)4 in aqueous solution and
its dependance on temperature and dielectric constant. Geochim Cosmochim Acta 69:283–
291
28. Kubicki DJ, Sykes D (1993) Molecular orbital calculations on H6Si2O7 with variable angle:
Implications for high-presure vibrational spectra of silicate glasses. Am Min 78:253–255
29. Jastrzębski W (2006) Spektroskopia oscylacyjna pierścieni krzemotlenowych w strukturach
krzemianów i siloksanów. AGH University of Sciecne and Technology, Kraków (in Polish)
30. Handke M (1984) Spektroskopia wibracyjna krzemianów a charakter wiązania Si–O w
krzemianach (in polish). Zeszyty Naukowe AGH, Ceramika 48, Kraków
31. Mikuła A, Król M, Mozgawa W, Koleżyński A (2018) New approach for determination of the
influence of long-range order and selected ring oscillations on IR spectra in zeolites.
32. Handke M, Sitarz M, Mozgawa W (1998) Model of silicooxygen ring vibrations. J Mol Struct
450:229–238
33. Marcolli C, Calzaferri G (1997) Vibrational structure of monosubstituted octahydrosi-
lasesquioxanes. J Phys Chem B 101:4925–4933
34. Bornhauser P, Calzaferri G (1996) Ring-opening vibrations of spherosiloxanes. J Phys Chem
100:2035–2044
35. Pechar F, Rykl D (1983) Study of the vibrational spectra of natural natrolit. Can Mineral
21:689–695
36. Król M, Mozgawa W, Jastrzbski W, Barczyk K (2012) Application of IR spectra in the
studies of zeolites from D4R and D6R structural groups. Microporous Mesoporous Mater
156:181–188
37. Mozgawa W, Jastrzębski W, Handke M (2005) Vibrational spectra of D4R and D6R
structural units. J Mol Struct 744–747:663–670
38. Mozgawa W, Jastrzębski W, Handke M (2006) Cation-terminated structural clusters as a
model for the interpretation of zeolite vibrational spectra. J Mol Struct 792–793:163–169
39. Pápai I, Goursot A, Fajula F (1994) Density functional calculations on model clusters of
zeolite-b. J Phys Chem 98:4654–4659
40. Pidko E, Hensen EJM, Zhidomirov GM, van Santen R (2008) Nonlocalized charge
compensation in zeolites: a periodic DFT study of cationic gallium-oxide clusters in
mordenite. J Catal 255:139–143
41. Hill JR, Sauer J (1995) Molecular mechanics potential for silica and zeolite catalysts based on
ab initio calculations. 2. Aluminosilicates. J Phys Chem 99:9536–9550
42. Ermoshin VA, Smirnov KS, Bougeard D (1996) Ab initio generalized valence force field for
zeolite modelling 2. Aluminosilicates. Chem Phys 209:41–51
43. Mozgawa W, Handke M, Jastrzębski W (2004) Vibrational spectra of aluminosilicate
structural clusters. J Mol Struct 704:247–257
44. Ermoshin VA, Smirnov KS, Bougeard D (1997) Ab initio force field for aluminosilicates;
molecular dynamics simulation of the infrared spectra of zeolites. J Mol Struct 410–411:371–
374
45. Hill JR, Sauer J (1997) Molecular mechanics potential for silica and zeolite catalysts based on
ab initio calculations. 1. Dense and microporous silica. J Phys Chem 98:1238–1244
46. Mozgawa W (2007) Spektroskopia oscylacyjna zeolitów. Uczelniane Wydawnictwa
Naukowo-Dydaktyczne AGH, Kraków (in Polish)
47. Mozgawa W, Król M, Barczyk K (2011) FT-IR studies of zeolites from different structural
groups. Chemik 65(7):667–674
48. Mozgawa W, Bajda T (2006) Application of vibrational spectra in the studies of cation
sorption on zeolites. J Mol Struct 792–793:170–175
49. Huang Y, Jiang Z (1997) Vibrational spectra of completely siliceous zeolite A. Microporous
Mater 12:341–345
50. Loewenstein W (1954) The distribution of aluminum in the tetrahedra of silicates and
aluminates. Am Miner 39:92–96
51. Mozgawa W, Sitarz M (2002) Vibrational spectra of aluminosilicate ring structures. J Mol
Struct 614:273–279
52. Fernández-Jiménez A, Palomo A (2005) Mid-infrared spectroscopic studies of alkali-activated
fly ash structure. Microporous Mesoporous Mater 86:207–214
53. Ma Y, Liu Z, Geng A, Vogt T, Lee Y (2016) Structural and spectroscopic studies of
alkali-metal exchanged stilbites. Microporous Mesoporous Mater 224:339–348
54. Yuan J, Yang J, Ma H, Liu C (2016) Crystal structural transformation and kinetics of NH4+/
Na+ ion-exchange in analcime. Microporous Mesoporous Mater 222:202–208
55. Mozgawa W, Bajda T (2005) Spectroscopic study of heavy metals sorption on clinoptilolite.
Phys Chem Minerals 31:706–713
56. Mozgawa W, Król M, Bajda T (2009) Application of IR spectra in the studies of heavy metal
cations immobilization on natural sorbents. J Mol Struct 924–926:427–433
57. Handke M, Jastrzębski W (2004) Vibrational spectroscopy of the ring structures in silicates
and siloxanes. J Mol Struct 704:63–69
58. Król M, Mozgawa W, Jastrzębski W (2016) Theoretical and experimental study of
ion-exchange process on zeolites from 5–1 structural group. J Porous Mater 23:1–9
59. Pedone A, Biczysko M, Barone V (2010) Environmental effects in computational
spectroscopy: accuracy and interpretation. Chem Phys Chem 11:1812–1832
60. Creighton JA, Deckman HW, Newsam JM (1994) Computer simulation and interpretation of
the infrared and Raman spectra of sodalite frameworks. J Phys Chem 98:448–459
61. Iyer KA, Singer SJ (1994) Local-mode analysis of complex zeolite vibrations: Sodalite.
J Phys Chem 98:12670–12678
62. Iyer KA, Singer SJ (1994) Local-mode analysis of complex zeolite vibrations: zeolite A.
J Phys Chem 98:12679–12686
63. Mikuła A, Król M, Koleżyński A (2015) Periodic model of an LTA framework. J Mol Model
21:275
64. Koleżyński A, Mikuła A, Król M (2016) Periodic model of LTA framework containing
various non-tetrahedral cations. Spectrochim Acta A 157:17–25
65. Blackwell CS (1979) Investigation of zeolite frameworks by vibrational properties. 2. The
double-six-ring in group 4 zeolites. J Phys Chem 83:3257–3261
66. Rodriguez A (1995) Vibrational spectroscopy and structural analysis of Na-Y zeolite. Vib
67. Falabella Sousa-Aguiar E, Camorim VLD, Zotin FMZ, Correa dos Santos RL (1998) A
Fourier transform infrared spectroscopy study of La-, Nd-, Sm-, Gd- and Dy-containing Y
zeolites. Microporous Mesoporous Mater 25:25–34
68. Król M, Mozgawa W, Barczyk K, Bajda T, Kozanecki M (2013) Changes in the vibrational
spectra of zeolites due to sorption of heavy metal cations. J Appl Spectrosc 80(5):662–668
69. Mikuła A, Król M, Koleżyński A (2015) The influence of the long-range order on the
vibrational spectra of structures based on sodalite cage. Spectrochim Acta A 144:273–280
70. Mikuła A, Król M, Koleżyński A (2016) Experimental and theoretical spectroscopic studies
of Ag-, Cd- and Pb-sodalite. J Mol Struct 1126:110–116
71. Schwartz M, Berry RJ (2001) Ab initio investigation of substituent effects on bond
dissociation enthalpies in siloxanes and silanols. J Mol Struct 538:9–17
72. Baker MD, Godber J, Ozin G (1985) Frequency and intensity considerations in the far-IR
spectroscopy of faujasite xeolites: experiment and theory. Metal cation vibrational assign-
ments, site locations, and populations. J Am Chem Soc 107:3033–3043
73. Baker MD, Godber J, Helwig K, Ozin GA (1988) Probing extra-framework cations in alkali-
and alkaline-earth-metal Linde type A zeolites by fourier transform far-infrared spectroscopy.
J Phys Chem 92(21):6017–6024
74. Mortier WJ (1992) Compilation of extra framework sites in zeolites. Butterworth & Co.,
Leuven, pp 41–48
332 M. Król et al.
75. Günther C, Richter H, Voigt I, Michaelis A, Tzscheutschler H, Krause-Rehberg R, Serra JM

(2015) Synthesis and characterization of a sulfur containing hydroxy sodalite without sulfur
radicals. Microporous Mesoporous Mater 214:1–7
76. Bonaccorsi E, Orlandi P (2003) Marinellite, a new feldspathoid of the cancrinite-sodalite
group. Eur J Mineral 15:1019–1027
77. Hackbarth K, Gesing TM, Fechtelkord M, Stief F, Buhl J (1999) Synthesis and crystal
structure of carbonate cancrinite Na8[AlSiO4]6CO3(H2O)3.4, grown under low-temperature
hydrothermal conditions. Microporous Mesoporous Mater 30:347–358
Chapter 11
In Situ and Operando Techniques
in Catalyst Characterisation and Design
Przemysław Jodłowski and Joanna Łojewska
Abstract This chapter intends to present the classical and modern techniques that
are used for in situ characterisation of catalytic materials. Determination of the
structure of the catalyst presents three main problems: (1) heterogeneous catalysis
phenomena are limited to the outer surface of the material where the molecules
adsorb and react, and for this reason, there are only a few methods able to assess
catalyst surface structure and composition; (2) the catalyst surface under reaction
conditions and upon the influence of the reacting agents is different from that
occurring under ambient conditions, which limits the application of the analytical
methods to those which operate at normal or elevated pressures and high temper-
atures, (3) catalytic materials are complex and heterogeneous, so many analytical
methods, including surface imaging, should be employed in order to understand the
structure–activity relationships. The remedy for the problems is the application of
in situ analyses that rely on several complementary spectroscopic methods and
utilise surface sensitive probe molecules. Different kinds of probe molecules are
described: from universal probes to specific ones that enable the determination of
acidic and basic activity. The IR, Raman and UV-Vis methods are presented here
and described using examples from the literature. New trends in in situ experi-
mentation involve time-resolved techniques for studying fast reactions, fluorescence
methods and coupled techniques for surface in situ imaging.
P. Jodłowski (&)
Faculty of Chemical Engineering and Technology, Cracow University
of Technology, Warszawska 24, 31-155 Kraków, Poland
e-mail: pjodlowski@pk.edu.pl
J. Łojewska
Faculty of Chemistry, Jagiellonian University, Gronostajowa 2,
30-387 Kraków, Poland

334 P. Jodłowski and J. Łojewska
11.1 Introduction
Understanding the structure of catalyst surfaces and the surface intermediates

appearing during catalytic reactions is one of the key steps in determining the
reaction mechanism. Modern catalyst design for any technological process con-
solidates three scales: microscale, mesoscale and macroscale. The simultaneous
development of the technology allows scientists to look deeper into catalyst
structure and the elementary activity on the catalyst surface. The golden age of the
development of surface techniques started in the 1950 s [1], when new techniques
that allowed the detection limit to be decreased to the molecular level were
developed. The possibility of preparing ultrafine electronic materials allowed the
development of techniques for which ultra-high vacuum conditions are necessary.
Nowadays, the wide variety of techniques for catalyst characterisation allows a
choice of complimentary techniques for obtaining information about the measured
system. Over the years, the palette of available techniques has been increasing
gradually. The techniques used for catalyst characterisation described in this chapter
are summarised in Table 11.1.
The classical approach to catalyst characterisation was based on the determi-
nation of catalyst activity and selectivity for a given process, while the determi-
nation of catalyst structure via spectroscopic and microscopic methods was
performed in the next stage. Process parameters such as composition of catalysts
and reacting mixture, reaction temperature and pressure, activity, selectivity, and
chemical and mechanical resistivity of catalytic material used to be determined
under external perturbation, for example by increasing reactor temperature while
keeping the other variables constant. Indeed, such a methodological approach
provided a lot of information about catalytic systems that could be used for catalyst
optimisation and design.
Table 11.1 Selected techniques used in catalyst characterisation

Technique Information provided Limitations
Atomic force Catalyst surface structure – Resolution
microscopy (AFM) – Single-crystal model catalyst
X-ray absorption Catalyst local surface – Semi-quantitative
spectroscopy structure, local geometry – Complex spectra
– Limited beam time
– Possible probe damage
Fourier transform Surface intermediates, – Possible overlapping of support
infrared spectroscopy acidity/basicity of vibrations with active site
(FTIR) catalysts vibrations
– Indirect determination of active
sites
Raman spectroscopy Surface intermediates, – Fluorescence may affect analysis
(Raman) catalyst structure – Possible probe damages
Ultraviolet–visible Catalyst electronic Broad spectrum, complex band
spectroscopy (UV-Vis) structure assignment
11 In Situ and Operando Techniques in Catalyst Characterisation … 335
In the past, a majority of techniques relied on the use of the UHV conditions or
reaction “freezing” to obtain structural information on the catalyst surface or active
sites. However, that approach is far from the working conditions of the catalysts,
which for a vast majority of processes entail elevated temperature and pressure. The
necessity of observing catalyst structure under real working conditions was the
crucial point in the development of the in situ experiments. According to Haw [2],
an ideal in situ investigation should look at catalyst structure and morphology, and
the reaction intermediates under real working conditions. However, in the literature,
the term in situ is also used for the spectroscopic measurements of catalysts or metal
oxide surface upon contact with the adsorbate or reacting species [3]. The definition
proposed by Daydov [3] does not exclude a whole palette of measurements under
HV/UHV, utilising quantitative experiments with probe molecules such as
ammonia, pyridine or CO for the determination of surface acidity. Thus, taking into
account that there are two different ideas of in situ analysis, the question arises as to
whether the application of the UHV technique really reflects the structure of the
surface of a working catalytic material. Another question is presented by Daydov’s
attitude to combining HV/UHV with in situ examinations in order to collect as
much information about the catalytic system as possible.
On the other hand, when considering the analytical methods for catalyst char-
acterisation, we have to take into account that the phenomenon of catalysis is
limited to the very surface (external surface) of a catalytic material, on which
reactants adsorb, form intermediates and then convert to products. Amongst the
analytical methods which are able to focus on this surface are the microscopic
methods (AFM and, to an extent, SEM), in situ spectroscopic methods (Raman,
FTIR and UV/VIS) and ion (LEIS) and light scattering (SAXS). Ambient XPS
(rarely available yet in laboratories) and Raman spectroscopy can aid surface
studies with only limited confidence, since the signal emanating from a sample,
depending on penetration depth, provides information on the external surface (a few
monolayers in XPS, a few microns in Raman) to only a limited extent. This
important drawback of a majority of the so-called surface methods can be overcome
by in situ and operando analysis utilising surface probe molecules.
Since in situ characterisation methods provide information on catalyst structure
under real reaction conditions, rather than yielding data on overall activity, the
development of a methodology able to monitor both parameters at the same time
was highly desirable for further assessment of structure–activity relationships. This
led to the development of the methodology of operando spectroscopic measure-
ments. This term was originally proposed by Prof. Miguel Bañares during a dis-
cussion with Eric Gaigneaux, Gerhard Mestl and Bert Weckhuysen at the 220th
ACS National Meeting in Washington, in 2002. The term operando refers to the
Latin word describing working or operating conditions [4, 5]. Since in situ and
operando methodologies have been established as conventional methods, the
number of scientific papers on their use and development for catalyst studies has
been growing. According to the Web of Science (Fig. 11.1) both in situ and
operando techniques have become unwritten standards for studies dealing with the
preparation and characterisation of catalysts for various heterogeneous reactions.
Fig. 11.1 Number of annual publications with a “in situ spectroscopy” and b “operando
spectroscopy” in the subject, according to Web of Science
Despite the real progress that has been noted in in situ and operando experimental
approaches, there are still several problems associated with them, including reaction
temperature/pressure influence, catalyst form, catalyst activity and selectivity deter-
mination, and in situ cell construction [6]. In fact, it is the appropriate construction of
an in situ cell that defines the quality of the acquired spectra. The construction of
reaction cells is most often optimised for a specific spectroscopic method and avail-
able analytical device. The in situ cell can be “home-made” or chosen from com-
mercially available solutions. Detailed examples of in situ cells for selected
spectroscopies can be found in the following exemplary references: IR [6, 7], Raman
[4, 6], UV-Vis [4, 6], XAS/XES [4, 8], and SAXS, AP XPS [9].
11.2 Methods for Real-Time Catalyst Investigation
11.2.1 Determination of Active Centres Using Probe

Molecules
The acidic properties of a catalyst’s surface play an important role in catalytic

processes. Many important technological processes, including catalytic cracking
[10], isomerization [11] and transesterification [12] rely on solid acid catalysts that
allow the use of fixed or fluidised-bed catalytic reactors [13]. The correct under-
standing of the nature, quantity and strength of the acid/base properties of the
catalysts used for technological processes has been the main topic of a vast array of
scientific papers over the decades [3, 14–18]. The term “active centre” was origi-
nally proposed by Taylor, who suggested that only a small percentage of a cata-
lyst’s surface is able to adsorb and promote the catalytic reaction.
Nowadays, active sites are classified into several categories, the boundary of
which is redox and acidic/basic sites. The first category distinguishes the sites
varying in oxidation states of the metal and the sites composed of the charge
transfer metal complexes. The latter follows the definition of Brønsted (protonic)
and Lewis (non-protonic) acids and bases. Real industrial catalysts may contain the
whole palette of active sites. Furthermore, the two categories overlap if we con-
sider, for example, metal cations which can be regarded as both redox and Lewis
non-protonic acidic sites.
The characterisation of active sites is performed at the molecular level. In order
to determine the acidic properties of the catalysts, several characterisation methods
have been developed. These include titration methods, sorption, adsorption and
temperature-programmed desorption of probe molecules, and spectroscopic tech-
niques [11, 19]. Amongst the listed methods, the spectroscopic techniques are most
commonly used, due to their versatility and availability.
The selection principles for a probe molecule used in spectroscopic methods to
determine desired acidic properties was proposed by Knözinger [13]. The selected
probe molecule used in spectroscopic studies must conform to several criteria that
cover the following aspects [13, 14]:
• The spectra response of the interaction between probe molecule with active sites
must allow for the distinction between Brønsted and Lewis sites.
• Extinction factors are sufficiently high and known to allow for the determination
of quality and strength of measured centres.
• The acidic strength of the probe molecule should be selected to achieve the
optimum response in acid–base interaction
• The selected probe molecule should be specific to differentiate between the sites
of different strengths.
• The size of the probe molecule should be selected for the material under
scrutiny, so that it can diffuse through the pores to cover the maximum number
of the active sites.
• The measurement conditions (temperature) during surface probing should be
selected to avoid reactions other than those with active centres.
In fact, selecting an ideal probe molecule is often a difficult task, because it is
impossible in practice to meet all of the above-mentioned criteria completely.
The probe molecules that are commonly used to determine catalyst surface
acidity can be divided into two groups, with weak and strong bases. The group of
strong bases includes pyridine, ammonia, piperidine, trimethylamine and n-buty-
loamine [13, 20, 21]. The interaction of these probe molecules provides information
about both Lewis and Brønsted acid centres. Amongst the group of the strong bases,
ammonia and pyridine are the most frequently used [13, 22, 23]. The group of weak
bases, which includes benzene and substituted benzenes, CO, N2, O2 and H2, is also
used to determine both Lewis and Brønsted acidity [14]. The principle that lies
behind the determination of the acidity by weak base probes is the formation of
weak hydrogen bonds with the hydroxyl group present on the catalyst surface [14].
An example of the application of the in situ FTIR method to studies of catalyst
active sites is a recent study by Ochońska et al. [22], devoted to preparation and
characterisation of Cu-exchanged Y, USY and ZSM-5 zeolites used as catalysts for
Fig. 11.2 FTIR spectra of zeolites with adsorbed probe molecules; a CO, b NO, c NH3 Reprinted
with permission from [22], copyright (2018) Elsevier
selective catalytic removal of NOx from stationary sources. In this study, the nature
of the active sites was studied using CO, NO and NH3 probe molecules. The CO
sorption at room temperature is able to distinguish two types of Cu+ Lewis sites of
different strengths, which give vibrations with CO at 2160 and 2146 cm−1. In turn,
the use of ammonia as a probe molecule is universal and captures both Lewis and
Brønsted active sites at the same time [21, 22, 24]. In this case, the characteristic
bands around 1620 cm−1 correspond to Lewis acidity, while those near
1450 cm−1—relate to Brønsted. The results provide additional information on the
existence of non-equivalent Cu active sites, as can be inferred from the appearance
of two maxima (Fig. 11.2c): the band around 1614 cm−1 attributed to NH3 bonded
to Cu+, and the band at 1674 cm−1, attributed to NH3 coordinated to Cu2+.
The characterisation of basic active centres seems far more complicated than that
of acidic ones, since probe molecules should contain the acidic part. The most
frequently used probe molecules used to study basic active sites on catalyst surfaces
are carbon monoxide [25–29]. The adsorption of carbon monoxide is used to
characterise metals and metal oxides. CO is also used to assess cationic sites, which
are in turn acidic in nature. Comprehensive reviews on carbon monoxide interaction
with catalyst surface active sites can be found in the references [14, 23].
The adsorption of carbon monoxide over various catalytic systems results in the
formation of carbonyls. The interaction of carbon monoxide with the basic oxygen
sites (373 °C) of metal oxide catalysts at decreased temperature leads to the for-
mation of carbonite species [23], which has been studied by many research groups
[24–26]. At room temperature, CO can interact exclusively with Cu+ surface spe-
cies (Fig. 11.2a), while lowered temperature (liquid nitrogen temperature) also
allows for the detection of Cu2+ species [14, 24, 30, 31]. In the work by Kryca et al.
[24], both room temperature and low-temperature adsorption of CO was used to
differentiate between Cu+ and Cu2+ species in the ZSM-5 catalysts (Fig. 11.3b, c).
Room temperature adsorption of carbon monoxide gives rise to the formation of
Cu+–CO (2157 cm−1) and Cu+(CO)2 (2157 cm−1) species [30]. On the other hand,
low-temperature CO adsorption results in the interaction with Cu2+ species causing
the appearance of several bands at 2150, 2170, 2177 and 2191 cm−1 (Fig. 11.3C).
Fig. 11.3 Sorption of probe molecules over zeolite catalysts; a NH3, b CO at room temperature,
c CO at 100 °C Reprinted with permission from [24], copyright (2018) Elsevier
Alternatively, a similar effect to that observed for CO probe molecules can be

achieved by NO, which has similar properties [22, 32]. However, the adsorption of
NO molecules over copper-exchanged ZSM-5 catalysts at room temperature pro-
vides a complex spectroscopic pattern (Fig. 11.2b). The bands shifted to the higher
wavenumber (1850–2000 cm−1) can be assigned to Cu2+–NO, whereas those below
1850 cm−1 can be assigned to both Cu+–NO and Cu2+–(NO)2.
Amongst the probe molecules, alcohols play a special role in the studies of the
nature and number of active sites at a catalyst’s surface [33–35]. The literature
provides the number of reports of probing catalyst surfaces using alcohols of
various acidities, including (CF3)3COH (pKa = 5.4) (CF3)2CHOH (pKa = 9.3),
CF3CH2OH (pKa = 12.3) and CH3OH (pKa = 15.1) [23], ethanol, 2-propanol and
tetr-butyl [33]. The adsorption of the alcohols proceeds irreversibly with the for-
mation of two species at the catalyst surface (Fig. 11.4). The dissociative
chemisorption of alcohols on weak Lewis acid sites results in the formation of
alkoxy species (Fig. 11.4a), whereas undissociated species are chemisorbed on
strong Lewis acid sites, resulting in the formation of structure B, as shown in
Fig. 11.4b [23].
The chemisorption of alcohols over metal oxide catalysts results in the formation
of two characteristic regions in the IR spectrum. In the range 3000–2800 cm−1,
characteristic –CH vibrations are shifted to higher wavenumber for species B
(Fig. 11.4b), whereas dissociative adsorption with the characteristic adsorbed
Fig. 11.4 Chemisorbed

species after chemisorption of
alcohols over metal oxide
catalysts
Fig. 11.5 Mechanisms of adsorption of alcohols over metal oxide catalysts
species A is represented by lower –CH frequencies. In the second region, the

bending modes of –OH vibrations are coupled with the d–CH2 and d–CH modes.
A number of studies have been devoted to the adsorption of halogeno-alcohols over
selected metal oxide supports such as alumina or titania, proving dissociative
adsorption [34, 36]. The adsorption of acidic alcohols proceeds over active centres
with predominant basic properties (Fig. 11.5a), whereas less acidic alcohols
including methanol adsorb at acidic Lewis active centres (Fig. 11.5b).
A detailed description of the use of methanol as a probe molecule to study the
nature and number of the active sites can be found in the work of Badlani and
Wachs [37]. In this work, complimentary studies of methanol adsorption over a
series of several dozens of different metal oxides were carried out. The authors
found that, for most of the considered metal oxides, the number of active sites for
methanol dissociative adsorption is approximately 3 µmol/m2. Additionally, it was
also concluded that the methanol chemisorption at a catalyst surface, and further
oxidation on redox active centres, results in the formation of formaldehyde and
formic acid. The reaction of methanol over acidic centres results in the formation of
dimethyl ether, whereas complete oxidation proceeds at centres of basic character.
11.2.2 in Situ FTIR Spectroscopy
Infrared spectroscopy analysing the vibrational patterns of solid and probe

molecule-active site complexes is commonly used as a technique for catalyst
characterisation. The development of technology facilitated the modification of
infrared spectrometers, allowing measurements not only of the catalysts but also the
reaction intermediates at the catalyst surface. The development of sensitive and fast
detectors and, more importantly, of interferometers, made IR a common technique
for the rapid analysis of low amounts of intermediates appearing during catalytic
reactions on solid surfaces. Amongst various available IR techniques, the trans-
mission, diffuse reflectance (DRIFT) [38–44], attenuated total reflectance (ATR)
[45–50] and photoacoustic (PAS) [51–55] ones are those most utilised for catalyst
studies. However, while transmission experiments are the most frequently used for
the determination of the active sites using probe molecules, the question arises as to
whether they are pure in situ experiments according to the definition proposed by
Haw [56], because UHV or HV are recommended for obtaining reproducible FTIR
spectra of adsorbed probe molecules? Since the use of probe molecules in trans-
mission mode to determine acidic properties is the most utilised method, this
question remains unanswered.
However, both transmission and DRIFT modes are also used to monitor the
surface-active species appearing at a catalyst’s surface during the catalytic reaction.
Transmission experimentation in FTIR is a direct method of measurement, as it
does not create as many secondary optical effects as reflectance-based methods.
However, if the simplicity of the catalyst preparation is considered, DRIFT appears
to be the method of choice for its ability to analyse the studied material directly, in
its original form. In the transmission mode, there is a limited selection of options as
KBr wafers, as the dispersal of solids practically prevents their interaction with
gaseous agents. The thin, self-supporting or Si wafers on which catalysts are
deposited are the method for transmission measurements. What is more, if we
consider the structure and mechanical stability of self-supported wafers, it has to be
taken into account that the pressure which secures the integrity of the pellets may
also change the material structure and properties [7]. To monitor a catalyst’s surface
in transmission-absorption special in situ cells were developed. Some details of
these can be found in several references [6, 7, 57, 58].
Transmission experiments are commonly used not only to determine the active
sites at a catalyst’s surface using probe molecules under HV/UHV conditions, but
also in transient-state measurements and examinations of surface intermediates
using operando methodology [57, 59–62]. However, the vibrational patterns of
surface intermediates are complex and often need additional analysis by compli-
mentary methods to be properly resolved. In the study by Rasmussen et al. [57], the
surface properties during NH3-SCR over V2O5/WO3/TiO2/sepiolite catalyst were
characterised by operando IR/MS. IR measurements of the catalyst surface under
NO + O2 dynamic conditions revealed several bands, including those from
adsorbed NH4+ appearing at 1435–1450 cm−1, those from ammonia coordinated at
the Lewis acid site at 1610–1620 cm−1 and those of molecular water adsorbed at
the catalyst surface, at 1620 cm−1 (Fig. 11.6). The assignment of both NH3 and OH
bands from vibrating adsorbates may be problematic, since both bands overlap. To
solve this problem, the chemometric methods are more and more frequently
engaged, as Multivariate Curve Resolution Alternating Least Squares (MCR-ALS)
methods, for example, can be used to separate the overlapping bands [57].
Since the preparation procedures in classical transmission-absorption mode
require fastidious preparation of self-supporting diluted pellets, the DRIFT method
has become one of most utilised in situ method for the characterisation of catalysts.
The variety of applications of DRIFT analysis in catalyst characterisation, from
surface analysis to determination of surface intermediates, is a result of the sim-
plicity of the method. Indeed, catalyst preparation is reduced to minimum, although
the impact of particle size and reflectivity of the sample on the spectra cannot be
Fig. 11.6 Operando FTIR spectra during SCR of NO at 250 °C with pre-adsorbed ammonia. As
obtained (left) and components after deconvolution by MCR-ALS: adsorbed ammonia on acid
sites (middle) and hydroxyl bending mode (right) Reprinted with permission from [57], copyright
(2018) Elsevier
marginalised [63]. Since the first heated DRIFT cell was presented by Hamadeh
et al. [64] the several in situ DRIFT cells have been proposed [4, 7]. The currently
commercially available high temperature in situ cells are able to achieve temper-
atures up to 910 °C under vacuum, which increases the applicability of DRIFT
analysis for catalyst characterisation.
In the literature, numerous reports describe the application of in situ and oper-
ando DRIFT spectroscopy to characterise catalyst structure–activity relationships
[7, 65–70]. The operando studies are not limited only to measuring the gas phase
by the detection of effluent gases using gas spectrometers or mass spectrometers,
but more detailed catalyst characterisation studies using combined DRIFT/XAS
[71, 72] and DRIFT/XANES/MS [73] can be also found.
The possible application of in situ DRFIT spectroscopy for the determination of
surface intermediates during the catalytic combustion of methane was presented in
work by Jodłowski et al., shown here as an example [66]. DRIFT analysis of the
catalyst’s surface indicated the different active site occupation under oxygen-rich
and oxygen-less conditions. It showed methoxy groups of characteristic bands at
2999 and 2907 cm−1, attributed to –OCH3, and formate HCOO− species, respec-
tively. TPSR experiments also show (cf. Figure 11.7) that methane is able to react
with the catalyst surface under oxygen-less conditions, indicating the Mars–
van-Kreveln mechanism which is involved in its oxidation. This resulted in more
intense bands in the 1650–1200 cm−1 range for oxygen-free conditions, which were
attributed to carbonates and formates. The combination of static and pulse together
with oxygen-rich and oxygen-free conditions brought about the recognition of
stable surface intermediates, which allowed the methane catalytic combustion
mechanism over supported oxide catalyst to be derived (Fig. 11.8).
The operando mode of DRIFT experiments seems to be a powerful tool for
determining the relations with the structure of active centres and the observed
catalytic activity. The detection of the effluent gases is carried out with connected
gas analysers, from which Gas chromatographs (GC) or quadrupole mass spec-
trometers (QMSs) are the most commonly used. GC or QMS allows precise
determination and quantitative analysis of the effluent gas composition. In a study
Fig. 11.7 In situ DRIFT experiments of catalytic combustion of methane. Column oxygen-free
conditions (left), column oxygen-rich conditions (right) Reprinted with permission from [66],
Fig. 11.8 Reaction mechanism determined by in situ DRIFT experiments Reprinted with per-
mission from [66], copyright (2018) Elsevier
by Jodłowski et al. [74], operando DRIFT/MS was used to prove that neither
formaldehyde nor formic acid were present in the gaseous phase, as previously
postulated in the literature.
11.2.2.1 Enhanced in Situ FTIR Techniques
The description of in situ FTIR spectroscopy as a tool for the characterisation of

catalysts would be incomplete without consideration of time-resolved techniques,
including step-scan and rapid-scan methods. Since the lifetime of the reacting
species at the catalyst surface in the form of transition state or intermediate species
varies 10−12–10−13 s, the development of rapid-scan techniques was highly desired.
The rapid-scan method was developed in the 1970 s. Its application during catalyst
characterisation allows the collection of spectra in a time of 10−3 s in modern FTIR
spectrometers. The step-scan method decreases the acquisition time to 10−9 s. The
significant shortening of the spectra recording using both rapid-scan and step-scan
methods made possible experimental spectroscopic studies including transient
absorption experiments, pulsed sorption–desorption studies and kinetic studies
[75–82]. Nevertheless, the use of the rapid and step-scan techniques is very
demanding, requiring highly stable and reproducible conditions during the exper-
iment. Additional factors such as pressure changes or temperature fluctuations may
also affect the experimental results [75]. The correct set-up of time-resolved tech-
niques, their complexity and sensitivity to even millisecond condition fluctuations
mean that the literature does not contain a great many reports on this matter [75–78,
83–85].
Monitoring fast reactions can be achieved with FTIR analysis by using time or
phase-resolved techniques. In a study by Iwamoto and Hoshino [76], the isotopic
exchange of 14NO and 15NO species over CoZSM-5 catalysts was measured by the
application of rapid-scan FTIR spectroscopy. For the first time in the literature, it
was shown that 15NO molecules are rapidly exchanged by 14NO from the gaseous
phase until equilibrium in gaseous NO was achieved. Achieving the equilibrium
state for the 15NO–14NO exchange was more rapid for mononitrosyl than dinitrosyl
adsorbates, and lasted 0.3 s.
In another study, by Thibault-Starzyk et al. [83], the deNOx process by CO over
Ag/Al2O3 catalysts using the step-scan FTIR spectroscopy was investigated. The
pulse heating of the reactants was performed by using a femtosecond laser with a
10 Hz repetition rate. The authors concluded that the rate-limiting step in the CO
deNOx reaction is the flipping of cyanide groups from Ag nanoparticles to the
alumina support. The lifetime of the rate-limiting step was estimated to be as short
as 2 µs. This was evidenced by the gradual decrease in the 2130 cm−1 band
attributed to AgCN, and the increase of the 2265 cm−1 band, attributed to iso-
cyanate species adsorbed on alumina.
Rapid recent development of synchrotron radiation has opened new opportuni-
ties to perform operando XANES/DRIFT analysis of heterogeneous catalysts. In
work by Brieger et al. [72], the structure and surface properties of the EuroPt-1
catalyst (6.3 wt% Pt/SiO2) during CO sorption studies were measured using
operando XANES/DRIFT. Since XANES is not able to differentiate between the
adsorbed CO species over the EuroPt-1 catalyst, the use of DRIFT analysis was
demanded. Indeed, the characteristic bands at 2070, 1850 and 1720 cm−1 assigned
to atop, bridge-bonded and triple-bonded CO species.
11.2.2.2 2D Correlation Spectroscopy
Two-dimensional correlation spectroscopy (2D COS) was originally developed to

analyse NMR spectra and untangle the information derived from many overlapping
signals [86]. The possibility of analysing the spectra as a function of the two
variables represented a breakthrough in spectral analysis. Although the application
of 2D COS to the analysis of NMR spectra is still dominant, there is an increasing
interest in connecting it with other spectroscopic techniques, including FTIR or
Raman. The first application of 2D COS to vibrational spectroscopy was proposed
by Noda in 1988 [87]. Since then, 2D COS has attracted a wide group of
researchers from various fields such as catalysis and surface science. For the studies
of catalysis, the application of 2D COS to spectroscopic studies may provide
several advantages [86]:
• simplification of complex spectra by adding a second dimension;
• increase in spectral resolution of derived spectra and separation of overlapping
peaks;
• determination of the sequential order during analysis of the individual bands at
the asynchronous spectra;
• heterospectral correlation by the correlation of results obtained using different
spectroscopies (such as FTIR and Raman spectroscopy);
• correlation between different methods, such as mass spectrometry, gas chro-
matography, etc.
The use of 2D COS in in situ techniques has been reported in references [88–92].
In work by Chlebda et al. [69], 2D COS was used to improve in situ characterisation
and band assignment of surface intermediates appearing on a series of metal oxide
catalysts during methane catalytic combustion. During the reaction, oxidative and
non-oxidative conditions were applied. Exemplary results of in situ 2D FTIR COS
for mixed Co/Pd/c-Al2O3 catalysts under non-oxidative conditions are presented in
Fig. 11.8. Synchronous spectrum analysis has revealed two groups of cross-peaks at
synchronous spectra (Fig. 11.9a), with positive and negative increase of bands at
increasing reaction temperature occurring for cross-peaks. Analysis of the corre-
sponding asynchronous spectrum (Fig. 11.9b) allowed the sequence of the band
changes during the experiment to be determined as follows: 1590 ! 1390 !
1380 ! 1620 ! 1304 ! 1520 ! 1505 ! 1540 ! 1420 ! 1560 ! 1460 !
1350 ! 1330 ! 1340. Further analysis of both synchronous and asynchronous
spectra achieved by the application of 2D COS to the in situ FTIR spectroscopies
allows the reaction mechanism to be determined, based on in situ 2D COS DRIFT
experiments.
The 2D COS spectroscopy can also be successfully applied for in situ µRaman
structural studies of cobalt catalysts deposited on metallic supports [70]. The
methodology presented in work by Chlebda et al. [70] shows that the use of in situ
DRIFT, µRaman and DR-UV-Vis methods in parallel with the application of 2D
COS provides complimentary information on the catalysts. The 2D COS in situ
Fig. 11.9 Synchronous (A) and asynchronous (B) 2D infrared correlation spectra Co/Pd/c-Al2O3
catalyst in methane oxidative conditions 1.6% CH4/20% O2/He 25 cm3∙min−1 Reprinted with
permission from [66], copyright (2018) Elsevier
Raman and in situ DR-UV-Vis experiments show that the surface of the cobalt
catalyst deposited at the metallic support oxidised during catalytic combustion on n-
nonane. Analysis of 2D correlation µRaman spectra reveals that the band at
520 cm−1, attributed to the cobalt spinel oxide, broadens with increased tempera-
ture. Additionally, analysis of the asynchronous DRIFT spectra reveals bands that
were not easily observable in standard and synchronous spectra.
A series of investigations devoted to 2D COS FTIR spectroscopy
transmission-absorption mode is reported in the references [88–90]. The in situ
FTIR studies were performed to characterise xylene isomerisation over hierarchical
zeolites. The complexity of the spectra for individual products of isomerisation and
by-products makes overall analysis and further insights into mechanistic studies
challenging. Analysis of 2D correlation spectra has also confirmed that the high
selectivity towards the p-xylene was also influenced by the microporous character
of the selected zeolites.
11.2.3 in Situ Raman Spectroscopy
Raman spectroscopy is based on an outstanding effect that occurs upon light

scattering in contact with matter. It happens that a part of the scattered light
undergoes inelastic effects as opposed to Rayleigh phenomena, which are dominant.
Depending on the distribution of bosons, representing the vibrating molecules in
ground and excited vibrational states, two effects occur. Notwithstanding the further
mechanism, the low-temperature effect called Stokes effect, leads to a change in the
occupancy of the states towards the excited vibrational state, as the majority of the
molecules are originally located on the ground state (as described by the Bose–
Einstein distribution function). The anti-Stokes effect is opposite, shifting the
distribution of the bosons from the excited to the ground state. It is obvious that the
described effects are analogical to the excitation and de-excitation induced by IR
radiation but occur for entirely different molecular structures. Stokes effects are
more intense under conditions of catalytic reaction, although they vanish in the
prevailing Rayleigh scattered light. For this reason, Raman spectroscopy was losing
favour for many years, until new instrumentation was developed around 30 years
ago, giving it new applications in the fields of catalysis.
Unlike IR spectroscopy, Raman does not require any optical arrangements to
collect light from the surface of the solids, or any special optical windows except
glass. The majority of set-ups use microscopes which are able to transfer laser
incident light and collect back-scattered light. The choice of incident laser light
wavelength depends on the catalytic system studied. The in situ Raman experiments
also require proper choice of the in situ cells. There are two temperature cells
available commercially that are designed to comply with Raman microscopes for
the purpose of catalyst surface investigation. The Linkam cell, although designed to
reach a temperature of 1000 °C, is equipped with a ceramic cup that glows at high
temperature, disturbing the weak Raman signal. The Harrick is made entirely of
steel, but allows a temperature no higher than 550 °C. Both are equipped with
heating and gas supplying systems.
In fact, both IR and Raman spectroscopy are perfect choices for the in situ
molecular characterisation of catalysts, as their selection rules depend on the
molecular structure of the studied material, and they can provide real-time infor-
mation on its changes. Additionally, Raman is entirely complementary to the IR
method. This is not only because of the opposite selection rules, but also because of
the specific absorption of typical materials used in heterogeneous catalysis. IR is
especially suitable for the analysis of the adsorbed probe molecules as they absorb
light above 1000 cm−1, and while the typical support oxides such as Al2O3 or SiO2
do not give any signals, Raman spectroscopy compensates for this. In Raman
spectroscopy, the majority of the support oxides do not scatter light above
100 cm−1, which gives perfect circumstances for the analysis of the majority of the
metal oxides used in catalysis and which give Raman signals, typically in the range
200–1200 cm−1. An important advantage of Raman spectroscopy over IR is that
water, a ubiquitous molecule in a great many reactions, is practically invisible in
Raman. However, we have to take into account that the structure of a working
catalyst exposed to high temperatures is different from that which occurs under
ambient conditions in which water molecules are incorporated to the structure of the
metal oxides both of the catalyst and the support materials. This is another
important reason why in situ methods are the only adequate approach for assessing
catalyst structure under working conditions.
Amongst the catalytic systems most intensely studied by Raman spectroscopy
are bulk metal oxides and supported metal oxides. There are also examples of the
application of Raman spectroscopy to zeolite and metallic catalysts. The bulk metal
oxides can be studied in terms of the existence of different crystalline phases, their
transformation, and the extent of their crystallisation. If aided with a microscope,
Raman spectroscopy also provides information on the special distribution of bulk

metal oxides.
As regards supported metal oxides, Raman spectroscopy is able to provide
useful data on the molecular structure of the metal oxide on the support, its specific
location and surface coverage and surface acidity. If used in an in situ mode upon a
catalytic reaction, the technique is able to trace the changes in the structure of the
metal oxide catalyst, and in particular the engagement of the metal-oxygen bonds in
the reaction mechanism [93, 94]. For this class of catalysts, great efforts were made
to understand the vanadium and chromium group of elements, due to their wide-
spread application in many chemical industrial processes, examples of which can be
found in numerous works by Bañares and Wachs. Raman is especially suitable for
the study of those systems because the double bond between the metal and oxygen
that is formed on the surface of a metal oxide support gives a strong signal in the
Raman spectrum, around 1000 cm−1 (see Fig. 11.10 showing the Raman spectrum
of an Mo-ZSM-5 catalyst).
Raman can also be utilised for the chemisorption studies, following the methods
used for IR studies described in Sect. 11.2.2. The information obtained is com-
plementary to that obtained using IR.
An example of this is the in situ Raman study of Mo-ZSM-5 catalysts (Si/
Al = 40), initially calcined at 500 °C. A characteristic feature of the spectrum is the
Raman band at 1000 cm−1, arising from dioxo ms(Mo(=O)2) species on the octa-
hedral framework. There is also a band characteristic of bridging Mo–O–Mo bands
at 835 (mas) and 501 (ms) cm−1 [95]. A very weak and broad Raman band at around
900 cm−1 can be assigned to the bridging Mo–O–Al support bond, with a corre-
sponding weak band at around 360 cm−1 for the d(O–Mo–O) mode, which inter-
feres with ZSM-5 support vibrations. These bands vanish upon adsorption of
methanol at 100 °C, which then desorbs at temperatures above 200 °C and restores
the original vibrational pattern of the Mo-ZSM-5 catalyst.
Fig. 11.10 In situ Raman

spectra of an Mo-ZSM-5 (Si/
Al = 40) catalyst upon
interaction with methanol
pre-adsorbed at 100 °C and
then exposed to increasing
temperature
11.2.4 In Situ UV-Vis Spectroscopy
Electron transition molecular spectroscopy is one of the oldest techniques used to

investigate catalyst materials, but its role is limited due to complexity of the
absorption signals. The instrumentation used for in situ catalyst investigations relies
mainly on the diffuse reflectance and fibre optic spectroscopy, and on microscopy in
a similar manner to IR. The same concerns the choice and arrangements of the
temperature cells.
A broad presentation of examples of the use of UV-Vis in situ techniques can be
found in the review [96]. The most important domains of the application of this
technique for solving catalytic problems are the determination of the coordination
of transition metal ions to surface oxygens and quantitative determination of the
oxidation states of transition metal ions. An interesting application of UV-Vis
analysis is for the determination of the degree of “polymerisation” or clusterisation
of surface metal oxides, as for example in Bañares et al. [97]. For this purpose, edge
energy is derived from the UV-Vis spectra of single supported metal oxides [98].
An example of the use of the in situ UV-Vis DRS technique to analyse a
dehydrated Mo-ZSM-5 catalyst under oxidative conditions is shown in Fig. 11.11,
as difference spectra possessing a single O (2p) to Mo (d) orbital centred at 242–
245 nm. The absence of any d-d transitions at *400 nm from crystalline MoO3
indicates that this phase is not present. The corresponding Eg values for the
molybdena species at *245 nm drop in the range 4.3 to 4.4 eV, which indicates
Fig. 11.11 In situ UV-Vis spectra of an Mo-ZSM-5 (Si/Al = 40) catalyst upon interaction with
methanol pre-adsorbed at 100 °C and then exposed to increasing temperature
the presence of O2Mo(=O)2 species coordinated to SiO2, exhibiting two charge

transfer transitions at *237 and 274 nm corresponding to an Eg value of 4.1 eV.
Precise analysis of the Eg values for molybdenum-supported catalysts can be found
in Lee and Wachs [99].
11.3 New Trends in in Situ Investigation of Catalysts
The new methods that have been developed are a logical response to the urgent
need for a new strategy of active catalyst design and process control. The strategy is
based on managing the three scales of catalytic reactor functioning from macro-
scale, where structured reactors are designed and modelled, through mesoscale,
where catalysts are formed on the surface of structured carriers, up to microscale,
where catalytic active centres are designed. On the real-time mesoscale, a few
methods should be considered for catalytic material heterogeneity, the majority of
which are based on UV-Vis and Raman, MRI and UV-Vis, and tomography and
Raman-conjugated systems and are thoroughly described by Weckhuysen in terms
of real-time imaging of catalysts [65]. At microscale, the majority of the methods
able to reach atomic scale are based on fluorescence effects.
11.3.1 Conjugated AFM/Raman Spectroscopy
Amongst the conjugated techniques, AFM coupled with Raman microscopy gives
the opportunity of simultaneous analysis of both structure and texture of the catalyst
surface, yielding chemical information at high spatial resolution (down to
nanometre scale). The method can be utilised for catalyst surface mapping generally
in two boundary modes, with (near-field) or without (far-field) the tip-enhanced
Raman effect (TERS), as shown by Łojewska et al. [100]. The principles of AFM-
Raman and its applications in catalytic systems can be found in recent review
papers [101–106].
The differences in spatial resolution between Raman and AFM are the main
problem encountered with far-field mode mapping. The spatial resolution of a
Raman microscope is related to the diffraction limit of the laser in the optical parts
of the microscope, which for a laser wavelength from a visible range reaches a
minimum of around 200 nm, while AFM can go down to atomic scale.
Amongst the vast array of near-field effects [96] is the surface or, more precisely,
tip-enhanced effect [101–103]. In theory, Raman enhancement can reach up to
106 but in practice only a 103 increase has been reported to date [107]. There
are several optical arrangements in which the interfaced AFM/Raman analyses can
be performed, as shown in Table 11.2.
Table 11.2 Optical arrangements of AFM/Raman-conjugated systems and their applicability to

different materials
Configuration Angled Aligned Inverted Inverted
Raman + ++ ++ +++
detection
TERS ++ +++ ++ ++
TERS ++ +++ ++ ++
mapping
Application Solids, Solids, opaque, Solids, Solids,
opaque, rough rough transparent, flat transparent, flat
In situ Only Temperature and Temperature Temperature
temperature gas phase control and liquid and liquid
control phase phase
For non-transparent samples such as catalyst materials, several different solu-

tions with an aligned [108] and angled [107] optical configuration have been
developed to obtain coupled AFM/Raman maps in reflection mode of solid mate-
rials, the advantages and disadvantages of which are discussed in the literature
[104]. The general observation of coupled AFM/Raman system users is that the
maps with molecular resolution utilising the TERS effect are difficult to obtain and
can be recorded using the aligned configuration [107].
An example of studying the heterogeneity of a cobalt palladium-doped catalyst
supported on alumina is given in Fig. 11.12 [109]. There, the system was adapted to
describe the catalyst surface at mesoscale, upon catalyst formation. It was
demonstrated that, on the catalyst surface, PdO grains of micrometre size were
formed, which was further used to optimise the preparation method.
Another important and equally interesting direction in which the AFM/Raman
technique is developing is the in situ investigation of catalyst surface changes with
simultaneous information on the texture and structure. First attempts at this kind of
analysis were made by Harvey [110] and show rhodamine oxidation on Ag
nanoparticles by in situ TERS analysis.
Fig. 11.12 Heterogeneity of pre-calcined Kanthal steel (Al sample) deposited with cobalt oxide
and doped with palladium, imaged by interfaced AFM/Raman microscopy. a AFM topography
map 30 30 µm; b Raman coupled map of the same size and the same area as in (a), obtained
using two marker bands of Co3O4 (685 cm−1) and PdO (654 cm−1); c exemplary Raman spectrum
of Co3O4; d exemplary Raman spectrum of PdO Reprinted with permission from [109], copyright
(2018) Elsevier
11.4 Summary
A golden rule of catalytic material engineering is that, to be able to design it, we

have to understand it. There are different scales at which a catalyst is approached,
designed and manufactured. At microscopic level, the focus of this chapter, fun-
damental relationships between catalyst structure and activity are surveyed with the
aid of advanced methods for catalyst surface characterisation. Since reactions occur
exclusively on the very surface of the material in heterogeneous catalysis, there are
in fact only a few methods able to analyse surface structure. However, any bulk
analytical method becomes a surface method when combined with probing with
surface sensitive molecules, providing information on real-time changes in the

structure of the catalyst surface. The in situ approach fills the so-called material and
process gaps in catalyst investigation when model simplified materials where
studied under conditions far removed from those inside industrial rectors.
In this chapter, the main spectroscopic techniques used for the in situ charac-
terisation of catalyst materials are described. These include FTIR, Raman and
UV-Vis spectroscopies. Notwithstanding the spectroscopic method under scrutiny,
the strategy for choosing probe molecules is the same. Depending on the type of
active centre, different chemical agents can be used for both qualitative and
quantitative analysis. Alcohols, however, can be used as universal probes, as they
are able to convert to various products upon interaction with different active centres,
in brief: on acidic centres—to eters, on basic—to carbonates, on redox—to alde-
hydes or carboxylic acids. For analysis of Brønsted and Lewis acidic sites, mole-
cules that are their basic pairs are used, and vice versa. Amongst the molecules with
basic properties, pyridine and ammonia are used most frequently, and amongst the
acidic molecules, carbon monoxide is most prevalent. Special devices are designed
for the titration of active sites on a catalyst surface, for the determination of their
number that is necessary for the derivation of turnover-frequency values (TOF).
Despite the fact that determination of TOF is not always obvious, especially if there
are many different kinds of active sites and if they are not fully accessible for a
given probe molecule, TOF is unquestionably the most universal gauge for mea-
suring and comparing catalyst activity.
Different spectroscopic methods that are utilised for in situ catalyst characteri-
sation are able to focus on different aspects of a catalyst surface during a chemical
process, and in this sense they are regarded as complementary. While IR can focus
on the structure of the probe molecule-active site complexes, Raman spots the
changes in the structure of metal oxides (the most commonly used catalytic
materials) upon these interactions, and UV-Vis captures the shifts in electronic
states that accompany the changes. For studies of reaction mechanisms, in situ
methods are used for tracking the above-mentioned changes during the catalytic
process. Additionally, operando techniques allow simultaneous analyses of reaction
products, giving insight into changes not only in the surface structure, but also in
catalyst activity. The spectroscopic in situ and operando methods are discussed in
terms of the used instrumentation, sampling and specific realms of their application,
and discussed using examples from the literature.
Finally, new trends in the development of in situ methods are presented, with
particular reference to:
• Time-resolved techniques for tracing fast reactions demonstrated for IR
spectroscopy.
• Coupled techniques, such as UV-Vis/Raman, used for imaging catalyst surfaces
at mesoscale and microscale, including in in situ mode. The AFM/Raman
technique is discussed in more detail.
• Fluorescence techniques at nanoscale, distinguishing changes on the catalyst
surface.
Acknowledgements This work was partially supported by National Centre for Research and
Development decision No. LIDER/204/L-6/14/NCBR/2015 and by the National Science
Foundation project 2013/09/B/ST8/00171.
References
1. Somorjai GA, Li Y (2010) Introduction to surface chemistry and catalysis. Wiley

2. Haw JF (2002) In-situ spectroscopy in heterogeneous catalysis. Wiley-VCH
3. Daydov A (2003) Molecular spectroscopy of oxide catalyst surfaces. Wiley
4. Weckhuysen BM (2002) Snapshots of a working catalyst: possibilities and limitations of
in situ spectroscopy in the field of heterogeneous catalysis. Chem Commun 2:97–110
5. Banares MA (2005) Operando methodology: combination of in situ spectroscopy and
simultaneous activity measurements under catalytic reaction conditions. Catal Today
100:71–77
6. Weckhuysen BM (2004) In-situ spectroscopy of catalysts. In: Weckhuysen BM (ed) In-situ
spectroscopy of catalysts. American Scientific Publishers, Stevenson Ranch, CA, pp 1–12
7. Ryczkowski J (2001) IR spectroscopy in catalysis. Catal Today 68:263–381
8. Niemantsverdriet JW (2007) Spectroscopy in catalysis. Wiley-VCH Verlag GmbH & Co,
KGaA, Weinheim
9. Frenken J, Groot I (eds) (2017) Operando research in heterogeneous catalysis. Springer
Nature, Cham, Switzerland
10. Tarach K, Góra-Marek K, Tekla J, Brylewska K, Datka J, Mlekodaj K, Makowski W,
Igualada López MC, Martínez Triguero J, Rey F (2014) Catalytic cracking performance of
alkaline-treated zeolite Beta in the terms of acid sites properties and their accessibility.
J Catal 312:46–57
11. Corma A (1995) Inorganic solid acids and their use in acid-catalyzed hydrocarbon reactions.
Chem Rev 95:559–614
12. Lee AF, Wilson K (2015) Recent developments in heterogeneous catalysis for the
sustainable production of biodiesel. Catal Today 242:3–18
13. Knozinger H (1993) Infrared spectroscopy as a probe of surface acidity. In: Joyner RW, van
Santen RA (eds) Elementary reaction steps in heterogeneous catalysis. Springer
Science + Business Media B.V., pp 267–285
14. Hadjiivanov KI, Vayssilov GN (2002) Characterisation of oxide surfaces and zeolites by
carbon monoxide as an IR probe molecule. Adv Catal 47:307–511
15. Taylor HS (1925) A theory of the catalytic surface. Proc R Soc A 108:105–111
16. Barzetti T, Selli E, Moscotti D, Forni L (1996) Pyridine and ammonia as probes for FTIR
analysis of solid acid catalysts. J Chem Soc, Faraday Trans 92:1401–1407
17. Lercher JA, Gründling C, Eder-Mirth G (1996) Infrared studies of the surface acidity of
oxides and zeolites using adsorbed probe molecules. Catal Today 27:353–376
18. Halasz I (2010) Silica and silicates in modern catalysis. Transworld Research Network
19. Boroń P, Chmielarz L, Gil B, Marszałek B, Dzwigaj S (2016) Experimental evidence of
NO SCR mechanism in the presence of the BEA zeolite with framework and
extra-framework cobalt species. Appl Catal B Environ 198:457–470
20. Barthomeuf D (1996) Basic zeolites: characterization and uses in adsorption and catalysis.
Catal Rev Sci Eng 38:521–612
21. Chlebda D, Stachurska P, Jędrzejczyk R, Kuterasiński Ł, Dziedzicka A, Górecka S,
Chmielarz L, Łojewska J, Sitarz M, Jodłowski PJ (2018) DeNOx abatement over sonically
prepared iron-substituted Y, USY and MFI zeolite catalysts in lean exhaust gas conditions.
Nanomaterials 8:21
22. Ochońska J, McClymont D, Jodłowski PJ, Knapik A, Gil B, Makowsk A, Łasocha W,

Kołodziej A, Kolaczkowski ST, Łojewska J (2012) Copper exchanged ultrastable zeolite Y
—a catalyst for NH3-SCR of NOx from stationary biogas engines. Catal Today 191:6–11
23. Lavalley JC (1996) Infrared spectrometric studies of the surface basicity of metal oxides and
zeolites using adsorbed probe molecules. Catal Today 27:377–401
24. Kryca J, Jodłowski PJ, Iwaniszyn M, Gil B, Sitarz M, Kolodziej A, Lojewska T, Lojewska J
(2016) Cu SSZ-13 zeolite catalyst on metallic foam support for SCR of NOx with ammonia:
catalyst layering and characterisation of active sites. Catal Today 268:142–149
25. Góra-Marek K, Gil B, Datka J (2009) Quantitative IR studies of the concentration of
Co2 + and Co3+ sites in zeolites CoZSM-5 and CoFER. Appl Catal A Gen 353:117–122
26. Góra-Marek K, Gil B, Śliwa M, Datka J (2007) An IR spectroscopy study of Co sites in
zeolites CoZSM-5. Appl Catal A Gen 330:33–42
27. Ciuparu D, Bensalem A, Pfefferle L (2000) Pd–Ce interactions and adsorption properties of
palladium: CO and NO TPD studies over Pd–Ce/Al2O3 catalysts. Appl Catal B Environ
26:241–255
28. Chin SY, Williams CT, Amiridis MD (2006) In situ FT-IR studies of CO adsorption on fresh
Mo2C/Al2O3 catalyst. J Phys Chem B 82:871–882
29. Feio LSF, Hori CE, Damyanova S, Noronha FB, Cassinelli WH, Marques CMP,
Bueno JMC (2007) The effect of ceria content on the properties of Pd/CeO2/Al2O3 catalysts
for steam reforming of methane. Appl Catal A Gen 316:107–116
30. Hadjiivanov K, Knözinger H (2000) FTIR study of low-temperature CO adsorption on
Cu-ZSM-5: Evidence of the formation of Cu2+(CO)2 species. J Catal 191:480–485
31. Datka J, Kozyra P (2005) TPD–IR studies of CO desorption from zeolites CuY and CuX.
J Mol Struct 744–747:991–996
32. Rejmak P, Broclawik E (2008) Nitrogen monoxide interaction with Cu (I)sites in zeolites X
and Y: quantum chemical calculations and IR studies. J Phys Chem C 112:17998–18010
33. Farneth WE, Staley RH, Sleight AW (1986) Stoichiometry and structural effects in alcohol
chemisorption/temperature-programmed desorption on MoO3. J Am Chem Soc 108:2327–
2332
34. Benaissa M, Travert J, Lavalley J (1985) Microcalorimetric and Fourier Transform infrared
spectroscopic studies of methanol adsorption on Al2O3. J Phys Chem 89:5433–5439
35. Busca G, Forzatti P, Lavalley JC, Tronconi E (1985) A TPD, FT-IR and catalytic study of
the interaction of methanol with pure and KOH doped TiO2 anatase. Stud Surf Sci Catal
20:15–24
36. Rossi PF, Busca G, Lorenzelli V, Saur O, Lavalley JC (1987) Microcalorimetric and FT-IR
spectroscopic study of the adsorption of isopropyl alcohol and hexafluoroisopropyl alcohol
on titanium dioxide. Langmuir 3:52–58
37. Badlani M, Wachs IE (2001) Methanol: a “smart” chemical probe molecule. Catal Lett
75:137–149
38. Mao C-F, Vannice MA (1995) Formaldehyde oxidation over Ag catalysts. J Catal 154:230–244
39. Busca G (1999) The surface acidity of solid oxides and its characterization by IR
spectroscopic methods. An attempt at systematization. Phys Chem Chem Phys 1:723–736
40. Vorob’eva MP, Greish AA, Ivanov AV, Kustov LM (2000) Preparation of catalyst carriers
on the basis of alumina supported on metallic gauzes. Appl Catal A Gen 199:257–261
41. Jacobs G, Williams L, Graham U, Thomas GA, Sparks DE, Davis B (2003) Low
temperature water-gas shift: in situ DRIFTS-reaction study of ceria surface area on the
evolution of formates on Pt/CeO2 fuel processing catalysts for fuel cell applications. Appl
Catal A Gen 252:107–118
42. Armaroli T, Minoux D, Gautier Euzen P (2003) A DRIFTS study of Mo/alumina interaction:
from Mo/boehmite solution to Mo/cAl2O3 support. Appl Catal A Gen 251:241–253
43. Liu X (2008) DRIFTS study of surface of c -alumina and Its dehydroxylation. J Phys
Chem C 112:5066–5073
44. Kung MC, Lin SSY, Kung HH (2012) In situ infrared spectroscopic study of CH4 oxidation
over Co-ZSM-5. Top Catal 55:108–115
45. Ferri D, Bürgi T, Baiker A (2001) Pt and Pt/Al2O3 thin films for investigation of catalytic
solid-liquid interfaces by ATR-IR spectroscopy: CO adsorption, H2-induced reconstruction
and surface-enhanced absorption. J Phys Chem B 105:3187–3195
46. Ferri D, Bürgi T, Baiker A (2002) Probing boundary sites on a Pt/Al2O3 model catalyst by
CO2 hydrogenation and in situ ATR-IR spectroscopy of catalytic solid–liquid interfaces.
47. Burgener M, Wirz R, Mallat T, Baiker A (2004) Nature of catalyst deactivation during citral
hydrogenation: a catalytic and ATR-IR study. J Catal 228:152–161
48. Grabow K, Bentrup U (2014) Homogeneous catalytic processes monitored by combined
in situ. ACS Catal 4:2153–2164
49. Rabeah J, Bentrup U, StoBer R, Bruckner A (2015) Selective alcohol oxidation by a copper
TEMPO catalyst: mechanistic insights by simultaneously coupled operando EPR/UV-Vis/
ATR-IR spectroscopy. Angew Chemie Int Ed 54:11791–11794
50. Ohlin L, Bazin P, Thibault-Starzyk F, Hedlund J, Mattias Grahn M (2013) Adsorption of
CO2, CH4, and H2O in zeolite ZSM-5 studied using in situ ATR-FTIR spectroscopy. J Phys
Chem C 117:16972–16982
51. Savara A, Weitz E (2014) Elucidation of intermediates and mechanisms in heterogeneous
catalysis using infrared spectroscopy. Annu Rev Phys Chem 65:249–273
52. Xu Z, Yu J, Jaroniec M (2015) Efficient catalytic removal of formaldehyde at room
temperature using AlOOH nanoflakes with deposited Pt. Appl Catal B Environ 163:306–312
53. Macina D, Piwowarska Z, Góra-Marek K, Tarach K, Rutkowska M, Girman V,
Blachowski A, Chmielarz L (2016) SBA-15 loaded with iron by various methods as
catalyst for DeNOx process. Mater Res Bull 78:72–82
54. Murakami N, Koga N (2016) In situ photoacoustic FTIR studies on photocatalytic oxidation
of 2-propanol over titanium(IV) oxide. Catal Commun 83:1–4
55. Qi L, Cheng B, Yu J, Ho W (2016) High-surface area mesoporous Pt/TiO2 hollow chains for
efficient formaldehyde decomposition at ambient temperature. Elsevier BV
56. Haw JF (2002) In-situ spectroscopy in heterogeneous catalysis. Wiley-VCH
57. Rasmussen SB, Portela R, Bazin P, Ávila P, Banares MA, Daturi M (2018) Transient
operando study on the NH3/NH4+ interplay in V-SCR monolithic catalysts. Appl Catal B
Environ 224:109–115
58. Thibault-Starzyk F, Mauge F (2012) Infrared Spectroscopy. In: Che M, Védrine J
(eds) Characterization of solid materials and heterogeneous catalysts. From structure to
surface reactivity. Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim, pp 3–48
59. Benaliouche F, Boucheffa Y, Thibault-Starzyk F (2012) In situ FTIR studies of propene
adsorption over Ag- and Cu-exchanged Y zeolites. Microporous Mesoporous Mater 147:10–
16
60. El-Roz M, Bazin P, Daturi M, Thibault-Starzyk F (2013) Operando IR coupled to SSITKA
for photocatalysis: reactivity and mechanistic studies. ACS Catal 3:2790–2798
61. El-Roz M, Kus M, Cool P, Thibault-Starzyk F (2012) New operando IR technique to study
the photocatalytic activity and selectivity of TiO2 nanotubes in air purification: Influence of
temperature, UV intensity, and VOC concentration. J Phys Chem C 116:13252–13263
62. Keturakis CJ, Zhu M, Gibson EK, Daturi M, Tao F, Frenkel AI, Wachs IE (2016) Dynamics
of CrO3–Fe2O3 catalysts during the high-temperature water-gas shift reaction: molecular
structures and reactivity. ACS Catal 6:4786–4798
63. Mitchell MB (1993) Fundamentals and applications of diffuse reflectance infrared fourier
transform (DRIFT) spectroscopy. In: Urban MW, Craver CD (eds) Structure-property
relations in polymers, spectroscopy and performance. American Chemical Society,
Washington, DC, pp 351–375
64. Hamadeh IM, King D, Griffiths PR (1984) Heatable-evacuable cell and optical system for
diffuse reflectance FT-IR spectrometry of adsorbed species. J Catal 88:264–272
65. Weckhuysen BM (2009) Chemical imaging of spatial heterogeneities in catalytic solids at

different length and time scales. Angew Chemie Int Ed 48:4910–4943
66. Jodłowski PJ, Jędrzejczyk RJ, Chlebda D, Gierda M, Łojewska J (2017) In situ
spectroscopic studies of methane catalytic combustion over Co, Ce, and Pd mixed oxides
deposited on a steel surface. J Catal 350:1–12
67. Becker E, Carlsson P, Kylhammar L, Newton MA, Skoglundh M (2011) In situ
spectroscopic investigation of low-temperature oxidation of methane over
alumina-supported platinum during periodic operation. J Phys Chem C 115:944–951
68. Rutkowska M, Díaz U, Palomares AE, Chmielarz L (2015) Cu and Fe modified derivatives
of 2D MWW-type zeolites (MCM-22, ITQ-2 and MCM-36) as new catalysts for DeNOx
process. Appl Catal B Environ 168–169:531–539
69. Chlebda DK, Jodłowski PJ, Jędrzejczyk RJ, Łojewska J (2018) Generalised two-dimensional
correlation analysis of the Co, Ce, and Pd mixed oxide catalytic systems for methane
combustion using in situ infrared spectroscopy. Spectrochim Acta, Part A 192:202–210
70. Chlebda DK, Jodłowski PJ, Jędrzejczyk RJ, Łojewska J (2017) 2D-COS of in situ l-Raman
and in situ IR spectra for structure evolution characterisation of NEP-deposited cobalt oxide
catalyst during n-nonane combustion. Spectrochim Acta, Part A 186:44–51
71. Marinkovic NS, Wang Q, Barrio L, Ehrlich SN, Khalid S, Cooper C, Frenkel AI (2011)
Combined in situ X-ray absorption and diffuse reflectance infrared spectroscopy: An
attractive tool for catalytic investigations. Nucl Instruments Methods Phys Res Sect A Accel
Spectrometers, Detect Assoc Equip 649:204–206
72. Brieger C, Melke J, Van der Bosch N, Reinholz U, Riesemeier H, Guilherme Buzanich A,
Kayarkatte MK, Derr I, Schökel A, Roth C (2016) A combined in-situ XAS-DRIFTS study
unraveling adsorbate induced changes on the Pt nanoparticle structure. J Catal 339:57–67
73. Parlett CMA, Gaskell CV, Naughton JN, Newton MA, Wilson K, Lee AF (2013) Operando
synchronous DRIFTS/MS/XAS as a powerful tool for guiding the design of Pd catalysts for
the selective oxidation of alcohols. Catal Today 205:76–85
74. Jodłowski PJ, Chlebda D, Piwowarczyk E, Chrzan M, Jędrzejczyk RJ, Sitarz M,
Węgrzynowicz A, Kołodziej A, Łojewska J (2016) In situ and operando spectroscopic
studies of sonically aided catalysts for biogas exhaust abatement. J Mol Struct 1126:132–140
75. Zaera F (2014) New advances in the use of infrared absorption spectroscopy for the
characterization of heterogeneous catalytic reactions. Chem Soc Rev 43:7624–7663
76. Iwamoto M, Hoshino Y (1995) Time-resolved FT-IR study on isotopic exchange of ditrosyl
species strongly adsorbed on Co-MFI zeolite. Chem Lett 24:729–730
77. Yeom YH, Frei H (2003) Step-scan FT-IR monitoring of transient HCO radicals in a room
temperature zeolite. J Phys Chem B 107:6286–6291
78. Zhang M, De Respinis M, Frei H (2014) Time-resolved observations of water oxidation
intermediates on a cobalt oxide nanoparticle catalyst. Nat Chem 6:362–367
79. Chang BK, Jang BW, Dai S, Overbury SH (2005) Transient studies of the mechanisms of
CO oxidation over Au/TiO2 using time-resolved FTIR spectroscopy and product analysis.
J Catal 236:392–400
80. Chen T, Feng Z, Wu G, Shi J, Ma G, Ying P, Li C (2007) Mechanistic studies of
photocatalytic reaction of methanol for hydrogen production on Pt/TiO2 by in situ fourier
transform IR and time-resolved IR spectroscopy. J Phys Chem C 111:8005–8014
81. Yao S, Mudiyanselage K, Xu W, Johnston-Peck AC, Hanson JC, Wu T, Stacchiola D,
Rodriguez JA, Zhao H, Beyer KA, Chapman KW, Chupas PJ, Martínez-Arias A, Si R,
Bolin TB, Liu W, Senanayake SD (2014) Unraveling the dynamic nature of a CuO/CeO2
catalyst for CO oxidation in Operando: a combined study of XANES (fluorescence) and
drifts. ACS Catal 4:1650–1661
82. Kang M, Perry D, Kim YR, Colburn AW, Lazenby RA, Unwin PR (2015) Time-resolved
detection and analysis of single nanoparticle electrocatalytic impacts. J Am Chem Soc
137:10902–10905
83. Thibault-Starzyk AF, Seguin E, Thomas S, Daturi M, Arnolds H, King DA (2016) Real-time
of cyanide infrared detection flip on silver-alumina catalyst NOx removal. Science
324:1048–1051
84. Johnson TJ, Simon A, Weil JM, Harris GW (1993) Applications of time-resolved step-scan
and rapid-scan FT-IR spectroscopy: dynamics from ten seconds to ten nanoseconds. Appl
85. Wille A, Fridell E (2007) Millisecond step-scan FT-IR transmission spectroscopy under
transient reaction conditions: CO oxidation over Pt/Al2O3. Appl Catal B Environ 70:294–
304
86. Noda I, Ozaki Y (2004) Two-dimensional correlation spectroscopy—applications in
vibrational and optical spectroscopy. Wiley, Chichester, UK
87. Noda I, Dowrey AE, Marcott C (1988) 2D IR spectroscopy a new tool for interpreting
infrared spectra. Mikrochim Acta 94:101–103
88. Gołąbek K, Tarach KA, Góra-Marek K (2017) Standard and rapid scan infrared
spectroscopic studies of o-xylene transformations in terms of pore arrangement of 10-ring
zeolites—2D COS analysis. Dalt Trans 46:9934–9950
89. Gołąbek K, Tarach KA, Góra-Marek K (2018) 2D COS analysis of m-xylene transformation
over medium-pore zeolites. Microporous Mesoporous Mater 266:90–101
90. Tarach KA, Gołąbek K, Choi M, Góra-Marek K (2017) Quantitative infrared spectroscopic
studies and 2D COS analysis of xylenes isomerization over hierarchical zeolites. Catal
Today 283:158–171
91. Thibault-Starzyk F, Vimont A, Gilson J-P (2001) 2D-COS IR study of coking in xylene
isomerisation on H-MFI zeolite. Catal Today 70:227–241
92. Fernandez C, Stan I, Gilson JP, Thomas K, Vicente A, Bonilla A, Pérez-Ramírez J (2010)
Hierarchical ZSM-5 zeolites in shape-selective xylene isomerization: role of mesoporosity
and acid site speciation. Chem A Eur J 16:6224–6233
93. Lewis IR, Edwards HGM, Dekker M (eds) (2001) Handbook of Raman spectroscopy. From
the research laboratory to the process line. Taylor & Francis Group, LLC, New York and
Basel
94. Bañares MA, Wachs IE (2010) Raman spectroscopy of catalysts. In: Encyclopedia of
analytical chemistry. American Cancer Society
95. Horsley JA, Wachs IE, Brown JM, Via GH, Hardcastle FD (1987) Structure of surface
tungsten oxide species in the tungsten trioxide/alumina supported oxide system from X-ray
absorption near-edge spectroscopy and Raman spectroscopy. J Phys Chem 91:4014–4020
96. Kim H, Kosuda KM, Van Duyne RP, Stair PC (2010) Resonance Raman and surface- and
tip-enhanced Raman spectroscopy methods to study solid catalysts and heterogeneous
catalytic reactions. Chem Soc Rev 39:4820–4844
97. Bañares MA, Martinez-Huerta MV, Gao X, Fierro JLG, Wachs IE (2000) Dynamic behavior
of supported vanadia catalysts in the selective oxidation of ethane. In situ Raman. UV-Vis
and reactivity studies. Catal Today 61:295–301
98. Delgass WN (1979) Spectroscopy in heterogeneous catalysis. Academic Press
99. Lee EL, Wachs IE (2007) In situ spectroscopic investigation of the molecular and electronic
structures of SiO2 supported surface metal oxides. J Phys Chem C 111:14410–14425
100. Łojewska J, Knapik Kołodziej, Jodłowski P (2013) Far field combined AFM and
micro-Raman imaging for characterisation of surface of structured catalysts: example of
Pd doped CoOx catalysts on precalcined kanthal steel. Top Catal 56:1088–1095
101. Pettinger B, Domke KF, Zhang D, Schuster R, Ertl G (2007) Direct monitoring of plasmon
resonances in a tip-surface gap of varying width. Phys Rev B Condens Matter Mater Phys
76:1–4
102. Bailo E, Deckert V (2008) Tip-enhanced Raman scattering. Chem Soc Rev 37:921–930
103. Yeo BS, Stadler J, Schmid T, Zenobi R (2009) Tip-enhanced Raman spectroscopy—its
status, challenges and future directions. Chem Phys Lett 472:1–13
104. Chan KLA, Kazarian SG (2011) Tip-enhanced Raman mapping with top-illumination AFM.
Nanotechnology 22:175701
105. Diskus M, Nilsen O, Fjellvåg H (2012) Combination of characterization techniques for

atomic layer deposition MoO3 coatings: From the amorphous to the orthorhombic a-MoO3
crystalline phase. J Vac Sci Technol A Vacuum, Surf Film 30:01A107
106. van Schrojenstein Lantman EM, Deckert-Gaudig T, Mank AJG, Deckert V,
Weckhuysen BM (2012) Catalytic processes monitored at the nanoscale with
tip-enhanced Raman spectroscopy. Nat Nanotechnol 7:583–586
107. Berweger S, Raschke MB (2010) Signal limitations in tip-enhanced Raman scattering: The
challenge to become a routine analytical technique. Anal Bioanal Chem 396:115–123
108. Novotny L, Stranick SJ (2006) Near-field optical microscopy and spectroscopy with pointed
probes. Annu Rev Phys Chem 57:303–331
109. Łojewska J, Knapik A, Jodłowski P, Łojewski T, Kołodziej A (2013) Topography and
morphology of multicomponent catalytic materials based on Co, Ce and Pd oxides deposited
on metallic structured carriers studied by AFM/Raman interlaced microscopes. Catal Today
216:11–17
110. Harvey CE, van Schrojenstein Lantman EM, Mank AJG, Weckhuysen BM (2012) An
integrated AFM-Raman instrument for studying heterogeneous catalytic systems: a first
showcase. Chem Commun 48:1742–1744
Chapter 12
Application of Spectroscopic Methods
in the Studies of Polysiloxanes,
Cubic Oligomeric Silsesquioxanes,
and Spherosilicates Modified by Organic
Functional Groups via Hydrosilylation
Aleksandra Chechelska-Noworyta, Jan Mrówka,

Maria Owińska and Magdalena Hasik
Abstract Organosilicon compounds whose structure is based on the stable Si–O

bonds show high chemical and thermal stability. Incorporation of organic groups into
their molecules results in the materials which combine the properties of the Si–O
skeleton and those of organic moieties. Hydrosilylation, i.e., catalytic addition of
Si–H groups to carbon–carbon double bonds, is a convenient route to prepare
organofunctional silicon compounds. In this chapter, the work that has been done
since the year 2000 on functionalization via hydrosilylation of polysiloxanes, cubic
oligomeric silsesquioxanes, and spherosilicates, i.e., the main classes of organosilicon
compounds containing Si–O bonds, is reviewed. The emphasis is put on spectroscopic
investigations of the functionalization process or its products.
12.1 Introduction
Organosilicon oligomers and polymers whose structure is based on Si–O bonds are
a unique class of chemical compounds. They show a number of advantageous
properties, such as high chemical and thermal stability or bioinertness. An attractive
feature is that such compounds of various molecular architectures can be prepared.
Oligo-/polysiloxanes and oligo-/polysilsesquioxanes are the main classes of
organosilicon compounds containing Si–O bonds. The molecules of the former can
be linear or cyclic, and their backbone is composed of repeating Si–O linkages
(general formula: [–R1R2SiO–]n) [1], whereas in the latter each Si atom is sur-
A. Chechelska-Noworyta J. Mrówka M. Owińska M. Hasik (&)

e-mail: mhasik@agh.edu.pl

362 A. Chechelska-Noworyta et al.
rounded by three O atoms (general formula: [–RSiO3/2–]n) and their molecules can
be of ladder or cage structure [2].
The most important representative of linear polysiloxanes is poly(dimethyl-
siloxane) (PDMS, R1 = R2 = CH3 in the general formula) which nowadays is a
commercial, fabricated on a large industrial scale, widely applied material. PDMS is
characterized by high chain flexibility, very low glass transition temperature, low
dependence of physical properties on temperature, high hydrophobicity; it is used
as an ingredient of waterproof sealants, a range of household goods, cosmetics,
medical products, etc. [3, 4].
Formation of regular, ladder oligo- or polysilsesquioxanes is rather difficult, and
it requires specific synthetic routes [5]. Therefore, these compounds are not fre-
quently investigated. In contrast, oligomeric silsesquioxanes of cage structures, the
so-called polyhedral oligomeric silsesquioxanes (POSS), especially the cubic ones,
can be synthesized in relatively high yields and give rise to a great scientific interest
[6]. In numerous studies, these uniform, nano-sized molecules have been covalently
bound to the main or side chains of or blended with organic polymers giving
composites [7–9]. The presence of POSS in the systems leads to the improved
thermal resistance and mechanical properties of the host materials. It should be
noted that apart from cubic oligomeric silsesquioxanes, compounds of similar cage
structure but containing four O atoms around each Si atom are also the topic of
many investigations. Although sometimes they are classified as POSS, they are in
fact silicates. Hence, correctly they should be named spherosilicates.
The most common substituents attached to Si atoms in organosilicon compounds
containing Si–O bonds are hydrocarbon groups (alkyl, phenyl, vinyl); in some
molecules, there are also Si–H bonds. Out of these moieties, vinyl and Si–H ones
are especially important as they ensure reactivity which allows introducing various
organic groups into the parent compounds and, in this way, altering their chemical
or physical properties. In particular, Si–H groups are necessary reagents in the
process called hydrosilylation which involves their catalytic addition to carbon–
carbon, carbon–heteroatom, or heteroatom–heteroatom multiple bonds [10]. As
discussed in a review article by Boutevin et al. [11], hydrosilylation is the most
often applied reaction in the preparation of side group-functionalized polysiloxanes.
The present chapter reviews the work that has been done since the year 2000 on
the use of hydrosilylation to modify polysiloxanes, cubic oligomeric silsesquiox-
anes, and spherosilicates by organic groups. We limit, however, the discussion to
the addition of Si–H group-containing compounds to the ones with carbon–carbon
double bonds in their structure. The emphasis is put on the implementation of
spectroscopic methods in the studies. Results reported by other researchers are
mainly presented here. Our team has the experience predominantly in the appli-
cation of hydrosilylation in the synthesis of various polysiloxane networks [12–15],
but recently we have started investigations on functionalization of polysiloxanes
with organic compounds by hydrosilylation. Some of their results are included in
the paper.
Organization of the chapter is as follows: First we describe briefly the
hydrosilylation process (Sect. 12.2), then the spectroscopic methods and the ways
12 Application of Spectroscopic Methods in the Studies … 363
of their application in the investigations of functional polysiloxanes, cubic oligo-

meric silsesquioxanes, and spherosilicates prepared by hydrosilylation are presented
(Sect. 12.3). Afterward, the details on incorporation of organic groups into par-
ticular compounds and application of spectroscopic methods in the studies are
discussed (Sects. 12.4 and 12.5). Section 12.6 contains concluding remarks.
We hope that this review will be a useful introduction to the chemistry and
spectroscopic properties of polysiloxanes, cubic oligomeric silsesquioxanes, and
spherosilicates for the readers who begin their work on these fascinating com-
pounds. As the summary of recent results, it may also be of interest to other, more
experienced researchers in this field. To the best of our knowledge, this is the first
such review in the literature.
12.2 Hydrosilylation Process
As has been already stated, hydrosilylation is a catalytic reaction which involves

addition of Si–H groups to carbon–carbon, carbon–heteroatom (oxygen, nitrogen),
or heteroatom–heteroatom (nitrogen–nitrogen, nitrogen–oxygen) multiple bonds.
Depending on the reagents, compounds containing Si–C, Si–O, or Si–N bonds are
formed [10].
The first hydrosilylation process was reported by Sommer et al. [16]. It was the
reaction between trichlorosilane (Cl3SiH) and oct-1-ene carried out with diacetyl
peroxide as catalyst and yielding trichlorooctylsilane (Cl3Si(CH2)7CH3).
At present, hydrosilylation of carbon–carbon multiple bonds is an important
method of generating Si–C bonds used in both laboratory and industry [10, 17–19].
Numerous studies employing various hydrosilanes and various alkenes, alkynes, or
their derivatives show that the process can be initiated in different ways. In par-
ticular, many early and late transition metals or their compounds are efficient cat-
alysts [10, 17–19]. Two platinum compounds, namely H2PtCl6 and Pt2[CH2=CHSi
(CH3)2O]3, should be primarily mentioned here. The former, dissolved in iso-
propanol, was first described as a hydrosilylation catalyst in 1957 [20], the latter,
prepared by treatment of H2PtCl6 with 1,1,3,3-tetramethyl-1,3-divinyldisiloxane—
in 1973 [21]. They are frequently referred to as Speier’s and Karstedt’s catalysts,
respectively. Addition of a small amount of any of these compounds (usually 5–
10 ppm of Pt) to the reaction medium makes it possible to conduct a great number
of hydrosilylation processes smoothly, with high selectivity, in solution or without
solvent. Because of this, Speier’s and Karstedt’s catalysts are very often used.
Addition of a Si–H group to a carbon–carbon multiple bond may result in two
products: Markovnikov and anti-Markovnikov (a and b, respectively, shown in
Fig. 12.1 for hydrosilylation of carbon–carbon double bonds in polysiloxanes with
hydrosiloxanes). In most cases, b adduct is the dominant one.
It should be noted that in the presence of some catalysts or for some compounds
participating in the process, side reactions may take place during hydrosilylation.
Fig. 12.1 Scheme of hydrosilylation of carbon–carbon double bond in a polysiloxane with a

hydrosiloxane
They include migration of the carbon–carbon double bond in the alkene derivative
or dehydrogenation of Si–H groups [17].
Hydrosilylation of carbon–carbon multiple bonds finds a wide range of appli-
cations. This process carried out in the presence of Karstedt’s catalyst is used to
cross-link PDMS with vinyl or Si–H terminal groups to form PDMS gels or
elastomers for which there is a big market [19]. Hydrosilylation of carbon–carbon
triple bonds leads to valuable, unsaturated products which can be further modified
[17]. This reaction is also a convenient way to synthesize polymers and to introduce
organic functional groups into organosilicon compounds [11, 17, 18].
12.3 Spectroscopic Methods Applied in the Studies

of Polysiloxanes, Oligomeric Silsesquioxanes,
and Spherosilicates Modified by Hydrosilylation
The main spectroscopic methods applied in the studies of polysiloxanes, cubic

oligomeric silsesquioxanes, and spherosilicates modified by hydrosilylation are
Fourier Transform Infrared (FTIR) and Nuclear Magnetic Resonance (NMR).
FTIR spectroscopy in mid-infrared range of 400–4000 cm−1, as a technique
sensitive to bonds between atoms, is primarily utilized for qualitative analysis, i.e.,
to identify chemical structure of the hydrosilylation products. The advantage of this
method in the studies of hydrosilylation is that the band corresponding to stretching
vibrations of Si–H groups, located between 2100 and 2200 cm−1 [22], is distinct
and does not interfere with any other bands in the spectra of organosilicon com-
pounds. Lowering in its intensity or its disappearance from the spectrum proves that
the reaction has occurred. The Si–H band is also sometimes used to follow
hydrosilylation quantitatively. Such analysis can be conducted in situ, i.e., during
the reaction [23, 24] or ex situ, i.e., using the spectra of the samples withdrawn from
the reaction medium after certain time periods [12, 13, 25]. In the former case, the
analysis is based on direct measurements of Si–H band integral intensities in the
spectra. In the latter, Si–H band calibration is performed using an external standard
[25], or the ratios of the integral intensities of the Si–H band and the band that stays
constant during the process (an internal standard, usually the band originating from
the asymmetric stretching vibrations of C–H bond in Si–CH3 groups at ca.
1250 cm−1 [22]) are calculated and compared as the reaction progresses [12, 13].
Due to low intensity or interference with other absorption lines, FTIR bands
characteristic for carbon–carbon multiple bonds in hydrocarbon groups connected
to Si atoms (C–H: doublet at 958 and 1006 cm−1, bands at 3016, 3056 cm−1 and
C=C: 1598 cm−1 [22]) are of limited utility in the studies of hydrosilylation. Their
presence in the spectrum of the reaction product shows, however, that conversion of
alkene derivative during the process has not been complete.
Hydrosilylation of carbon–carbon double bonds is accompanied by the forma-
tion of ethylene (–CH2CH2–) linkages between Si atoms (Fig. 12.1). The bands due
to C–H bonds of these moieties are located in FTIR spectra of the reaction products
at 1140, 2850, and 2900 cm−1 [22].
Additionally, in all FTIR spectra of polysiloxanes, oligomeric silsesquioxanes,
and spherosilicates modified by hydrosilylation, a strong band due to Si–O–Si
vibrations in the range of 1000–1200 cm−1 [22] is always visible.
In the studies of hydrosilylation processes or their products, FTIR spectroscopy
is very often complemented by NMR, particularly 1H and 29Si NMR techniques. 1H
NMR spectra, similarly to FTIR, make it possible to follow the progress of
hydrosilylation by monitoring the changes in intensity of the line due to protons of
Si–H groups, located at d = 4.7–4.8 ppm [26]. Completeness of the reaction can be
also established by the disappearance of the signals in the range of d = 5–6 ppm
attributed to protons of vinyl/allyl groups [26]. Moreover, 1H NMR spectra allow
discriminating between b and a adducts formed in hydrosilylation (Fig. 12.1) as
they give rise to the signals at d = 1.63 ppm (Si–CH(CH3) and 1.03 ppm (Si–CH
(CH2)), respectively. Integral intensities of the appropriate lines visible in the 1H
NMR spectra provide information on the degree of the initial compound func-
tionalization or on the selectivity of the hydrosilylation process.
The signals corresponding to the units containing Si–H bond in the 29Si NMR
spectra are observed at d = −35 ppm ([CH3Si(O2)H] units in polysiloxanes) , at
d = −85 ppm ([SiO3H] units in cage silsesquioxanes), and at d = 0.5 ppm ([(CH3)2
SiO(H)] units in spherosilicates) [27]. Formation of the hydrosilylation product is
confirmed by the appearance of the lines at d = −20 to −24 ppm ([CH3Si(O2)CH2]
units in polysiloxanes), d = −65 ppm ([Si(O3)CH2] units in cage silsesquioxanes),
or d = 10–11 ppm ([(CH3)2 Si(O)CH2] units in spherosilicates). 29Si NMR spectra
are also applied to verify if the cage structure of spherosilicates is preserved after
functionalization. Its retention is manifested by the clear line at d = −108–110 ppm
([SiO4] units [27]).
12.4 Polysiloxanes Modified by Organic

Functional Groups
There are three classes of polysiloxanes modified by organic functional groups. In

the first one, organic moieties are attached to Si atoms of the polymer backbone;
these are the so-called side group-functionalized or pendant group polysiloxanes.
The other two classes contain terminal functional groups located either at one or at
both chain ends. Due to limited space, only the first class of modified polysiloxanes
will be considered in this review.
Organic moieties incorporated into polysiloxanes can modify physical, chemical,
or both types of properties inherent to the parent polymers. Epoxy and other
oxygen-containing groups, amine and other nitrogen-containing groups, and
fluoroalkyl groups are the most important substituents that have been introduced to
polysiloxanes by hydrosilylation.
In principle, functionalization by hydrosilylation can be performed using the
polymers that contain either unsaturated hydrocarbon, such as vinyl groups, or H
atoms linked to Si atoms in the siloxane backbone. Survey of the literature shows,
however, that there are just a few examples of polyvinylsiloxanes subjected to this
process [28, 29]. The majority of work is devoted to functionalization of polyhy-
dromethylsiloxane (PHMS) or its copolymers with various fractions of dimethyl-
siloxane units in the chain (PHMS-DMS). This is most probably due commercial
availability and relatively low cost of PHMS and the copolymers, whereas
polyvinylsiloxanes are not commercial products and have to be synthesized.
Structural formulae of polysiloxanes modified by hydrosilylation are presented in
Fig. 12.2.
12.4.1 Modified Polyvinylsiloxanes
The polymers containing one or two vinyl groups at Si atoms in a siloxane chain
(Fig. 12.2, polymers I, II), prepared by anionic ring-opening polymerization
(ROP) of either 1,3,5-trimethyl-1,3,5-trivinylcyclotrisiloxane or 1,3,5,7-
tetramethyl-1,3,5,7-tetravinylcyclotetrasiloxane (Fig. 12.2, polymer I) or hex-
avinylcyclotrisiloxane (Fig. 12.2, polymer II), were hydrosilylated by
3,3,3-trifluoropropyldimethylsilane or 1H,1H,2H,2H-perfluorooctyldimethylsilane
in the presence of Karstedt’s catalyst [28]. Chemical structures of the products
(Fig. 12.3, compounds 1a and 1b) were verified by 1H, 13C, 19F, and 29Si NMR as
well as IR spectroscopies. Although the given spectral data were not discussed in
detail in the paper, based on 1H NMR the authors concluded that upon modification
of both polymers, a significant amount of hydrosilylation a product was formed
Fig. 12.2 Polysiloxanes subjected to modification by hydrosilylation: polyvinylsiloxanes (I–III),

PHMS (III), and PHMS-DMS copolymers (V)
Fig. 12.3 Compounds obtained by hydrosilylation of polyvinylsiloxanes
which was explained by the presence of strongly electronegative fluorine atoms in

the substituents. Integration of the 1H NMR signals, in turn, made it possible to
estimate that about 15% of the Si-vinyl groups in polymer II remained unreacted,
independently of the fluorosilane applied in the modification process. For
3,3,3-trifluoropropyldimethylsilane, this was confirmed by IR spectroscopy: The
spectrum of the hydrosilylation product showed the band at 1597 cm−1 corre-
sponding to the vibrations of C=C bonds.
In another work [29], a block copolymer obtained by sequential anionic ROP of
hexamethylcyclotrisiloxane and tetramethyltetravinylcyclotetrasiloxane (PDMS-b-
PMVS—Fig. 12.2, polymer III) was hydrosilylated by dimethyloctadecylsilane,
triethoxysilane or (1,2)-epoxypropoxypropyldimethylsilane to give octadecylsilyl,
triethoxysilyl, and epoxypropoxysilyl group-modified copolymer, respectively
(Fig. 12.3, compounds 2a, 2b, and 2c). All the reactions were carried out in
chloroform at 60 °C, and Karstedt’s catalyst was applied. The authors claimed that
conventional 1H NMR spectroscopy was not sufficient to analyze the modified
copolymer. Therefore, they used one of the 2D methods, namely diffusion-ordered
spectroscopy (DOSY-NMR). This technique relates chemical shifts of protons with
the translation diffusion coefficient of the molecule that contains them. It was found
that all the proton signals in the NMR spectra of the modified samples exhibited the
same diffusion coefficient which meant that the protons belonged to the same
molecule and thus proved successful PDMS-b-PMVS copolymer functionalization.
12.4.2 Modified PHMS and PHMS-DMS Copolymers
General chemical formulae of PHMS and PHMS-DMS copolymers functionalized

with organic groups by hydrosilylation, discussed in this section, and the moieties
incorporated to their molecules, are presented in Table 12.1.
Table 12.1 General structures of functionalized PHMS (A) and PHMS-DMS (B) copolymers and
substituents (S1–S37) present in their molecules (denoted in the general formulae as R1, R2)
A B
S1 S2 S3
S4 S5 S6
S7 S8 S9
S10 S11 S12
S13 S14 S15
S16 S17 S18
S19 S20 S21
S22 S23 S24
(continued)
Table 12.1 (continued)
S25 S26 S27
S28 S29
S30 S31 S32
S33 S34
S35 S36 S37
12.4.2.1 Polymers with Epoxy Side Groups
The interest in polysiloxanes with incorporated epoxy side groups stems from the
fact that, owing to high reactivity of a 3-membered oxirane ring, they can be quite
easily converted to or combined with other moieties. In the studies, epoxy groups
were introduced into PHMS and its copolymers by hydrosilylation of

1-allyloxy-2,3-epoxypropane (allyl glicydyl ether, AGE). In some cases, this pro-
cess was conducted in the presence of another unsaturated organic compound to
afford a polymer containing two types of pendant groups. Spectroscopic studies of
the functionalized polymers allowed for verification of their chemical structures,
determination of the functionalization degrees or served as the basis for kinetic
investigations.
In several studies, polysiloxanes were modified exclusively by 3-(2,3-
epoxypropoxy)propyl moieties (Table 12.1, S1) originating from AGE and trans-
formed to other ones in the next step. Srividhya et al. [30] and Safa et al. [31]
hydrosilylated AGE with PHMS (toluene, 80–90 °C, Speier’s catalyst). Then the
obtained polymers grafted with epoxy groups were converted to the ones containing
imide [30] or bulky tris(trimethylsilyl)methyl [31] substituents in the reactions with
amine-terminated imides or tris(trimethylsilyl)methyl lithium, respectively. Racles
and Cozan [32] prepared glucose-modified and El-Sukkary et al. [33]—hydrox-
yamide functionalized polysiloxanes by treatment of the epoxy groups from AGE
incorporated into PHMS-DMS copolymers (Karstedt’s catalyst, toluene, 75 or
120 °C) with D-glucose in the presence of a strongly acidic gel-type resin
(Amberlite IR-120 plus) or with the excess of various a,x-alkanediamines followed
by the reactions of the amine-modified polysiloxanes thus generated with d-gluco-
nolactone, respectively. In all investigations, the synthesized materials were exam-
ined by FTIR and 1H NMR spectroscopies, in [30, 32]—additionally by 13C NMR
spectroscopy. FTIR spectra of the polymers modified by AGE moieties showed the
bands at 907 and 1237 cm−1 (symmetric and asymmetric stretching vibrations of
epoxy groups, respectively), 1H NMR spectra—the lines at d = 2.5, 2.7, 3.1 ppm
(protons of epoxy groups), and at d = 0.5, 1.5, 3.4 ppm (protons of silylpropyl
groups), 13C NMR spectra—the signals at d = 44.2 and 50.8 ppm (carbon atoms of
epoxy ring). Spectral features of the epoxy ring disappeared after the next processes
conducted and the new ones, confirming the presence of new moieties in the
polymers, showed up. Final functionalized materials were applied as the components
of cross-linked PDMS-based membranes that exhibited high thermal stability and
high tensile strength [30] or studied as potential surfactants [33].
Kowalewska and Stańczyk [34], Rutnakornpituk [35], and Min et al. [36] pre-
pared polysiloxanes with two types of side groups: tris(trimethylsilyl)
hexyl (Table 12.1, S2) and glycidoxypropyl (Table 12.1, S1) [34], cyanopropyl
(Table 12.1, S3) and glycidoxypropyl (Table 12.1, S1) [35] or dodecyl
(Table 12.1, S4) and glycidoxypropyl (Table 12.1, S1) [36] by hydrosilylating the
mixture of appropriate alkene derivatives (6,6,6-tris(trimethylsilyl)hex-1-ene and
AGE, allyl cyanide and AGE, dodec-1-ene and AGE) with PHMS [34] or
PHMS-DMS copolymers [34–36] using Karstedt’s [34] or Speier’s catalyst [35,
36]. Progress of the hydrosilylation process was monitored by 1H NMR [34] or
FTIR spectroscopy [35]. NMR spectra (1H, 13C and 29Si) proved modification of
the polymers by the moieties originating from the alkene derivatives applied; in the
work by Rutnakornpituk [35], the functionalized polysiloxanes were additionally
characterized by FTIR spectroscopy. 29Si NMR spectra of polysiloxanes with
tris(trimethylsilyl)hexyl side chains contained the line at d = 2.1 ppm corre-

sponding to [CSi(CH3)3] units [34]. 1H NMR revealed that some amounts (ca. 2%
in [34] and 10–15% in [35]) of Markovnikov products were formed in the reactions.
Kowalewska and Stańczyk [34] attributed generation of this adduct to hydrosily-
lation of AGE, whereas Rutnakornpituk [35]—to hydrosilylation of allyl cyanide.
The functionalized polymers prepared by Kowalewska and Stańczyk [34] were
subjected to photocrosslinking in order to obtain solid membranes. 13C NMR
spectra of the cross-linked materials contained no signals assigned to epoxy groups
which meant that they participated in the cross-linking process. Rutnakornpituk
[35] applied polysiloxanes modified by cyanopropyl and glycidoxypropyl pendant
groups as reactive additives in the fabrication of epoxy-novolac resin networks.
Min et al. [36] reacted epoxy groups in their modified polysiloxanes with N-
aminoethylpiperazine which resulted in the polymers functionalized by dodecyl and
piperazine groups (Table 12.1, S4 and S5). They were characterized by FTIR (the
bands at 1260 and 3430 cm−1 due to C–N and N–H vibrations, respectively) and 1H
NMR (signals in the d range of 2.1–3.8 ppm assigned to protons in piperazine
moieties) spectroscopies and used to prepare emulsions from which coatings on
cotton fabrics were fabricated [36].
The already described investigations on epoxy-substituted polysiloxanes were
application-oriented. Spectroscopic methods were employed in these studies to
characterize the synthesized materials. Work by Cancouët et al. [25] demonstrates
that spectroscopy can be a valuable tool to characterize the course of the process of
incorporation of epoxy groups into polysiloxanes as well. These researchers con-
ducted hydrosilylation of AGE by PHMS and PHMS-DMS copolymers of various
compositions at 70 °C, in toluene, with Speier’s catalyst. Kinetic studies were based
on determination of changes in concentrations of Si–H groups using FTIR spectra of
the samples withdrawn from the reaction mixtures after certain time periods (cali-
bration of the Si–H band at 2160 cm−1 with the solutions of 1,1,3,
3-tetramethyldisiloxane in toluene of known concentrations was made first). Results
obtained led to the conclusion that a given [SiO2(CH3)H] unit in a polymer chain is
significantly more reactive when it is a part of larger similar units than when it is
isolated between adjacent dimethyl-substituted units. This finding was corroborated
by the 29Si NMR spectra of the initial PHMS-DMS copolymer and the same
copolymer after incomplete reaction with AGE. They clearly showed that the units
of the [SiO2(CH3)H]3 triad (signal at d = *−35 ppm) were consumed first [25].
12.4.2.2 Polymers with Ester, Polyether, and Other

Oxygen-Containing Side Groups
Oxygen-containing substituents, being polar, can turn hydrophobic polysiloxanes

into more hydrophilic materials or change their other physical properties by
inducing strong intermolecular interactions, absent in non-functionalized polymers.
They can also provide reactivity that allows for, e.g., polymer cross-linking or (in
the case of coatings) ensures its good adhesion to the substrate material. Among
polysiloxanes modified by oxygen-containing side groups, the ones with acrylic,

methacrylic ester or acid and polyether moieties seem to be the most important
because potentially they can be applied as biomaterials, surfactants, or polymer
electrolytes. In the literature there are, however, also examples of polysiloxanes
functionalized with other oxygen-containing pendant groups via hydrosilylation.
In several investigations, acrylic and methacrylic ester or acid moieties were
introduced to PHMS and its copolymers as the only side groups [23, 37–39]. Hao
et al. [23] hydrosilylated allyl methacrylate with PHMS-DMS copolymers, called
by the authors prepolymers, containing various amounts of [Si(CH3)H] units
(60 °C, toluene, Karstedt’s or Speier’s catalyst). Hydrosilylation products
(Table 12.1, polymer B-S6), in the paper referred to as macromonomers, were
polymerized in the presence of a photo-initiator under irradiation of blue light
which resulted in the formation of soft polysiloxane gels. The aim of the work was
to obtain injectable, curable in situ materials suitable for fabrication of artificial,
accommodating intraocular lenses that would mimic the performance of natural
lenses in eyes of a young person. Spectroscopic methods were applied in the studies
to determine the average molecular weight of prepolymers and macromonomers
(based on the end group analysis by 1H NMR), to follow the progress of the
hydrosilylation process (in situ FTIR measurements), to establish the chemical
structure of the prepolymers and macromonomers (FTIR and 1H NMR spectro-
scopies), and to check optical properties of the obtained gels (UV-Vis spec-
troscopy). The real-time 3D FTIR spectra (Fig. 12.4a) showed the decrease in
intensity of the Si–H band at 2152 cm−1 during functionalization of prepolymers.
They also allowed concluding that Karstedt’s catalyst was more efficient in the
processes than Speier’s: modification with methacrylate groups of the copolymer
Fig. 12.4 Reaction of PHMS-DMS copolymers and allyl methacrylate: a in situ 3D FTIR spectra
showing the decrease of the Si–H absorption band at 2152 cm−1 in time; b in situ FTIR profiles of
the reactions carried out with Karstedt’s (solid dots) and Speier’s (circles) catalyst; c 1H NMR
spectra of the starting and functionalized copolymer (prepolymer and macromonomer,
respectively). Reproduced from [23] with permission from Elsevier
containing 1 mol% of Si–H moieties was complete after 80 min independently of

the catalyst applied; however, only with Karstedt’s catalyst their 100% conversion
was attained (Fig. 12.4b). This conclusion was confirmed by 1H NMR spectra
(Fig. 12.4c) which contained no signal corresponding to protons of Si-H groups
(d = 4.7 ppm) when Karstedt’s catalyst was used. Moreover, the spectra proved
incorporation of methacrylic moieties into the macromonomers: They showed the
signals at d = 6.1 and 5.5 ppm (C=CH2), d = 4.1 ppm (OCH2), and d = 1.9 ppm
(CH3). The gels exhibited high optical transparency. According to UV-Vis spectra,
their transmittance in the visible range exceeded 95%. They absorbed UV light
below 310 nm which was advantageous since ensured their built-in UV protection.
Wang et al. [37] subjected PHMS-DMS copolymers (containing 0.18 wt% or
0.50 wt% of Si–H groups) to the reactions with methyl methacrylate (MMA) using
dichloro(dicyclopentadienyl)platinum(II) complex as catalyst and tetrahydrofuran
(THF) as solvent. Polysiloxanes with incorporated side groups originating from
MMA (Table 12.1, polymer B-S7) were obtained in this way. Then, oxidative
polymerization of aniline was performed in their presence to yield electroconduc-
tive, pseudo-graft copolymers in which polyaniline (PANI) and functionalized
polysiloxane were linked by hydrogen bonds between amine/imine groups of PANI
and carbonyl groups of the modified polysiloxane. Chemical structures of the
prepared materials were investigated by FTIR, 1H, and 29Si NMR spectroscopies.
FTIR spectra of the functionalized copolymers contained no band at 2156 cm−1,
whereas a strong band at 1737 cm−1 was visible evidencing complete consumption
of copolymer’s Si–H groups and incorporation of carbonyl moieties, respectively,
upon hydrosilylation. 1H NMR spectra proved formation of b product in the process
(signal at d = 1.2 ppm, SiCH2CH(CH3)COOCH3) and confirmed modification of
the polymers by the units derived from MMA (signals at: d = 2.6 ppm, SiCH2CH
(CH3)COOCH3 and d = 3.7 ppm, COOCH3). 29Si NMR spectra, similarly to
FTIR, contained no signal corresponding to the units with Si–H bond but showed,
like 1H NMR, formation of the expected [OSi(CH3)(CH2)] units (signal at
d = −24 ppm). Additionally, using 1H NMR spectra fractions of MMA units in the
modified siloxane copolymers were calculated.
It should be noted that hydrosilylation of alkyl acrylates and methacrylates, due
to the conjugated nature of their molecules (p–p conjugation), can lead to the
products of 1,2- and 1,4-addition. The first case involves the reaction of Si–H
groups and C=C bonds generating anti-Markovnikov (b) and Markovnikov (a)
products (Fig. 12.1). In the second one ,both C=C and C=O bonds participate in the
process and the unsaturated compound with an internal C=C bond is formed.
Lin et al. [38] found that hydrosilylation of MMA and octadecyl acrylate
(OA) with PHMS-DMS copolymers (H2PtCl6 catalyst, toluene, 100–130 °C) was a
b 1,2-addition. This was manifested by a strong IR band at *1740 cm−1 due to
C=O group in saturated esters and the 1H NMR signals at d = 1.2, 2.6, 3.7 ppm
(assignments as in the work [37]) in the spectra of the obtained compounds. No 1H
NMR lines at d = 1.1 ppm or 1.6 ppm attributed to protons of methyl groups
bonded to CHCOOR in the a 1,2-addition or linked to C=C in the 1,4-addition,
respectively, were observed. Hydrolysis of incorporated MMA moieties
(Table 12.1, S7) resulted in polysiloxanes with carboxyl side groups. This was
demonstrated by significant decrease in the intensity of the band at *1740 cm−1
with simultaneous appearance of the band at 1711 cm−1 and the broad peak at
*3300 cm−1 (carbonyl and OH groups in organic acids, respectively) in the IR
spectra.
Mukbaniani et al. [39] investigated the reactions of acrylic and methacrylic acids
with PHMS catalyzed by H2PtCl6. Based on the analysis of FTIR and NMR (1H,
13
C) spectroscopic data, they concluded that apart from the 1,2- and
1,4-hydrosilylation, dehydrogenative condensation ( Si–H + HO(O)C– ! Si–
O(O)C– + H2) takes place in the systems. According to the results presented in the
paper, reactions between PHMS and acrylic acid involve predominantly
1,2-addition with the formation of anti-Markovnikov (b) product and dehydro-
condensation. Reactions of PHMS with methacrylic acid proceed in more direc-
tions: They lead to 1,2-Markovnikov and anti-Markovnikov, 1,4-addition and
dehydrocondensation products. Some of these compounds undergo further trans-
formations which result in a very complex product mixture [39].
Polysiloxanes modified by other oxygen-containing substituents, namely ester or
benzylamide-terminated long-chain hydrocarbon groups (Table 12.1, S8–S9), were
prepared by the reaction between PHMS-DMS copolymer and methyl undecylenate
or undecylenic acid benzylamide (Karstedt’s catalyst, toluene, 20 °C [40]).
Completeness of the reactions was established by the disappearance of the signals
from the Si–H and CH=CH2 groups in 1H NMR spectra which were also used to
confirm the presence of functional groups in the polymer structures. Their signals
were seen at d = 3.69 ppm (C(O)OCH3, d = 4.47 ppm (C(O)NHCH2Ph),
d = 5.90–6.30 ppm (C(O)NHCH2Ph), and d = 7.25 ppm (C(O)NHCH2Ph). In the
work, the influence of the type and content of pendant groups on the
low-temperature and rheological properties of the functionalized copolymers as
compared to PDMS was established.
In a number of studies, polysiloxanes were modified by polyether side groups.
PHMS-DMS copolymers were applied as hydrosilylating agents for terminal allyl
poly(ethylene oxide) derivatives bearing different groups at the other chain end:
hydroxyl [41, 42], acetate [43], methyl [44–46] (Table 12.1, S10: R=OH, CH3CO,
CH3, respectively). Into PHMS, in turn, functional groups originating from tri
(ethylene oxide) allyl methyl ether (Table 12.1, S10—x = 3, R = CH3) [47],
allyltrimethoxysilane and, subsequently, tetra(ethylene oxide) allyl methyl ether
[48] (Table 12.1, S11 and S10—x = 4, R = CH3) or, simultaneously, poly(ethylene
oxide) allyl methyl ether and N,N-dimethylallylamine (Table 12.1, S10—R = CH3,
S12) [49] were introduced. Processes were conducted with various rhodium cata-
lysts and (for comparison) with Karstedt’s catalyst [41], various Pt catalysts [45],
Speier’s catalyst [42–44, 46], Karstedt’s catalyst [47, 48], and dichloro(dicy-
clopentadienyl)platinum(II) complex, Cp2PtCl2 [49]. Progress of the reactions was
followed by FTIR [41, 44–46] or 1H NMR spectroscopy [48]. Some of the prepared
polymers containing polyether groups were then cross-linked using the incorpo-
rated methoxy groups [48] or in the second hydrosilylation reaction that involved
the Si–H groups remaining in the modified polymer and a,x-diallyl(polyethylene
oxide) [47]. In the work by Wang et al. [42], hydroxyl groups terminating the
polyether moieties were subsequently transformed into the quaternary ammonium
salts (QAS) in the reaction with SOCl2 followed by the one with trimethylamine.
Kang and Fang [49] converted the dimethylamine groups present in their func-
tionalized polymer to QAS by treatment with methyl iodide. Incorporation of
polyether groups into the studied polysiloxanes was corroborated by FTIR (the
band at 1095 cm−1, C–O–C bond stretching [42]), 1H NMR (signal at d = 3.8 ppm,
protons of polyether blocks [43]), and 13C NMR (lines at d = *70 ppm, carbon
atoms in polyether blocks [48]) spectra. The amounts of the polyether moieties in
the systems were calculated based on the 1H NMR spectra [41, 43]. These spectra
were also used to determine the shares of the Markovnikov, anti-Markovnikov
hydrosilylation adducts as well as side products (cis and trans isomers of the
internal alkene derivative derived from the hydrosilylated allyl compound, propanal
resulting from its decomposition) formed during the reactions [41, 45]. Moreover,
FTIR and 1H NMR spectroscopic studies allowed Ganicz et al. [41] to conclude that
only one of the studied rhodium catalysts showed slightly higher activity than
Karstedt’s complex, but it gave significantly more undesired a-adduct. Among the
Pt catalysts investigated by Chung and Kim [45]—Karstedt’s, Speier’s,
Lamoreaux’s, i.e., Pt-octanal/octanol complex, and Pt/C, Speier’s turned out to be
the best. Surface properties [43, 44], ionic conductivity [47–49], and a model drug
(tocopherol) loading capacity [42] of the resultant materials were examined in view
of their potential use as surfactants, polymer electrolytes, and drug carriers,
respectively.
Methoxy-substituted aromatic pendant fragments (Table 12.1, S13—R = CH3,
R3 = H, R4 = H or OCH3) were introduced into PHMS by its reaction with
4-allyl-1-methoxybenzene or 4-allyl-1,2-dimethoxybenzene (Karstedt’s or Speier’s
catalyst, toluene, 60 °C) [50]. FTIR and 1H NMR spectra of the synthesized poly-
mers showed that Karstedt’s catalyst was more efficient: in its presence, all Si–H
groups of the polymer were substituted. The Si–H groups which remained in the
polymer after the process performed using Speier’s catalyst were responsible for
partial gelation of the reaction mixture giving rise to the IR band at 1185 cm−1
characteristic for silsesquioxane bonds. Presence of methoxy-substituted aromatic
fragments in the modified polymer was evidenced by the IR bands at: 835 cm−1
(1,4-disubstituted benzene ring, polymer with S13 in Table 12.1, R3 = R4 = H), 728,
841 cm−1 (1,2,4-trisubstituted benzene ring, polymer with S13 in Table 12.1,
R3 = H, R4 = OCH3), 1505–1600 cm−1, and 3020–3060 cm−1 (vibrations of C=C
and C–H bonds in aromatic ring, respectively). 1H NMR spectra indicated that the
hydrosilylation proceeded according to both Markovnikov and anti-Markovnikov
rules; ratios of the anti-Markovnikov and Markovnikov adducts were equal to 8:1
(polymer with S13 in Table 12.1, R3 = R4 = H) and 6:1 (polymer with S13 in
Table 12.1, R3 = H, R4 = OCH3). Moreover, 1H NMR spectra confirmed the pres-
ence of phenyl rings (multiple signal in the range of d = 6.78–7.29 ppm) and
methoxy groups (signal at d = 3.81 ppm, OCH3) in the polymer structures.
In another work [51], polysiloxanes modified by methoxy- and
hydroxy-substituted aromatic pendant fragments were prepared by the reactions
between PHMS and either exclusively eugenol, i.e., 4-allyl-2-methoxyphenol or

eugenol and vinyltrimethoxysilane mixture (Speier’s catalyst, without solvent,
room temperature). The obtained polymers (with S13 groups in Table 12.1,
R=R4 = H, R3 = OH) were tested as additives for anticorrosion epoxy coatings
which, after curing, exhibited improved thermal stability and hydrophobicity. Full
spectroscopic characterization of the products was carried out using FTIR, 1H, and
13
C NMR. The authors showed that the hydrosilylation reaction proceeded with the
formation of only one, b-addition (anti-Markovnikov) product.
12.4.2.3 Polymers with Fluoroalkyl and Other Fluorine-Containing

Side Groups
Fluoroalkyl-functional polysiloxanes are a class of high-performance materials

which combine unique properties of both polysiloxanes and fluorocarbon polymers.
They can be applied to generate superhydrophobic and icephobic surfaces [52],
coatings resistant to hydrocarbon solvents [53], corrosive acidic and oxidative
environments [54], and antifoaming materials [55].
Hydrosilylation of terminal fluoroalkenes that bear some or no linkers between
the carbon–carbon double bond and fluoroalkyl group can lead to undesired side
products [56]. Successful post-functionalization of siloxane polymers with
long-chain fluoroalkyl [57], fluoroalkylether [24, 58], lithiumsulfonate-terminated
perfluoroether [59], fluoroalkylester [52, 60], fluoroalkylhydroxyl [61, 62] side
groups by hydrosilylation was, however, reported. Since fluorine-containing groups
are unreactive, quite often the second type of organic, reactive side group was
introduced into the modified polymer in the next hydrosilylation process.
Maciejewski et al. [24] functionalized PHMS-DMS copolymers of different
chain lengths and Si–H group contents using hydrosilylation of
1,1,2,2,3,3,4,4-octafluoropentyl allyl ether to yield polymers containing fluoroalk-
ylether side groups (Table 12.1, polymer B-S14). Rhodium siloxide complex [{Rh
(OSiMe3) (cod)}2] or Karstedt’s served as catalysts. FTIR-ATR analysis conducted
in situ during syntheses was used to compare their activity. As judged by the
decrease in the intensity of the IR band at 2180 cm−1, reactions in the presence of
the rhodium compound were faster than when Karstedt’s catalyst was applied. In
the former case, the process carried out at 80 °C in toluene resulted in complete
conversion of Si–H groups within 10–15 min, whereas in the latter, still 62% of Si–
H groups remained after this time. Similarly, according to FTIR, all Si–H groups
reacted in the consecutive hydrosilylation of 1,1,2,2,3,3,4,4-octafluoropentyl allyl
ether and either vinyltrimethoxysilane or AGE with fluoroalkylether-modified
PHMS-DMS copolymers catalyzed by rhodium complex [24]. This was corrobo-
rated by 1H NMR spectra of polysiloxanes containing mixed functional groups
(fluoroalkylether and either trimethoxysilyl or glycidoxypropyl) obtained in this
way. No signals corresponding to protons of Si–H (d = 4.8 ppm) or vinyl/allyl
(d = 5.0–6.0 ppm) groups were observed which showed that also alkene deriva-
tives reacted completely. The presence of fluoroalkylether, glycidoxypropyl, or
trimethoxysilyl moieties in the hydrosilylation products was confirmed by their 1H

and 13C NMR spectra (1H NMR: signals in the range of d = 5.9–6.2 ppm—
CF2H terminal group, d = 2.50–2.80, *3.00 ppm—protons of epoxy groups,
d = 3.47–3.50 ppm—protons of OCH3 group; 13C NMR: d = 105–118 ppm—
carbon atoms adjacent to fluorine atoms, d = 44, 51 ppm—carbon atoms in epoxy
groups, d = 50 ppm OCH3 groups). Formation of the CH2CH2 linkages upon
hydrosilylation was manifested by the appearance of the IR bands in the range of
2800–3000 cm−1 and the 29Si NMR signal at −22 ppm. Coatings fabricated using
copolymers containing mixed groups exhibited good adherence to the glass sub-
strate provided by glycidoxypropyl or trimethoxysilyl moieties and hydrophobic or
superhydrophobic surface properties ensured by fluoroalkylether ones [24].
Consecutive hydrosilylation of the appropriate allyl alkene derivatives
(fluorine-containing compound first) by PHMS (Karstedt’s catalyst, toluene, 80 °C)
made it possible to prepare siloxane polyelectrolytes with other mixed, i.e., tetra-
glyme and lithiumsulfonate-terminated perfluoroether side chains (Table 12.1, S10
—x = 4, R = CH3 and S15) [59]. Significant reduction in the FTIR Si–H band at
2161 cm−1 after the processes indicated high degree of PHMS functionalization;
based on 1H NMR spectra, it was established that about 3% unreacted Si–H groups
were preserved in the materials. 1H NMR spectra were also applied to determine the
ratios of tetraglyme and perfluoroether side chains in the samples. Additionally,
polyelectrolytes prepared in the work were characterized by 13C and 19F NMR
spectroscopies. In their FTIR spectra, characteristic bands in the range of 1200–
1350 and 1080–1200 cm−1 corresponding to CF2 group asymmetric and symmetric
stretching vibrations, respectively, appeared [59].
Fluoroester group-functionalized polysiloxanes (Table 12.1, S16) were prepared
by the reaction between 2,2,3,4,4,4-hexafluorobutyl acrylate (HFBA) and PHMS
(Karstedt’s catalyst, p-xylene, 100–130 °C) [60]. In another work, polysiloxanes
with different grafting densities of fluoroester groups were synthesized by hy-
drosilylation of tridecafluorooctyl methacrylate (H2C=C(CH3)COCH2CH2
(CF2)5CF3) with PHMS-DMS copolymer; they were subsequently reacted with
vinyltriethoxysilane (H2C=CHSi(OCH2CH3)3) to form bifunctional polysiloxanes
(Table 12.1, polymer B-S17 and B-S18) (Speier’s catalyst, trifluorotoluene as
solvent, 80 °C) [52]. FTIR spectra of all fluoroester-modified polymers contained a
characteristic, strong band at 1756 cm−1 [60] or 1730 cm−1 [52] assigned to car-
bonyl (C=O) group stretching vibrations. In the spectrum of the monofunctional
polymer, there was also a maximum at 1180 cm−1 due to the vibrations of the C–F
bond [60]. The Si–H band at *2160 cm−1 either was absent [52] or showed very
low intensity [60] which proved complete or high degree of polymer functional-
ization. These results were in concert with the 1H NMR spectra: No [52] or weak
[60] signal corresponding to the proton of the Si–H group (d = 4.7 ppm) was
observed. Comparison of its integral intensities in the spectra of the initial and
functionalized PHMS allowed finding out that in the monofunctional polymer
grafting ratio of 92.1% was reached [60]. Moreover, 1H NMR spectra confirmed the
presence of both organic groups in the bifunctional polymer—their signals were
seen at d = 4.4 ppm (OCH2CH2CF2), 2.5 ppm (OCH2CH2CF2), d = 0.6 ppm (Si
(CH2)2Si(OCH2CH3)3, and d = 3.8 ppm (Si(CH2)2Si(OCH2CH3)3 [52]. The pre-

pared fluoroester group-modified polysiloxanes were applied to fabricate icephobic
coatings on Al plates [52] or as additives in paint formulations for generation of
antigraffiti coatings on Sn plates [60]. Using X-ray photoelectron spectroscopy
(XPS), it was revealed that in both cases the surface of the cured coatings was
enriched in fluorine which ensured their desired properties.
PHMS functionalized by a perfluoroalkylhydroxyl, namely hexafluoroiso-
propanolyl group (Table 12.1, polymer A-S19), was synthesized via hydrosilylation
of 1,1,1,3,3,3-hexafluoro-2-vinylpropan-2-ol with PHMS (Karstedt’s catalyst,
toluene, room temperature) [61]. In another approach, PHMS was applied as the
hydrosilylating agent for 2-allylphenol (reflux in toluene, Karstedt’s catalyst); then the
polymer with incorporated hydroxyphenyl rings was subjected to the reaction with
hexafluoroacetone which yielded polysiloxane with introduced hexafluoroiso-
propanolyl groups attached to hydroxyphenyl ring (Table 12.1, polymer A-S20) [62].
Structures of the polymer functionalized in both ways were confirmed by FTIR and 1H
NMR spectroscopies. IR bands at 3444 and 1224 cm−1, corresponding to stretching
vibrations of O–H and C–F bonds, respectively, demonstrated the presence of
hexafluoroisopropanolyl moieties, whereas the one at 3026 cm−1 originating from
stretching vibrations of aromatic C–H bonds—the presence of aromatic rings in the
polymer A-S20 (Table 12.1). In neither case, the IR Si–H band at *2100 cm−1 was
observed showing that complete functionalization of PHMS occurred. 1H NMR
spectra contained all the expected signals: at d = 3.76 ppm (OH), d = 1.26 ppm
(SiCH2), d = 1.83 ppm (CCH2), d = 0.17 ppm (SiCH3), and d = 5.11 ppm (protons
of hydroxyphenyl group, the signal seen in the spectrum of polymer A-S20). Sensing
properties of the materials toward organic vapors were evaluated [61, 62].
Photocrosslinkable polysiloxanes bearing either perfluoroalkyl (Table 12.1, S21)
or perfluoroalkylether (Table 12.1, S22) and methacrylate side groups (Table 12.1,
S6) were obtained by a two-step hydrosilylation [57, 58]. First, allylperfluorooctane
[57] or allyl 2-(perfluorohexyl)ethyl ether [58] reacted with PHMS-DMS copoly-
mer; then allyl methacrylate was added to the product formed. Speier’s catalyst was
applied in the reactions which were performed at 90 °C without solvent. IR and 1H
NMR spectra of the polymers functionalized with two organic groups showed no
maxima corresponding to Si–H moieties indicating their complete conversion.
Incorporation of methacrylate groups was manifested by an intense IR band at
*1720 cm−1 and 1H NMR signals at d = 1.9 ppm (C(CH3)=CH2) as well as at
d = 5.5 and d = 6.1 ppm (C(CH3) = CH2). In the 29Si NMR spectra of the polymer
containing perfluoroalkyl groups (Table 12.1, polymer B-S21), the expected
maxima in the range of −23.3 to −20.0 ppm ([Si(O2)C3H6] units formed upon
hydrosilylation) and at 8.0 ppm ([Si(CH3)3O], polymer terminal units) were visible.
A small, unexpected signal at *−60 ppm ([SiO3C] units) was attributed to the
formation of a by-product that resulted from the elimination of propene from the
allyl methacrylate during hydrosilylation reaction [57]. Coatings on poly(vinylidene
fluoride) or poly(ether sulfone), prepared from polymer B-S21, S6 (Table 12.1),
after cross-linking, were evaluated for organic vapor permeation [57]. High UV
transparency and high resistance to YAG laser exposure at 355 nm make the
coatings obtained from polymer B-S6, S22 (Table 12.1) promising for use as
optical fiber gratings [58].
12.4.2.4 Polymers with Nitrogen-Containing Side Groups
Nitrogen-containing organic side groups can add new chemical or physical prop-
erties to polysiloxanes. In particular, amine-functionalized polysiloxanes can act as
ligands for metal ions or particles and thus can be used, e.g., for extraction of toxic
metal ions from wastewater [63], as supports for metallic catalysts [64, 65] or as
stabilizers for metal nanoparticles [66]. After transformation into quaternary
ammonium salts (QAS), they exhibit antimicrobial properties [67].
Unfortunately, direct hydrosilylation of unsaturated amines in the presence of Pt
catalysts may be accompanied by the undesired side processes that result in
polymer cross-linking due to the reaction between Si–H and N–H groups or catalyst
poisoning [68]. These problems may be avoided if protected amines are hydrosi-
lylated. Thus, Binet et al. [69] reacted PHMS-DMS copolymer simultaneously with
N-tertbutoxycarbonyl (t-Boc) allylamine and dodec-1-ene at 70 °C using
(Et2S)2PtCl2 as catalyst and toluene as solvent. The reaction product contained
t-Boc aminopropyl side group. Its deprotection with trimethylsilyliodide (solution
in dichloromethane, strictly anhydrous conditions −10 °C) afforded aminopropyl
and dodecyl side group-modified polysiloxane. The expected structures of the
functionalized polymers before and after deprotection of amine groups were
established by 1H and 13C NMR spectroscopies.
In the literature there are, however, also reports on successful direct hydrosily-
lation of amines. Kanjilal et al. [70] performed the reaction of PHMS with various
amounts of allylcyclohexylamine (H2PtCl6 catalyst, toluene, 70 °C). The authors
found by 1H NMR spectroscopy that it was possible to quantitatively substitute the
targeted amount of Si–H groups in PHMS. 1H NMR spectrum corresponding to the
polymer of the highest functionalization level showed no signals due to protons
attached to carbon atoms of C=C bond (d = *5–6 ppm) or protons of Si–H groups
(d = 4.8 ppm), whereas several lines of chemical shift values below 3 ppm,
attributed to cyclohexylamine moieties (Table 12.1, S23), were visible. Similarly,
in the FTIR spectrum there was no band at *2100 cm−1 (Si–H bond), but there
were bands proving incorporation of cyclohexylamine groups. Using
cyclohexylamine-functionalized polymers, membranes on porous polyethylene
supports were fabricated. They exhibited good separation factors in the pervapo-
rative enrichment of propane-1,3-diol concentration from binary aqueous solutions.
Guerra-Contreras et al. [71] and Lei et al. [72] hydrosilylated N,N-dimethy-
lallylamine (DMAA), i.e., a tertiary amine, with PHMS-DMS copolymers
(Karstedt’s catalyst, anhydrous toluene, 90–100 °C) which afforded
amine-substituted polysiloxanes (Table 12.1, polymer B-S12 and B-S24). The
processes were continued until the Si–H band disappeared from the FTIR spectrum
of the reaction mixture. Then Lei et al. [72] introduced PDMS blocks at both
termini of the modified copolymer to obtain PDMS/amine-modified PHMS-DMS/

PDMS triblock copolymer. Afterward, in both studies the synthesized polysilox-
anes with amine pendant groups were converted to the QAS by treatment with
1-iodooctane (Table 12.1, S25) [71] or benzyl chloride (Table 12.1, S26 and S27)
[72]. 1H NMR spectra revealed that hydrosilylation of DMAA with PHMS-DMS
copolymers led to both b (signals at d = 0.49 ppm, 1.51 ppm, 2.21 ppm) and a
(lines at d = 0.98, 0.99 ppm, 1.25 ppm, 2.26 ppm) addition products [71].
Quaternization of incorporated amine groups was proved by the presence of new
signals at d values above 3 ppm: 3.35 and 3.49 ppm [71] or 3.14 ppm [72].
Guerra-Contreras et al. [71] confirmed the chemical structures of the synthesized
products also by 13C and 29Si NMR spectroscopies, Lei et al. [72]—by FTIR
spectroscopy. The prepared QAS showed good solubility in polar organic solvents,
such as ethanol [71] and high antibacterial activity [72].
Another approach to the preparation of polysiloxanes with introduced QAS
groups was presented by Zhao et al. [73] and Hou et al. [74]. They synthesized
polysiloxanes containing epoxy groups first, then the epoxy ring was opened using
an amine which was quaternized during this reaction or in the next step. Thus, Zhao
et al. [73] hydrosilylated allyl polyether epoxy compound (APEE500) and Hou
et al. [74]—AGE with PHMS-DMS copolymers (Speier’s catalyst, 80–85 °C,
isopropanol [73] or 90 °C, toluene [74]). The obtained polysiloxanes with epox-
yether pendant groups were then reacted with N,N-dimethyldodecylamine
(DMDA)/acetic acid mixture [73] or diethyl amine and, subsequently, benzyl
chloride [74] yielding polymers with QAS moieties (Table 12.1, S28 or S29,
respectively). The presence of epoxy groups in the intermediate products obtained
by Zhao et al. [73] was evidenced by the 1H NMR signals in the d range of 2.56–
3.06 which were replaced by the new signals at d = 3.22 ppm (N+(CH3)2),
3.33 ppm (N+CH2CHOH), and 3.58 ppm (N+CH2CH2) in the spectra of the final
compounds. FTIR spectra showed the bands at 2870 and 1110 cm−1 ascribed to the
vibrations of C–H and C–O bond in the polyether group, respectively. After the
reaction with amine, the IR band at 3412 cm−1, due to the OH group, appeared in
the spectrum [73]. 1H NMR spectra of polysiloxanes with introduced QAS moieties
prepared by Hou et al. [74] contained additionally the signals at d = 4.71 ppm and
7.36–7.53 ppm corresponding to benzyl groups attached to nitrogen atoms.
Polysiloxanes with quaternary ammonium groups were applied to prepare coatings
on fabrics of antimicrobial properties [73] and as cationic surfactants for waterborne
polypropylene oxide emulsions [74].
Among other nitrogen-containing side groups that were used to modify
polysiloxanes, carbazolyl (Table 12.1, S30–S33), phthalocyanine (Table 12.1,
S34), and cyclam (Table 12.1, S35) should be mentioned. Their specific electronic
structures bring about interesting optical properties of these moieties [75, 76]. Thus,
their incorporation into polysiloxanes may lead to the materials suitable for use in
various optical devices.
Nordin et al. [77, 78] prepared polysiloxanes with various amounts of carbazolyl
side groups (Table 12.1, polymer B-S30) by hydrosilylation of 9-vinylcarbazole
with PHMS-DMS copolymer (Karstedt’s catalyst, toluene, 120 °C). This process
was followed by the reaction (conducted at 60 °C) with vinyltrimethoxysilane in

order to introduce methoxy side groups that allowed polymer’s cross-linking in the
next step. Products of the reactions were studied by FTIR-ATR, 1H (1D as well as
2D COSY technique) and 29Si NMR spectroscopies. In the FTIR-ATR spectra,
there was no Si–H absorption band at 2156 cm−1 and new bands at 3056 and
1600 cm−1, representing aromatic units of carbazole, occurred. The band at
2840 cm−1 of C–H vibrations in methoxy groups manifested the incorporation of
trimethoxysilyl moieties into the polymer. Successful hydrosilylation of the two
vinyl compounds was also confirmed by 1H NMR spectra in which the signals due
to protons of C=C bonds (d = 5.0–6.5 ppm) were absent and those attributed to
aromatic protons of carbazolyl units (d = 7.0–8.0 ppm) were visible. A 2D 1H–1H
COSY dataset provided information on mutually coupled protons and thus made
the complete structural characterization of the synthesized compounds possible
[77]. 1H NMR data and 1H-1H COSY correlations are presented in Table 12.2. In
the 29Si NMR spectra, the expected lines were observed: at d = −22.319 ppm
([SiO2C2] unit of the linear polysiloxane), at d = −69.834 ppm ([SiO3C] of the
cross-linked polymer), and at d = 7.119 ppm (terminal [SiOC3] groups of
polysiloxane chains). It was found that the amount of the introduced carbazolyl
moieties as well as the cross-linking degree influenced the values of refractive index
and thermo-optical coefficient of the materials. This may lead in future to fabri-
cation of the compounds exhibiting tunable optical properties suitable for use as
components in polysiloxane-based optical devices.
Hydrosilylation of 9-allylcarbazole [79], 9-allylcarbazole and simultaneously
1-allylindole [80] and 11-allyl-11H-benzo[a]carbazole or 13-allyl-13H-dibenzo[a,i]
carba-zole [81] with PHMS performed in toluene, at 60 or 80 °C, in the presence of
dichloro(dicyclopentadienyl)platinum(II) complex [79, 80] or Speier’s catalyst [81]
yielded polysiloxanes containing carbazolyl (Table 12.1, S30), carbazolyl and
Table 12.2 1H NMR data and 1H–1H COSY correlations of carbazole substituted polysiloxane
prepared by Nordin et al. [77]
1
Structure Proton Chemical Multiplicity H–1H
shift (ppm) COSY
c
d d
c Ha 7.25 Hb
Hb 7.3 m Ha
a
a
Hc 7.1 Hd
b N b Hd 8.0 Hc
e He 4.3 t Hj
CH3 j CH3 CH3 CH3 CH3
H3C Hf 0.5 t Hj
O O g O g O g O CH3
g Si
Si Si Si Si Si g Hg 0–0.2 m Hi, Hj
CH3 CH3
g j H CH3 CH3 Hh 3.5 s –
f i
OCH 3 Hi 4.6 s Hg
Si Hj 1.1 t Hf, He,
H3CO OCH 3
h Hg
indolyl (Table 12.1, S30 and S31) or benzo-/dibenzocarbazolyl (Table 12.1, S32
and S33) side groups. Those with incorporated exclusively carbazolyl units were
further functionalized by azo chromophores containing p-sulfonyl or p-nitrophenyl
electron acceptors [79]. Indolyl side groups in the hydrosilylation products were
subsequently modified by azo chromophores with introduced p-sulfonylphenyl
moieties [80]. Products of the hydrosilylation reactions and the final functionalized
compounds were characterized by FTIR and 1H NMR spectroscopies which cor-
roborated their expected chemical structures. Based on 1H NMR spectra, degrees of
polymer functionalization and the amounts of chromophores in carbazolyl/indolyl
moieties were calculated [79, 80]. Additionally, their UV/Vis absorption [80, 81]
and photoluminescence [81] spectra were measured. Polysiloxanes functionalized
by the proposed carbazolyl or indolyl side groups were found to be good candidates
for photorefractive [79, 81] or nonlinear optical [80] applications.
Polysiloxane with phthalocyanine side groups (Table 12.1, S34) was prepared in
the reaction between PHMS and a phthalocyanine derivative containing alkoxy
groups and a single terminal alkenyloxy substituent in the molecule (Karstedt’s
catalyst, toluene, 60 °C [82]). Progress of the PHMS functionalization process was
monitored by IR spectroscopy (disappearance of the Si–H absorption band at
2150 cm−1). As determined by the spectra, the maximum conversion of Si–H
groups equal to 94.4% could be attained in the adopted experimental conditions.
Modification of the polymer by the phthalocyanine group was evidenced by 1H and
13
C NMR spectroscopies which also showed that the b hydrosilylation product was
formed. From thermo-optical measurements, it was concluded that attachment of
phthalocyanine moieties to polysiloxane chain causes slower crystallization of the
compound as compared with the corresponding free phthalocyanine [82].
Chemical membrane sensors for detection of Fe3+ ions in water based on cyclam
(1,4,8,11-tetraazacyclotetradecane)-functionalized polysiloxanes were investigated
by Mefteh et al. [83]. One of the materials studied was prepared by hydrosilylating
N-10-undecenylcyclam (synthesized in the work) with PHMS in a chloroform
solution, at 60 °C using Karstedt’s catalyst. FTIR, 1H, and 13C NMR spectroscopies
proved formation of the expected product (Table 12.1, polymer A-S35). The
presence of azamacrocycles attached to Si atoms was manifested by the IR bands at
1127 cm−1 (stretching C–N vibrations in secondary amines), 1573 and 3185 cm−1
(deformational and stretching, respectively, N–H vibrations in secondary amines),
1
H NMR signals at d = 1.25, 1.4, 1.7, 2.5, 2.55, 2.75, and 2.85 ppm from the
protons of the CH2 groups of azamacrocycle and the undecamethylene spacer, and
at d = 0.5, 1.95, 1.13 ppm assigned to methylene protons in the linker formed upon
hydrosilylation (b adduct), 13C NMR lines at d = 14.57, 18.8, and 31.7 ppm from
carbon atoms of methylene groups. No spectral signatures of the preserved vinyl or
Si–H groups were observed.
Our group performed the reactions of PHMS with N-allylmorpholine (NAM,
Karstedt’s catalyst, solvent-free conditions, room temperature, 24 h) and 4-
vinylpyridine (4VP, Karstedt’s catalyst, toluene, or solvent-free conditions, 60 or
110 °C, 72 h) with the aim to obtain new catalysts or catalytic supports containing
NAM or 4VP moieties (Table 12.1, S36 and S37). Products were solid materials
Fig. 12.5 FTIR (a) and 29Si MAS-NMR (b) spectra of PHMS reacted with N-allylmorpholine
(NAM) or 4-vinylpyridine (4VP). For comparison, FTIR spectrum of PHMS is shown
which indicated that cross-linking of PHMS occurred during the processes. FTIR
spectra (Fig. 12.5a) showed that in both systems Si–H groups were consumed—
decrease in the intensity of the band at *2100 cm−1 with respect to that in the
spectrum of the starting PHMS was observed. However, 29Si MAS-NMR spectra
(Fig. 12.5b) proved unequivocally that hydrosilylation occurred only in the system
with NAM—this was manifested by the signal at d = −20.5 ppm visible exclu-
sively in the spectrum of the product of the reaction between PHMS and NAM.
Additionally, as evidenced by a distinct signal at d =*−65 ppm due to [Si(CH3)
O3], cross-linking took place at the expense of Si–H groups of the polymer units.
This explains lowering in intensity of the Si–H band in the FTIR spectrum of
PHMS treated with 4VP. These results show complementarity of both spectroscopic
methods in the studies of hydrosilylation products.
12.5 Organofunctional Cubic Oligomeric

Silsesquioxanes and Spherosilicates
Hydrosilylation is also a widely used method of introducing organic moieties to

silsesquioxanes and silicates of cubic cage molecular structure. Functionalization of
such molecules results in the materials that can serve as nano-sized building blocks in
the preparation of hybrid organic-inorganic systems.
Similarly to polysiloxanes, cage molecules were modified by organic groups
mainly via reactions between cubic silsesquioxanes or silicates containing Si–H
bonds and unsaturated organic compounds. There are only a few reports on
hydrosilylation of cage molecules with vinyl groups in their structure. Manson et al.
[84] described successful hydrosilylation of octavinyloctasilsesquioxane
(R1 = vinyl group in Fig. 12.6a) with 2-(4-(dimethylsilyl)phenyl)-1,3-dioxolane
and 2-(3-(dimethylsilyl)phenyl)-1,3-dioxolane (Karstedt’s catalyst, diethyl ether,
Fig. 12.6 Structures of cage compounds modified by hydrosilylation: a T8 H (R1 = H), b Q8 M8 H

(R2 = H) and those formed in this process (substituents R1, R2 are given in Table 12.3)
reflux). Products were transformed to octaaldehyde molecules proved to yield imine

moieties-containing ones upon the reaction with an aniline derivative. All the
synthesized compounds were fully characterized by 1H, 13C NMR, and FTIR
spectroscopies [84]. Herrero et al. [85], in turn, prepared dendrimers with intro-
duced 16 or 32 ferrocenyl units by hydrosilylation of octavinyloctasilsesquioxane
or octasilsesquioxane containing 16 vinyl groups (synthesized in the work) with bis
(ferrocenyl)methylsilane (Karstedt’s catalyst, toluene, 75 °C). 1H NMR spectra
evidenced completeness of reactions (no Si-vinyl lines in the d = 5.7–6.1 ppm
region, decrease in the d = 4.92 ppm line of Si–H protons). 1H and 13C NMR
spectroscopies demonstrated that only the anti-Markovnikov (b) addition products
were formed. In the 29Si NMR spectra, all signals could be easily assigned to the
corresponding atoms, and no signals indicating cleavage of the cage structure were
seen. The materials with incorporated ferrocenyl units are electroactive and can be
potentially used as biosensors [85].
In the following parts of this section, studies of octahydrooctasilsesquioxane
(T8 H ) and octakis(dimethylsiloxy)octasilsesquioxane (Q8 M8 H ), i.e., silsesquioxane
and silicate, respectively, containing Si–H groups, modified by organic moieties in
hydrosilylation processes will be reviewed. Chemical structures of the initial and
modified molecules are presented in Fig. 12.6, whereas those of organic sub-
stituents in the modified compounds—in Table 12.3.
12.5.1 T8 H and Q8 M8 H Modified by Epoxy Groups
Liu et al. [86] and Zeng et al. [87] synthesized octa(propylglycidyl ether) func-
tionalized silsesquioxane (Table 12.3, R1 = S1) by the reaction between T8 H and
AGE (Karstedt’s catalyst, toluene, 95 °C). In [87], additionally AGE was
hydrosilylated with the prepared in the work hepta(3,3,3-trifluoropropyl)
Table 12.3 Structures of substituents (R1, R2 in Fig. 12.6) introduced to T8 H and/or Q8 M8 H via
hydrosilylation
Substituents R1, R2
S1 S2 S3
S4 S5 S6
S7 S8 S9
S10 S11 S12
S13 S14 S15
S16 S17 S18
S19 S20 S21
S22 S23
(continued)
Table 12.3 (continued)
S24 S25
S26 S27 S28
hydrooctasilsesquioxane (Karstedt’s catalyst, THF, 60 °C) which generated cage

molecules with introduced seven 3,3,3-trifluoropropyl and one propylglycidyl ether
unit. Formation of the expected products was confirmed by IR (bands at 906 and
1255 cm−1—symmetric and asymmetric stretching vibrations of C–O–C in epoxide
ring, respectively) and NMR (epoxide ring—1H: d = 2.60 and 2.79 ppm, OCH2CH
and d = 3.16 ppm, OCH2CH, 13C: d =44.4 ppm, CH2, d =51.0 ppm, OCH2CH)
spectroscopies. As concluded based on 29Si NMR spectra, reaction between Si–H
groups of T8 H and C=C bond of AGE took place mainly as b-addition with only a
small amount of a-addition product (signals at d = −65.2 ppm and d = −67.6 ppm
of high and low intensity, respectively). The synthesized materials were used to
prepare POSS-containing nanocomposites of epoxy resin [86] or poly(ethylene
imine) hybrids [87].
Szubert et al. [88] tested two rhodium siloxide complexes immobilized on SiO2
as heterogeneous catalysts in the hydrosilylation of AGE, oct-1-ene,
4-allyl-1,2-dimethoxybenzene or allyl 1,1,2,2,3,3,4,4-octafluoropentyl ether with
Q8 M8 H which resulted in spherosilicate modified by the relevant groups
(Table 12.3, S1, S2, S3, and S4). One of the catalysts was also applied in the
consecutive hydrosilylation of two compounds: AGE and oct-1-ene or allyl
1,1,2,2,3,3,4,4-octafluoropentyl ether and AGE with Q8 M8 H leading to bifunctional
(Table 12.3, S1 and S2 or S4 and S1) spherosilicate molecules. Conversion degrees
of Si–H bonds in the systems were determined by FTIR spectroscopy (ATR
technique with a diamond crystal—monitoring the Si–H band at 903 cm−1).
Chemical structures of the hydrosilylation products were established using 1H, 13C,
and 29Si NMR spectroscopies. 13C and 29Si CP/MAS-NMR spectra of the catalysts
allowed concluding that the immobilized Rh complexes were stable after reuse in
ten hydrosilylation cycles.
Bifunctional spherosilicates with incorporated epoxide and methacrylate or
carboxylic acid moieties were prepared by Soh et al. [89] and Kastner et al. [90].
They were obtained by a one-pot process in which Q8 M8 H was treated with an
epoxy group-containing compound first: 4-vinylcyclohexene-1,2-epoxide [89] or

1,2-epoxyhex-5-ene [90]; then to the reaction mixture di(propylene glycol) allyl
ether methacrylate [89] or pent-4-enoic acid or 2-methylpent-4-enoic acid [90] was
added. In this way, the molecules containing S5 and S7 or S6 and S8, S9 groups
(Table 12.3) were obtained. Reactions were conducted in toluene, with dichloro
(dicyclopentadienyl)platinum(II) complex at 55 °C [89] or Speier’s catalyst at 80–
70 °C [90]. For comparison, Soh et al. [89] treated Q8 M8 H exclusively with di
(propylene glycol) allyl ether methacrylate and Kastner et al. [90]—with each
unsaturated compound used in the work. It was revealed by 1H NMR that the
reaction between Q8 M8 H and 1,2-epoxyhex-5-ene was quantitative as the features
corresponding to Si–H and C=C groups were absent in the spectrum of function-
alized spherosilicate (Fig. 12.7a). Progress of all reactions was monitored by the
disappearance of the Si–H IR band at 2140 cm−1 (Fig. 12.7b). Based on spectro-
scopic studies (FTIR, 1H, 13C NMR), it was found that in the hydrosilylation
process, alkene C=C bonds of 4-vinylcyclohexene-1,2-epoxide and di(propylene
glycol) allyl ether methacrylate were involved [89]. Moreover, analysis of 1H and
13
C NMR spectra showed that hydrosilylation of 4-vinylcyclohexene-1,2-epoxide
resulted in a mixture of b and a products in the ratio of ca. 50:50 [89]. In contrast,
selective formation of b adduct in the hydrosilylation of 1,2-epoxyhex-5-ene and
significant degree of dehydrocondensation of carboxylic acids upon reaction with
Q8 M8 H (FTIR and 1H spectra—presence of C=C bonds) was observed [90]. In both
investigations, 29Si NMR spectra confirmed that hydrosilylation occurred (signal at
d = 13.0 ppm) and that the cage molecules stayed intact in this process (signal at
d = −108.8 ppm). The hybrid materials obtained were intended for use in dentistry
[89] or as stamps in microcontact printing [90].
Fig. 12.7 ATR-IR spectra of a Q8 M8 H (denoted as octasilane POSS), 1,2-epoxyhex-5-ene and

the product of their reaction (denoted as epoxy-POSS) and b disappearance of the Si–H vibration
band during a 2.5 h hydrosilylation reaction of 1,2-epoxyhex-5-ene with Q8 M8 H with a ATR-IR
measurement taken every 10 min. Reproduced from [90] with permission from Elsevier
12.5.2 T8 H and Q8 M8 H Modified by Other

Oxygen-Containing Groups
Similarly to polysiloxanes (Sect. 12.4.2.2), cubic cage molecules were function-

alized with substituents containing hydroxyl, ester, and polyether groups in their
structure via the appropriate hydrosilylation reactions.
Hydroxyaliphatic moieties (Table 12.3, S10 and S11) were introduced to
Q8 M8 H by direct hydrosilylation of allyl alcohol [91, 92], 2-allyloxyethanol or
trimethylsiloxyprop-2-ene followed by desilylation of the product via the reaction
with methanol [91]. T8 H with hydroxyaliphatic substituents (Table 12.3, S10 and
S12) was obtained by hydrosilylation of allyl alcohol or hex-5-ene-1-ol [92]. 1H
NMR spectra showed that in the hydrosilylation of allyl alcohol with Q8 M8 H
conducted in toluene Karstedt’s was more active than Speier’s catalyst or
(dichlorodicyclopentadienyl)platinum(II) complex (judged by the rate of disap-
pearance of the Si–H signal); it also gave exclusively the product of C=C
hydrosilylation, whereas in the presence of the other catalysts, significant amounts
of the compound resulting from the so-called O-silylation (dehydrogenative con-
densation: –OH + H–Si ! –O–Si + H2), with the C=C bond preserved in its
structure, were formed [91]. Spectroscopic studies confirmed that hydrosilylation of
allyl alcohol, trimethylsiloxyprop-2-ene followed by the reaction with methanol and
2-allyloxyethanol with Q8 M8 H in toluene in the presence of the proper concen-
tration of Karstedt’s complex, led to the expected cage molecules containing eight
hydroxyaliphatic groups (FTIR: 3339 cm−1—OH stretching, 1H NMR: d =3.76
ppm—OH, d =3.54 ppm—CH2–OH, 13C NMR: d = 65.2 ppm—CH2–OH, 29Si
NMR: d =13.9 ppm, d = −108.3 ppm). Moreover, as corroborated by FTIR, 1H,
13
C, and 29Si NMR spectra, product of allyl alcohol hydrosilylation treated with
methacrylolyl chloride transformed to the cage molecules with methacrylate sub-
stituents, an UV-curable precursor for polyester nanocomposites [91]. Kim et al.
[92] characterized the materials prepared from Q8 M8 H (Karstedt’s catalyst,
1,2-dichloroethane as solvent, reflux) and T8 H (Speier’s catalyst, 1,2-dichloroethane
as solvent, reflux) by 1H and 29Si NMR spectroscopies and blended them with
organic polymers: poly(2-methyl-2-oxazaline), poly(N-vinylpyrrolidone), poly(N,
N-dimethylacrylamide). Existence of hydrogen bonds in the composite systems was
evidenced by the shift in the band due to carbonyl groups of the polymers to lower
wavenumbers in the FTIR spectra of the blends as compared to its position in the
spectra of the polymer matrices [92].
In a series of papers [93–99], a group from Taiwan applied Q8 M8 H as a
hydrosilylating agent for 4-acetoxystyrene (Karstedt’s catalyst, toluene, 80 °C).
FTIR spectrum proved that the product (denoted in the work as AS-POSS) con-
tained ester moieties (Table 12.3, S13) (C=O band at 1765 cm−1). Both FTIR and
1
H NMR spectra showed that complete hydrosilylation took place (FTIR: no bands
at 1650 and 2200 cm−1, 1H NMR: no signals at d = 5.8 and 4.7 ppm). According
to 1H NMR, hydrosilylation resulted in a mixture of b and a adducts [93].
AS-POSS was subsequently hydrolyzed to afford cage molecules with incorporated

eight 4-hydroxyphenyl groups (OP-POSS, FTIR: hydrogen-bonded and free OH
bands at 3340 and 3540 cm−1 and no C=O band). The prepared functionalized
POSS were blended with phenolic resins [93, 94], poly(4-acetoxystyrene) and poly
(4-vinylpyridine) [94], poly(ethylene oxide) [95], diblock styrene-4-vinylpyridine
copolymer [96–98], or diblock styrene-2-vinylpyridine, styrene-methyl methacry-
late copolymers [98]. In the studies, conclusions on hydrogen bonds between
polymer matrices and functionalized spherosilicate molecules were drawn based on
positions of C=O and/or OH bands in the 1D FTIR spectra [93, 95–98]; in several
cases [93, 95], they were supported by the 2D FTIR correlation spectra analysis.
Cage molecules with introduced acetoxy groups (Table 12.3, S14) were also
obtained in the reaction of allyl acetate (AA) and T8 H carried out in toluene, at 60 °
C catalyzed by Karstedt’s complex to which, in a series of experiments, triph-
enylphosphine (PPh3) was added [100]. Other ester, namely unsaturated pho-
tocurable propyl cinnamate moieties (Table 12.3, S15), was incorporated to Q8 M8 H
by hydrosilylation of allyl cinnamate performed in toluene, at 70 °C using Pt/C as
catalyst [101]. Star-shaped molecules with polyester arms (Table 12.3, S16) were
synthesized in the reaction between methacrylate-terminated poly(L-lactide)
(mPLLA) and Q8 M8 H conducted in THF, at 60 °C, with Karstedt’s catalyst [102].
NMR (1H, 13C, 29Si) spectra revealed that hydrosilylation of AA with T8 H resulted
in the products which contained, besides acetoxypropyl, also acetoxy, propyl and
allyl units as well as hydrogen atoms [100]. Integration of 1H NMR signals using
biphenyl as an internal standard showed that addition of PPh3 to the reaction
medium increased process selectivity to the desired hydrosilylation product, but
decreased conversion degree of Si–H groups [100]. The expected structure of the
cage molecules with introduced propyl cinnamate groups was established using 1H
and 13C NMR as well as FTIR spectroscopies [101]. Spectral features attributed to
C=C (UV: absorption maximum at 275 nm, FTIR: the band at 1637 cm−1) and
C=O bonds (FTIR: the band at 1716 cm−1) gradually changed upon exposure of the
Fig. 12.8 UV-Vis (a) and FTIR (b) spectra of Q8 M8 H (denoted as POSS) and Q8 M8 H
functionalized with cinnamate moieties before (denoted as pristine) and after UV irradiation with
increasing exposure time. Reproduced from [101] with permission from Elsevier
propyl cinnamate group-containing material to UV radiation (Fig. 12.8) which was

connected with photopolymerization of these moieties [101]. Formation of poly-
lactide arms and completeness of hydrosilylation in the mPLLA-Q8 M8 H system
was verified by FTIR and 1H NMR spectroscopies [102].
It should be noted that Igarashi et al. [100] treated T8 H with allyl ethyl ether
(AEE) and N,N-dimethylallylamine (DMAA) in the same conditions as those in the
reactions of T8 H and AA already described. The molecules modified by groups S17
and S18 (Table 12.3) resulted from these processes. Based on spectroscopic studies,
it was found that addition of PPh3 to the reaction mixture had no effect on selec-
tivity and conversion of Si–H groups in the case of AEE while for DMAA both
parameters decreased.
Spherosilicate/silsesquioxane-polyether hybrids were investigated by Mya et al.
[103], Markovic et al. [104], and Pozza et al. [105]. They were synthesized by
hydrosilylation of allyl poly(ethylene oxide) with Q8 M8 H in toluene using
Karstedt’s catalyst (at 80 °C [103]), Q8 M8 H or T8 H (at room temperature [104]) or
hydrosilylation of a-methoxy-x-undecenyl poly(ethylene oxide) with Q8 M8 H (at
40 °C [105]). Products of the reactions were characterized by FTIR and 1H NMR
[103], FTIR and 1H, 13C, 29Si NMR [104, 105] which confirmed the presence of
polyether units (Table 12.3, S19 or S20) in the systems. Incomplete functional-
ization of Q8 M8 H was found in [103, 105]. Ratios of integral intensities of the 1H
NMR signals at d = *0.6 ppm (protons of CH2 groups in PEO) and at
d = *0.1 ppm (protons of CH3 groups in Q8 M8 H ) showed that the average number
of polyether chains attached to one cubic core was equal to 6 [103] or 7.2 [105].
12.5.3 T8 H and Q8 M8 H Modified by Fluorocarbon Groups
Although cage molecules with introduced fluorocarbon substituents exhibit

attractive properties, reports on their preparation via hydrosilylation are scarce.
Dutkiewicz et al. [106] applied a rhodium siloxide complex as a homogeneous
catalyst in the hydrosilylation of 1,1,2,2,3,3,4,4-octafluoropentyl allyl ether with
Q8 M8 H conducted in toluene at 110 °C—this process led to the molecules with
eight fluorocarbon substituents (Table 12.3, S4). Then, spherosilicates with various
amounts of fluorocarbon and epoxy groups (Table 12.3, S1 and S4) were synthe-
sized by successive hydrosilylation of AGE and 1,1,2,2,3,3,4,4-octafluoropentyl
allyl ether at various ratios with Q8 M8 H in the same conditions. Spectroscopic
analyses (1H, 13C, 29Si NMR) unambiguously corroborated formation of the tar-
geted, octasubstituted cage molecules.
Wang et al. [107] prepared a series of spherosilicates with incorporated fluori-
nated ether, fluorinated oligoether, or perfluoroaromatic units (Table 12.3, S21–
S23) by the reactions between Q8 M8 H and various fluorinated alkene derivatives
which were synthesized in the work (Karstedt’s catalyst, toluene, room tempera-
ture). Progress of the hydrosilylation processes was followed by monitoring the
disappearance of the Si–H maxima in the FTIR and 1H NMR spectra of the reaction
mixtures. 1H, 13C, 29Si, and 19F NMR spectroscopies were applied to characterize
the products. 1H combined with 13C and 29Si NMR spectra confirmed exactly the
eightfold substitution of hydrogen atoms in Q8 M8 H . Moreover, the 1H NMR spectra
allowed to conclude that b adduct was the dominant one in most cases; only
incorporation of perfluoroaromatic units (Table 12.3, S23) resulted in ca. equal
amounts of b and a adducts. Blends of poly(methyl methacrylate) (PMMA) and
fluorinated POSS materials obtained in the studies showed increased hydropho-
bicity as compared to pure PMMA [107].
12.5.4 T8 H and Q8 M8 H Modified by Nitrogen-Containing

Groups
Cage molecules with introduced nitrogen-containing substituents were obtained by

direct hydrosilylation of a nitrogen-containing alkene derivative or indirectly, by
successive chemical transformations in which hydrosilylation was the first step.
Direct hydrosilylation was applied to incorporate amine moieties originating
from N,N-dimethylallylamine (DMAA) (Table 12.3, S18) to Q8 M8 H (Karstedt’s
catalyst, THF, reflux [108, 109]), variously substituted azobenzene units (e.g.,
Table 12.3, S24) to T8 H (dichlorodicyclopentadienyl)platinum(II) complex, diox-
ane, 80 °C [110]) or Q8 M8 H (Karstedt’s catalyst, toluene, 40–80 °C [111–113]),
carbazole groups (Table 12.3, S25) to Q8 M8 H via its reaction with 9-vinylcarbazole
(Karstedt’s catalyst, toluene, room temperature [114]), moieties-containing iso-
cyanate groups (Table 12.3, S26) to Q8 M8 H by its treatment with m-isoprenyl-a,a′-
dimethylbenzyl isocyanate, m-TMI (Karstedt’s catalyst, 80 °C, solvent-free con-
ditions [115]). In most cases, 1H NMR and/or FTIR spectra of hydrosilylation
products showed complete functionalization of Q8 M8 H . Incomplete modification
was found for T8 H [110] and Q8 M8 H in the work by Ledin et al. [112]. 1H NMR
spectra analysis allowed establishing that conducted hydrosilylation reactions were
exclusively b additions [111, 113, 115], but when T8 H was functionalized, sig-
nificant amounts of a addition product were formed [110]. 29Si NMR spectra
measured in several studies [110, 111, 113–115] confirmed that hydrosilylation
occurred and that the cage structure stayed intact during the performed processes.
FTIR (the band at 2243 cm−1 of –N=C=O) and NMR (1H: multiplet at
d = *7.5 ppm of aromatic protons, 13C: signals in the range of d = 121.8–
149.7 ppm—carbon atoms of phenyl rings, d = 25–35 ppm—aliphatic carbon
atoms) spectroscopies showed incorporation of the moieties originating from m-
TMI (Table 12.3, S26) to the spherosilicate [115].
Product of the reaction of Q8 M8 H and DMAA was subsequently converted to
QAS by treatment with alkyl iodides of various chain lengths [108] or long-chain
alkyl iodides, bromides, and chlorides [109]. Quaternization of amine groups
incorporated into spherosilicate molecules was proved by the appearance of new

signals in the 1H NMR spectra (d = 3.3 ppm: –N+–(CH3)2–, d = 3.5 ppm: –N+–
CH2–) and the decrease in the intensity of the line due to the initial amine group
(d = 2.2 ppm: –N(CH3)2). Additionally, based on 1H NMR spectra, degrees of
quaternization were calculated. Antimicrobial properties of aqueous solutions of the
prepared compounds as well as of PDMS coating containing them were tested [108,
109].
Nonlinear optical [110, 113], liquid crystalline [111] properties and photoiso-
merization of the introduced azobenzene moieties [111, 112] of the prepared
materials with azobenzene-based substituents were investigated. Electrochemical
properties of spherosilicate containing carbazole units were evaluated using cyclic
voltammetry, whereas optical properties—by measuring the electronic absorption
and photoluminescence spectra [114]. Spherosilicate modified by isocyanate groups
was applied as the monomer in the synthesis of polyurethane with poly(ethylene
glycol) of various molecular weights and dibutyltin dilaurate serving as the second
monomer and catalyst, respectively [115].
Multistep processes in which hydrosilylation was the first one were used to
incorporate amine groups [116] or adenine, i.e., one of nucleobases, moieties [117]
to Q8 M8 H .
The procedure developed by Schäffer and Kickelbick [116] consisted of three
steps. In the first one, Q8 M8 H was reacted with x-alkenyl bromides of various chain
lengths (Karstedt’s catalyst, toluene, 80 °C). The resultant cage molecules modified
by bromine-terminated alkyl groups (Table 12.3, S27) were then transformed in the
reaction with NaN3 to the ones with terminal azide moieties which—after hydro-
genation over Pd/C in ethanol—gave quantitatively amine functions. All successive
transformations were accompanied by changes in FTIR and 1H NMR spectra
(FTIR: no Si–H band at *2200 cm−1 after hydrosilylation, the band at 2090 cm−1
due to azide groups after the reaction with NaN3 which disappeared after hydro-
genation; 1H NMR: no signals at d = 5.74, 4.89 ppm and at d = 4.7 ppm of vinyl
and Si–H groups, respectively, after hydrosilylation, shifts in the signals assigned to
CH2 unit next to the transformed group—d = 3.41 ppm for CH2Br, d = 3.26 ppm
for CH2N3, d = 2.66 ppm for CH2NH2, Fig. 12.9a). As revealed by FTIR
(Fig. 12.9b) and 29Si NMR spectroscopies (disappearance of the band at 557 cm−1
and the signal at d = −108 ppm, respectively), aminoalkylspherosilicates easily
degraded upon storage with cleavage of their cage structure. They could be,
however, quantitatively converted in situ by treatment with propyl isocyanate to the
hydrogen bond forming spherosilicates with introduced ureido groups. In the
authors’ opinion, these final compounds can interact with appropriate polymers to
yield the materials showing shape memory effect or self-healing properties [116].
Hu et al. [117] reacted Q8 M8 H with vinyl benzyl chloride first (Karstedt’s cat-
alyst, toluene, 80 °C). Then, the generated spherosilicate with incorporated benzyl
chloride moieties (Table 12.3, S28) was treated with adenine (DMF, 70 °C, K2CO3
added) to form adenine-substituted cage molecules. 1H NMR spectra of the reaction
product showed that in the hydrosilylation step, all Si–H groups in Q8 M8 H were
Fig. 12.9 1H NMR spectra of spherosilicate modified by alkylbromide, azide, amine and ureido
groups (a) and FTIR spectra showing degradation of cage structure of the amine-functionalized
spherosilicate upon storage (b). Reproduced from [116] with permission from The Royal Society
of Chemistry
converted to benzyl chloride units (no signal at d = 4.8 ppm, presence of the signal
at d = 4.50 ppm, CH2Cl); the latter were subsequently transformed to
adenine-containing benzyl groups (line at d = 4.31 ppm instead of that at
d = 4.50 ppm). Similar conclusions were drawn based on FTIR and 13C NMR
spectra. Additionally, 1H NMR spectra revealed that hydrosilylation of vinyl benzyl
chloride with Q8 M8 H yielded both b and a adducts in the ratio of 1.64:1 (b:a). The
obtained adenine-containing spherosilicate was applied to prepare nanocomposites
with thymine-functionalized polybenzoxazine. Existence of multiple hydrogen
bonds between adenine and thymine moieties present in the systems was estab-
lished by FTIR spectroscopy.
12.6 Conclusions
Review of the literature shows that spectroscopic methods play an important role in
the studies on modification of polysiloxanes, cubic oligomeric silsesquioxanes,
and spherosilicates by organic functional groups via hydrosilylation.
FTIR and 1H NMR spectroscopies are very often applied to monitor the progress
of the hydrosilylation process by establishing changes in the spectral features
characteristic for the Si–H bonds during the reaction. Based on FTIR and 1H NMR
spectra, conclusions on completeness of functionalization of the starting compounds
are also drawn. Conversion degrees of Si–H bonds within a certain time period,
determined by these spectroscopic methods, allow evaluating activity of a given
catalyst and comparing activities of various catalysts applied in the process. In the
case of heterogeneous catalysts, changes in their activity as well as their stability
after several reaction cycles (i.e., the “lifetime” of the catalyst) can be determined
by spectroscopic investigations.
FTIR, 1H, 13C, and 29Si NMR spectroscopies provide detailed information on
chemical structures of the compounds formed during hydrosilylation. Therefore,
they make it possible to confirm incorporation of organic functional groups into the
initial compound and to find out the main as well as side reaction routes (if the latter
occur). Moreover, spectroscopic data permit revealing decomposition, particularly
cleavage of the cage structure, of reacting or functionalized compounds during or
after hydrosilylation.
It is worth noting that in the studies of polysiloxanes, cubic oligomeric
silsesquioxanes, and spherosilicates modified by organic groups via hydrosilylation,
apart from conventional spectroscopic techniques, modern 2D correlation methods
are utilized. They ensure precise structural characterization of the functionalized
compounds.
Careful analysis of the spectroscopic data may be a means to optimize the
conditions of hydrosilylation processes so that they proceeded smoothly and with
high selectivity to the desired products.
References
1. Noll W (1968) Chemistry and technology of silicones. Academic Press, pp 1–9

2. Baney RH, Cao X (2000) Polysilsesquioxanes. In: Jones RG, Ando W, Chojnowski J
(eds) Silicon-containing polymers. Kluwer Academic Publishers, pp 157–184
3. Kuo ACM (1999) Poly(dimethylsiloxane). In: Polymer data handbook. Oxford University
Press, Inc., pp 411–435
4. Graiver D, Fearon G (2000) Polysiloxanes: directions of applications and perspectives. In:
Jones RG, Ando W, Chojnowski J (eds) Silicon-containing polymers. Kluwer Academic
Publishers, pp 233–243
5. Zhou Q, Yan S, Han CC, Xie P, Zhang R (2008) Promising functional materials based on
ladder polysiloxanes. Adv Mater 20:2970–2976
6. Cordes DB, Lickiss PD, Rataboul F (2010) Recent developments in the chemistry of cubic
polyhedral oligosilsesquioxanes. Chem Rev 110:2081–2173
7. Pielichowski K, Njuguna J, Janowski B, Pielichowski J (2006) Polyhedral oligomeric
silsesquioxanes (POSS)-containing nanohybrid polymers. Adv Polym Sci 201:225–296
8. Kuo SW, Chang F-C (2011) POSS related polymer nanocomposites. Prog Polym Sci
36:1649–1696
9. Zhang W, Müller AHE (2013) Architecture, self-assembly and properties of well-defined
hybrid polymers based on polyhedral oligomeric silsesquioxane (POSS). Prog Polym Sci
38:1121–1162
10. Marciniec B, Guliński J, Urbaniak W, Kornetka ZW (1992) Comprehensive handbook on
hydrosilylation. Pergamon Press
11. Boutevin B, Guida-Pietrasanta F, Ratsimihety A (2000) Side group modified polysiloxanes.
In: Jones RG, Ando W, Chojnowski J (ed) Silicon-containing polymers. Kluwer Academic
Publishers, pp 79–112
12. Nyczyk A, Paluszkiewicz C, Pyda A, Hasik M (2011) Preceramic polysiloxane networks
obtained by hydrosilylation of 1,3,5,7-tetravinyl-1,3,5,7-tetramethylcyclotetrasiloxane.
13. Nyczyk A, Paluszkiewicz C, Hasik M, Cypryk M, Pospiech P (2012) Cross-linking of linear
vinylpolysiloxanes by hydrosilylation—FTIR spectroscopic studies. Vib Spectrosc 59:1–8
14. Nyczyk-Malinowska A, Wójcik-Bania M, Gumuła T, Hasik M, Cypryk M, Olejniczak Z

(2014) New precursors to SiCO ceramics derived from linear poly(vinylsiloxanes) of regular
chain composition. J Europ Ceram Soc 34:889–902
15. Olejarka J, Łącz A, Olejniczak Z, Hasik M (2018) Non-porous and porous materials
prepared by cross-linking of polyhydromethylsiloxane with silazane compounds. Eur
Polym J 99:150–164
16. Sommer LH, Pietrusza EW, Whitmore FC (1947) Peroxide-catalyzed addition of
trichlorosilane to 1-octene. J Am Chem Soc 69:188–188
17. Marciniec B, Maciejewski H, Pawluć P (2017) Hydrosilylation of carbon–carbon multiple
bonds—application in synthesis and materials science. In: Lee VY (ed) Organosilicon
compounds, vol 2. Experiment (physico-chemical studies) and applications. Elsevier,
pp 169–217
18. Marciniec B, Maciejewski H, Pietraszuk C, Pawluć P (2009) Hydrosilylation: a compre-
hensive review on recent advances. Springer Science & Business Media BV
19. Troegel D, Stohrer J (2011) Recent advances and actual challenges in late transition metal
catalyzed hydrosilylation of olefins from an industrial point of view. Coord Chem Rev
255:1440–1459
20. Speier JL, Webster JA, Barnes GH (1957) The addition of silicon hydrides to olefinic double
bonds. Part II. The use of group VIII metal catalysts. J Am Chem Soc 79:974–979
21. Karstedt BD (1973) Platinum complexes of unsaturated siloxanes and platinum containing
organopolysiloxanes. US 3 775 452
22. Smith AL (1960) Infrared spectra-structure correlations for organosilicon compounds.
Spectrochim Acta 16:87–105
23. Hao X, Jeffery JL, Wilkie JS, Meijs GF, Clayton AB, Watling JD, Ho A, Fernandez V,
Acosta C, Yamamoto H, Aly MGM, Parel JM, Hughes TC (2010) Functionalised
polysiloxanes as injectable, in situ curable accommodating intraocular lenses. Biomaterials
31:8153–8163
24. Maciejewski H, Karasiewicz J, Dutkiewicz A, Dutkiewicz M, Dopierała K, Prochaska K
(2014) Synthesis and properties of polysiloxanes containing mixed functional groups. React
Funct Polym 83:144–154
25. Cancouët P, Pernin S, Hélary G, Sauvet G (2000) Functional polysiloxanes. II. Neighboring
effect in the hydrosilylation of poly(hydrogenmethylsiloxane-co-dimethylsiloxane)s by
allylglycidylether. J Polym Sci A 38:837–845
26. Taylor RB, Parbhoo B, Fillmore D (1991) Nuclear magnetic resonance spectroscopy. In:
Smith AL (ed) The analytical chemistry of silicones. Wiley, pp 347–419
27. Marsmann H (1981) 2Si-NMR spectroscopic results. In: Diehl P, Fluck E, Kosfeld R
(eds) NMR basic principles and progress, vol 17. Springer
28. Cai G, Weber WP (2002) Synthesis and chemical modification of poly(divinylsiloxane).
Polymer 43:1753–1759
29. Boehm P, Mondeshki M, Frey H (2012) Polysiloxane-backbone block copolymers in a
one-pot synthesis: a silicone platform for facile functionalization. Macromol Rapid Commun
33:1861–1867
30. Srividhya M, Madhavan K, Reddy BSR (2006) Synthesis of novel soluble poly
(imide-siloxane)s via hydrosilylation: characterization and structure property behavior. Eur
Polym J 42:2743–2754
31. Safa KD, Bahadori A, Tofangdarzadeh S, Nasirtabrizi MH (2008) Trisyl modification of
epoxy- and chloromethyl-polysiloxanes. J Iran Chem Soc 5:37–47
32. Racles C, Cozan V (2011) Synthesis of glucose-modified siloxanes by a simplified
procedure. Soft Materials 10:413–425
33. El-Sukkary MMA, Ismail DA, Rayes SMEl, Saad MA (2013) Synthesis, characterization
and surface properties of amino-glycopolysiloxane. J Ind Eng Chem 20:3342–3348
34. Kowalewska A, Stańczyk WA (2003) Highly thermally resistant UV-curable poly(siloxane)s
bearing bulky substituents. Chem Mater 15:2991–2997
35. Rutnakornpituk M (2005) Modification of epoxy–novolac resins with polysiloxane

containing nitrile functional groups: synthesis and characterization. Eur Polym J 41:1043–
1052
36. Min C, Pu Q, Yang L, Fan H (2014) Synthesis, film morphology, and performance on cotton
substrates of dodecyl/piperazine functional polysiloxane. J Appl Polym Sci 131:40186
37. Wang H, Ni T, Li G, Lu Y (2013) Conducting pseudo graft copolymers based on polyaniline
and carbonyl-functionalized polysiloxanes: preparation and properties. Synth Met 177:52–
59
38. Lin F, Song M, He Z, Zhang T (2008) Synthesis and structural characterization of
methacrylic acid/octadecyl acrylate-graft-poly(methylhydrosiloxane) by hydrosilylation.
J Appl Polym Sci 107:3773–3780
39. Mukbaniani O, Zaikov G, Pirckheliani N, Tatrishvili T, Meladze S (2007) Hydrosilylation
and dehydrocondensation reactions of methylhydridesiloxane to acrylic and methacrylic
acids. J Appl Polym Sci 103:3243–3252
40. Pryakhina TA, Shragin DI, Kononevich YN, Vasi’ev GV, Buzin MI, Papkov VS,
Muzafarov AM (2015) Synthesis, rheological, and thermal properties of polydimethylsilox-
anes modified with long-chain hydrocarbon substituents with polar fragments. Russ Chem
Bull Int Ed 64:605–612
41. Ganicz T, Mizerska U, Moszner M, O’Brien M, Perry R, Stańczyk WA (2004) The
effectiveness of rhodium (I), (II) and (III) complexes as catalysts in hydrosilylation of model
olefin and polyether with triethoxysilane and poly(dimethylsiloxane-co-methylsiloxane).
Appl Catal A 259:49–55
42. Wang Z, Li R, Chen H, Ma F, Zhang X, Cheng Y, Gu X, Qi Z (2017) Cationic amphiphilic
copolymers: synthesis, characterization, self-assembly and drug-loading capacity. Polym Int
66:1199–1205
43. Pricop L, Hamciuc V, Marcu M, Ioanid A, Alazaroaie S (2005) Graft copolymers
polydimethylsiloxane–polyethyleneoxide. Synthesis and characterization. High Perform
Polym 17:303–312
44. Chung DW, Kim TG (2007) Solvent Effect on the hydrosilylation reactions for the synthesis
of polydimethylsiloxane grafted with polyoxyethylene catalyzed by Speier’s catalyst. J Ind
Eng Chem 13:979–984
45. Chung DW, Kim TG (2007) Study on the effect of platinum catalyst for the synthesis of
polydimethylsiloxane grafted with polyoxyethylene. J Ind Eng Chem 7:571–577
46. Chung DW, Lim JC (2009) Study on the effect of structure of polydimethylsiloxane grafted
with polyethylene oxide on surface activities. Colloids Surf A 336:35–40
47. Zhang Z, Sherlock D, West R, West R (2003) Cross-linked network polymer electrolytes
based on a polysiloxane backbone with oligo(oxyethylene) side chains: synthesis and
conductivity. Macromolecules 36:9176–9180
48. Karatas Y, Kaskhedikar N, Burjanadze M, Wiemhöfer HD (2006) Synthesis of cross-linked
comb polysiloxane for polymer electrolyte membranes. Macromol Chem Phys 207:419–425
49. Kang J, Fang S (2002) Synthesis and ionic conductivity of comb-like polysiloxanes with
pendant oligo(oxyethy1ene) side chains and quaternary ammoniurn groups. Polym Bull
49:127–134
50. Mukbaniani O, Titvinidze G, Dundua A, Doroshenko M, Tatrishvili T (2008) Synthesis of
polysiloxanes with pendant methoxy-substituted aromatic fragments. J Appl Polym Sci
107:2567–2571
51. Chen G, Feng J, Qiu W, Zhao Y (2017) Eugenol-modified polysiloxanes as effective
anticorrosion additives for epoxy resin coatings. RSC Adv 7:55967–55976
52. Li X, Li Y, Ren L, Zhu K, Zhao Y, Yuan X (2017) Self-crosslinking coatings of fluorinated
polysiloxanes with enhanced icephobicity. Thin Solid Films 639:113–122
53. Indulekha K, Behera PK, Rajeev RS, Gouri C, Ninan KN (2017) Polyfluoroalkyl siloxanes
with varying trifluoropropyl content: synthesis, characterization and solvent resistance
studies. J Fluorine Chem 200:24–32
54. Indulekha K, Mathew A, Rajeev RS, Ninan KN, Gouri C (2017) A facile route to
fluoroalkylsiloxane polymers having resistance to corrosive acidic conditions: synthesis and
characterization. Eur Polym J 97:94–103
55. Zhan-Xiong L (2008) Preparation, characterization and antifoaming property of fluorosil-
icone oils with fluoroalkyloxypropyl group substitution. e-Polymers 003:1–12
56. Fukurawa Y, Kotera M (2002) Synthesis of fluorosilicone having highly fluorinated alkyl
side chains based on the hydrosilylation of fluorinated olefins with polyhydromethylsiloxane.
J Polym Sci Part A Polym Chem 40:3120–3128
57. Boutevin B, Guida-Pietrasanta F, Ratsimihety A (2000) Synthesis of photocrosslinkable
fluorinated polydimethylsiloxanes: direct introduction of acrylic pendant groups via
hydrosilylation. J Polym Sci A 38:3722–3728
58. Colomines G, Andre S, Andrieu X, Rousseau A, Boutevin B (2003) Synthesis and
characterization of ultraviolet-curable fluorinated polydimethylsiloxanes as ultraviolet-
transparent coatings for optical fiber gratings. J Appl Polym Sci 90:2021–2026
59. Snyder JF, Hutchison JC, Ratner MA, Shriver DF (2003) Synthesis of comb polysiloxane
polyelectrolytes containing oligoether and perfluoroether side chains. Chem Matter
15:4223–4230
60. Liu H, Fu B, Li Y, Shang Q, Xiao G (2013) Antigraffiti polyurethane coating containing
fluorocarbon side chains grafted polymethylsiloxane. J Coat Technol Res 10:361–369
61. Wang Y, Du X, Li Y, Long Y, Qiu D, Tai H, Tang X, Jiang Y (2013) A simple route to
functionalize siloxane polymers for DMMP sensing. J Appl Polym Sci 130:4516–4520
62. Huang J, Jiang Y, Du X, Bi J (2010) A new siloxane polymer for chemical vapor sensor.
Sens Actuators B 146:388–394
63. El-Ashgar NM, El-Nahhal IM, Chehimi MM, Babonneau F, Livage J (2010) Extraction of
Co, Ni, Cu, Zn and Cd ions using 2-aminophenylaminopropylpolysiloxane. Environ Chem
Lett 8:311–316
64. Zawartka W, Pośpiech P, Cypryk M, Trzeciak AM (2015) Palladium supported on
aminopropyl-functionalized polymethylsiloxane microspheres: simple and effective catalyst
for the Suzuki-Miyaura C–C coupling. J Mol Catal A Chem 407:230–235
65. Zawartka W, Pośpiech P, Cypryk M, Trzeciak AM (2016) Carbonylative Suzuki-Miyaura
coupling catalyzed by palladium supported on aminopropyl polymethylsiloxane micro-
spheres under atmospheric pressure of CO. J Mol Catal A Chem 417:76–80
66. Wang Z, Beletskiy EV, Sungsik L, Hou X, Wu Y, Li T, Kung MC, Kung HH (2015)
Amine-functionalized siloxane oligomer facilitated synthesis of subnanometer colloidal Au
particles. J Mater Chem A 3:1743–1751
67. Sauvet G, Fortuniak W, Kazmierski K, Chojnowski J (2003) Amphiphilic block and
statistical siloxane copolymers with antimicrobial activity. J Polym Sci A 41:2939–2948
68. Brook MA (2000) Silicon in organic, organometallic, and polymer chemistry. Wiley,
pp 404–406
69. Binet C, Dumont M, Fitremann J, Gineste S, Laurent E, Marty JD, Mauzac M,
Mingotaud AF, Moukarzel W, Palaprat G, Zadoina L (2008) Hydrosilylation of
polymethylhydrogenosiloxanes in the presence of functional molecules such as amines,
esters or alcohols. In: Ganachaud F, Boileau S, Boury B (eds) Silicon based polymers.
Springer, Dordrecht, pp 135–150
70. Kanjilal B, Noshadi I, McCutcheon JR, Asandei AD, Parnas RS (2015)
Allylcyclohexylamine functionalized siloxane polymer and its phase separated blend as
pervaporation membranes for 1,3-propanediol enrichment from binary aqueous mixtures.
J Membr Sci 486:59–70
71. Guerra-Contreras A, Villegas A, Ramirez-Oliva E, Cervantes J (2015) Characterization and
study of properties in a polar solvent of a functionalized and quaternized poly
(dimethylsiloxane-co-methyl-hydridosiloxane). Silicon 9:525–533
72. Lei Y, Zhou S, Dong C, Zhang A, Lin Y (2018) PDMS tri-block copolymers bearing
quaternary ammonium salts for epidermal antimicrobial agents: synthesis, surface adsorption
and non-skin-penetration. React Funct Polym 124:20–28
73. Zhao J, An QF, Li XQ, Huang LX, Xu X (2015) A comblike polysiloxane with pendant
quaternary ammonium polyether groups: its synthesis, physical properties and antibacterial
performance. J Polym Res 22:174
74. Hou Z, Kan C (2015) Polysiloxanes with quaternary ammonium groups for SPPO aqueous
emulsions. J Surfact Deterg 18:517–522
75. Grazulevicius JV, Strohriegl P, Pielichowski J, Pielichowski K (2003) Carbazole-containing
polymers: synthesis, properties and applications. Prog Polym Sci 28:1297–1353
76. Claessens CG, Hahn U, Torres T (2008) Phthalocyanines: from outstanding electronic
properties to emerging applications. Chem Rec 8:75–97
77. Nordin NH, Othman N, Ahmad Z (2013) Synthesis and thermal stability of cross-linked
carbazole-substituted poly(dimethylsiloxane) for LED encapsulation. Adv Mater Res
747:733–736
78. Nordin NH, Ramli MR, Othman N, Ahmad Z (2015) Synthesis of carbazole-substituted poly
(dimethylsiloxane) and its improved refractive index. J Appl Polym Sci 41654:1–9
79. Hua J, Li Z, Long K, Qin J, Li S, Ye C, Lu Z (2005) Synthesis and second order nonlinear
optical properties of multifunctional polysiloxanes with sulfonyl based chromophores.
J Polym Sci A 43:1317–1324
80. Li Z, Li J, Qin J, Qin J, Ye C (2005) Synthesis and characterization of polysiloxanes
containing carbazolyl and sulfonyl-indole based chromophore as side chains. Polymer
46:363–368
81. Moon IK, Oh JW, Kim N (2008) Synthesis and optical properties of benzocarbazole-
substituted polysiloxanes for polymeric photorefractive materials. J Photochem Photobiol A
194:351–355
82. Ganicz T, Makowski T, Stanczyk WA, Tracz A (2012) Side chain polysiloxanes with
phthalocyanine moieties. Express Polym Lett 6:373–382
83. Mefteh WB, Touzi H, Chevalier Y, Ouada HB, Othmane A, Kalfat R, Jaffrezic-Renault N
(2014) Comparison of polysiloxane films substituted by undecenyl-cyclam and by
naphthyl-cyclam for the design of ISFET devices sensitive to Fe3+ ions. Sens Actuators B
204:723–733
84. Manson BW, Morrison JJ, Coupar I, Jaffrès PA, Morris RE (2001) Synthesis of aldehyde
functionalised polyhedral oligomeric silsesquioxane. J Chem Soc Dalton Trans: 1123–1127
85. Herrero M, Alonso B, Losada J, Garcıaě-Armada P, Casado CM (2012) Ferrocenyl
dendrimers based on octasilsesquioxane cores. Organometallics 31:6344–6350
86. Liu Y, Zheng S, Nie K (2005) Epoxy nanocomposites with octa(propylglycidyl ether)
polyhedral oligomeric silsesquioxane. Polymer 46:12016–12025
87. Zeng K, Liu Y, Zheng S (2008) Poly(ethylene imine) hybrids containing polyhedral
oligomeric silsesquioxanes: Preparation, structure and properties. Eur Polym J 44:
3946–3956
88. Szubert K, Marciniec B, Dutkiewicz M, Potrzebowski MJ, Maciejewski H (2014)
Functionalization of spherosilicates via hydrosilylation catalyzed by well-defined rhodium
siloxide complexes immobilized on silica. J Mol Catal A: Chem 391:150–157
89. Soh MS, Yap AUJ, Sellinger A (2007) Methacrylate and epoxy functionalized nanocom-
posites based on silsesquioxane cores for use in dental applications. Eur Polym J 43:315–327
90. Kastner J, Lorret O, Rank A, Schwarzinger C, Dittert B, Mühlberger M (2015) Nanocontact
printing stamp material via bi-functionalization of polyhedral oligomeric silsesquioxane. Eur
Polym J 65:221–231
91. Zhang C, Laine RM (2000) Hydrosilylation of Allyl Alcohol with [HSiMe2OSiO1.5]8: octa
(3-hydroxypropyldimethylsiloxy)octasilsesquioxane and its octamethacrylate derivative as
potential precursors to hybrid nanocomposites. J Am Chem Soc 122:6979–6988
92. Kim KM, Inakura T, Chujo Y (2001) Organic-inorganic polymer hybrids using
octasilsesquioxanes with hydroxyl groups. Polym Bull 46:351–356
93. Kuo SW, Lin HC, Huang WJ, Huang CF, Chang FC (2006) Hydrogen bonding interactions
and miscibility between phenolic resin and octa(acetoxystyryl) polyhedral oligomeric
silsesquioxane (AS-POSS) nanocomposites. J Pol Sci Pol Phys 44:673–686
94. Lin HC, Kuo SW, Huang CF, Chang FC (2006) Thermal and surface properties of phenolic
nanocomposites containing octaphenol polyhedral oligomeric silsesquioxane. Macromol
Rapid Commun 27:537–541
95. Huang KW, Tsai LW, Kuo SW (2009) Influence of octakis-functionalized polyhedral
oligomeric silsesquioxanes on the physical properties of their polymer nanocomposites.
Polymer 50:4876–4887
96. Lu YS, Kuo SW (2014) Functional groups on POSS nanoparticles influence the
self-assembled structures of diblock copolymer composites. RSC Adv 4:34849–34859
97. Lu YS, Yu CY, Ln YC, Kuo SW (2016) Hydrogen bonding strength of diblock copolymers
affects the self-assembled structures with octa-functionalized phenol POSS nanoparticles.
Soft Matter 12:2288–2300
98. Yu CY, Kuo SW (2018) Phenolic functionality of polyhedral oligomeric silsesquioxane
nanoparticles affects self-assembly supramolecular structures of block copolymer hybrid
complexes. Ind Eng Chem Res 57:2546–2559
99. Sheen YC, Lu CH, Huang CF, Kuo SW, Chang FC (2008) Synthesis and characterization of
amorphous octakis-functionalized polyhedral oligomeric silsesquioxanes for polymer
nanocomposites. Polymer 49:4017–4024
100. Igarashi M, Kobayashi T, Sato K, Ando W, Matsumoto T, Shimada S, Hara M, Uchida H
(2013) Selective hydrosilylation of allyl esters with octahydridosilsesquioxane. J Organomet
Chem 725:54–59
101. Kim Y, Roh J, Kim JH, Kang CM, In-NK Jun B, Lee C, Hwang DH (2013) Photocurable
propyl-cinnamate-functionalized polyhedral oligomeric silsesquioxane as a gate dielectric
for organic thin film transistors. Org Electron 14:2315–2323
102. Zhang X, Wang C, Fang S, Sun J, Li C, Hu Y (2013) Synthesis and characterization of
well-defined star PLLA with a POSS core and their microspheres for controlled release.
Colloid Polym Sci 291:789–803
103. Mya KY, Li X, Chen L, Ni X, Li J, He C (2005) Core-corona structure of cubic
silsesquioxane-poly(ethylene oxide) in aqueous solution: fluorescence, light scattering, and
TEM studies. J Phys Chem B 109:9455–9462
104. Markovic E, Ginic-Markovic M, Clarke S, Matisons J, Hussain M, Simon GP (2007) Poly
(ethylene glycol)-octafunctionalized polyhedral oligomeric silsesquioxane: synthesis and
thermal analysis. Macromolecules 40:2694–2701
105. Pozza GME, Crotty S, Rawiso M, Schubert US, Lutz PJ (2015) Molecular and structural
characterization of hybrid poly(ethylene oxide)—polyhedral oligomeric silesquioxanes
star-shaped macromolecules. J Phys Chem B 119:1669–1680
106. Dutkiewicz M, Maciejewski H, Marciniec B, Karasiewicz J (2011) New fluorocarbofunc-
tional spherosilicates: synthesis and characterization. Organometallics 30:2149–2153
107. Wang X, Ye Q, Song J, Cho CM, Heab C, Xu J (2015) Fluorinated polyhedral oligomeric
silsesquioxanes. RSC Adv 5:4547–4553
108. Majumdar P, Lee E, Gubbins N, Stafslien SJ, Daniels J, Thorson CJ, Chisholm BJ (2009)
Synthesis and antimicrobial activity of quaternary ammonium functionalized POSS
(Q-POSS) and polysiloxane coating containing Q-POSS. Polymer 50:1124–1133
109. Majumdar P, He J, Lee E, Kallam A, Gubbins N, Stafslien SJ, Daniels J, Chisholm BJ
(2010) Antimicrobial activity of polysiloxane coatings containing quaternary
ammonium-functionalized polyhedral oligomeric silsesquioxane. J Coat Technol Res 7
(4):455–467
110. Su X, Guang S, Xu H, Liu X, Wang X, Deng Y, Wang P (2009) Controllable preparation
and optical limiting properties of POSS-based functional hybrid composites with different
molecular architectures. Macromolecules 42:899–8976
111. Białecka-Florjańczyk E, Sołtysiak JT (2011) Synthesis and characterization of liquid
crystalline silsesquioxanes containing azobenzene groups. Mol Cryst Liq Cryst Sci Technol
548:28–36
112. Ledin PA, Tkachenko I, Xu W, Choi I, Shevchenko VV, Tsukruk VV (2014) Star-shaped
molecules with polyhedral oligomeric silsesquioxane core and azobenzene dye arms.
Langmuir 30:8856–8865
113. Miniewicz A, Girones J, Karpinski P, Mossety-Leszczak B, Galina H, Dutkiewicz M (2014)
Photochromic and nonlinear optical properties of azo-functionalized POSS nanoparticles
dispersed in nematic liquid crystals. J Mater Chem C 2:432–440
114. Imae I, Kawakami Y (2005) Development of a novel POSS-based material having carbazole
moiety. In: Proceedings of SPIE, vol 5937
115. Markovic E, Nguyen K, Clarke S, Constantopoulos K, Matisons J, Simon GP (2013)
Synthesis of POSS-polyurethane hybrids using octakis(m-isoprenyl-a,a′-dimethylbenzyli-
socyanato dimethylsiloxy) octasilsesquioxane (Q8M8 TMI) as a crosslinking agent. J Polym
Sci A 51:5038–5045
116. Schäfer S, Kickelbick G (2017) Simple and high yield access to octafunctional azido, amine
and urea group bearing cubic spherosilicates. Dalton Trans 46:221–226
117. Hu WH, Huang KW, Chiou C, Kuo SW (2012) Complementary multiple hydrogen bonding
interactions induce the self-assembly of supramolecular structures from heteronucleobase-
functionalized benzoxazine and polyhedral oligomeric silsesquioxane nanoparticles.
Chapter 13
Spectroscopic Aspects
of Polydimethylsiloxane (PDMS)
Used for Optical Waveguides
Dengke Cai and H. Michael Heise
Abstract Polydimethylsiloxane polymers were studied for their suitability in

optical waveguides. Optical constants of core and clad materials have been mea-
sured within the visible and near-infrared spectral range. The absorption loss of
PDMS in the datacom region of 600–1600 nm is mainly caused by scattering loss
and attenuation by overtone and combination band vibrations of CH containing
groups. Based on observed positions of those fundamental, overtone and combi-
nation bands and their vibrational assignment, normal vibration frequencies and
anharmonicity constants were determined. An empirical correlation between inte-
gral band strength and intrinsic absorption loss was derived based on the vibrational
spectra of the two materials, providing also estimates for deuterated and halo-
genated PDMS materials with even lower absorption losses. Besides intrinsic
material attenuation, also extrinsic losses from waveguide fabrication have been
investigated. Furthermore, multimode waveguides were fabricated with an insertion
loss of 0.05 dB/cm at the most frequently applied data communication wavelength
of 850 nm. Effects of accelerated ageing have been assessed for the optical char-
acteristics of PDMS materials.
13.1 Introduction
Next-generation Internet switches and high-end computers are expected to process

aggregate data rates in the order of Tbit/s. Consequently, the interconnections
between the processing units will have to handle data rates in the order of 10–40
D. Cai
Laboratory of Micro-Structure Technology, Department of Electrical
Engineering and Information Technology, TU Dortmund University,
44227 Dortmund, Germany
H. M. Heise (&)
Interdisciplinary Center for Life Sciences, South-Westphalia University
of Applied Sciences, Frauenstuhlweg 31, 58644 Iserlohn, Germany
e-mail: heise.h@fh-swf.de

402 D. Cai and H. M. Heise
Gbit/s. It is, however, well known from basic physical laws that electrical inter-
connections will suffer from high transmission losses and severe signal integrity
problems at such data rates. In order to overcome the evident high-speed inter-
connection bottleneck, optical interconnects are considered as the preferred option
[1].
In the optical communication field, polymethylmethacrylate (PMMA), poly-
styrene (PS), polycarbonate (PC), epoxy, polyimide, perfluorocyclobutyl (PFCB)-
based polymers, one-component polysiloxane and UV-cured polysiloxane have been
widely used as multimode waveguide materials for short-distance datacom applica-
tions, e.g. in combination with polymer optical fibres (POFs) in filtering, routing
devices and others. However, their implementation in applications for the datacom
market is still critical and restricted due to various factors such as high cost,
shrinkage, thermal and environmental instability, etc., despite the fact that all these
materials do not show the required optical transparency in the datacom and telecom
regions (600–1600 nm) [2–5]. Recently, polysiloxane materials, which were pro-
duced, for example, by addition thermal curing from two-component PDMS,
received much interest for application in waveguide-based devices due to simple and
low cost ( 50 €/L) thermal casting techniques, avoiding by-product formation
(except for the catalyst) and material shrinkage [5–9]. In addition, such
siloxane-based polymers, in which the thermally robust molecular SiO units are the
material backbone granting thermal stability up to 300 °C, show refractive indices
that range from 1.39 to 1.60 for a wavelength of 589 nm (20 °C) [7]. Therefore, these
polymers can be employed as core and cladding materials for optical waveguides.
The bands in the spectral VIS-NIR region are primarily overtones and combinations
of the fundamental bands found in the classical mid-infrared region. Most chemical
bonds have negligible harmonic absorption compared with so-called X–H subgroup
bonds, i.e. C–H, N–H and O–H. Among the latter, the absorption losses from N–H and
O–H subgroups are usually 10–20 times larger than those arising from C–H vibrations
[2, 3], so for the pre-selection of optical siloxane-based polymers, N–H and O–H bonds
should not exist. With the growing interest in two-component siloxane-based materials
with addition thermal curing for optical datacom applications, however, until now only
very few reports have been devoted to the characterization of their optical characteristics
in the visible and near-infrared (VIS-NIR) region.
In this chapter, PDMS materials were selected for this study with regard to their
bandwidth applicability. For this, refractive indices and the absorption as attenua-
tion in dB/cm related to the power budget have been determined. An empirical
equation for estimating the most significant intrinsic absorption losses due to the
corresponding vibration bands was formulated. It provides a relationship between
the integral band strength and the intrinsic absorption loss, thus furnishing us with
estimates of the loss limit in such materials for implementation in datacom and
telecom devices. For such purpose, multimode waveguides were fabricated and
their respective optical insertion loss was measured at 850 nm with tests simulating
also accelerated ageing, as well their thermal stability verified. Tolerances with
regard to mould inner surface roughness and waveguide interlayer from the fabri-
cation have also been investigated.
13 Spectroscopic Aspects of Polydimethylsiloxane (PDMS) Used for … 403
13.2 Experiment
13.2.1 Polymer Sample Preparation
The key functional group of the PDMS polymers is –SiO(CH3)2, which can be
considered as an organically modified quartz; i.e. two oxygen atoms attached to
every silicon atom have been replaced by methyl groups. By this, the
three-dimensional backbone structure of quartz is changed to give linear polymer
molecules. Here, the methyl groups are free to rotate around the –Si–O–Si– chain.
As derivatives, other groups instead of the methyl subunit can be attached to the
silicon-oxygen backbone. The linear silicone polymers can be cross-linked to each
other by different extents, i.e. covalently via groups containing different atoms,
which explains the wide range of their optical, thermal, mechanical and chemical
properties.
For this work, a two-component commercial PDMS polymer called RT 601
(Wacker Chemie, Burghausen, Germany; http://www.wacker.com) was used as
cladding material. In close cooperation with our group, the PDMS core material
was especially developed by Wacker Chemie (provisional product name SLM
77522), with a few methyl groups substituted by phenyl groups for obtaining a
higher refractive index as required for waveguiding. Both materials were prepared
from two components by means of a cross-linking procedure through a hydrosi-
lylation reaction by thermal curing using a platinum catalyst. The addition reaction
occurs mainly on the terminal carbon and is catalysed by organo-metallic Pt colloid
complexes. The mechanism, as illustrated in Fig. 13.1, has been proposed after the
oxidative addition of the SiH groups in the coupling agent by means of the Pt
catalyst with the following hydrogen transfer onto the double bond and a reductive
elimination in the product (for simplification, other Pt ligands and Si substituents
are omitted).
To study the optical properties of amorphous PDMS cladding and core poly-
mers, it was necessary to purify the raw materials before the curing process to avoid
extrinsic losses, e.g. by particle scattering. The base material and coupling agent
were both filtered using a cellulose-mixester membrane filter of pore size of
0.2 µm. After purification, they were mixed in the ratio of 19:1 for the cladding and
9:1 for core materials, respectively, and deaerated through vacuum application.
Subsequently, the mixed polymers were cured in a clean oven for 2 h at 70 °C to
obtain samples for the optical characterization. All experiments were performed
under clean room (class 10) conditions.
13.2.2 Waveguide Fabrication
For waveguide fabrication using the PDMS materials, casting is adopted as the
suitable technology [5, 10, 11]. The first basic prerequisite for the replication
Fig. 13.1 a Scheme of platinum catalysed hydrosilylation (symbol “” represents two methyl
groups and the rest of the polymer chain); b PDMS polymerization scheme based on a
two-component set
process is a robust mould or stamp. Further requirement for optical waveguide

fabrication is an extremely low sidewall surface roughness of the mould or stamp
(<30 nm) to guarantee low scattering losses. Taking advantage of the high
mechanical and chemical resistance, SU-8 photoresist patterns were employed
directly as mould providing high structure fidelity. In this process, no special
anti-adhesion layer was required to separate the PDMS material from the mould.
The fabrication started by filling the waveguide grooves with the liquid core
polymer, which was done using a blading process (see Fig. 13.2a). For reliably
controlled process conditions, a stencil printer was used, which was originally
implemented in surface-mount technology. The blading velocity is motor con-
trolled. The choice of PMMA or polyoxymethylene (POM) as blading material was
guided by the demands of the blade softness, flexibility and long-lasting
hard-wearing. To prevent the core material from creeping into the areas outside
the waveguides, laser cut metal squeegee shims are used during blading. The first
process step is finalized with thermal curing of the core polymer. Finally, the
squeegee shim was removed from the mould. For the second step, liquid cladding
material is poured into the whole mould area (see Fig. 13.2b). After curing at 70 °C
for 30 min, cladding and core were separated from the mould (Fig. 13.2c). This
half-open structure is then filled with a second cladding material (Fig. 13.2d). With
a final curing step at 70 °C for 60 min, the production process of the waveguiding
laminate was terminated, and after this, the optical layer can be packaged.
Fig. 13.2 Schematics of the waveguide fabrication using PDMS
13.2.3 Optical Measurements
Refractive index measurements were carried out using an “Abbe Refractometer”

from Zeiss (Jena, Germany) to determine the PDMS refractive indices by the limit
angle of total reflection at different wavelengths. For this, a sodium emission lamp
with filter at 589 nm and a halogen tungsten lamp with different filters at 515, 632,
852, 1300 and 1550 nm, respectively, under a constant temperature of 20 °C were
used.
Near-infrared and visible spectroscopic measurements were carried out using a
FT-NIR spectrometer Vector 22 N from Bruker Optics (Ettlingen, Germany)
equipped with a Ge diode as detector. Further measurements including the
long-wave NIR region and the visible range were performed with a scanning double
monochromator-based spectrometer Cary 5 from Varian (Darmstadt, Germany).
PDMS samples were prepared with optically flat surfaces, and their thickness varied
between 3 and 30 mm for reaching appropriate band absorbance values measured in
transmission. For FT-NIR measurements, 1500 scans were co-added (spectral
resolution 64 cm−1, Norton–Beer apodization), whereas spectra were recorded
using the Cary spectrometer at a scan speed of 100 nm/min and a slit width of
5 nm. The measurements were carried out in a temperature-controlled environment
of 23 ± 2 °C.
The optical insertion loss of PDMS waveguides was measured at 850 nm, which
is generally used for commercial optical datacom transmission in multimode
waveguides. For this, the instrument OMS150 of the company “Wavetek Wandel
Goltermann” (Eningen u.A., Germany) was used. For our experiments, a dual laser
arrangement with wavelengths of 650 nm and 850 nm and one detector were
applied. The spectral range was independent of the operating temperature with
tolerances of ± 10 nm. The laser radiation was launched into the waveguides
through a butt-in-coupling GI fibre of 50-µm diameter and a 200-µm diameter SI
fibre for butt-out-coupling.
Further optical measurements on materials and waveguides were carried out after
the application of different stress conditions. For reliable fabrication and operation of a
product, it is of crucial importance to test it against relevant environmental parameters.
Based on potential applications of PDMS multimode waveguides, e.g. in splitters and
EOCBs, we combined general PCB processing and quality test methods (e.g.
IPC-TM-650 and IEC 68) together with standard test methods for optical components
in telecommunications (Telcordia Standards GR-1220 and GR-1209). From these
standards, the following main test conditions have been selected:
• Lamination (180 °C for 2 h);
• Soldering process (300 °C for 1 min in soda bath);
• Dry heat: 150 °C for 4 days (depending on the specific applications);
• Damp heat: 40 °C, 93% humidity, 96 h; and
• Thermal shock: −50–125 °C, 30 °C/min, 100 cycles.
13.3 Characterization of the Optical PDMS Materials
13.3.1 Refractive Index and Bandwidth
Guided wave optics confines radiation in the optical waveguides through the
phenomenon of total internal reflection (TIR), where the core material is surrounded
by a cladding material with a lower refractive index. The coupling and propagation
characteristics of the waveguide are thus defined by the core–cladding index
difference.
The refractive index of polymers mainly depends on the molecular polarizability,
the molecular volume and the difference between the used optical wavelength and
the absorption band wavelength of polymers [12]. In terms of the PDMS polymer
described in this chapter, the core material has an intrinsic higher refractive index
than the cladding material through the introduction of phenyl groups that are partly
substituting methyl groups in the side chain of the PDMS backbone. The ratio of
methyl to phenyl units can be varied to obtain a broad refractive index interval.
With the refractive indices measured at several wavelengths, there are various
dispersion formulas available that could be used for the fitting refractive index data,
e.g. equations provided by Hartman, Conrady, Herzberger, Schott Optical Glass,
Inc. and Sellmeier [13]. Some of the formulas, e.g. by Conrady, Schott or
Herzberger, are linear in their fitting constants, which means they can be fitted
easily using a classical least squares procedure, but it should be clear that the fitting
accuracy should be the dominant aspect in choosing the most appropriate dispersion
formula. From this point of view, the Sellmeier equation is especially suited for the
coverage of the refractive index in the wavelength range from the UV through the
visible to the mid-IR range (up to 2.3 µm) due to its high accuracy and applicability
[7]. The Sellmeier formula—see Eq. (13.1)—is derived from the classical disper-
sion theory and allows the description of the progression of refractive index
throughout the total transmission region with one set of parameters and the cal-
culation of accurate interpolated values. Presently, it has been found that a
three-term Sellmeier formula is generally a necessary and sufficient descriptor for
fitting satisfactorily almost all refractive index data of materials in practical use
within their main transparent region:
B 1 k2 B2 k2 B3 k2
n2 ¼ 1 þ þ þ ð13:1Þ
k2 C1 k2 C2 k2 C3
where n is the optical index at wavelength k, and B1, B2 and B3 and C1, C2 and C3
are the constants to be determined by the fitting process.
In Fig. 13.3a, the refractive index measurement results of the cladding and core
PDMS materials with the Sellmeier fitting curves and the calculated NA values
(numerical aperture):
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
NA ¼ n2core n2cladding ð13:2Þ
at different wavelengths are shown, respectively. In addition, by varying the tem-

perature from 20 to 55 °C, the respective thermo-optic coefficients (TOCs) of core
and cladding were obtained and are displayed in Fig. 13.3b, where the core TOC is
−3.59 10−4 K and that of the cladding material is −3.65 10−4 K. The NA
Fig. 13.3 a Wavelength dependence of the refractive indices of the PDMS core and cladding
materials (left ordinate scale) and the numerical aperture calculated for this combination of
materials (right ordinate scale). b Temperature dependence for the refractive indices of both
materials and of the corresponding numerical aperture at 589 nm. c Refractive index and numerical
aperture stability of PDMS multimode waveguides measured at 20 °C (for the core material, a
different charge as used for a and b had been supplied by the manufacturer)
values are quite constant with a rise in temperature, although the temperature is
possibly not stable enough for the PDMS waveguides due to environmental
changes and heating from active elements, e.g. the optical laser and electronics.
In order to test the material refractive indices under different heat stresses sim-
ulating accelerated ageing, the same conditions as used for optical insertion loss
measurements as discussed later had been applied. The refractive indices measured
at 589 nm changed only within the measurement uncertainties (PDMS RT 601
cladding (mixture 19:1) n = 1.4108 ± 1 and PDMS SLM 77522 (mixture 9:1)
n = 1.4405 ± 2). From those measurements, the stability of the numerical aperture
has been calculated, which is also illustrated in Fig. 13.3c.
In optical waveguides, mode dispersion (different propagation constants for each
mode) and material dispersion are the main effects, which limit the bandwidth of the
optical waveguide transmission. Based on the obtained Sellmeier fitting curves, as
shown in Fig. 13.3a, the material dispersion curves of the PDMS cladding and core
were both obtained using the material dispersion equations below (c/n is the fibre
light velocity):
2
nk d n
M ðkÞ ¼ ð13:3Þ
c dk2
As seen in Fig. 13.4, the material dispersion of the core PDMS is −144.9 ps/
(nm km) and −45.01 ps/(nm km) and of the clad material −136.87 ps/(nm km) and
−45.96 ps/(nm km) at 850 and 1300 nm wavelengths, respectively. The latter value
is especially useful for single-mode waveguide application.
For multimode waveguides, apart from the material dispersion, there also exists
mode dispersion, which is the maximal time delay Dtgmax between the longest and
Fig. 13.4 Calculated wavelength-dependent dispersion for PDMS clad and core materials
shortest rays through the core of the PDMS waveguide. This can be expressed by
Eq. 13.4 [14]:
L NA2
Dtgmax ¼ ð13:4Þ
2c ncore
where L is the length of the polymer waveguide layer, NA is the numerical aperture
of the step-index waveguide, c is the light velocity in vacuum, and ncore is the
refractive index of the core material at the work wavelength.
Also, the relation between dispersion and bandwidth may be simplified with
Eq. 13.5 [14]:
B 0:443=Dtgmax ð13:5Þ
According to the Nyquist rule, the maximum data rate is Rmax ¼ 2B log2 V bps,
where B is the bandwidth and V is the number of states encoded. In our case is
V = 2, so Rmax = 2B, with B and Dt substituted from Eqs. 13.5 and 13.4, respec-
tively, leading to:
1:772 c ncore
Rmax ¼ ð13:6Þ
L NA2
For such a PDMS waveguide system, assuming a transmission speed of 10 Gbps

at 850-nm wavelength and 20 °C work temperature, it can be concluded that its
optical pathway should be below 1.14 m, which is the precondition for using such
PDMS polymers for optical waveguides. If longer optical interconnects are
required, the material design can be achieved through varying the ratio of phenyl
and methyl groups in PDMS for decreasing the index difference between core and
cladding material.
13.3.2 Optical Loss Phenomena
In general, all optical waveguide devices need to show low optical loss, in particular
at the major communication wavelengths of 850 and 980 nm as common for dat-
acom application. Normally, five kinds of attenuation phenomena contribute to the
propagation loss in optical polymer materials within the VIS-NIR region: absorp-
tion corresponding to the excitation of outer electrons and additionally by vibra-
tional overtones and combination bands, scatter by material inhomogeneities, pores
and dust, and scatter from waveguide structuring technology leading, e.g. to side-
wall roughness, and loss from the waveguide structure. For obtaining the total loss,
the optical insertion loss has also been measured for the PDMS waveguides.
13.3.2.1 Intrinsic Absorption Loss by Electronic Transitions
In the far ultraviolet region with a wavelength of less than 200 nm, the molecules
with single bonds (predominantly molecular groups with aliphatic hydrogen atoms)
can undergo r ! r* transitions, but this is not of interest to us. In addition, there
are absorptions arising from p ! p* or n ! p* electronic transitions caused by
double-bond structural units, e.g. >C=C< or >C=O. These usually appear in the
ultraviolet wavelength region and influence the propagation loss of optical polymer
materials still in the visible wavelength region through their absorption tails
observed in the interval of 200–400 nm.
At longer optical path length, the influence of material electronic transitions can
have a significant influence on the optical absorption even at wavelengths in the
short-wave NIR. PDMS as studied in this paper belongs to the category of silicone
polymers obtained from two components as starting material with thermal addition
curing. Due to polymerization, the C = C double-bond concentrations have been
significantly decreased. To investigate the influence of the Pt compound on the
optical scatter and absorption loss, a commercial Pt catalyst complex (Catalyst EP
from Wacker Chemie) and a pure filtered PDMS polymer (RT 601, Wacker,
including 30–40 ppm of Pt in the base component), not containing other impurities
like silicone oil and inhibitors, were selected for studying its contribution to the loss
budget.
In Fig. 13.5, the lowest difference spectrum shows the curing effect on pure
PDMS RT 601 at 150 °C with a time difference of 115 min. Furthermore, it was
found that addition of Pt colloids (0.2 wt%) to the pure PDMS RT 601 under the
same curing conditions changed peak intensity at *220 nm with significant
attenuation up to 325 nm (see middle spectrum). A very similar difference spectrum
was obtained for the same mixture with a curing duration of 120 min at 150 °C
Fig. 13.5 Difference absorbance spectra for PDMS prepared under different conditions
versus the spectrum of the pure PDMS RT 601 compound under the same reaction
conditions, which indicates that the addition of the catalyst without reaction
changed the optical characteristics only marginally. Adding a larger amount of Pt
catalyst (1 wt%) to the PDMS RT 601 changes the absorption spectrum after curing
significantly (see top trace in Fig. 13.5) including scattering loss. The PDMS
sample with the increased addition of Pt catalyst shows yellowness and brownness
with time. In order to improve the PDMS optical stability and decrease absorption
and scattering losses, the Pt concentration should be controlled in a preliminary step
before reaching its final applications.
13.3.2.2 Intrinsic Absorption Loss by Molecular Vibrations
Molecular vibration absorption exists due to the higher harmonics (overtones) of

fundamental stretching vibrations and combination bands with corresponding
bending vibrations, which are main contributors to the intrinsic absorption losses as
found in amorphous polymers at longer wavelengths beyond 600 nm, as the
influence due to scattering is not too strong except after further thermal material
treatment. In the following part, the wavelengths with the most significant vibra-
tional absorption and the loss limits in PDMS polymers will be determined.
13.3.2.3 Assignment and Wavenumber Calculation of Vibrational

PDMS Bands
Mid-infrared spectra of PDMS materials have been studied already 50 years ago
[15]. In addition, Bokobza et al. investigated also NIR spectra of such polymers
[16]. Previously, fundamental bands of Si–O and Si–C bond vibrations of
PDMS RT 601 and SLM 77522 (core material), respectively, have been recorded
and partly assigned [17–19]. Overtone absorption bands due to these chemical
substructures can be neglected in the VIS-NIR wavelength region when compared
with vibrations of the C–H group, e.g. as found in methyl, methylene, vinyl or
phenyl groups. NIR spectral assignments of those bands for cladding PDMS have
already been published [18, 19].
The general formula (without degenerate vibrations) for combination and
overtone vibrational energy levels of a polyatomic group (the term scheme
described by the vibration quantum numbers is given in wavenumber units) is [20]:
X 1
XX
1

1

Gðv1 ; v2 ; . . .Þ ¼ mi vi þ þ Xik vi þ vk þ þ... ð13:7Þ
i
2 i ki
2 2
where mi is the corresponding wavenumber of the harmonic symmetric stretching

and bending, asymmetric stretching and deformation fundamental vibrations, etc.
(also called normal vibrations), measured in units of cm−1, and the Xik are their
anharmonicity constants.
In this chapter, the methyl groups were selected for our study, although the
approach for the determination of anharmonicity constants is also valid for other
molecular group vibrations. The NIR spectrum and a band assignment for methyl
group vibrations in RT 601 cladding PDMS have been recently reported by us [18].
However, the anharmonicity constants reported there were reassessed with slightly
different values now for some constants. In normal waveguide systems, especially
for multimode operation, most of the radiation will travel through the core material
attenuated by the optical absorption losses. Therefore, the results from calculations
for the NIR overtone and combination bands of methyl group vibrations are
especially important with PDMS SLM 77522 as core material, but also results for
the cladding material will be listed.
According to the general term scheme as given in Eq. (13.7), the vibrational
energy levels of CH3 group vibrations, taking into account the two stretching and
two deformation vibrations in SLM 77522 PDMS, can be described by:

1 1 1 1
Gðv1 ; v2 ; v3 ; v4 Þ ¼ m1 v1 þ þ m2 v2 þ þ m3 v3 þ þ m4 v4 þ
2 2 2 2
2 2 2
1 1 1 1 2
42:5 v1 þ 26:5 v2 þ 11 v3 þ 14 v4 þ
2 2 2 2

1 1 1 1 1 1
143 v1 þ v2 þ 5 v1 þ v3 þ 4:5 v1 þ v4 þ
2 2 2 2 2 2

1 1 1 1 1 1
10 v2 þ v3 þ 18 v2 þ v4 þ 5 v3 þ v4 þ
2 2 2 2 2 2
ð13:8Þ
The wavenumbers of the four fundamental methyl group vibrations are:

m1 = 2905 cm−1, m2 = 2963 cm−1, m3 = 1261 cm−1 and m4 = 1412 cm−1, respec-
tively; see also reference [19]. Using the combination and overtone band positions,
anharmonicities have been determined as given in Eq. (13.8). The values for
m1 = 3066 cm−1, m2 = 3102 cm−1, m3 = 1295 cm−1 and m4 = 1454 cm−1 are the
harmonic vibration wavenumbers. From the equation above, we can now calculate
the wavenumbers of overtone and combination bands of CH3 group vibrations from
600 to 2400 nm by taking into account the ground state level. The calculated results
with experimentally observed band positions and assignments for core and cladding
materials are summarized in Table 13.1.
Furthermore, the higher overtone and combination bands of other vibrations with
the involvement of CH containing groups in SLM 77522, i.e. methylene, vinyl and
phenyl groups, cannot easily be assigned from its NIR spectrum because of their
weak absorption bands, although their band positions can be found in reference
[20]. Only by methyl group deuteration shifting the methyl group vibrations to
lower frequencies, those overtone and combination bands contributing to the NIR
spectrum can be visualized. The assignment of the band at 850 nm, the most
Table 13.1 Calculated and experimental spectral band positions of CH3 group vibrations in the
VIS-NIR region; s, m and w indicate strong, medium or weak absorption for a given band
Assignment Cladding Core Intensity
k (nm) kexp (nm) k (nm) kexp (nm)
v1 + v3 2402 2402 2402 2402 s
v2 + v3 2372 2372 2372 2372 s
v2 + v4 2295 2295 2295 2295 s
v1 + 2v3 1852 1849 1852 1851 m
v2 + 2v3 1836 1839 1836 1835 m
v2 + v3 + v4 1784 1783 1784 1784 m
2v1 1747 1746 1746 1746 s
v1 + v2 1747 1746 1744 1744 s
2v2 1702 1703 1703 1703 s
v1 + 3v3 1512 1539 1512 1539 m
v2 + 3v3 1502 1526 1503 1526 m
2v1 + v4 1403 1401 1403 1403 m
2v2 + v4 1379 1371 1378 1369 m
3v1 1182 1185 1183 1185 m
3v2 1145 1149 1146 1149 m
2v1 + 3v3 1062 1094 1061 1094 w
3v1 + v3 1030 1019 1029 1019 w
4v1 900 908 901 908 w
2v1 + 2v2 907 897 906 895 w
4v2 867 879 868 877 w
5v1 731 742 731 742 w
3v1 + 2v2 742 730 742 728 w
5v2 700 715 701 715 w
6v1 619 632 620 632 w
frequently used wavelength for datacom transmission in multimode waveguide

systems, cannot be assigned with confidence.
13.3.2.4 Absorption Loss for PDMS Core and Cladding due

to Molecular Vibrations
Estimations of the absorption loss in amorphous acrylic-based polymers have been

carried out by several authors [2–4], providing detailed procedures for such cal-
culations. However, their empirical equations were deduced for acrylic polymers,
using additionally the assumption of same shape of the absorption bands. Thus, it
was our intention to derive an optical loss equation that is applicable to
siloxane-based materials by considering integrated band intensities.
In order to evaluate the optical losses Dmax in terms of dB/cm, the integral band
strengths have to be correlated with the corresponding absorption loss. An equation
for the integral band strength (in cm/mol) for C–X stretching vibrations (X = H, D,
Cl and F, and aromatic C–H) in dependence of the integral band absorbance, as well
as for the ratio of band strengths of overtones to the related fundamental vibration,
was derived by Groh [2]. Under the assumption that the same band shape, but
different half-widths for the calculated overtone vibrations existed, his equations
were used by us, giving the following expression:
q nc EvCX
Dmax ¼ a in ðdB=cmÞ ð13:9Þ
MG D~v E1CH
where a is a constant, q (g/cm−3) is the density of the polymer, MG (g/mol) is the

molecular mass of the polymer repeat unit, nc is the number of C–X bonds in this
repeat unit, D~v is the band half-width in (cm−1), and ECXv /E1
CH
is the ratio of the
band strength of the C–X vibration overtone and that of the fundamental C–H
vibration band.
According to the near-infrared loss spectrum of core PDMS, we selected two
similar shape overtone bands for comparison and calculation, i.e. the 4m1 and 5m1
overtones of the CH3 group vibrations (see Fig. 13.6a, b). In the following, values
for the cladding material are given in brackets (see also Fig. 13.6c, d). The maxi-
mum optical loss of Dmax (4m1) and Dmax (5m1) can be measured from these spectra,
for which values of about 0.601 and 0.049 dB/cm (0.599 and 0.0504 dB/cm) were
found. The estimated band half-widths are 86.07 and 208.6 cm−1, respectively,
from its curve fitting spectra ranging from 600 to 1250 nm as shown in Fig. 13.6a,
b (corresponding half-widths for the cladding: 138.8 and 214 cm−1; see also
Fig. 13.6c, d). The detailed assignment and band half-widths in the wavelengths are
also summarized in Tables 13.2 and 13.3, respectively.
The density of cured core material, PDMS SLM 77522, is q = 0.98 g/cm3,
whilst the density of cured cladding material, PDMS RT 601, is q = 1.02 g/cm3;
the number of C–H bonds is six and the molecular mass of the PDMS repeat unit
74 g/mol. Referring to the corresponding theoretically calculated values of ECH v /
−4
1 for the 4m1 and 5m1 overtones of the CH3 group vibration, i.e. 7.2 10
ECH and
9.1 10−5, respectively, and then substituting these parameters together with the
experimentally derived optical loss values within the last equation, the constant a is
calculated to be 0.906 106 and 1.41 106 (1.40 106 and 1.43 106) pro-
viding an average value of 1.15 106 (1.415 106). The latter value is used for
defining the relationship between the maximum optical loss and the band strength
ratio, which can be summarized as:
q nc EvCH
Dmax ¼ 1:15 106 dB=cm ð13:10Þ
MG Dv E1CH
where the first factor for the cladding material reads 1.415 106.
Fig. 13.6 Band deconvolution by curve fitting of PDMS core (a, b) and cladding (c,
d) absorbance spectra ranging from 600 to 1250 nm (the insets provide the intrinsic loss in dB/
cm for the band at 740 nm)
Table 13.2 Peak report of core material (from 600 to 1250 nm)
Assignment Area FIT Peak centre (nm) Height Half-width (cm−1)
6v1(CH3) 0.47 631 0.010 399.8
5v2(CH3) 0.21 709 0.0056 349.8
5v1(CH3) 1.00 742 0.045 208.6
5v1(CH2) 0.23 807 0.00087 245.4
4v1(Ar) 0.57 853 0.0023 229.5
4v2(CH3) 2.57 877 0.01 239.4
2v1 + 2v2(CH3) 6.94 904 0.030 218.3
4v1(CH3) 3.07 908 0.034 86.1
4v1(CH2) 0.65 918 0.0071 86.2
n.a. 0.64 994 0.0046 132.0
n.a. 0.44 1003 0.0048 85.7
3v1 + v3(CH3) 4.63 1019 0.023 186.0
2v1 + 3v3(CH3) 1.11 1091 0.0070 149.1
3v2(CH3) 36.38 1148 0.19 184.9
3v2(CH2) 46.27 1181 0.24 177.9
3v1(CH3) 45.77 1186 0.49 87.2
3v1(CH2) 6.26 1214 0.055 106.3
Table 13.3 Peak report of cladding material (from 600 to 1250 nm)
Assignment Area FIT Peak centre (nm) Height Half-width (cm−1)
6v1(CH3) 1.79 633 0.0033 514.4
5v1(CH3) 1.34 741 0.0059 214.0
4v1(Ar) 0.37 847 0.0017 202.1
4v2(CH3) 0.69 880 0.0036 178.2
4v1(CH3) 9.88 907 0.067 138.8
n.a. 1.97 963 0.0034 545.7
n.a. 2.18 998 0.0093 220.7
3v1 + v3(CH3) 5.94 1020 0.028 196.6
3v2(CH3) 23.33 1152 0.14 158.4
3v1(Ar) 0.54 1160 0.011 45.0
3v2(CH2) 29.04 1176 0.20 137.0
3v1(CH3) 76.32 1186 0.75 116.1
2v1 + v2(CH3) 2.10 1200 0.45 44.1
3v1(CH2) 8.39 1213 0.072 110.3
To check the equation’s applicability, optical loss due to the 3m1 overtone of the
CH3 group vibration at 1185 nm is measured and identified to have a band
half-width of 87.2 cm−1 (116.1 cm−1) Based on the equation above and the theo-
retical value of ECH CH
v /E1 , the calculated value is 7.10 dB/cm (6.8 dB/cm) compared
with the measurement result of 6.98 dB/cm (6.34 dB/cm). This result means that
the agreement between experiment and theoretical calculation is astonishingly
satisfactory [2].
Additionally, with respect to the PDMS core material, the optical loss at 850 nm
is measured to be 0.019 dB/cm from its loss spectra shown in Fig. 13.7, and the
loss through molecular vibration at 853 nm is measured to be 0.014 dB/cm from
curve fitting of spectra shown in Fig. 13.6b. Whilst for this wavelength the elec-
tronic absorption can be neglected, since it is far away with its absorption bands in
the UV region, we can conclude that the Rayleigh scattering at this wavelength is
around 0.005 dB/cm.
13.3.2.5 Estimation of Absorption Loss for PDMS Derivatives
Based on the measured optical loss and its theoretical calculation, PDMS will be a
very good choice, owing to its low cost, easy mass production and high thermal
stability, for short haul data communication application (from several centimetres to
one metre) like its use in electrical–optical circuit boards (EOCBs) or some
short-distance interconnections. However, once also relatively longer distances
need to be served, PDMS has to be modified for decreasing the relatively high
optical loss from overtone and combination band absorption in the visible/
near-infrared (VIS-NIR) region. Presently, deuteration and halogenation are the
Fig. 13.7 NIR-VIS spectra of core and clad PDMS with optical loss given in dB/cm
main two options to decrease the optical loss, since the CD and CX (halogen
elements, e.g. F, Cl) fundamental absorption bands (in particular, the stretching
vibrations) in the mid-infrared region are shifted to longer wavelengths (m(CD):
2230 cm−1, m(CF): 1250 cm−1 and m(CCl): 770 cm−1). Notably, also smaller
anharmonicity constants exist compared with CH bond vibrations, so that finally a
much lower optical loss will result for the VIS-NIR region.
After obtaining the correlation equation above and in combination with the
detailed calculation results, as previously published for the ratios of ECX CH
v /E1 , we
can roughly estimate also the optical loss in deuterated or halogenated core PDMS
within the VIS-NIR region. For example, 4m1 of the CD stretching vibration is
located at 1160 nm and the ECD CH
4 /E1 ratio is 1.3 10−4, and assuming its band
−1
half-width of 150 cm being similar to that of 4m1 of the C–H bond vibration, we
can find the corresponding CD stretch overtone loss value to be only 0.109 dB/cm.
This must be compared with the 3m1 of the CH stretching vibration located at
1183 nm, as discussed above for the normal isotopic material with its optical loss
found about 60–70 times larger than for the 4m1 CD stretching overtone loss. Whilst
from a theoretical view deuteration and halogenation in siloxane-based materials
are proving their applicability to render even more transparent materials, practical
aspects about the use of such materials must still be investigated.
In addition, in terms of core PDMS material and with respect to the siloxane
polymer design, assuming a ratio of q/MG = 0.01 mol/cm−3 based on possible
deuteration or halogenation and nc = 6, an absorption loss of 0.001 dB/cm could be
achieved when referring to a quotient of ECX v /(E1 D
CH
v) of 1.09 10−8.
Furthermore, specifically to 850 nm, the optical loss at 850 nm of polysiloxanes
will then be mainly from Rayleigh scattering, e.g. 0.005 dB/cm as found for PDMS
SLM core material, if the absorption peaks near 850 nm are shifted due to
deuteration or halogenation. Therefore, the inhomogeneities in polysiloxanes
should be also controlled to obtain a higher optical transparency, especially opti-

mizing Pt catalyst concentration and curing conditions and others. After knowing
the optical loss limit, it is very important and helpful for the optical material
formulation to decrease the loss by appropriate hydrogen substitution, but even
amplification (which can be accomplished by doping the PDMS material with some
rare-earth metal materials showing luminescence at the requested signal wave-
length) may still be necessary.
13.3.2.6 Intrinsic Scattering Loss
A significant loss factor is due to Rayleigh scattering, caused by fluctuations of the

density and the refractive index in such polymer materials. Extrinsic loss arising
from scattering in polymers can also result from unfiltered particles, dust, dissolved
bubbles and non-reacting monomers. To eliminate this source of extrinsic scatter-
ing, it is necessary to follow rigorously ultimate procedures in the preparation of
polymer formulations, e.g. by filtering and deaeration, and to perform all prepa-
rations in a clean room environment.
As scattering often arises from a number of sources, the experimental scattering
data often fit with an empirical law of the form [12]:
ascatter ¼ A þ B=k2 þ C=k4 ð13:11Þ
where A is the contribution from large particle scattering (i.e. with their diame-
ters k), B the contribution from inhomogeneities in the order of k in size (Mie
scattering), and C the contribution from smaller inhomogeneities (size
k,
Rayleigh-like scatter).
The material of RTV 2 PDMS used here is a kind of elastomer polymer with a
very low glass transition temperature (around-100 °C), which means that it is in the
liquid state at the normal operation temperature. In addition, its base components
have been filtered and deaerated before thermal curing, so Rayleigh scattering loss
will be negligible compared with other intrinsic losses, e.g. from vibrational
overtone absorptions beyond 600 nm.
As a second step, RTV 2 PDMS is polymerized through a hydrosilylation
reaction by thermal curing using a platinum catalyst (see also Fig. 13.1), and after
hydrosilylation, the Pt catalyst complex will be a by-product (see Fig. 13.5). The Pt
catalyst complex is possibly bonded with some leftover polar components in cured
PDMS, e.g. vinyl groups as found in the base component to form a more stable and
larger size colloid with increasing temperature and continuous curing. This will
result in PDMS compositional inhomogeneities in even larger length scales that
give rise to optical scattering. In addition to absorption effects (see next subsection)
within the UV/VIS range, certainly a loss contribution from scatter exists.
13.3.2.7 Scattering Loss from Mould Roughness
Besides the intrinsic scattering losses, core–clad interface roughness is another

source of extrinsic loss. Actually, except scattering loss, roughness and inhomo-
geneity can also generate unwanted mode coupling and modulation of mode cou-
pling in devices and beating/coupling length. Thus, the extrinsic loss in a
waveguide system typically results from core–clad interface roughness, which
needs to be analysed. Referring to the casting method for PDMS waveguide fab-
rication, actually the interface roughness is the copy of grooves and defects from the
sidewalls of the casting moulds. Additional attenuation (scattering loss), which is
from continuous coupling from bound modes to the radiation field, may be caused
due to these surface defects and roughness. A white-light interferometer, model
Veeco NT 1000, was used to quantify the roughness of our casting moulds (see
Fig. 13.8a), and scanning electron microscopy (SEM) was adopted for direct
topography observation (Fig. 13.8b).
From the NIR spectra of core and cladding PDMS materials, the optical losses at
850 nm were measured and calculated based on Lambert law. Optical intrinsic loss
of PDMS core is 0.0192 dB/cm, and the respective cladding loss is 0.0176 dB/cm;
the core material has a relatively higher loss than the cladding. For such a
cross-section PDMS waveguide system (refractive index of cladding and core:
1.4041 and 1.4275 at 589 nm and 20 °C, respectively), we simplify the calculation
of the waveguide loss attributed to PDMS. It is considered that 10% of the radiation
will be transmitted in the cladding layer, whilst for the other 90% the waveguide
loss is from core material. Hence, we can assume a waveguide loss of 0.0190 dB/
cm.
Different qualified SU-8 casting moulds had been fabricated from different
masks, which directly determinate the sidewall surface qualities of moulds. Their
roughness was quantified with the Veeco NT 1000 device, where ten spots were
Fig. 13.8 a Roughness of the casting mould sidewall; b topography of the casting mould sidewall
obtained by scanning electron microscopy (SEM). Reprinted with permission from [21], copyright
(2010) IEEE
Fig. 13.9 Relationship between sidewall roughness and optical waveguide loss at 850 nm.
Adapted with permission from [21], copyright (2010) IEEE
measured per waveguide; in addition, PDMS optical waveguides of 8 cm length

were fabricated based on these different quality moulds. The optical insertion loss
of these PDMS waveguides obtained with different core–clad interface roughness
was measured at 850 nm wavelength. Compared with sidewall roughness, the top
and bottom interfaces between core and cladding can be neglected. Hereby, we
consider the interface roughness is mainly from the two sidewalls between them; in
addition, it is also assumed that the edge qualities of lines in every mask are
identical. Theoretically, the optical propagation loss due to scattering will be
increased evidently when average roughness of waveguide sidewalls is more than k
(wavelength)/20 [14], and hence to 850 nm wavelength, the critical transition
average roughness Ra is 42.5 nm. From the measured diagram shown in Fig. 13.9,
the experimental critical transition Ra is about 50 nm. Also, the diagram can help us
to determine the quality of mask based on different optical quality requirements to
PDMS waveguides. For example, suppose the optical waveguide loss is less than
0.05 dB/cm, then the average roughness of the casting mould should be controlled
below 40 nm, and thus, a relevant high-quality mask should be selected.
13.3.2.8 Optical Loss Due to Interlayer
Based on the PDMS casting technique, the waveguide structure is defined by the
casting mould and blade (see Fig. 13.10). The interlayer will be formed by
overfilling of PDMS core and incomplete blading. The standard waveguide
cross-section is shown in Fig. 13.11a. The interlayer formed can be divided into
two kinds: One is the structure shown in Fig. 13.11b, where two waveguides or
more are interconnected through the interlayer, which can cause optical crosstalk
Fig. 13.10 Doctor blading by hand (a) and by controlled machine (b). Reprinted with permission
from [21], copyright (2010) IEEE
Fig. 13.11 Cross-section of fabricated PDMS multimode waveguides and interlayers
and leads to extra optical loss, and the other is a “muffin”-like structure (see
Fig. 13.11c), which can lead to some extra mode dispersion and some radiation into
the cladding layer and ultimately results in optical loss by insufficient coupling.
By modifying the flatness of the casting mould, the blading pressure and angle,
etc., interlayers of different thickness were formed and their thickness was mea-
sured by an ellipsometer. Additionally, the fabricated waveguides were character-
ized with an optical laser measurement at 850 nm; the relationship between
interlayer thickness and optical loss is illustrated in Fig. 13.12. For example, the
interlayer thickness should be less than 500 nm when a loss value of 0.1 dB/cm is
requested.
13.3.3 Optical Insertion Loss of PDMS Waveguides

after Fabrication and Ageing
A top view of a series of fabricated PDMS multimode waveguides is illustrated in

Fig. 13.11a, which have a cross-section of 70 70 µm2. The measured optical
insertion loss of PDMS waveguides after curing and different temperature
Fig. 13.12 Relationship between optical loss and waveguide interlayer thickness
treatments is shown in Fig. 13.13, and the insertion loss of PDMS waveguides is
below 0.05 dB/cm after temperature treatment up to 300 °C (soldering conditions).
Further climatic changes had been applied as described in the experimental
section. A similar stability was found for the waveguide loss at 850 nm as described
in Fig. 13.13. According to the identified test conditions, results of the PDMS
waveguide loss measurements at 850 nm are shown in Fig. 13.14. In summary, it
has been proven that the investigated PDMS materials are well compatible with
Fig. 13.13 Optical insertion loss of PDMS multimode waveguides with standard deviations as
measured from eight waveguides under different heat treatment conditions
Fig. 13.14 Environmental stability of realized PDMS multimode waveguides obtained with
accelerated ageing (for each test condition, eight different waveguides had been measured; for the
measurement, a different PDMS core charge was supplied by the manufacturer compared to the
material tested with results shown in Fig. 13.13)
PCB manufacturing and suited for short-distance optical data transmission even in
harsh environments replacing other optical polymers such as acrylates and
epoxy-based materials.
13.4 Conclusions
The absorption loss of PDMS in the spectral datacom and telecom regions of 600–
1600 nm was analysed and summarized to result from scattering loss and vibra-
tional overtone and combination bands of the CH groups of the materials. In terms
of the observed positions of the fundamental, overtone and combination bands of
CH containing group vibrations in the mid-IR and NIR spectra, anharmonicity
constants including normal harmonic vibration frequencies have been determined
and an empirical equation for estimating the significant intrinsic absorption loss
wavelengths has been presented. The results agreed well with the experimental
data. Furthermore, we derived an empirical equation between integral band strength
and the intrinsic absorption loss based on the NIR spectral data of PDMS, which
can be exploited also for the limit estimation of the optical loss in deuterated or
halogenated siloxane-based polymers (0.005 dB/cm at 850 nm) applied in different
datacom devices.
Furthermore, the PDMS waveguides were fabricated with low optical insertion
loss at 850 nm and with high thermal stability (with temperatures > 250 °C after
heating). This indeed has proven that two-component thermally cured PDMS is a
very promising candidate for the production of highly reliable datacom devices.
Acknowledgements The authors gratefully acknowledge the continued support by Dipl.-Ing.

Rüdiger Kuckuk (Leibniz-Institut für Analytische Wissenschaften—ISAS) and Prof. Andreas
Neyer (TU Dortmund University). The fruitful cooperation with Wacker Chemie, Burghausen
(Germany), is also acknowledged.
References
1. Miller DAB (1997) Physical reasons for optical interconnections. Int J Optoelectronics
11:155–168
2. Groh W (1988) Overtone absorption in macromolecules for polymer optical fibers. Macromol
Chem 189:2861–2874
3. Tanio N, Koike Y (2000) What is the most transparent polymer? Polym J 32:43–50
4. Ballato J, Foulger S, Smith DW Optical properties of perfluorocyclobutyl polymers. J Opt Soc
Am B 20:1838–1843
5. Cai DK, Neyer A, Kuckuk R, Heise HM (2007) Estimation of absorption loss in
siloxane-based materials implemented as passive optical interconnects. In: Proceedings of
optical fiber communication and exposition (OFC) and national fiber optic engineers
conference (NFOEC), Anaheim, CA, paper JWA27
6. Flipsen T, Pennings AJ, Hadziioannou G (1998) Polymer optical fiber with high thermal
stability and low optical losses based on novel densely crosslinked polycarbosiloxanes. J Appl
Polym Sci 67:2223–2230
7. Yasuda N, Yamamoto H, Minami S, Nobutoki H, Wada Y, Yanagia S (2002) Novel silicone
polymeric material with high thermal stability for optical waveguides. Jpn J Appl Phys
41:624–630
8. Usui M, Hikita M, Amano TM, Sugawara S, Hayashida S, Imamura S (1996) Low-loss
passive polymer optical waveguides with high environmental stability. J Lightwave Technol
14:2338–2343
9. Su K, DeGroot JV Jr, Norris AW, Lo PY (2006) Siloxane materials for optical applications.
Proc SPIE 6029:318–325
10. Xia Y, Whitesides G (1998) Soft lithography. Angew Chem 37:550–575
11. Cai DK, Neyer A (2007) Realization of Electrical-circuit-board self-packaging. In:
Proceedings of IEEE conference on electronic components and technology conference,
pp 1368–1374
12. Ma H, Jen AK, Dalton LR (2002) Polymer-based optical waveguides: materials, processing,
and devices. Adv Mater 14:1339–1365
13. Poelman D, Smet PF (2003) Methods for the determination of the optical constants of thin
films from single transmission measurements: a critical review. J Phys D Appl Phys 36:1850–
1857
14. Voges E, Petermann K (2002) Optische Kommunikationstechnik. Springer, Heidelberg
Chapter 3
15. Smith AL (1960) Infrared spectra-structure correlations for organosilicon compounds.
16. Bokobza L, Buffeteau T, Desbat B (2000) Mid-and near-infrared investigation of molecular
orientation in elastomeric networks. Appl Spectrosc 54:360–365
17. Cai DK (2008) Optical and mechanical aspects on polysiloxane based
electrical-optical-circuits-board. Ph.D. thesis, TU Dortmund University. http://dx.doi.org/10.
17877/DE290R-8242
18. Cai DK, Neyer A, Kuckuk R, Heise HM (2008) Optical absorption in transparent PDMS
materials applied for multimode waveguides fabrication. Opt Mater 30:1157–1161
19. Cai D, Neyer A, Kuckuk R, Heise HM (2010) Raman, mid-infrared, near-infrared and
ultraviolet-visible spectroscopy of PDMS silicone rubber for characterization of polymer
optical waveguide materials. J Mol Struct 976:274–281
20. Siesler H, Ozaki Y, Kawata S, Heise HM (2002) Near-infrared spectroscopy. Wiley-VCH,
Weinheim Chapter 2
21. Cai D, Neyer A (2010) Quality standardization for polydimethysiloxane-based optical
interconnect production from optical aspects. IEEE Photonic Tech L 22:996–998
Chapter 14
The Luminescent Properties of Photonic
Glasses and Optical Fibers
Jacek Zmojda, Marcin Kochanowicz, Piotr Miluski

and Dominik Dorosz
Abstract In this chapter, the visible luminescent properties of antimony-germanate

and gallo-germanate glasses, glass-ceramics and optical fibers co-doped with dif-
ferent rare-earth (RE) systems (e.g., Sm3+, Eu3+, Yb3+/Tm3+, Yb3+/Ho3+, Yb3+/
Tm3+/Ho3+, Yb3+/Tb3+, Yb3+/Eu3+) have been analyzed. Emission was obtained by
direct excitation of RE ions or in case of co-doped systems by donor-acceptor
energy transfer and upconversion processes. Novel constructions of double-clad,
offset core, double-core, and RE triply doped optical fibers have been presented.
Moreover, the effect of the structure modification of antimony-germanate glasses on
the luminescence shaping has been also presented. Another, presented scientific
area is investigation of plasmon effect in Ag0/Eu3+ co-doped glasses and optical
fibers. The optimization of Ag0/Eu3+ content and formation of silver nanoparticles
process (thermal annealing) in order to enhance luminescence has been presented.
14.1 Introduction
Excellent optical properties of glasses pretended them as the only materials for the
construction of optical fibers. The advantage of glasses is variety of their chemical
compositions and different methods of synthesis. They tend to shape optical
properties in a very wide range and design advanced photonic structures. In par-
ticular, rare-earth (RE) doped glasses offer new possibilities for constructing optical
fiber radiation sources which are characterized by unique features. One of the
current directions in optical fiber photonics is research on functional materials
characterized by luminescent properties within the range from ultraviolet to
J. Zmojda M. Kochanowicz P. Miluski

Faculty of Electrical Engineering, Bialystok University of Technology,
Wiejska 45D Street, 15-351 Białystok, Poland
D. Dorosz (&)
e-mail: ddorosz@agh.edu.pl

428 J. Zmojda et al.
mid-infrared, enabling construction both laser and broadband emission sources for
the application in integrated systems, precise sensors and medicine. Glasses, in
contrary to crystalline and polymer materials, play an important role here, as they
offer features that are important from the photonic applications point of view. The
most important glass feature is its transparency with a diverse spectral range, as well
as amorphousness, which makes it possible to present a variety of compositions
with properties different from crystals. Although structural amorphousness usually
has a negative effect on optical properties of glass, it makes the technology of
fabricating glass much easier and cheaper, which implies greater functionality. It is
worth to note that the ease of the glass composition modification allows for the
introduction of higher concentrations of active elements (lanthanides) even several
at the same time, thus enabling the manipulation of its luminescent properties and
developing new functional photonics materials.
The operation of the mentioned above systems occurs through the interaction of
optical radiation with lanthanide ions, as a result of excitation and emission in a
strictly defined spectral range, which is dependent on the quantum transitions
within the energy structure of an active element. In practice, we have encountered
several major limitations related to obtaining efficient emissions, i.e.,
• narrow range of spectral transmission of glass;
• high value of phonon energy, which reduces the probability of radial transitions;
and
• low thermal stability of glass, which makes it difficult to use in fiber optic
technology.
A good example is silica glass used in telecommunications (e.g., Erbium Doped
Fiber Amplifier) as well as sources of laser radiation in the range of 1–2 µm
realized by fiber optics doped with Yb3+, Nd3+, Tm3+ ions [1–4]. Despite silica
unquestionable advantages, it has limitations related to the absorption edge located
in the infrared range at a wavelength of 2 µm and high-phonon energy of
1100 cm−1, which prevents lanthanide emission transitions in the visible range
using a donor-acceptor energy transfer. Therefore, new functional glasses are being
sought that enable efficient emission in the infrared (above 2 µm) and visible range.
The most promising optical glasses which could serve as an alternative in the ranges
mentioned are multi-component systems based on oxides, fluorides and chalco-
genides [1, 5–8]. Non-oxide glasses, in particular, have been viewed as a good
solution for many years because of their wide transmission window (up to 20 µm)
and one of the lowest phonon energies (350–500 cm−1). However, the expense of
the fabrication technology, the high toxicity of compounds used, its low mechanical
strength and low thermal stability, strongly limit the practical use of these glasses as
optical fibers [8, 9]. As a result, many research groups around the world have been
focusing their research on oxide and oxide-fluoride glass [10–13].
Special attention has been paid to the issues of fabrication the glassy matrices with
different phonon energy of glass-forming elements. From the one side, low-phonon
energies of glass-forming elements indicate higher efficiency of quantum transitions.
14 The Luminescent Properties of Photonic Glasses … 429
On the other hand, high-phonon energy compounds ensure thermal stability and
mechanical rigidity of glass. Moreover, the combination of different glass-forming
elements drives to greater separation of RE ions and enables an effective broadening
of the spontaneous emission spectrum [14]. These investigations require an
unconventional approach to shaping the glass matrix with the complex phonon
distribution, in such a way that the radiation emission can be achieved as a result of
optical transitions within the energy structure of lanthanides. It is a new approach at
the construction of glass in a confrontation with the generally accepted classification
of the glassy matrix in regard to the maximum of phonon energy.
Among oxide glasses, the antimony-based matrices have been widely investi-
gated due to their combination of chemical durability, low-phonon energies
(*600 cm−1) and high transparency in a wide range [15–20]. However, the low
field strength (0.73) of Sb3+ makes it a poor glass former; hence, it does not exist in
bulk monolithic form which is very much essential for practical applications. The
main advantage of antimony glasses is the ability to create a wide glassy area with
glass-forming components (Ge4+, Si4+, P5+, B3+) which in consequence increase
their thermal stability, needed for fiber production. The phenomenon of antimony
matrix relays on the fact that despite the significant concentration of other
glass-forming components (creating high phonon network), they conserve strong
luminescence form RE elements. On the other hand, peculiar choice in modified
components (Ga3+, Te4+, Bi3+, Ba2+, W6+) allows to form antimony glass com-
positions with no tendency to crystallization and adjust optical parameters like
transmission and refractive index. Another important fact is that Sb2O3 is a mild
reducing agent of noble metal ions. This mild reduction property enables in situ
reduction of Ag+ to Ag0 in a one-step during the melting process, thereby providing
a simple, low-cost method for the preparation of bulk photonic materials [21, 22].
Germanate glasses represent a group of the glassy matrix which is characterized
by favorable physicochemical and optical properties as: reduced phonon energy
compared with silica glass, transmission from the visible (Vis) to mid-infrared
(mid-IR) approx. 5.5 µm and possibility of incorporation relatively high concen-
tration of rare-earth ions without the crystallization or phase separation effect (up to
2 mol%). Regarding optical fiber technology, the high thermal stability is a key
parameter and allows its use in active optical fibers and planar waveguides struc-
tures thus makes them the object of scientific interest in the leading research centers
[23–27]. Another important aspect is the selection of a modifier and its optimal
concentration. Introduction to the glassy matrix various modifiers (oxide and/or
fluoride) leads to partial depolymerization of the structure of glasses resulting in a
positive or negative impact on their optical and thermal properties. For this reason,
there are many studies on the optimal choice of modifier and its concentration while
maintaining the high thermal stability of the glass.
This chapter contains the results of studies on shaping the luminescence prop-
erties of antimony-germanate and germanate glasses with regard to their structural
features. Research on the developed matrices was planned in order to achieve the
following functional properties:
• correlation of the effect of modifying the structure of antimony-germanate

glasses on the shape of the luminescence while maintaining their optical and
thermal properties;
• obtaining a plasmonic effect by thermochemical reduction of silver ions in
antimony-germanate glasses and glass fibers doped with lanthanides,
• analysis of correlation of structural changes of germanate glasses for detailed
characterization of luminescence properties.
14.2 The Effect of the Structure of Antimony-Germanate

Glasses Modification on the Luminescent Properties
According to the mentioned above assumptions for photonic applications, the

fabricated glass should be characterized by low phonon energy and high thermal
stability. Thus, influence of the substitution of germanium oxide with antimony
oxide on the spectroscopic and thermal properties of glass with chemical compo-
sition (50 − x)Sb2O3–xGeO2–45SiO2–5Al2O3 doped with Sm3+ ions with a molar
content of 0.5% was analyzed [28]. In this case, the content of germanium oxide
was changed from 0 to 45 mol%. Introduction of a samarium ion into the matrix
allowed obtaining a strong visible emission of a characteristic orange color, due to
high emission efficiency resulting from a large energy gap (ca. 12,000 cm−1)
between the 4G5/2 metastable level, and multiplet located below the 6HJ, an ion that
is readily used in phosphors [29].
The reason is that Sm3+ ion has absorption bands located at 405 and 470 nm
wavelengths, which can be efficiently excited by high power lasers diodes. As a
result of excitation in the spectra of the presented glasses, four characteristic emis-
sion bands have been observed at wavelengths of 566, 603, 647, and 709 nm cor-
responding to radiative relaxation from the level 4G5/2 to the lower sub-levels of the
6
HJ multiplet where J = 5/2, 7/2, 9/2, and 11/2 (Fig. 14.1). Analysis of the intensity
changes of individual luminescence bands, as a function of the content of GeO2
oxide, shows that all fabricated glass samples were characterized by the strongest
luminescence at 603 nm. In contrast, the maximum level of luminescence has been
achieved for glasses at 15 mol% Ge2O content, and a minimum at 30 mol%.
Samarium ions belong to the group of lanthanides characterized by radiative tran-
sitions that depend on the environment (structure) in which they are found.
According to the theory of the selection rule DJ = 0, ±1, the first two bands are
described as magnetic-dipolar transitions (MD) as well as partially electro-dipole
transitions, while the second two are described as electro-dipole transitions (ED).
Analyzing the ratio of the transition intensity 4G5/2 ! 6H9/2 (ED) to the transition
intensity 4G5/2 ! 6H5/2 (MD) the value was greater than 1, which proves that all the
samples produced can be characterized by asymmetry in structure around the
Sm3+ ion.
Fig. 14.1 Luminescence

spectra of
antimony-germanate glasses
doped with Sm3+ ions for the
different molar concentration
of GeO2 compounds. (inset)
Luminescence decay of 4G5/2
level
The shape of the luminescence decay curve for each of the sample was of a
two-exponential nature, resulting from two relaxation cross-processes between
the energy levels of dopant ions (4G5/2, 6H5/2) ! (6F5/2, 6F11/2) and (4G5/2,
6
H5/2) ! (6F11/2, 6F5/2) (Fig. 14.2). In addition, as the molar content of GeO2
increased, the lifetime of the 4G5/2 level increased, which indicates the positive effect
of germanium oxide on the reduction of non-bridging oxides, which are unfavorable
from the point of view of glassy structure stability.
Another aspect in modifications of antimony-germanate glass is related to the
creation of transparent glass-ceramics [30]. This problem is a part of a new area of
study on the use of active optical nanostructures in the optical fiber technology. For
this purpose, the glass with a molar composition: 25Sb2O3–25Ge2O3–10Al2O3–
5Na2O–35SiO2 can be modified with phosphorus pentoxide P2O5 in the range of
0.5–10 mol%. All fabricated glasses were doped with 0.5 mol% Eu2O3. The
selection of the dopant was also dictated, as in the case of the samarium ion, by
visible emission efficiency resulting from a large energy gap (of about
12,000 cm−1) between the excited level 5D0 and the lower 7FJ multiplet. As a
particularly special interest of the developed matrix, experimental results showed
that modifying SGS glass by a small amount of P2O5 (up to 10 mol%) it is possible
to obtain glass-ceramic material directly from the synthesis process. This is an
innovative approach to the issue of nucleation, as well as to the growth of
nanocrystals (in this case EuPO4) during the cooling of the vitreous mass. In the
case of excitation of the fabricated glasses with laser radiation at the wavelength of
394 nm, the luminescence spectra with different intensity distribution depending on
P2O5 content were presented in Fig. 14.3. Observed changes in the shape and
intensity of two significant emission transitions at 594 and 613 nm indicate changes
in the structure of the glasses around the Eu3+ ions. Analyzed changes showed that
Fig. 14.2 Simplified energy levels scheme of samarium ions with possible transitions and
cross-relaxation mechanisms
for the modified glass of 5 mol% P2O5, the local asymmetry field of the dopant,
defined by the ratio of (5D0 ! 7F2)/(5D0 ! 7F1) transitions, was 0.92. The
obtained result confirms unequivocally that for the analyzed SGS5P sample, the
europium ions surroundings are characterized by a high degree of order, charac-
teristic for crystals.
Also, the formation of EuPO4 nanocrystals in the XRD pattern was observed
(Fig. 14.4). After the reference analysis of peak positions calculated reflections are
in good accordance with monoclinic EuPO4 crystalline phase (ICSD:01-083-0656).
A visual confirmation of the glass-ceramic material are measurements of the
morphology of the samples made using SEM microscope where EuPO4 micro- and
nanocrystals on the amorphous SGS glass structure were observed (Fig. 14.5).
Changes in the local environment were also confirmed based on the analysis of
luminescence decay time from the level of 5D0. The double-exponential nature of
the decay curves indicates the presence of two glass structures: amorphous and
crystalline (Fig. 14.6).
By retaining the ability to reproduce the technological process and controlling
the content of P2O5 compounds, authors determined that it is possible to create a
double-shell optical fiber with a glass-ceramic core made of EuPO4 nanoparticles
[31]. On the basis of luminescence research, the shape of the spectrum obtained for

spectra of
antimony-germanate glasses
doped with Eu3+ ions for
different P2O5 contents
optical fiber was different from the spectrum obtained when the core material is
antimony-germanate glass modified with 0.5 mol% P2O5 (Fig. 14.7). Changes
resulting from the additional thermal treatment of the core glass in the process of
fiber drawing are related to the partial ordering of the glass structure surrounding
the europium ions.
As a result, a decrease in the luminescent FIR asymmetry ratio, a narrowing of
the emission band at 616 nm (5D0 ! 7F2) and significant splitting of the emission
bandwidth corresponding to the 5D0 ! 7F1 transition were observed. The obtained
results clearly indicate that in optical fiber produced with a core made of
antimony-germanate glass, the nanocrystallization of EuPO4 takes place during the
fiber drawing process.
Fig. 14.4 XRD pattern of antimony-germanate glass modified by 5 mol% P2O5
Fig. 14.5 SEM images of SGS glasses morphology with visible (a) nano- and (b) micro-crystals
of EuPO4
14.3 Plasmon Effect in Antimony-Germanate Glasses
Plasmon resonance is a novel area of research in the field of optoelectronics and

nanophotonics, which is based on the luminescence signal formation by controlling
local field enhancement (LFE) and energy transfer between noble metal nanopar-
ticles and lanthanide elements. The main purpose of introducing noble metal
nanoparticles into the glass is the shaping of the visible luminescence resulting from
energy transfer between metal nanoparticles and rare-earth elements, obtained
through electric field amplification in the vicinity of rare-earth ions caused by
surface plasmon resonance. Studies concerning interaction mechanisms of noble
metal nanoparticles on luminescent properties of active glasses are currently the

decay curves of 5D0 obtained
upon excitation at the
wavelength of 394 nm
Fig. 14.7 Comparison of the luminescence spectrum of Eu3+ ions obtained for glass-ceramics
core, glass fiber, and double-clad optical fiber
subject of research being carried out by many scientific centers around the world.
This will lead to new possibilities in the production of optoelectronic devices,
including optical fibers sensors and sources currently being used in medicine, optics
and nanophotonics (SERS—Surface-Enhanced Raman Scattering, LSPR—Local
Surface Plasmon Resonance sensors).
In addition to the aforementioned thermal and structural properties of
antimony-germanate glasses, the antimony ions contained therein have another
special feature. Namely, due to their chemical potential (Sb5+/Sb3+, E0 = 0.649 V),
they enable selective reduction of noble metal ions directly in the synthesis process.
This phenomenon, described by thermochemical reduction, was included in
research in order to obtain the effect of enhancing the luminescence signal through
the Local Field Effect. For the study, the SGS glass was fabricated with the
chemical composition: 25Sb2O3–25Ge2O3–5Al2O3–5Na2O–40SiO2, co-doped with
0.8 mol% Er2O3 and AgNO3 ions successively from 0; 0.1; 0.2; and 0.4 mol% [32].
Due to the complex structure of their energy levels, erbium ions are one of the most
recognizable and known elements, which in addition to their significant attribute of
offering optical amplification in the III telecommunications window (EDFA
amplifier), is characterized by a strong visible emission about 540 nm created
during excitation conversion. Introducing silver ions into the fabricated glasses and
controlling the reduction process to form nanoparticles of Ag0 tends to observe an
additional (not counting the absorption bands of Er3+ ions) band with a maximum at
the wavelength of 450 nm (Fig. 14.8).
A specific confirmation of the plasmon effect is the observation of glass mor-
phology carried out by means of transmission electron microscopy (TEM), where
silver nanoparticles had already formed at the stage of glass synthesis (Fig. 14.9a).
Then, the process of glass annealing at a slightly higher temperature (30 °C) than
the transformation temperature (Tg) caused further thermo-reduction of silver ions,
increasing their partial contribution to the glass structure (Fig. 14.9b), forming
agglomerates of these nanoparticles (Fig. 14.9c). The fabricated nanoparticles are
mostly spherical and their diameter was approximately 6 nm.
The luminescence analysis showed that it is possible to obtain both amplification
and quenching of the luminescence signal depending on the concentration of Ag+
ions and annealing time. Figure 14.10 presents changes in the spectrum of emission
of Er3+ ions in the visible range obtained in antimony-germanate glasses induced by
Fig. 14.8 Absorption spectra of antimony-germanate glasses co-doped with Er3+ and Ag+ ions.
Reprinted with permission from [32], copyright (2018) Elsevier
Fig. 14.9 TEM images of SGS glasses doped with 0.1Ag0 (a) heated for 1 h (b) and for 2 h (c) at
the temperature of 400 °C Reprinted with permission from [32], copyright (2018) Elsevier
radiation at the wavelength of 976 nm, with different concentrations of silver ions.
The erbium luminescence bands in the visible range, successively at 532, 546, and
660 nm are assigned to appropriate radiative transitions from the upper energy
levels 2H11/2, 4S3/ 2, and 4F9/2 to the basic energy level 4I15/2 (Fig. 14.11). As a
result of the analysis, the obtained signal gain is a consequence of the plasmon
resonance effect of silver nanoparticles, surrounded by erbium ions, and partly from
the Ag0 ! Er3+ energy transfer. In addition, the increase in the concentration of
Ag+ ions causes the reduction of the luminescence signal, which indicates an
unfavorable phenomenon of the backward Er3+! Ag0 energy transfer
(Fig. 14.10a). Bearing in mind the future applications of the fabricated glasses in
the form of optical fibers, an analysis of the influence of additional thermal treat-
ment of glasses and its impact on the profile of luminescence of Er3+ ions were
carried out (Fig. 14.10b). The process of heating the glasses at a temperature
slightly above the transformation temperature Tg for 1 h caused an increase of the
luminescence signal intensity at 546 nm by about 4% in relation to the as melted
sample. Further annealing (duration 2 h) led to quenching of the luminescence
Fig. 14.10 Luminescence spectrum of the “up-conversion” of SGS antimony-germanate glasses,

co-doped with Er3+ and Ag0 (a) and the influence of the annealing time on the level of
luminescence signal (b) under excitation at 976 nm. Reprinted with permission from [32],
Fig. 14.11 Scheme of energy

levels of Er3+ and Ag0 ions
with marked mechanisms of
quantum transitions and
energy transfer Reprinted
with permission from [32],
intensity due to energy transfer between Er3+ ions and agglomerates of Ag0
nanoparticles (Fig. 14.9c). The obtained results clearly show that both, the con-
centration of Ag+ ions, as well as annealing time, are key in shaping the lumi-
nescence signal in antimony-germanate glasses co-doped with Er3+/Ag0.
Continuing research on plasmon resonance and, due to the location of the
“plasmon” band in SGS (silica–germanium–antimony) glasses (kSPR = 450 nm),
europium ions were used, which, due to the location of their energy levels, are
characterized by effective emission at 612 nm and a good overlap between the 5L6
and 5D3 metastable levels and the “plasmonic” band [33]. At the same time, the
previously described SGS glasses were modified with B2O3 and P2O5 compounds
as follows: 25Sb2O3–25Ge2O3–5Al2O3–5Na2O–40B2O3/P2O5. The analysis of the
asymmetry ratio (R) of the Eu3+ ion environment depending on the introduced
modifiers showed that the B2O3 has the greatest influence on shape of luminescence
spectra (Fig. 14.12).
Hence, further research on the surface plasmon resonance has been directed
toward those glasses co-doped with Eu3+ and Ag+ ions and their additional thermal
treatment. Once more, the luminescence measurements showed a strong depen-
dence both on doping with silver ions and further annealing (Fig. 14.13). In the
study, the fabricated glass characterized by the highest level of luminescence signal
was the sample without Ag ions. However, in both co-doped with 0.6 mol%
AgNO3/0.5 Eu2O3 and after the annealing process, a decrease in emission in all
emission bands occurred, suggesting a reversal role of Ag0 nanoparticles (absorp-
tion of luminescence) and an increase in energy transfer from Eu3+ ! Ag0.
In order to explain the mechanisms responsible for shaping luminescence by
interactions of silver nanoparticles with rare-earth ions, the SGB antimony-
germanate-borate glasses and glass fibers were investigated to obtain a lumines-
cence enhancement [34]. The following SGB glass composition: 25Sb2O3–25Ge2O3–
5Al2O3–5Na2O–40B2O3 was selected and co-doped with Eu2O3/AgNO3. In the first
Fig. 14.12 Luminescence spectra of antimony-germanate glasses doped with Eu3+ ions (a) and
modified with high-intensity glass-forming elements
Fig. 14.13 Luminescence spectra of antimony-germanate-borate (SGB) glasses, co-doped with

Eu3+ and Ag+ ions
Fig. 14.14 Luminescence spectra of SGB glasses for various concentrations of silver ions
step, the effect of Ag+ ion concentration on the excitation spectra and emission of SGB
doped with 0.2Eu2O3 glasses was examined. As a result of the excitation of the glasses
with radiation at 394 nm, an 8% enhancement of europium ion luminescence signal
for 5D0 ! 7F2 (616 nm) transition was obtained in glasses doped with 0.1 mol% of
AgNO3 with respect to non-Ag0 glasses resulting from plasmon resonance phe-
nomenon and energy transfer (Fig. 14.14). In other cases, the quenching of lumi-
nescence associated with the backward energy transfer Ag0 ← Eu3+ was observed. In
the next step, the glass with the highest level of luminescence underwent a heating
process at 360 °C at various time intervals (Fig. 14.15). As a result, an increase in the
luminescence intensity in the glasses annealed for 2 h, suggesting a progressive
thermal reduction of silver ions in the glass and thus increasing the effect of the
resulting Ag0 nanoparticles with neighboring Eu3+ ions by plasmon resonance was
observed. Finally, to produce the optical fibers, the classic method of drawing of the
preform was applied.
Figure 14.11 inset shows a comparison of changing luminescence spectra for
glass fiber co-doped with 0.1AgNO3/0.2Eu2O3 annealed under analogous condi-
tions as a bulk sample. Analyzing the obtained spectra, it is worth to notice that the
transformation of glass into fiber did not bring significant differences in the char-
acter of luminescence, which suggests good thermal stability of the fabricated SGB
glasses. The only difference is a larger change in luminescence for glass samples
annealed for 1 h, for which similar value was observed in the optical fiber for a
sample annealed for 2 h. This effect is related to the additional process of thermal
influence on the glass during the fiber drawing process. The experiment with
annealing the glass fiber also brought an unexpected result. Namely, when
annealing fibers with a molar content of 0.6%, a thin film with a characteristic
amber coloration formed on their surface. Performing measurements of absorption
Fig. 14.15 Luminescence spectra of SGB glasses, doped with 0.1Ag/0.2Eu and optical fibers
(inset) for different annealing times
and observation of fibers under the electron microscope allowed to clearly state that
the observed phenomenon is the effect of reduction of Ag+ ions on the surface of
the glass fiber. This phenomenon is related to the mechanism of self-organization of
silver molecules inside the glass structure under thermal conditions above the
transformation temperature Tg. The absorption spectrum (Fig. 14.16) clearly shows
the band at a wavelength of 450 nm obtained for a fiber annealed for 2 h. The inset
in Fig. 14.16 presents optical fiber images acquired before and after the annealing
process.
A visual confirmation of the described effect is shown in the SEM images
(Fig. 14.17). Based on observations, the described effect occurred only in glasses

spectra of SGB glasses, doped
with 0.1Ag/0.2Eu and optical
fibers (inset) for different
annealing times
Fig. 14.17 SEM images of the surface of fabricated glass fibers (a) before and (b) after
heat-treatment process at 360 °C
with the highest concentration of Ag+ ions. A specific addition to research is the
extensive characterization of the structural and spectroscopic properties of the
fabricated antimony-germanate-borate SGB glass co-doped with ions of 0.2 mol%
Eu2O3 and 0.6 mol% AgNO3 [35]. In this case, the effect of modification of glass
structure with composition 25Sb2O3–25Ge2O3–5Al2O3–(40 − x)B2O3 by Na2O
compounds on thermal stability and correlation of glass structural features with the
thermochemical reduction of silver ions due to the annealing process at 400 °C. The
analysis of DSC curves (Fig. 14.18a) confirmed that the fabricated glasses are
thermally stable. In addition, the observed “atypical” shift of temperature trans-
formation Tg toward higher temperatures, together with an increase of Na2O con-
tent, is characteristic of the so-called germanate anomaly [36]. This effect was also
confirmed in the studies of infrared absorption spectra (FT-IR), on the basis of
which authors observed the transformation of [GeO6] groups into [GeO4] groups
resulting in an increase in the proportion of sodium compounds (Fig. 14.18b).
Fig. 14.18 DSC curves of fabricated glasses (a) and FTIR spectra (b) of SGB glasses for various
contents of Na2O compounds
Fig. 14.19 Luminescence spectra of SGB glasses co-doped with Ag+ and Eu3+ ions for the
different content of Na2O compounds a before and b after heat-treatment process
As part of the study on luminescence, an increase in the content of Na2O

compounds led to increasing the degree of disorder of the local active dopant field,
due to the increased proportion of non-bridging oxygen (NBO) in the glass struc-
ture. Figure 14.19a shows the luminescence spectra of SGB glasses co-doped with
0.6AgNO3/0.2Eu2O3 for different concentrations of Na2O, obtained as a result of
their excitation with laser radiation at a wavelength of 397 nm. Analysis of the
luminescent asymmetry coefficient showed that for both annealed and as melted
glass the effect of Na2O is analogous (inset Fig. 14.19b), which proves their high
thermal stability, which is significant for their processing into optical fibers.
On the one hand, too much non-bridging oxygen leads to depolymerization of
the glass structure, which hinders their practical use. But, this increases the effi-
ciency of the reduction of Ag+ ions introduced into the matrix, which strongly
depends on material characteristics and on concentration [35].
Spectroscopic studies show that apart from self-assembly, silver nanoparticles
form a thin film on the surface of the glass fiber; hence, there is an increase in the
ordering of the structure around the Eu3+ ion (Fig. 14.20). The results regarding
reduction on the fiber surface and self-assembly of silver nanoparticles in SGB
glasses and optical fibers have a unique character, necessary for functionalization of
optical materials used in optical biosensors based on the Surface-Enhanced Raman
Scattering (SERS) phenomenon [37, 38].
14.4 Upconversion Luminescence in the Vis Range

of Optical Fibers from Low-Phonon Energy Glasses
Efficient radiation emission in the visible range by energy transfer between lan-
thanide ions and upconversion processes (ESA—Excited State Absorption, ETU—
Energy Transfer Upconversion) requires the development of glass with low phonon
Fig. 14.20 Comparison of

the luminescence spectra of
SGB glasses and glass fibers
versus Na2O content before
and after the annealing
process (inset)
energy. Therefore, a chemical composition of antimony-germanate glasses (SGS) ,

consisting of glass-forming elements Sb2O3–GeO2–SiO2, as well as modifiers
Al2O3 and Na2O, were proposed. In order to generate emission through energy
transfer, glasses were doped with lanthanide ion in following systems: Yb3+/Tm3+,
Yb3+/Ho3+ and Yb3+/Tm3+/Ho3+ [39–42]. In each analyzed system, the ytterbium
ions functioned as a donor, which by transferring excitation energy to the remaining
acceptor ions (Tm3+, Ho3+) in energy transfer processes (donor ! acceptor) led to
the energy levels of thulium and holmium ions. Co-doping has many benefits, one
being that we can use commercially available high-energy laser diodes that generate
radiation in the near-infrared (976 nm) range to achieve the conversion of excita-
tion, which simplifies the future use of the fabricated glasses in photonic structures.
Also, the upconversion process enables radiative transitions to be obtained with
energy higher than the excitation energy, which is only observed in materials with
low-phonon energy, where the adverse impact of non-radiation transitions is min-
imized. In addition, by co-locating more than two rare-earth ions, it is possible to
obtain the emission of radiation simultaneously in several luminescence bands in
the visible range (multicolor emission).
The mentioned phenomena became the starting point for the research on new
antimony-germanate glasses, which combine high thermal stability, low phonon
energy, and optical properties necessary to carry out the emission of radiation in the
processes of converting excitation energy to higher energy levels. As cores of
optical fibers glass from the set of (25 − 30)Sb2O3–(25 − 30)GeO2–24SiO2–
10Al2O3–5Na2O, co-doped with oxides 1Yb2O3–xTm2O3, yHo2O3 (x = 0, 0.1; 0.2;
0.5; 0.75, y = 0, 0.1; 0.2; 0.5; 0.75; mol%) were used [39, 41, 43]. In the first stage,
optimization of the concentrations of rare-earth ions was accomplished, to obtain a
maximum of luminescence intensity in the visible range. The analysis of the
intensity of luminescence in the band of 478 nm (1G4 ! 3H6) and 650 nm
(1G4 ! 3F4) as a function of Tm2O3 content indicated that the maximum intensity
of luminescence obtained as a result of the Yb3+ ! Tm3+ energy transfer and
three-photon upconversion (kexc= 976 nm) was obtained in glass co-doped with
1Yb2O3/0.1Tm2O3. Based on the analysis of the luminescence decay time from the
Yb3+:2F5/2 level, the efficiency of the donor-acceptor energy transfer was calculated.
The lifetime of this level reduces from 781 µs (1Yb2O3) to 71 µs in the presence of
Tm3+ ions (1Yb2O3/0.75Tm2O3), which implies the maximum efficiency of the
Yb3+ ! Tm3+energy transfer at 90%. In turn, the highest value of luminescence
intensity in the visible range was obtained in glass co-doped with 1Yb2O3/
0.1Tm2O3. Therefore, this glass was used as the core of the fabricated double-clad
optical fiber. This optical fiber was drawn using modified “rod-in-tube” method. Its
basic parameters are as follows: outer diameter = 340 µm, core diameter = 10 µm,
NAcladding = 0.58, NAcore = 0.2. Attenuation of amplified spontaneous emission
(ASE) signal at the wavelength of 481 nm was 3.29 dB/m.
Another scientific problem, which has not been analyzed in the literature yet,
was a comparison of the luminescence spectra of core glass and fabricated optical
fiber, and next the analysis of mechanisms that have the impact on these differences.
The analysis was conducted on the basis of two fabricated optical fibers doped with
Yb2O3/Tm2O3 and Yb2O3/Ho2O3. According to the analysis of luminescent
properties of the optical fiber of the 1–2.5 m length under classical pumping of one
end type, the shift of the maximum of the emission has been observed in the range
of blue color (481 nm) of 3 nm toward longer wavelengths (Fig. 14.21). This effect
is the result of the optical trapping resulting from the Tm3+:3H6 ! 1G4 transition
occurring in the shortwave part of the ASE signal propagating in the core. In the
above process, radiation emitted by ions located at the beginning of optical fiber is
partially reabsorbed by ions of active admixture during the propagation along the
optical fiber.
Fig. 14.21 Comparison of the emission spectra of core glass and double-clad optical fiber
co-doped with 1Yb2O3/0.1Tm2O3 mol%, intensity I481mn/I653nm ratio as a function of the length of
the fiber and the photograph of the optical fiber (inset)
Moreover, another important effect which is decreasing of the luminescence

intensity ratio I481mn/I653nm with the increase of the length of optical fiber was
observed. In fabricated optical fiber with a length of 1 m, this value I481mn/I653nm is
17.9 and it decreases to 13.8 with the increase of the length optical fiber to 2.5 m.
Such a big change of the intensity of luminescence band I481mn/I653nm as a function
of optical fiber length results from the three-level quantum scheme for the emission
band in the range of blue color and simultaneously quasi four-level quantum
scheme in the case of red emission band. Due to the fact that the 3F4 level is
separated from the ground level of 3H6, the luminescence in the band of 653 nm
(1G4 ! 3F4) is not reabsorbed in the optical fiber in a process of excitation from the
ground state absorption (GSA) in Tm3+ions.
Analogously, antimony-germanate-silicate glass co-doped with Yb3+/Ho3+ions
was investigated to obtain luminescence in the visible range [39]. Holmium ions do
not absorb radiation in the range of commercially available semiconductor laser
diodes of high power operating in the range of 976 nm. The process of their
excitation results in the Yb3+ ! Ho3+ energy transfer with the participation of
phonons. Carried out optimization of the concentration of rare-earth has shown that
the maximum concentration of luminescence, obtained in the process of two-photon
upconversion (at 546 nm, (Ho3+:5S2(5F4) ! 5I8) and 655.5 nm (Ho3+:5F55I8), in
glass co-doped with 1Yb2O3/0.5Ho2O3 mol%. Based on measurements of a lifetime
of the Yb3+:2F5/2 donor level in glass doped with Yb3+ions and co-doped with
Yb3+/Ho3+, the analysis of the Yb3+ ! Ho3+energy transfer was conducted
(Fig. 14.22). The most efficient energy transfer of η = 91% was obtained in glass
co-doped with 1Yb2O3/0.75Ho2O3. The increase of holmium ions over 1 mol%
results in concentration quenching of luminescence.
Fig. 14.22 The efficiency of Yb3+ ! Ho3+ energy transfer in antimony–germanate–silicate glass
co-doped with 1Yb2O3/(0.1–0.75) Ho2O3 mol% and a lifetime of Yb3+:2F5/2, luminescence decay
curves (inset)
As in case of fibers described above, the double-clad optical fiber with offset core
co-doped with 1Yb2O3/0.5Ho2O3 was fabricated by the modified method of
“rod-in-tube”. The basic parameters of optical fiber are outer diameter = 400 µm,
diameter of the core = 15 µm, NAcladding = 0.58, NAcore = 0.4, ASE signal atten-
uation 3.68 dB/m at 658 nm. Further investigation enables to characterize mech-
anisms that have the impact on the profile of emission spectrum obtained in the
optical fiber, with respect to the luminescence of glass used as the core of this
optical fiber [39]. This aspect is new and has not been studied yet. Developed
optical fiber similarly as in glass was characterized by emission bands in the range
of green (Ho3+:5S2(5F4) ! 5I8) and red (Ho3+:5F5 ! 5I8) colors. It was explicitly
stated that the upconversion intensity ratio I658nm/I546nm in optical fiber is a few
times higher than in bulk glass sample. This ratio (I658nm/I546nm) increases with the
increase of the length of optical fiber and in fibers of 50 cm and 180 cm length is
5.25 and 8.9, respectively. What is more, with the increase of the length of optical
fiber, there is a decrease of the intensity of pumping radiation at the points more
distant from the source of exciting radiation and as a result, the lower contribution
of absorption phenomenon from the ESA2 (Excited State Absorption 2)
(Fig. 14.23b) in a population of the 5S2(5F4) level.
Furthermore, in analyzed optical fiber, as a result of different geometry of the
impact of the pump radiation (contrary to the glass sample), with the increase of the
length of fiber the dominating phenomena responsible for the population of the 5F5
level (emission in the range of red color, Ho3+:5F5 ! 5I8) are ETU and ESA1
processes [39]. At the same time, the reabsorption of propagated amplified spon-
taneous emission (ASE) signal along the optical fiber, which was the result of
Ho3+:5S2(5F4) ! 5I8, 5F5 ! 5I8 transitions causes the shift of the maximum of
luminescence bands of 2.5 nm toward longer wavelengths.
The results of above research indicated the potential possibilities of obtaining the
emission of white color in the optical fiber. Novel constructions of the double-clad
Fig. 14.23 Emission spectra (upconversion) of (a) antimony-germinate-silicate optical fiber

co-doped with 1Yb2O3/0.5Ho2O3, (inset) Luminescence spectra of glasses co-doped with 1Yb2O3/
(0.1–0.75)Ho2O3 mol% (b) energy scheme with indicated quantum transitions, kexc= 976 nm
two cores optical fiber (with offset) made of antimony-germanate-silicate glass, in

which the first one was co-doped with 1Yb2O3/0.1Tm2O3, and the second one was
co-doped with 1Yb2O3/0.5Ho2O3 was proposed [40, 41]. The aim of the carried out
research was the analysis of possibilities of obtaining the emission of white color in
optical fiber, resulting in the superposition of luminescence bands, which resulted in
radiative transitions in Tm3+and Ho3+ ions. Also, the studies were focused on the
impact of the length of optical fiber and the pumping power on the possibility of
profiling the emission signal of optical fiber indicating their impact on obtained
colorimetric parameters. The basic parameters of fabricated double-core optical
fiber were as follows: outer diameter = 350 µm, diameters of cores = 10 µm,
distance between the cores = 30 µm, offset of the cores = 60 µm,
NAcladding = 0.58, NAcore = 0.29, ASE signal attenuation = 3 dB/m (at 481 nm).
Applying the same offset of the cores provides equal pumping of the two cores.
Measurements of the luminescence spectra (kexc= 976 nm) as a function of the
length of the optical fiber showed that the ratio of the emission intensity in the band
of 481 nm to the band of 545 nm and 655 nm decreases with the increase of the
length of optical fiber (Fig. 14.24).
In the range of the fiber length up to 100 cm and the single watts power of
pumping radiation, the effective three-phonon conversion of NIR radiation to VIS
in the core co-doped with Yb3+/Tm3+ was obtained. The lack of the reabsorption
effect of radiation resulting in the Tm3+:3H6 ! 1G4 transition was possible to
obtain due to applying a relatively low concentration of acceptor ions (0.1 mol%
Tm2O3). At the same time, in the second core co-doped with Yb3+/Ho3+ the higher
concentration of holmium ions (0.5 mol% Ho2O3) causes the increase of contri-
bution of the energy migration process between Ho3+ions and leads to lumines-
cence quenching in the visible range. In optical fibers with the length over 100 cm,
the speed of changes in the ratio of luminescence intensity I481nm/I655nm is lower.
Fig. 14.24 Emission spectra

(upconversion) of double-core
antimony-germanate-silicate
optical fiber co-doped with
1Yb2O3/0.5Ho2O3 and
1Yb2O3/0.1Tm2O3,
Luminescence I481nm/I655nm
and I481nm/I545nm (inset)
intensity ratio, kexc= 976 nm
As a result, obtained emission spectrum in the visible range (kexc = 976 nm) at the
end of optical fiber consists of the dominant band in the range of blue (481 nm,
Tm3+:1G4 ! 3H6), green (Ho3+:5S2(5F4) ! 5I8), and red (655 nm) colors.
Luminescence in the range of the red color is the result of the superposition of
transitions among Ho3+:5F5 ! 5I8 and Tm3+:1G4 ! 3F4.
According to the CIE 1931 model, the analysis of chromatic coordinates was
conducted based on obtained luminescence spectra. The increase of the length of
studied optical fiber causes the shift of chromatic coordinates from the region of
blue color toward the point corresponding to the emission of white color. What is
more, changing the power of pumping radiation in the range from 0.5 to 1.5 W
showed the possibility of changing the luminescence spectra. The increase of power
of the pump up to 1.5 W causes the shift of chromatic coordinates toward the center
of the triangle of CIE 1931.
Other developed construction of active optical fiber enabling the emission of
white light is double-clad one core optical fiber co-doped with Yb3+/Tm3+/Ho3+ ions
[40]. The antimony-germanate glass, in which luminescence of white color was
obtained (0.5Yb2O3/0.1Tm2O3/0.2Ho2O3, mol%), has been used as a core of the
fabricated optical fiber. Chromatic coordinates in this case were x = 0.35, y = 0.32,
which locates the point in the center of the CIE 1931 diagram. Optical fiber was
characterized by the following parameters: outer diameter = 330 µm, diameter of
the core = 10 µm, offset of the core = 60 µm, NAcladding = 0.58, NAcore = 0.29.
Analysis of mechanisms that have the impact on the differences of luminescence
properties of core glass and optical fiber shows that the luminescence spectrum
(kexc = 976 nm) of fabricated optical fiber consists of—similarly to glass—three
emission bands: 479 nm (Tm3+:1G4 ! 3H6), 545 nm (Ho3+:5S2(5F4) ! 5I8) and
656 nm (Ho3+:5F5 ! 5I8 + Tm3+:1G4 ! 3F4).
However, the strong luminescence quenching was observed in the optical fiber
in the range of blue color and decrease of the luminescence intensity in the band of
656 nm relatively to the intensity on the band of 545 nm. It is caused by phe-
nomena of optical trapping of ASE signal propagating in optical fiber resulting from
the presence of Tm3+:1G4 ! 3H6, Ho3+:5I8 ! 5G6 and Ho3+:5I8 ! 5F5 transitions.
Furthermore, in the set of Yb3+/Tm3+/Ho3+ ions, the resonance energy transfer
Tm3+ ! Ho3+ limits the population of higher metastable levels of Tm3+. Finally,
the change of geometry of the interaction of radiation of the pump (glass sample vs.
optical fiber) caused a significant shift of the coordinates on the CIE diagram (inset
Fig. 14.25).
The calculated chromatic coordinates in the case of optical fibers are x = 0.37,
y = 0.49, which locates the point of the color in the area of green color [40]. The
practical aspect of carried out research is the description of mechanisms that have
the impact on the shape of luminescence spectrum of optical fiber co-doped with
Yb3+/Tm3+/Ho3+ ions, enabling to change obtained chromatic coordinates and
simultaneously the color temperature of the emission spectrum of fiber.
Fig. 14.25 Emission spectra (upconversion) of double-core antimony-germanate-silicate optical

fiber co-doped with 1Yb2O3/0.5Ho2O3 and 1Yb2O3/0.1Tm2O3, Luminescence I481nm/I655nm and
I481nm/I545nm (inset) intensity ratio, kexc= 976 nm
14.5 Optical Fibers from GeO2–Ga2O3–BaO System
Barium gallo-germanate glass system has been recently studied and proved to be an
excellent candidate as a host material as possessing good RE ions solubility,
comparatively low-phonon energy and favorable mechanical characteristics [44,
45]. Investigations on the optical fibers from the GeO2–Ga2O3–BaO glass system
related with low-phonon energy (805 cm−1) resulting in vibrations Ge–O(Ge)
enables obtaining the luminescence in the visible range as a result of the
donor-acceptor energy transfer and the upconversion process [4, 46–48]. As in the
previous antimony-germanate glass, the analysis of concentrations of lanthanide
dopants conducted in co-doping systems (Yb3+/Tm3+, Yb3+/Tb3+, Yb3+/Eu3+)
enables optimization of the luminescence properties in the visible range. For further
research on optical fibers, sets of Yb3+/Tb3+ ions and Yb3+/Eu3+ ions were selected.
In this case, the emission in the visible range was obtained as a result of the
cooperative energy transfer with the efficiency η = 12.92%—Tb3+, η = 27.64%—
Eu3+.
Proposed sets of co-dopants enable to develop new constructions of active
double-clad optical fibers, in which there is the two-photon conversion of the pump
radiation (kexc = 976 nm) via the cooperative energy transfer and emission in the
visible range. Based on these studies, two types of double-clad optical fibers with
“offset” core co-doped with 0.7Yb2O3/0.7Tb2O3 and 0.5Yb2O3/0.75Eu2O3 have
been fabricated. Next, a comparative analysis of luminescence spectra of core glass
and optical fiber was made and mechanisms that have the impact on these differ-
ences were discussed [47–49].
Fig. 14.26 Emission spectra (upconversion) of germanate glass and optical fiber co-doped with
0.7Yb2O3/0.7Tb2O3, Decomposition of green emission band (Tb3+:5D4 ! 7F5) (inset), pho-
tograph of the fiber (inset), kexc= 976 nm
In optical fiber co-doped with Yb3+/Tb3+ ions, luminescence bands corre-

sponding to Tb3+:5D4 ! 7F3–6 transitions have been observed in a good accordance
with bulk glass sample under kexc= 976 nm excitation. However, the main emission
band of fiber in the range of 543 nm is shifted about 4 nm toward the longer
wavelengths. In addition, the luminescence intensity ratio (I547nm/I621nm) is greater
in the optical fiber (I547nm/I621nm = 21.6, the length of optical fiber = 130 cm) than
in the glass sample (I547/I621 = 16.3). There were no changes at the same time in the
value of the I547nm/I587nm parameter. Detailed analysis of the shape of the main
luminescence band which was made by Gaussian decomposition (inset Fig. 14.26),
showed that the emission band consists of two sub-bands of 543 nm
(FWHM = 2.8 nm) and 547 nm (FWHM = 9.5 nm), and the parameter I547nm/
I543nm in the case of the glass sample is 3.95. However, the following value of
I547nm/I543nm = 12.3 parameter was achieved in the optical fiber.
Furthermore, in fabricated double-clad optical fiber with offset core the main
ASE emission band in the range of green color is amplified along propagation in
fiber and, the value of I547nm/I621nm parameter increases linearly with the increase of
optical fiber. In an optical fiber, radiative transitions from lower, Tb3+sublevel are
dominant and as a result, the main ASE emission band is 4 nm shifted toward
longer wavelengths (compared to the glass sample) [47].
In the case of glass co-doped with Yb3+/Eu3+ ions, the highest intensity of the
luminescence in the bands of 616 nm (5D0 ! 7F2) and 591 nm (5D0 ! 7F1) was
obtained in glass co-doped with 0.5Yb2O3/0.75Eu2O3 (mol%) and this glass was
used as a core of double-clad optical fiber [48]. Used method of fabrication of
optical is similar to previously modified “rod-in-tube” method, which enables to
obtain the ASE signal attenuation at the level of 2.18 dB/m, which is a better
Fig. 14.27 Comparison of the luminescence spectra (upconversion) glass, “glass fiber” and a
double-clad optical fiber co-doped 0.5Yb2O3/0.75Eu2O3 (mol%), kexc = 976 nm signal power vs.
a function fiber length (inset)
achievement in comparison to known tellurite or fluorogermanate optical fibers. In

the next stage, the impact of the construction of optical fiber (glass fiber and
double-clad fiber), its length and pumping method (kexc = 405 nm and kexc =
976 nm) on the shape of obtained emission spectra was investigated (Fig. 14.27).
Comparative analysis shows that the length of double-clad optical fiber has no
impact on obtained luminescence spectrum that arises by upconversion. In com-
parison to previously described optical fibers doped with Yb3+/Ho3+ or Yb3+/Tm3+
ions, where was the reabsorption of the ASE signal that has the impact on the shape
of luminescence, in studied optical fiber co-doped with Yb3+/Eu3+ this phenomenon
does not occur [48]. The intensity ratio of the luminescence I616nm/I591nm parameter
in the studied range of the length of optical fiber (up to 5 m) remained unchanged.
However, in the glass fiber type with the length of 0.1, the ratio of intensities of
luminescence in the range of 616 nm (Eu3+:5D0 ! 7F2) and 591 nm is greater
(I616nm/I591nm = 4.17). In this case, the whole radiation of the pump is absorbed in a
short segment of glass fiber. As a result, the radiative transition corresponding with
the luminescence band of the highest ratio of luminescence branching begins to be
visibly promoted during the propagation along the fiber. Hence, the scheme of
pumping Eu3+ ions in an optical fiber has the important impact on the shape of the
luminescence spectrum (Fig. 14.28).
The maximum of the emission obtained in the band of 616 nm that arises by
upconversion process (kexc = 976 nm) is shifted by 2 nm toward longer wave-
lengths, and its half width is about 5 nm greater than in the same optical fiber, in
which Eu3+ ions were excited directly by the radiation at the wavelength of 405 nm.
Furthermore, if the emission in the visible range is realized in the process of
Fig. 14.28 Luminescence spectra (upconversion), double-clad fiber co-doped 0.5Yb2O3/

0.75Eu2O3 (% mol), kexc = 976 nm or kexc= 405 nm, CIE diagram with marked chromatic
coordinates (inset)
cooperative energy transfer and upconversion the parameter I616nm/I591nm is equal

3.4, but in the case of excitation by radiation of 405 nm, it is equal 4.2 [48].
14.6 Summary
Development of fiber amplifier systems, lasers, and broadband sources of amplified

spontaneous emission is closely connected with research on novel glass composi-
tions and constructions of active fibers, doped with rare-earth ions. In the fields
mentioned above, investigations on new types of glasses and optical fibers doped
with lanthanides ions have been carried out in two directions:
• non-silicate, non-oxide glasses, and
• oxide and oxyfluoride glasses co-doped with lanthanides.
Compared to high silica glass, reduced phonon energy, good thermal stability,
acceptance of higher concentration of lanthanides and possibility of their processing
into fiber structures make antimony-germanate and gallo-germanate glasses very
attractive material for the construction of new types of active optical fibers. In
particular, presented solutions of antimony-germanate and gallo-germanate glasses
and optical fibers enabled to emit radiation in the visible spectral range using
donor-acceptor energy transfer between lanthanide ions and conversion processes
(downconversion and upconversion). Moreover, another presented aspect in mod-
ifications of antimony-germanate glass is related to the creation of transparent
glass-ceramics. Nanocomposite, antimony-germanate glass core with EuPO4
nanocrystals fabricated during the fiber drawing process gives possibility of

emission spectra profiling. On the other hand, enhancement of luminescence of
RE-doped glasses can be obtained by co-doping with noble particles and
RE-nanoparticles. The enhancement results from surface plasmon resonance, local
field enhancement in the vicinity of RE or noble metal nanoparticles to RE energy
transfer. It should be noted that the formation of metallic nanoparticles in glasses
and optical fibers with specific size distribution embedded in the glass matrix is a
challenging issue. Presented Ag0/Er3+ or Ag0/Eu3+ glasses and optical fibers are
promising materials to their use in novel fiber sources of radiation or fiber sensors.
To summarize, fields of application of presented glasses and optical fibers doped
with lanthanides emitting in the VIS range rare systems of HD recording and
storing information, 3D displays, and optical sensors. Another area of application is
widely understood biophotonic, in particular, there is a continuous need to intro-
duce new sources of radiation from the range of the therapeutic window (0.6 up to
0.9 µm) used in the processes of low energy optical biostimulation. In authors
opinion, developed glasses, glass-ceramics and optical fibers doped with lanthanide
ions presented in this chapter, can be used in the construction of novel optical fiber
sources in the visible spectral range. These sources, compared to LED-based
structures, are characterized by higher power density.
Acknowledgements The research activities were supported by the National Science Centre
(Poland) granted on the basis of the decision No. DEC-2016/21/D/ST7/03453 and the project
No. 2016/23/B/ST8/00706.
References
1. Tanabe S (2002) Rare-earth-doped glasses for fiber amplifiers in broadband telecommuni-

cation. C R Chim 5:815–824
2. Bünzli J-CG, Eliseeva SV (2010) Lanthanide NIR luminescence for telecommunications,
bioanalyses and solar energy conversion. J Rare Earths 28:824–842
3. Jacquier B (1997) Rare earth-doped fiber lasers and amplifiers. In: Rotman SR (ed) Wide-gap
luminescent materials: theory and applications. Springer, Boston, pp 303–365
4. Kochanowicz M, Dorosz D, Zmojda J, Dorosz J, Pisarska J, Pisarski WA (2014)
Up-conversion luminescence of Tb3+ ions in germanate glasses under diode-laser excitation
of Yb3+. Opt Mater Express 4:1050–1056
5. Gu SQ, Ramachandran S, Reuter EE, Turnbull DA, Verdeyen JT, Bishop SG (1995) Novel
broad-band excitation of Er3+ luminescence in chalcogenide glasses. Appl Phys Lett 66:670–672
6. Heo J, Chung WJ (2014) Rare-earth-doped chalcogenide glass for lasers and amplifiers. In:
Adam J-L, Zhang X (eds) Chalcogenide glasses. Woodhead Publishing, Sawston, pp 347–380
7. Li L, Bian J, Jiao Q, Liu Z, Dai S, Lin C (2016) GeS2–In2S3–CsI chalcogenide glasses doped
with rare earth ions for near- and mid-IR luminescence. Sci Rep 6:37577
8. Lucas J, Adam J-L (1992) Rare-earth doped fluoride glass fibres. J Alloy Compd 180:27–35
9. Tang Z, Furniss D, Fay M, Sakr H, Sójka L, Neate N, Weston N, Sujecki S, Benson TM,
Seddon AB (2015) Mid-infrared photoluminescence in small-core fiber of praseodymium-ion
doped selenide-based chalcogenide glass. Opt Mater Express 5:870–886
10. Fedorov PP, Luginina AA, Popov AI (2015) Transparent oxyfluoride glass ceramics.
J Fluorine Chem 172:22–50
11. Jha A, Richards B, Jose G, Teddy-Fernandez T, Joshi P, Jiang X, Lousteau J (2012) Rare-earth
ion doped TeO2 and GeO2 glasses as laser materials. Prog Mater Sci 57:1426–1491
12. Polishchuk SA, Ignat’eva LN, Marchenko YV, Bouznik VM (2011) Oxyfluoride glasses (a
review). Glass Phys Chem 37:1–20
13. Yonezawa S, Kim J-H, Takashima M (2010) Preparation and Properties of
Rare-Earth-Containing Oxide Fluoride Glasses. In: Tressaud A (ed) Functionalized inorganic
fluorides: synthesis, characterization & properties of nanostructured solids. Wiley, New York
14. Dorosz D, Zmojda J, Kochanowicz M (2013) Investigation on broadband near-infrared
emission in Yb3+/Ho3+ co-doped antimony–silicate glass and optical fiber. Opt Mater
35:2577–2580
15. Hedden WA, King BW (1956) Antimony oxide glasses. J Am Ceram Soc 39:218–222
16. Dorosz D, Zmojda J, Kochanowicz M, Miluski P, Jelen P, Sitarz M (2015) Structural and
optical study on antimony-silicate glasses doped with thulium ions. Spectroch Acta A
134:608–613
17. Nalin M, Poulain M, Poulain M, Ribeiro SJL, Messaddeq Y (2001) Antimony oxide based
glasses. J Non Cryst Solid 284:110–116
18. Ouannes K, Soltania MT, Poulain M, Boulonc G, Alombert-Goget G, Guyot Y, Pillonnet A,
Lebbou K (2014) Spectroscopic properties of Er3+-doped antimony oxide glass. J Alloy
Compd 603:132–135
19. Som T, Karmakar B (2010) Structure and properties of low-phonon antimony glasses and
nano glass-ceramics in K2O–B2O3–Sb2O3 system. J Non-Cryst Solid 356:987–999
20. Som T, Karmakar B (2011) Nephelauxetic effect of low phonon antimony oxide glass in
absorption and photoluminescence of rare-earth ions. Spectrochim Acta A 79:1766–1782
21. Som T, Karmakar B (2011) One-step synthesis and properties of monolithic photoluminescent
ruby colored cuprous oxide antimony oxide glass nanocomposites. J Alloy Compd 509:4999–
5007
22. Som T, Singh SP, Karmakar B (2016) Plasmonic antimony and bismuth oxide glass
nanocomposites: synthesis and enhanced photoluminescence. In: Karmakar B, Rademann K,
Stepanov AL (eds) Glass nanocomposites. William Andrew Publishing, Boston, pp 215–238
23. Balda R, Fernández J, Pablos Ad, Fdez-Navarro JM (1999) Spectroscopic properties of Pr3+
ions in lead germanate glass. J Phys Condensed Matt 11:7411–7421
24. Cai M, Zhou B, Tian Y, Zhou J, Xu S, Zhang J (2016) Broadband mid-infrared 2.8 lm
emission in Ho3+/Yb3+-codoped germanate glasses. J Lumin 171:143–148
25. Pisarski WA, Pisarska J, Dorosz D, Dorosz J (2015) Rare earths in lead-free oxyfluoride
germanate glasses. Spectrochimi Acta A 134:587–591
26. Rachkovskaya GE, Zakharevich GB (2012) Germanate lead-tellurite glasses for optical light
filters. Glass Ceram 68:385–388
27. Wang R, Yang Z, Zhou D, Song Z, Qiu J (2014) Structure and luminescent property of Er3+-
doped germanate glasses. J Non Cryste Solid 383:200–204
28. Zmojda J, Kochanowicz M, Miluski P, Leśniak M, Sitarz M, Pisarski W, Pisarska J, Dorosz D
(2016) Effect of GeO2 content on structural and spectroscopic properties of antimony glasses
doped with Sm3+ ions. J Mol Struct 1126:207–212
29. Zhao M, Liu Y, Ma S, Liu D, Wang K (2018) Investigation of energy transfer mechanism and
luminescence properties in Eu3+ and Sm3+ co-doped ZnWO4 phosphors. J Lumin 202:57–64
30. Zmojda J, Kochanowicz M, Miluski P, Baranowska A, Pisarski WA, Pisarska J, Jadach R,
Sitarz M, Dorosz D (2017) Optical characterization of nano- and microcrystals of EuPO4
created by one-step synthesis of antimony-germanate-silicate glass modified by P2O5. Mater
10:1059
31. Zmojda J, Kochanowicz M, Miluski P, Lukowiak A, Pisarski WA, Pisarska J, Marciniak M,
Ferrari M, Sitarz M, Dorosz D (2016) Rare-earth doped optical fibers with nano-phase
glass-ceramic structures. In: 18th International conference on transparent optical networks
(ICTON), 10–14 July 2016, pp 1–4
32. Zmojda J, Kochanowicza M, Miluski P, Baranowska A, Basa A, Jadach R, Sitarz M,

Dorosz D (2018) The influence of Ag content and annealing time on structural and optical
properties of SGS antimony-germanate glass doped with Er3+ ions. J Mol Struct 1160:428–
433
33. Zmojda J, Miluski P, Baranowska A, Jadach R, Kochanowicz M (2017) Local field effect in
antimony-germanate glasses co-doped with rare-earth ions and silver nanoparticles. Photonics
Lett Poland 9:107–109
34. Zmojda J, Miluski P, Kochanowicz M (2018) Nanocomposite antimony-germanate-borate
glass fibers doped with Eu3+ ions with self-assembling silver nanoparticles for photonic
applications. Appl Sci 8:790
35. Zmojda J, Kochanowicz M, Miluski P, Baranowska A, Pisarski WA, Pisarska J, Jadach R,
Sitarz M, Dorosz D (2018) Structural and optical properties of antimony-germanate-borate
glass and glass fiber co-doped Eu3+ and Ag nanoparticles. Spectrochim Acta A 201:1–7
36. Henderson GS (2007) The germanate anomaly: what do we know? J Non-Cryst Solid
353:1695–1704
37. Candeloro P, Iuele E, Perozziello G, Coluccio ML, Gentile F, Malara N, Mollace V, Di
Fabrizio E (2017) Plasmonic nanoholes as SERS devices for biosensing applications: an easy
route for nanostructures fabrication on glass substrates. Microelectron Eng 175:30–33
38. Schneider R, Schneider R, de Campos EA, Santos Mendes JB, Felix JF, Santa-Cruz PA
(2017) Lead-germanate glasses: an easy growth process for silver nanoparticles and their
promising applications in photonics and catalysis. RSC Adv 7:41479–41485
39. Kochanowicz M, Dorosz D, Zmojda J, Miluski P, Dorosz J, Pisarska J, Pisarski WA (2015)
Upconversion emission in antimony–germanate double-clad optical fiber co-doped with Yb3+/
Tm3+ ions. Opt Mater 41:108–111
40. Zmojda J, Kochanowicz M, Miluski P, Dorosz J, Pisarska J, Pisarski WA, Dorosz D (2016)
Investigation of up conversion luminescence in antimony–germanate double-clad two cores
optical fiber co-doped with Yb3+/Tm3 + and Yb3+/Ho3+ ions. J Lumin 170:795–800
41. Zmojda J, Kochanowicz M, Miluski P, Righini GC, Ferrari M, Dorosz D (2016) Investigation
of upconversion luminescence in Yb3+/Tm3+/Ho3+ triply doped antimony-germanate glass
and double-clad optical fiber. Opt Mater 58:279–284
42. Zmojda J, Dorosz D, Kochanowicz M, Miluski P, Dorosz J (2013) White upconversion in
Yb3+/Tm3+/Ho3+co-doped antimony-germanate glasses. Acta Phys Polon A 124:598–601
43. Kochanowicz M, Zmojda J, Miluski P, Pisarska J, Pisarski WA, Dorosz D (2015) NIR to
visible upconversion in double-clad optical fiber co-doped with Yb3+/Ho3+. Opt Mater
Express 5:1505–1510
44. Jadach R, Zmojda J, Kochanowicz M, Miluski P, Pisarska J, Pisarski WA, Sołtys M,
Lesniak M, Sitarz M, Dorosz D (2018) Investigation of the aluminum oxide content on
structural and optical properties of germanium glasses doped with RE ions. Spectrochim
Acta A 201:143–152
45. Kochanowicz M, Zmojda J, Miluski P, Ragin T, Pisarski WA, Pisarska J, Jadach R, Sitarz M,
Dorosz D (2017) Structural and luminescent properties of germanate glasses and double-clad
optical fiber co-doped with Yb3+/Ho3+. J Alloy Compd 727:1221–1226
46. Dorosz D, Kochanowicz M, Zmojda J (2014) Blue upconversion emission in germanate glass
co-doped with Yb3+/Tm3+ ions. Int J Appl Glass Sci 5:393–400
47. Kochanowicz M, Zmojda J, Miluski P, Sitarz M, Pisarska J, Pisarski WA, Dorosz D (2016)
Analysis of upconversion luminescence in germanate glass and optical fiber codoped with
Yb3+/Tb3+. Appl Opt 55:2370–2374
48. Kochanowicz M, Zmojda J, Ragin T, Miluski P, Jelen P (2016) Up-conversion luminescence
in germanate glass and double-clad optical fibre co-doped with Yb3+/Eu3+ ions. Opto-Electron
Rev 24:155–161
49. Kochanowicz M, Zmojda J, Dorosz D (2014) Fluorosilicate and fluorophosphate
superfluorescent multicore optical fibers co-doped with Nd3+/Yb3+. Opt Fiber Technol
20:245–249
Chapter 15
Spectroscopic Characterization
of Silicate Amorphous Materials
Włodzimierz Mozgawa, Maciej Sitarz and Magdalena Król
Abstract In spite of a large amount of the literature on amorphous silicate structure

studies and a significantly long time that has passed since the pioneer works of
Lebiedev, Zachariasen, and Warren, there are still many doubts concerning the
description of their structure. The selection of the correct research method allowing
for a description of glass structure, specifically one that yields information on
short-range order present in the amorphous state, seems to be especially important.
In the framework of this manuscript, vibrational spectroscopy was proposed as a
fundamental research method for describing the amorphous structure of glass
materials. Interpretation procedures were also presented that enable obtaining the
maximum amount of information on glass build on the basis of oscillation spectra.
15.1 Glassy State Theory
Glass is one of the oldest known materials. Until recently (second half of the
twentieth century), it was believed that only certain substances called glass-forming
can be obtained in the amorphous state [1]. However, it is currently accepted that
practically every substance that does not disintegrate at melting temperature can be
obtained in the amorphous form as long as it is strongly cooled down at a suffi-
ciently rapid rate [2]. Thus, the ability to create glassy state is almost a universal
property of matter [3]. The amorphous state is a very particular state of matter that
practically cannot be included in the traditional classification of states: solid
(understood as crystalline), liquid, and gas. Those three physical states differ from
one another mainly due to the degree of order inside the materials in a given state,
i.e., lack of order in the gas state, local order in microareas in the case of liquids,
and complete order (so-called long-range order) in crystalline substances.
W. Mozgawa M. Sitarz M. Król (&)

e-mail: mkrol@agh.edu.pl

458 W. Mozgawa et al.
Of course, the consequence of these different orders is completely different prop-

erties of matter in the respective physical states.
Taking into account its mechanical properties (hardness, fragility, rigidity, etc.),
glass can certainly be classified as a solid material. However, in contrast to
crystalline bodies, glass does not exhibit long-range order characteristic for the
solid state. Moreover, the presence of local density fluctuation and lack of order in
its structure resulting in isotropy of all vector physical properties indicate similarity
between glass and liquids. Still, it should be noted that the position of atoms, ions,
or molecules in glass is not completely random. Glass, much like crystalline solids,
possesses a certain short-range order as a consequence of chemical bonds
responsible for the cohesion of solid materials [3]. Therefore, it can be stated that
the glass assumes a state between crystalline and liquid and is classified as
amorphous.
It is obvious that structural differences in glass result in physical and chemical
property differences. Finding a correlation between the glass structure and
physicochemical properties of the materials has become the reason to perform
studies on a large scale for determining the structures of various glasses. Becoming
familiar with the internal structure of glass undoubtedly has key importance in
designing new materials with planned out and controlled properties.
The first considerations on the topic of glass structure pertained to silicate glass
(vitreous SiO2; m-SiO2). The interest in silicate glass is mainly the result of silicates
becoming widespread, as well as the belief that understanding m-SiO2 structures is
the starting point for explaining the structures of more chemically complex silicate
glasses. In the literature, there are two predominant concepts concerning m-SiO2
internal structure, i.e., Lebedev theory [4] and Zachariasen theory [5].
Historically, the first hypothesis on m-SiO2 structure was formulated by Lebedev
in 1921 [4]. This gave birth to the so-called crystallite theory of glass structure.
Lebedev assumed that silicate glass is an aggregation of microregions with a very
high degree of fragmentation, connected with each other in a statistically disordered
manner (random, stochastic) (Fig. 15.1a). These crystallites constitute around 25%
of the glass volume, whereas the remaining part is in a disordered state. The
crystallite theory was experimentally supported by Randall et al. [6]. Based on the
line width in powder diffractograms, they assumed that crystallites built of a
high-temperature cristobalite structure with 15–20 Å dimensions are present in
m-SiO2. On the other hand, Valenkow and Porai-Koshitz [7] assumed that crys-
tallites with 10–12 Å dimensions are more ordered on the inside than on the surface
and they connect with one another using those surfaces in a continuous disordered
network.
In summary, the crystallite theory assumes the existence of submicroscopic
regions (domains, crystallites, paracrystals) in glass with a very high degree of
fragmentation, in which there is structural order equivalent to that present in crystals
with the same chemical composition—high-temperature cristobalite (a-cristobalite)
in the case of silicate glass. The aforementioned crystals connect with one another
through regions with a looser packing of structural elements (voids).
15 Spectroscopic Characterization of Silicate Amorphous Materials 459
Fig. 15.1 Schematic representation of m-SiO2 structure according to crystallite theory (a), and
continuous random network theory (b)
An alternate theory on the internal silicate glass structure was presented by

Zachariasen in 1932 [5], who assumed the existence of a continuous, uninterrupted
construction formed by [SiO4]4− tetrahedrons, statistically disordered and con-
nected with one another. This assumption is called the continuous random network
theory (Fig. 15.1b). Thus, a m-SiO2 glass network, according to this concept,
consists of an uninterrupted frame of randomly connected [SiO4] tetrahedrons, in
which every silicon ion is tetrahedrally surrounded by oxygen ions and every
oxygen ion is located between two silicon ions—thus, broken bonds are not present.
This theory has been experimentally supported by Warren [8–10], who determined
on the basis of radical atom density distribution the presence of proper tetrahedrons
arranged in a completely random manner and ordered regions in silicate glass not
exceeding 7–8 Å (with 5.7 Å tetrahedron dimensions). Therefore, the continuous
random network model limits the range of structural order to the first coordination
sphere, i.e., to one [SiO4] tetrahedron in silicate glass.
According to Görlich [1], the differences between these theories in fact come
down to the sizes of ordered regions—15–20 Å in the case of crystallite theory and
a single tetrahedron in the case of the continuous random network theory.
A graphical comparison of both concepts is illustrated in very idealized form of a
two-dimensional schematic representation in Fig. 15.1. After analyzing them, the
hypothesis of a continuous random network can be described as the hypothesis of a
homogeneous glass structure. On the other hand, the crystallite hypothesis is the
hypothesis of a heterogeneous m-SiO2 structure.
It can be easily noticed that experimental proof supporting one or the other
concepts of the internal glass structure is mainly based on X-ray diffraction
methods. However, taking into account that standard diffraction methods yield
results on average atom distribution, the application of these methods for studying
glass structures is very limited [11]. On the basis of these methods, only the lack of
long-range order can be definitively determined in the studied glasses; however,
determining homogeneity (Zachariasen–Warren theory) or heterogeneity (Lebedev
theory) inside glass is very difficult or even impossible. The development of other
diffraction techniques, e.g., with the utilization of neutron beams, X-ray fluores-
cence analysis, etc., would enable the determination of several glass network
parameters. Mozzi and Warren [12] determined on the basis of X-ray fluorescence
analysis that the length of the Si–O and O–O distance in m-SiO2 are 1.62 and
2.65 Å, respectively. Si–O–Si bond angles exhibit a normal statistical distribution
with a maximum around 152°. However, there is less variety in the values of O–Si–
O angles around the tetrahedron angle equal to 109.5°. These values are comparable
to those determined for several crystalline silicates [13, 14]. The initial advantage of
the continuous uninterrupted network model was due to its support from more
precise and more graphic experiments [8–10]. On the other hand, many authors
[15–17] have called attention to the presence of characteristic points in property–
chemical composition or property–property curves, at which these curves change
their course. These points are usually associated with the boundaries of a certain
phase existing, or as in the case of the refractive index–temperature curve, the
temperature of cristobalite polymorphic changes. These phenomena, much like
many others, such as the presence of voids in glass (greater permeability of gases in
glass than in the respective crystal), cannot be explained by the Zachariasen–
Warren theory. All these phenomena can be explained under the assumption that
regions with chemical composition and structure similar to respective crystalline
substances are present in the glass structure. The existence of such regions (do-
mains) has been unambiguously confirmed by means of vibrational spectroscopy
studies (e.g., [18–22]), which have demonstrated a great similarity between m-SiO2
MIR spectra and those obtained for high-temperature cristobalite. Therefore, current
views on glass structure lean more in the direction of the crystallite theory; how-
ever, the predominant opinion is that the structure of real glasses is somewhere in
between, with characteristics corresponding to one of the theories [16]. The range
of order is currently the subject of the most controversy; however, most authors
agree that short-range order is close to respective crystalline structures. Assuming
that middle-range order is present in silicate glass, there remains the question of the
type of order within domains and whether or not the domains only have one
characteristic (homogeneous build) or different kinds of order (heterogeneous build)
(Fig. 15.2).
In the literature pertaining to m-SiO2 structure, many pseudo-crystalline models
can be found postulating the existence of ordered regions (domains) with different
SiO2 polymorphs (e.g., [11, 23–29]). These are most common models that assume
homogeneous build, i.e., with domains exhibiting only one characteristic order.
Still, there is no universal consent among scientists pertaining to the characteristics
Fig. 15.2 Schematic glass structure containing only one type of domain order (a), and with a
heterogeneous build (b)
of the assumed order in domains. The most common assumption is that the ordered
domains more [23] or less [24] resemble the order in high-temperature cristobalite.
There are also theories that postulate the existence of ordered domains with tri-
dymite structure [25–27]. The model proposed by Goodman [28] can be
acknowledged as a compromise for the controversy concerning the type of domain
order. This model postulates the existence of domains exhibiting the structure of
several SiO2 polymorphs—the so-called strained mixed cluster model. At the
boundaries of such domains, compressive and tensile stresses are present due to the
structural mismatch of individual silicate polymorphs.
A large problem when describing a silicate glass network is also determining the
structure of disordered intermediate regions located between the domains. The
disorder present in these regions constitutes the difference between glass and a
polycrystalline structure. The disorder is so large that it eliminates long-range
translation symmetry. There are doubts concerning the way to carry out mutual
topological fitting of the domains through disordered regions. It is known that in
this range the network is more relaxed (voids) and chaotic. Up to now, it has not
been determined whether there is only one type of deformed atom arrangement
sequence, or if there are more (Fig. 15.3).
It seems that the most consistent model of silicate glass structure, developed on
the basis of detailed XRD and spectroscopic studies, was presented by Görlich
[1, 30–32]—domain model of glass structure. According to this concept, m-SiO2
consists of a relaxed and randomly mixed texture of tetragonal domains in
high-temperature cristobalite, between which empty spaces are located, creating a
Fig. 15.3 Ways to connect ordered regions: in one type of arrangement sequence (a), and in
different arrangement sequences (b)
microporous material. The number and length of silicon–oxygen (Si–O) bridges

between domains are random. Stresses are present between domains, thanks to
which the bridges can be easily broken, which explains the very wide interval of
glass softening. During m-SiO2 crystallization, topological matching of domains
takes place, which orient themselves relative to each other, correctly creating
a-cristobalite crystals—crystallization proceeds without diffusion. Thus, differences
between a high-temperature cristobalite crystal (ordered domain arrangement) and
silicate glass (statistically distributed domains) in fact come down to differences in
the mutual arrangement of domains and not individual atoms. A picture of silicate
glass structure, according to this theory, is illustrated in Fig. 15.4. A similar view
on glass structure can be found, inter alia, in the works of Verwej and Konijnendijk
[33]. It can be stated that the domain theory of silicate glass build proposed by
Görlich is a certain compromise between Lebedev’s crystallite theory and
Zachariasen–Warren’s continuous random network theory—ordered domains exist
in glass, but there is also a continuous lattice (Fig. 15.4).
General concepts pertaining to silicate glass structures can be adopted for
describing the structures of more complex silicate glasses, in which cation substi-
tutions (in the place of silicon ions—so-called network formers) and/or anion
substitutions (in the place of oxygen ions) are present. Additionally, so-called
network modifiers, i.e., cations that break Si–O bridges, can also be present in glass
structures.
The introduction of other cations (network formers or modifiers) or anions into a
silicate glass structure leads to rapid changes in the physicochemical properties of
the glass, which, of course, is the consequence of glass structure changes.
Fig. 15.4 m-SiO2 structure

according to Görlich’s
domain theory.
High-temperature cristobalite
domains are connected with
each other in a statistically
ordered way
15.2 Vibrational Spectroscopy as a Tool to Study

of Disordered Structure
Due to the lack of long-range order in amorphous structure, the choice of one
specific experimental method that constitutes a fundamental tool for their structural
analysis is very difficult. In the case of crystalline materials, X-ray diffraction is
seen as that method and is most frequently used for the aforementioned analysis.
However, it requires, as in the case of most diffraction methods, order that extends
over the area of several hundred interatomic distances. Thus, it cannot be directly
used to determine glass structures, in which such order is not present.
Regardless of the glass structure model taken under consideration, the only order
that can be stated with certainty in this material is so-called short-range order, i.e.,
order that extends only to the closest coordination zones around a given atom or
ion. As previously mentioned, some theories also assume an order that extends
further than polyhedron dimensions, which constitutes intermediate-range order
[18, 34]. It can generally be assumed that the dimensions of such areas do not
exceed several dozen Å. Therefore, the fundamental conditions for using a given
experimental method in studies on glass structures should be that it yields infor-
mation on structure fragments with such small dimensions.
There are some measurement methods and techniques that meet the above-
mentioned condition; however, the most commonly used are spectroscopic meth-
ods. Spectroscopic methods are primarily molecular spectroscopy, which includes
vibrational, nuclear magnetic resonance and electron paramagnetic resonance
spectroscopy. Among them, a very important, if not the most important, one is
vibrational spectroscopy, i.e., infrared (IR) spectroscopy and Raman spectroscopy.
Using each of these methods, vibration frequencies of constituent atoms or their
coordination groups can be determined. On that basis, conclusions can be made
about the arrangement of atoms in a molecule, as well as the type of chemical bonds
between the atoms [35], and as a consequence about the aforementioned
short-range order. Thus, vibrational spectroscopy is considered a very important
tool for glass structure studies.
15.2.1 Decomposition of IR Spectra of Glasses
One of the difficulties in interpreting IR spectra is coincidence of bands, which

causes the formation of indistinct signal envelopes consisting of overlapping
maxima. This is the result of the large full width at half maximum (FWHM) of
certain bands characteristic for spectra received from analyzing amorphous bodies.
IR spectra decomposition is a mathematical procedure, in which spectra obtained
during measurement are subjected to the appropriate procedure in order to receive
individual component bands with precisely defined parameters. From this, much
more information can be obtained that allows for an in-depth interpretation.
Aside from several factors characteristic for all solids, the complex form of
amorphous body spectra is influenced by a number of additional phenomena
associated with the specific structure of the glassy state. Silicate glass spectra, while
similar to spectra of their respective crystalline substances, exhibit several separate
features only present in the glass phase. Compared to analogous crystalline spectra
bands originating from lattice, vibrations disappear and mainly those associated
with short-range order remain. The bands have a significantly larger full width at
half maximum and an altered shape. This is the result of, inter alia, a large dis-
tribution of lengths and bond angles, as well as disorder in mutual orientation of the
structure fragments. Widening the bands leads to their coincidence and at the same
time a decrease in the number of well-defined maxima. According to Griffiths and
Pariente [36], the number of separated bands N can be expressed using the
following correlation:
ðm1 m2 Þ
N¼ ð15:1Þ
Rci
where (m1 − m2) is the considered spectrum range and ci—full width at half maxi-
mum of individual bands. Thus, as ci increases, the number of clearly defined bands
decreases and their overlapping occurs. This leads to indistinct peaks in a complex
envelope. The interpretation of such spectra often becomes difficult or even
impossible.
In order to fix this problem, it is necessary to separate complex envelopes into
individual component bands, i.e., spectra decomposition, which must then be
ascribed a certain physical meaning. Such elaboration of a measured spectrum often
requires subjective decisions of the people carrying out the operation, which can
lead to interpretation errors difficult to estimate. Therefore, spectra decomposition
cannot be a random operation and should be carried out in accordance with a
unified and logical procedure. The separation procedure for glass spectra proposed
by Handke and Mozgawa [19] consists of the following operations:
• deconvolution and second derivative of the initial spectrum, initial parameters,
and preliminary number of peaks;
• simulation of pseudo-spectra for an analogous crystalline material and sub-
traction from the glass spectrum, final number, and peak parameters;
• proper decomposition;
• minimalization of the number of component bands with the largest degree of
fitting complex spectra with the original spectrum;
• comparison with the initial spectrum of the crystalline material;
• ascribing a physical meaning to the bands obtained as a result of decomposition.
Mathematical procedures dedicated to the issue of spectrum decomposition were
developed in the form of several algorithms and corresponding, constantly devel-
oped, computer programs [37, 38]. In this work, the Spectra-Calc (SC) program of
Galactic Industries Corp. was used.
In order to decompose spectra, the following output parameters are entered:
number of peaks, their position, full width at half maximum, relative intensity, and
shape expressed by the appropriate function. This data is initially determined on the
basis of two mathematical procedures: calculation of the spectrum’s second
derivative and deconvolution carried out in accordance with the method proposed
by Griffith and Pariente [36]. The spectrum’s second derivative allows for deter-
mining inflection points and for precisely determining the positions of band max-
ima. Similar information along with subsequent spectra parameters—full width at
half maximum, absolute intensity—is obtained with the help of the deconvolution
operation. As a result of this action, separate peaks are formed by the program in
maxima positions and inflection points of the spectrum. The algorithm of this
operation is based on the theory of Fourier cosine transform [36], in which the
function established after deconvolution is described by the following formula:
exp½2pðki c0 Þx
F 0 ð xÞ ¼ F ð xÞ expð2pc0 xÞ ¼ 0:5A0i ci cos 2pm0 x ð15:2Þ
Rci
This formula consists of three parts:

• 0:5A0i ci —is directly proportional to the band area;
• cosð2pm0 xÞ—depends on the frequency of band maximum;
• exp½2pðki c0 Þx—affects band width after deconvolution.
During the process, the degree of deconvolution is determined by the change in
parameters c′ and x present in the equation provided above. Parameter c′ is directly
associated with the width of formed peaks and as a consequence with the number of
bands. On the other hand, x determines the degree of Fourier filtration and enables
smoothing of the obtained spectrum.
Both procedures, i.e., deconvolution and second derivative, can be used together
to determine output parameters for the actual spectrum decomposition. It should,
however, be noted that mechanically transferring values obtained in their result can
sometimes provide misleading results. This mostly pertains to the number of peaks
used for decomposition. Mysen et al. [39] propose that when establishing the
number of bands in the decomposition of silicate glass spectra, the rule of mini-
malization of the number of peaks should be followed. Having initial data
describing spectrum topology, begin with the least possible number of bands and
then increase the number during the following cycle. After subsequent steps, the
degree of fitting between the calculated spectrum (i.e., the superposition of spec-
trum components obtained as a result of decomposition) and the original spectrum
should be checked, using the v2 coefficient (least squares method). If that coefficient
does not significantly change, return to the number of bands from the previous
step. The algorithm of the operation of checking the degree of fitting is based on the
following correlation [40]:
X
n
v2 ¼ ½wr f ðxr =qÞ ð15:3Þ
r¼1
where: n—total number of points in the spectrum; wr—weight function, which in

this case denotes y−1r ; xr and yr—values of measurement point coordinates of the
initial spectrum; f ðxr =qÞ—value of a dependent variable of the spectrum fitted in
point x for parameter q.
After every calculation step, the band parameters should be corrected so that the
value of v2 coefficient is as low as possible, which means better compatibility of
both curves.
15.3 Glass Structure in the Light of Spectroscopic Studies
15.3.1 Model Silica Glass
Silicate glass (m-SiO2) is a typical representative of amorphous bodies. It is rec-

ognized as the most important, even model example of a solid with a disordered
structure. Thus, the m-SiO2 structure often constitutes the starting point for studies
on glasses with more complex compositions.
As previously mentioned, it is assumed that, within the short and intermediate-
range order, glass structure corresponds to the order of respective crystalline forms.
In Fig. 15.5, infrared spectra of silicate glass and SiO2 in crystalline polymorphic
forms were compared. They exhibit a clear similarity, making it difficult to
unambiguously state which crystalline SiO2 form is closest to the structure of
silicate glass. However, analysis of IR spectra [19, 20, 41] proves that the
Fig. 15.5 MIR spectra of silicate glass (m-SiO2) and crystalline polymorphic forms of SiO2
collected at 300 K (a) and at 10 K (b). Adapted with permission from [19], copyright (1993)
Elsevier
high-temperature crystalline cristobalite polymorphic form should be recognized as

the analog of silicate glass. Still, in spite of clear similarity, element characteristic
for only a disordered structure can be pointed out. It can be demonstrated that
m-SiO2 structure is close to the model proposed by Görlich [1].
According to [42], the silicate skeleton can be viewed as a lattice of structural
fragments similar to molecules, connected with one another, in which the smallest
of these fragments are tetrahedrons [SiO4] or Si–O–Si bridges (Fig. 15.6). Of
course, this is a considerable approximation, because in the case of solids, espe-
cially with a mesodesmic structure, there is no such thing as an isolated molecule.
In most theoretical works, aiming to describe the structure of SiO2 polymorphs, a
[SiO4] tetrahedron is accepted as an elementary monomer unit (e.g., [43–45]).
Using such an approach, silicate is considered as a kind of non-organic polymer
created from tetrahedrons connected with one another by means of bridging oxygen
atoms, where the differences in mutual orientation of tetrahedrons are caused by
different structure symmetries that change into subsequent SiO2 polymorphic forms.
On the other hand, the similarity of spectra of individual SiO2 polymorphic
forms (Fig. 15.5) can be understood by assuming another monomer structural unit
than [SiO4] tetrahedron in silicon dioxide [46]. In all spectra illustrated in
Fig. 15.5a, three groups of bands in the same wavenumber ranges with comparable
intensity relations repeat themselves. In accordance with the proposition of Bell and
Dean [47] (supported, inter alia, in theoretical works of Laughlin and Joannopoulos
[48]), assuming separation of energy of individual oscillations, these bands can be
attributed to three types of normal vibrations:
• bands in the range of 1200–1000 cm−1 originating from stretching vibrations of
asymmetric Si–O bonds (mas Si–O);
• bands in the range of 850–600 cm−1 originating from symmetrical stretching
vibrations of a Si–O(Si) bond (ms Si–O);
• bands around 500 cm−1 originating from bending vibrations of O–Si–O bonds
(d O–Si–O).
Fig. 15.6 Six-membered

ring from the structure of
high-temperature cristobalite.
[SiO4] tetrahedron and Si–O–
Si pseudo-molecule have been
highlighted
From this perspective, a Si–O–Si (Fig. 15.6) bridge in all analyzed structures is
the result of the silicon–oxygen spatial connection with an adjacent polyhedron and
the consequence of its presence is mutual elements in IR spectra. In the geometric
sense, such a unit is close to the well-known in spectroscopy literature angular H2O
molecule [49] with heavier (compared to hydrogen) outermost silicon atoms and
with a smaller angle between bonds. As in the case of H2O molecules, three
oscillations (3N−6) are present in a Si–O–Si unit. These types of molecules can be
described using a C2v point group, in which the symmetry elements are a double
axis and two symmetry planes perpendicular to one another. A reducible repre-
sentation of the oscillation is expressed as:
Cosc ¼ 2A1 þ B1 ð15:4Þ
All three vibrations, characteristic for this group (Fig. 15.7), are active in
infrared and can be assigned to the aforementioned three main bands present in the
discussed spectra (Fig. 15.5a).
The correctness of the Si–O–Si molecule model was verified on the basis of
spectra measurements of several silicate polymorphic forms, which contained iso-
tope substitutions of silicon and oxygen atoms [50, 51]. It follows from these
studies that in the SiO2 structure oxygen is the more mobile atom. This means that
spatial lattice vibrations of silicate are predominantly due to oxygen sublattice
vibrations. In this case, all oxygen atoms are bridging atoms. That does not con-
tradict the proposed model, in the center of which is bridge oxygen. This reasoning
can also be generalized for other framework silicates.
As can be seen in Fig. 15.5, spectra of each of the polymorphic forms of silicate
vary in the presence (or lack) of bands associated with characteristic structural
elements of a given form—bands, which unambiguously identify only a given
phase and exist only in the spectrum of this form, can be indicated. After assuming
a Si–O–Si bridge as a quasi-molecular elementary unit, all the remaining vibrations
(aside from the three previously mentioned) can be treated as external lattice
vibrations:
• typical crystalline lattice vibrations yielding bands below 400 cm−1, i.e., in the
far infrared range (indicating the existence of long-range order—not present in
m-SiO2 spectrum; Fig. 15.5a);
Fig. 15.7 Si–O–Si bridge with three fundamental types of vibration highlighted
• pseudo-lattice vibrations (bands associated with these vibrations much more

clearly change with temperature than internal Si–O–Si unit vibrations;
Fig. 15.5b).
In order to differentiate between bands associated with lattice vibrations and
molecular-type vibrations, spectra measurements in helium temperatures (around
10 K) can be used (Fig. 15.5b). In accordance with the Rakow effect [52], the first
type of vibrations exhibits significant sensitivity to temperature changes. Bonds
associated with those vibrations significantly vary in shape, integral intensity, and
position in individual spectra (Fig. 15.5a), which is the result of the different
geometries of individual polymorphic forms. Such a classification of vibrations is
confirmed by the lack of clear and well-defined bands of this type in a silicate glass
spectrum, which does not exhibit long-range order. Still, in the glass spectrum in
the range of 800–500 cm−1, weak bands are noticeable, which are most likely
associated with larger than molecular structure fragments and indicate the existence
of a certain kind of order in a range greater than a single coordination polyhedron
[21].
As previously mentioned, the crystalline structural model, most commonly
assumed when interpreting silicate glass spectra, is the structure of a high-
temperature cristobalite polymorph [19, 30]. Figure 15.8 illustrates the
decomposition of cristobalite MIR spectra. As a result of decomposition, three
dominant component bands were obtained (1115, 789, and 487 cm−1), which can
be attributed to the described internal vibrations of an Si–O–Si bridge—mas, ms, and
d, respectively. Aside from the specified, spectrum decomposition revealed the
presence of two additional bands: 1208 and 701 cm−1. The first enters the envelope
of the most intensive band around 1100 cm−1 causing its asymmetry relative to the
maximum. Higher vibration frequency relative to the position of band 1115 cm−1
can mean that it originates from the stretching vibrations of a double bond between
silicon and oxygen [53]. The presence of such bonds in cristobalite is probable, and
its existence can be considered as a type of defect in SiO2 structure. Because this
band exhibits a much lower integral intensity, much lower concentration of that
type of bonds compared to single bonds can be estimated. The second band may be
associated with pseudo-lattice vibrations of six-membered rings [20].
For comparison, Fig. 15.8 presents the decomposition of a low-temperature
cristobalite polymorph. Aside from bands at 1202, 1099, 795, and 491 cm−1, the
positions of which are close to respective bands in the spectrum of the
low-temperature cristobalite and can be attributed to identical vibrations. In a
high-temperature cristobalite spectrum, bands at 1159, 621, and 522 cm−1 are
present. The presence of a band at 621 cm−1 is especially characteristic for this
polymorphic form and is not observed in any other SiO2 spectrum. Its
disappearance at higher temperatures is equated to the polymorphic
high- ! low-temperature cristobalite and the change from tetragonal to higher
(pseudo-regular) symmetry [54]. All three previously mentioned bands clearly
become sharper in the spectrum obtained at low temperatures [19] and are therefore
associated with lattice vibrations. Decomposition of a quartz spectrum (Fig. 15.8c)
Fig. 15.8 Decomposition of MIR spectra of various polymorph of SiO2: high-temperature

cristobalite (a), low-temperature cristobalite (b), quartz (c), and m-SiO2 (d)
can be analyzed in a similar way. A characteristic trait of this spectrum is the

presence of doublet bands at 799 and 779 cm−1. Due to decomposition of quartz
spectrum, a band at 1014 cm−1 is also revealed, the presence of which can indicate
the existence of broken oxygen bridges in its structure [55].
In Fig. 15.8d, the decomposition of a silicate glass spectrum is presented. As a
result of decomposition, five bands were obtained (1196, 1106, 800, 733, and
470 cm−1), the positions of which are close to respective bands in crystalline
polymorph spectra and subsequently are identically assigned as in the
high-temperature cristobalite spectrum (Fig. 15.8a). However, integral intensities of
individual component bands clear change; for example, the intensity of the band at
1196 cm−1 associated with the presence of Si=O bonds increases, which can be
easily explained by the fact that the probability of double bonds in glass is much
greater than in a respective crystalline structure.
Another important difference revealed by m-SiO2 spectrum decomposition

compared to spectra of SiO2 crystalline forms (Figs. 15.8a–c) is the appearance of
an intensive band at 1044 cm−1, the presence of which can indicate great con-
centration of Si–O− bonds (formed by breaking Si–O–Si bridges), which, in turn, is
associated with regions exhibiting a more relaxed structure density that include
intermediate regions between ordered domains. Moreover, the decomposition
demonstrated the presence of another band at 591 cm−1 in the m-SiO2 spectrum,
which like the band at 733 cm−1, can originate from vibrations of larger than
tetrahedron elements in the glass structure [21].
15.3.2 The Influence of Cationic Substitutions on Glass

Structure
Due to the model nature of the silica glass structure, its description is the starting point
for the study of more complex systems. The introduction of modifying ions (alkali
ions) or various glass-forming ions (e.g., Al3+ ions) into the glass structure signifi-
cantly impedes the interpretation of infrared spectra. The presence of alkali ions in the
glass structure leads to breaking of Si–O–Si bridges, resulting in the appearance of a
large number of Si–O− bonds. On the other hand, the [SiO4] ! [AlO4] substitution
leads to the appearance in the spectrum of bands associated with the vibration of Al–O
bonds.
Based on the analysis of IR spectra obtained for silicate glasses with an
increasing contribution of alkali, it is possible to follow the influence of modifying
ions on the degree of polymerization (its lowering) of a silicate spatial lattice.
Figure 15.9 illustrates MIR spectra of potassium–silicate glasses with increasing
K2O content and the spectrum of pure silica glass. As the K+ ion addition increases,
the band at around 1100 cm−1 in the m-SiO2 spectrum gradually becomes a doublet,
in which the intensity of bands with maxima at lower wavenumbers increases. The
band at around 1100 cm−1 in the silicate glass spectrum originates from mas Si–O
(Si) vibrations. The presence of potassium ions in the structure causes the breaking
of Si–O–Si bridges, and thus, bands at lower wavenumbers originating from mas Si–
O− begin to appear [56]. Simultaneously, as the K2O content increases, a shift in
both bands occurs in the direction of lower wavenumbers, which also indicates a
decrease in the degree of polymerization in silicates [19].
Interpretation of MIR spectra of glass containing alkaline ions in its composition
is much more difficult than in the case of m-SiO2 due to the complexity of the
envelopes. Figure 15.10 presents the decomposition of spectra obtained for crys-
talline layered lithium, sodium, and potassium silicates, as well as their respective
amorphous counterparts. A clear increase in the number of component bands can be
observed in comparison with the m-SiO2 spectrum (Fig. 15.8d), especially in the
case of lithium glass. Aside from depolymerization of the glass network, Li+ cation
has the tendency to substitute silicon in tetrahedral positions, which can be
Fig. 15.9 MIR spectra of glasses from the K2O–SiO2 system. Adapted with permission from [19],
explained by the large number of bands in the spectrum and is also reflected in the
crystalline form of this silicate.
It cannot be unambiguously determined which ion has the greatest impact on the
degree of glass network depolymerization; however, an interesting observation can
be the complete lack of bands associated with the presence of Si=O bonds. It can be
stated that alkaline ions eliminate the presence of multiple bonds in the glass
structure.
From the analysis of the spectra decomposition performed for selected silicate
glasses (Fig. 15.10), it can be concluded that compared to m-SiO2, larger than
Fig. 15.10 Decomposition of MIR spectra of glassy and crystalline Li2[Si2O5] (a and b),
Na2[Si2O5] (c and d), and K2[Si2O5] (e and f). Reprinted with permission from [20], copyright
(1999) Elsevier
tetrahedron order also exists in more complex systems. Based on the results pre-
sented, inter alia, in the work of Sitarz et al. [20], it can be stated that this order is
analogous to that inside the respective crystalline form. IR spectra decomposition of
glasses with compositions corresponding to that of Na2[Si2O5] (Fig. 15.10c) and
K2[Si2O5] (Fig. 15.10d) reveals two bands in the first case at 671 and 587 cm−1,
and 661 and 599 cm−1 in the second case. These bands as in the case of m-SiO2
(Fig. 15.10d) can be associated with ring vibrations. In the case of glass composed
of Li2[Si2O5] (Fig. 15.10a), in the pseudo-lattice range, an entire series of bands are
visible, the presence of which can indicate the existence of deformed rings with
different elements due to the small size of lithium and its above-mentioned capa-
bility of inserting itself into tetrahedron positions [57].
Similar conclusions can be derived from analysis of IR spectra obtained from
aluminosilicate glasses. Crystalline structures of potassium tecto-aluminosilicate
(orthoclase K[AlSi3O8], leucite K[AlSi2O6], and calsilite K[AlSiO4]) possess
aluminum atoms in tetrahedron positions. The existence of [AlO4]5− tetrahedrons is
associated with the presence of potassium ions K+ in the structure (they compensate
the excess negative charge), which locate themselves in free areas of the alumi-
nosilicate network [58]. MIR spectra obtained for glasses composed of these
aluminosilicates are illustrated in Fig. 15.11. They have a similar course as a
silicate glass spectrum (Fig. 15.6a); however, due to the presence of aluminum in
the structure and subsequently Si–O–Al bridges, there are no single bands, only
groups of bands that overlap with one another. Glass spectra can be divided into
three characteristic ranges:
Fig. 15.11 MIR spectra of

glasses from the K2O–Al2O3–
SiO2 system. Adapted with
permission from [19],
• bands in the range of 1200–800 cm−1 originating from asymmetrical stretching

vibrations of Si–O bonds taking place inside Si–O–Si (mas Si–O–Si) and Si–O–
Al (mas Si–O–Al) bridges;
• bands in the range of 850–600 cm−1 originating from symmetrical stretching
vibration of Si–O–(Si, Al) bonds (ms Si–O–Si and ms Si–O–Al);
• bands around 500 cm−1 originating from bending vibrations of Si–O–Si and
Si–O–Al bonds (d Si–O–(Si, Al)).
The first range is associated with (as in the case of m-SiO2) stretching vibrations
of the Si–O bond. These vibrations take place inside Si–O–Si or Si–O–Al bridges.
However, it should be noted that at high aluminum concentration, it is highly
possible that there is a third kind of bridge bond, i.e., Al–O–Al [59]. The existence
of such bridge should be considered as a type of defect in the aluminosilicate
network. Depending on the contribution of individual bonds in the structure, i.e.,
depending on the Si:Al ratio, the envelope of the spectra changes in that range. In
basic m-SiO2 glass, there are only Si–O–Si bridges; however, in calsilite (with the
Si:Al ratio equal to 1), only Si–O–Al bridges are present. Therefore, the band at
1107 cm−1 should be assigned to mas Si–O–Si, and the band at 997 cm−1 in the
spectrum obtained from a K[AlSiO4] assigned to mas Si–O–Al vibrations. In the
spectra of glasses with orthoclase or leucite composition, both bands (and both
types of bridges in their structures) are present, the positions of which are located
between those previously mentioned. The greater number of Si–O–Al bridges
(lower Si:Al ratio) in leucite than in orthoclase leads to a change in the band
position toward lower wavenumbers.
Of course, the presence of broken Si–O− bonds and multiple Si=O bonds cannot
be ruled out in the discussed structures, which is made visible in the spectra after
decomposition process (Fig. 15.12a). As in the case of silicate glass analysis
(Fig. 15.10), Fig. 15.12 compares the spectrum obtained from crystalline calsilite
(K[AlSiO4]) with the spectrum of glass with the same chemical composition. The
similarity in the spectra can be noticed; however, the individual bands in the glassy
spectrum exhibit a much greater full width at half maximum. On this basis, it can be
stated that in glass composed of K[AlSiO4], there are domains with calsilite
structure. It can also be confirmed that potassium ions cause the disappearance of
Si=O bonds (no band at around 1200 cm−1).
Based on the spectra of glasses from the K2O–Al2O3–SiO2 system (Fig. 15.11),
interesting information can be derived from analyzing the pseudo-lattice range, i.e.,
800–650 cm−1. In this range, there are bands with relatively small integral inten-
sities; however, in this range, the greatest changes associated with differences in the
composition of the analyzed glasses can be observed. There are at least two bands: a
band assigned to ms Si–O–Si vibrations in the m-SiO2 spectrum at 800 cm−1 and a
band assigned to ms Si–O–Al vibrations in the calsilite spectrum at 694 cm−1. The
lower aluminum mass leads to a change in the band position in the direction of
lower wavenumber values. m-SiO2 and glass with calsilite composition constitute
two extreme cases of the network. In the case of m-SiO2, there are only Si–O–Si
Fig. 15.12 Decomposition of MIR spectra of glassy (a) and crystalline (b) K[AlSiO4]. Reprinted
with permission from [20], copyright (1999) Elsevier
bridges; however, in the case of calsilite glass, there are only Si–O–Al bridges. In
leucite and potassium feldspar, both connections are formed simultaneously, which
is why two maxima are observed in their spectra. The correctness of attributing
bonds in this region to bridge vibrations is indicated by the change in the ratio of
their intensities along with aluminum concentration increase, i.e., as the number of
Si–O–Al bridges increased.
When discussing the spectra of aluminosilicate glasses, it should be remembered
that the influence of aluminum ions, present in material in an octahedral coordi-
nation, on these spectra is relatively difficult to determine. Tarte [60] suggests that
band characteristic for [AlO6] should appear in the range of 530–400 cm−1 in the
case of isolated octahedrons and in the range of 680–500 cm−1 for connected ones.
However, these are regions where bands associated with ring vibrations are present.
Thus, on the basis of IR spectra measurements, it cannot be unambiguously
determined to what degree such aluminum coordination exists in the glass structure
or if it is present at all. Of course, its presence cannot be ruled out.
Another operation that facilitates interpretation of spectra with complex envel-
opes is the possibility of creating their arithmetic sum or difference. Summation
should be performed on spectra comprising an identical absorbance scale, after first
aligning them. This alignment consists of correcting the baseline and using the
appropriate weights for added spectra, proportional to the percentage share of
individual components. Of course, the spectrum obtained as a result of summation
cannot be considered as a kind of averaged spectrum, inter alia, due to different
integral radiation absorption coefficients of vibrations in the initial spectra.
However, after verifying the result with an experimentally obtained spectrum, the
structural heterogeneity of multi-component glasses can be confirmed on the basis
of that operation.
Fig. 15.13 Summation of spectra of glasses (a) and their devitrificates (b) with the composition of
K[AlSi2O6] and K[AlSiO4]
Figure 15.13 illustrates the summation of spectra obtained from glasses with
leucite K[AlSi2O6] and calsilite K[AlSiO4] compositions with a weight of 1:2. The
result of such an operation is spectra almost identical to those obtained from
measurements performed on samples of glass composed of K3[Al3Si4O14]. Thanks
to this operation, bands originating from individual summation components, i.e.,
from leucite or calsilite, can be once again identified in a complex spectrum, as well
as in a K3[Al3Si4O14] spectrum.
Concepts pertaining to silicate glass structure and the method of spectra inter-
pretation can also be adopted for description of more complex systems, which
include phosphosilicate (e.g., [61–63]) or borosilicate glass (e.g., [64–66]).
15.4 Conclusions
The application of a triatomic Si–O–Si quasi-molecule model as a monomeric unit

to replicate the crystalline and glassy silicates structure—better than a [SiO]4
tetrahedron model—explained their IR spectra. In this approach, three basis groups
of bands, repeatedly in all spectra in similar integral intensity relations and similar
wavenumber ranges, originate from internal Si–O–Si bridge vibrations—mas, ms, and
d Si–O–(Si, Al), respectively. In the case of aluminosilicates, additional bands
appear in the spectrum as a result of bending vibrations of Si–O–Al pseudo-
molecule (695 cm−1).
With the help of correctly performed mathematical procedures (inter alia
decomposition, subtraction, or summation of spectra), important information can be
obtained, which allows for drawing conclusions on the structure of silicate glasses.
On the basis of the presented considerations, it can be stated that in glass structure,
there are regions with different structural orders. This order is similar to that present
in the respective crystalline phases. In addition to bands corresponding to crys-

talline analogs, bands associated with the presence of a large number of Si=O
multiple bonds and broken Si–O− oxygen bridges also appear. They can be con-
sidered as a type of structural defect in comparison with the crystalline form. On the
basis of MIR spectra, the influence of alkali on lowering the degree of polymer-
ization of a spatial silicate lattice can be followed.
In spite of the fact that in glassy spectra bands from lattice vibrations associated
with long-range order of a crystalline lattice disappear, bands indicating the pres-
ence of order in a range exceeding the scale of a single coordination polyhedron
appear in these spectra. The range 700–550 cm−1 in infrared spectra is most often
associated with bands originating from pseudo-lattice vibrations in silicates. They
can be found in both spectra of crystalline forms, as well as in spectra of glassy
forms. They are characteristic for order larger than a molecule and indicate the
existence of so-called intermediate-range order. They mostly originate from
vibrations of ring systems present in the structure.
IR spectra can reveal the heterogeneous character of glass structure. In the case
of one-component m-SiO2 glass, they are caused by density fluctuations resulting
from large defect concentration. Next to regions with more relaxed structure
(voids), there is also region with very compact build (large number of Si=O bonds).
In the case of multi-components glasses, IR spectra can provide information about
the existence of regions with different forms of order. This order corresponds to
phases included in their composition. Determining heterogeneity in the glass
structure is an important argument that contradicts the existence of an uninterrupted
network inside glass. At the same time, it suggests the concept of a submicro-
heterogeneous texture of glasses, on the basis of which lies the crystallite theory.
References
1. Görlich E (1989) Stan szklisty. Skrypt uczelniany AGH nr 1155, Kraków (in Polish)
2. Turnbull D (1969) Under what conditions can a glass be formed? Contempt Phys 10:473–488
3. Zallen R (1994) Fizyka ciał amorficznych. PWN, Warszawa (in Polish)
4. Lebedev AA (1921) The polymorphism and annealing of glass. Trans Gos Opt Inst 2:1–20
5. Zachariasen WH (1932) The atomic arrangement in glass. J Am Ceram Soc 17:3841–3857
6. Randall JT, Rooksby HP, Cooper BS (1930) The structure of glasses: the evidence of X-ray
diffraction. J Soc Glass Technol 14:219–228
7. Valenkow N, Porai-Koshitz E (1936) X-ray investigations of the glassy state. Nature
137:273–274
8. Warren BE (1933) X-ray diffraction of vitreous silica. Z Kristallogr 86:349–358
9. Warren BE (1934) X-ray determination of the structure of glass. J Am Ceram Soc 17:249–254
10. Warren BE, Krutter H, Morningstar O (1936) Fourier analysis of X-ray patterns of vitreous
SiO2 and B2O2. J Am Ceram Soc 19:202–206
11. Wright AC, Leadbetter AL (1976) Diffraction studies of glass structure. Phys Chem Glasses
17:122–145
12. Mozzi RL, Warren BE (1969) The structure of vitreous silica. J Appl Cryst 2:164–172
13. Tossell JA, Gibbs GV (1978) The use of molecular-orbital calculations on model systems for
the prediction of bridging-bond-angle variations in siloxanes, silicates, silicon nitrides and
silicon suffides. Acta Crystallogr A 34:463–472
14. Baur WH (1980) Straight Si–O–Si bridging bonds do exist in silicates and silicon dioxide
polymorphs. Acta Crystallogr B 36:2198–2202
15. Demkina LI (1958) Issledowania zavisimosti swoistw stekoł ot ich sostawa. Izd Obr Prom,
Moskwa
16. Babcock CL (1977) Silicate glass technology methods. Wiley, New York
17. Babcock CL (1968) Substructures in silicate glasses. J Am Ceram Soc 51:163–169
18. Elliott SR (1991) Medium-range structural order in covalent amorphous solids. Nature
354:445–452
19. Handke M, Mozgawa W (1993) Vibrational spectroscopy of the amorphous silicates. Vib
Spectrosc 5:75–84
20. Sitarz M, Mozgawa W, Handke M (1999) Rings in the structure of silicate glasses. J Mol
Struct 511–512:281–285
21. Sitarz M, Mozgawa W, Handke M (2000) Identification of silicooxygen rings in SiO2 based
on IR spectra. Spectrochim Acta A 56:1819–1823
22. Sitarz M (2011) The structure of simple silicate glasses in the light of middle infrared
spectroscopy studies. J Non-Cryst Solid 357:1603–1608
23. Phillips JC (1986) Comments on “the J.C. Phillips model for vitreous SiO2: a critical
appraisal”. Solid State Commun 60:299
24. Hosemann R, Hentschel MP, Schmeisser U, Bruckner R (1986) Structural model of vitreous
silica based on microparacrystal principles. J Non-Cryst Solids 83:223–234
25. Konnert JH, Karle J (1973) The computation of radial distribution functions for glassy
materials. Acta Crystallogr 429:702–710
26. Konnert JH, Ferguson GA, Karle J (1973) Crystalline ordering in silica and germania glasses.
Science 179:177–179
27. Konnert JH, D’Antonio P, Karle J (1982) Comparison of radial distribution function for silica
glass with those for various bonding topologies: use of correlation function. J Non-Cryst
Solids 53:135–144
28. Goodman CHL (1975) Strained mixed-cluster model for glass structure. Nature 257:370–372
29. Gaskell PH (1991) Glasses and amorphous materials. In: Zarzycki J (ed) Materials science
and technology, vol 9. VCH, Weinheim, Germany, p 175
30. Görlich E (1977) Structure and phase transformations in glass. Rev Int Hautes Temp Refract
14:201–206
31. Görlich E, Błaszczak K (1977) Polymorphic transition in silica glass. Nature 265:39–40
32. Görlich E (1982) The structure of SiO2—current views. Ceram Int 8:3–16
33. Verwej H, Konijnendijk WL (1976) Structural units in K2O–PbO–SiO2 glasses by Raman
spectroscopy. J Am Ceram Soc 59:517–521
34. Gaskell PH, Eckersley MC, Barnes AC, Chieux P (1991) Medium-range order in the cation
distribution of a calcium silicate glass. Nature 350:675–677
35. Wilson EB, Decius JC, Cross PC (1955) Molecular vibrations. McGraw-Hill, New York
36. Griffiths PR, Pariente GL (1986) Introduction to spectral deconvolution. Trends Anal Chem
5:209–215
37. Zlokazov VB (1978) UPEAK—spectro-oriented routine for mixture decomposition. Comput
Phys Commun 13:389–398
38. von Meerwall E (1975) A general-purpose routine for the analysis of spectroscopic peak
shapes. Comput Phys Commun 10:145–154
39. Mysen BO, Finger LW, Virgo D, Seifert FA (1982) Curve-fitting of Raman spectra of silicate
glasses. Am Mineral 67:686–695
40. von Meerwall E (1975) A fortran code for automatic spectrum analysis on medium-scale
computers. Comput Phys Commun 9:351–1359
41. Handke M, Mozgawa W, Nocuń M (1994) Specific features of the IR spectra of silicate
glasses. J Mol Struct 325:129–136
42. Lazarev AN, Mirgorodsky AP (1991) Molecular force constants in dynamical model of
a-quartz. Phys Chem Miner 18:231–243
43. Dowty E (1987) Vibrational interactions of tetrahedra in silicate glasses and crystals. Phys
Chem Miner 14:80–93
46. Handke M, Mozgawa W (1995) Model quasi-molecule Si2O as an approach in the IR spectra
description glassy and crystalline framework silicates. J Mol Struct 348:341–344
47. Bell RJ, Dean P (1970) Atomic vibrations in vitreous silica. Discuss Faraday Soc 50:55–61
48. Laughlin RB, Joannopoulos JD (1977) Phonons in amorphous silica. Phys Rev B 16:2942–
2952
49. Barrow GM (1962) Introduction to molecular spectroscopy. MCGraw-Hill Book Company
Inc, New York
50. Galeener FL, Mikkelsen JC (1981) Vibrational dynamics in 18O-substituted vitreous SiO2.
Phys Rev B 23:5527–5530
51. Sato RK, McMillan PF (1987) An infrared and Raman study of the isotopic species of
alpha-quartz. J Phys Chem 91:3494–3498
52. Rakow AV (1962) Temperature dependence of the line width of infrared absorption spectra.
Opt Spektr 13:369–373
53. Khanna RK, Stranz DD, Donn B (1981) A spectroscopic study of intermediates in the
condensation of refractory smokes: matrix isolation experiments of SiO. J Chem Phys
74:2108–2115
54. Görlich E, Błaszczak K, Sieminska G (1974) Infra-red studies of vitreous silica at elevated
temperatures. J Mat Sci 9:1926–1932
55. Taylor WR (1990) Application of infrared spectroscopy to studies of silicate glass structure:
examples from the melilite glasses and the systems Na2O–SiO2 and Na2O–Al2O3–SiO2. Proc
Indian Acad Sci 99:99–117
56. Shiraishi Y, Kusabiraki K (1990) Infrared spectrum oh high temperature melts by means of
emission spectroscopy. High Temp Sci 28:67–77
57. Handke M, Nocuń M (1997) Vibrational spectroscopy of lithium silicates and aluminosil-
icates in crystalline form. Mikrochim Acta 14:507–510
58. Sharma SK, Philpotts JA, Matson DW (1985) Ring distributions in alkali- and alkaline-earth
aluminosilicate framework glasses—a Raman spectroscopic study. J Non-Cryst Solids
71:403–410
59. Tossel JA (1993) A theoretical study of the molecular basis of the Al avoidance rule and of
the spectral characteristics of Al–O–Al linkages. Am Mineral 78:911–920
60. Tarte P (1967) Infra-red spectra of inorganic aluminates and characteristic vibrational
frequencies of AlO4 tetrahedra and AlO6 octahedra. Spectrochim Acta A 23:2127–2143
61. Sitarz M (2008) Influence of modifying cations on the structure and texture of silicate–
phosphate glasses. J Mol Struct 887:237–248
62. Sitarz M, Rokita M, Handke M, Galuskin EW (2003) Structural studies of the NaCaPO4–SiO2
sol-gel derived materials. J Mol Struct 651–653:489–498
63. Li H, Liu S, Zhang T, Wu H, Guo S (2018) The evolution of the network structure in
tin-fluoro-phosphate glass with increasing temperature. J Non-Crystal Solids 492:84–93
64. Sitarz M, Bulat K, Olejniczak Z (2012) Structure and microstructure of glasses from
NaCaPO4–SiO2–BPO4 system. Vib Spectrosc 61:72–77
65. Bułat K, Sitarz M, Wajda A (2014) Influence of aluminium and boron ions on the
crystallization of silicate-phosphate glasses from NaCaPO4–SiO2 system. J Non-Cryst Solids
401:2007–2012
66. Kaur R, Singh S, Pandey OP (2013) Absorption spectroscopic studies on gamma irradiated
bismuth borosilicate glasses. J Mol Struct 1049:386–391
Chapter 16
Spectroscopy in the Analysis
of Artworks
Tomasz Łojewski and Barbara Łydżba-Kopczyńska
Abstract This chapter is dedicated to the application of selected spectroscopic

techniques to investigations on cultural heritage objects. The rapid technical
advancement of Raman instrumentation, observed in the recent years, positioned
this spectroscopy as an outmost tool in this field. The use of Raman spectroscopy in
the analysis of chemical composition is presented for several classes of heritage
materials: manuscripts, painting, ceramics, minerals, and amber. In Sect. 16.3, Vis
fiber optic reflectance spectroscopy is presented as a tool allowing one to obtain
information important for selecting proper preventive measures, in this case,
exhibition policies safeguarding artifacts against photodegradation. The technique
discussed here—the microfade testing (MFT)—allows monitoring color changes as
induced by the action of light on a selected spot on the artifact in real time, thus
giving the most direct, empirical clues to a possible future alteration of the objects’
appearance when it is exposed to light on a museum wall.
16.1 Introduction
Analysis of art and heritage objects using spectroscopic methods can provide
exceptionally valuable information on the chemical composition and morphology
of the object as well as technology of its creation. Information gathered in this way
is of a fundamental importance to conservators planning and executing their work
as well as for art historians studying ancient techniques. Using methods that allow
surface imaging, one can reach under the layers of paint and reveal the artist’s
earlier ideas or read the hidden content of the palimpsest or charred papyri.
T. Łojewski (&)
e-mail: lojewski@agh.edu.pl
B. Łydżba-Kopczyńska
Faculty of Chemistry, University of Wroclaw, F. Joliot-Curie 14,
50-383 Wroclaw, Poland

484 T. Łojewski and B. Łydżba-Kopczyńska
Knowledge gathered through spectroscopic analysis deepens our understanding of

degradation processes, past and present, and thus provides scientific bases for
proper preservation measures to ensure safe future for our heritage. The fact that
most spectroscopic methods can be used on art objects directly, without any
sampling, makes art and heritage artifacts a particularly fruitful field of activity for
spectroscopists.
The first study of artworks encompassing chemical analysis of pottery, glass, and
porcelain was accomplished by Rene-Antoine Ferchault de Réaumur (1683–1757)
between 1719 and 1739 in France [1–4]. Then at the beginning of the nineteenth
century, Sir Humphry Davy (1778–1829) published the study on pigments from the
wall paintings discovered during the excavation in Pompeii. In both studies, the
specimens had to be destructed in order to achieve the chemical identification of the
pigments used by the artists [5]. From the early beginnings mentioned above, we
observe the contribution by chemists, applying chemical knowledge to solving
various questions referring to the wide diversity of the heritage objects. Martin
Heinrich Klaproth (1743–1817) determined the composition of some Greek and
Roman coins [6], previously mentioned Sir Humphry Davy identified synthetic
pigment (later called Egyptian Blue) among the ancient pigments collected in Rome
and Pompeii [6]. Jöns Jakob Berzelius (1778–1848) investigated the composition of
ancient bronzes [6], and Michael Faraday (1791–1867) studied Roman pottery
bronze [6]. Unfortunately, at that time the chemical analysis was associated with the
undesired necessity of complete destruction of the analyzed objects.
At the beginning of the twentieth century, the rapid development of modern
instrumentation and analytical techniques was accompanied by the growth of sci-
entific investigations of art and archeological objects. However, the destructive
approach to study of the artworks has not changed significantly when in 1922
Eccles and Rackham published their work [7] on the wet chemical analyses of
English, Welsh, Chinese, and continental porcelain from the Victoria & Albert’s
Museum collection. The valuable specimens were destroyed in order to establish
their chemical composition [5].
In this chapter, authors present in details only a selection of spectroscopic
techniques applied in heritage studies, focusing on the ones with a particular
importance for identification of artwork constituents (Raman spectroscopy) and a
technique where visible light spectroscopy is used with a unique purpose of pre-
dicting extent of possible future degradation of the artifact upon exposure to light
(microfading tests).
16.2 Raman Spectroscopy
In 2016, the Philosophical Transactions of the Royal Society A: Mathematical,

Physical and Engineering Sciences, 201 years after publishing Davy’s paper,
published a special issue presenting the diversity of the cultural heritage materials
and objects studied by means of the Raman spectroscopy alone or with the
16 Spectroscopy in the Analysis of Artworks 485
conjunction with other analytical techniques [5]. Over last two decades, the Raman
spectroscopy became one of the leading techniques in the investigations of the
cultural heritage objects like frescos, mortars, paintings (including wall and rock
paintings), plates, plasters and textiles, pigments, dyes, paints, fibers and inks,
glasses, ivory, resins, porcelain, pottery, bronzes, fossils, seashells, gemstones, and
pearls. The extensive knowledge gathered during the study of such versatile
materials and objects resulted in publishing of several reviews [8–13], databases
[14–20] dealing with the Raman spectroscopy, and its application combined with
other techniques used in this field [5, 21]. Most of the cutting-edge developments
were presented as communications to the International Congress on the
Applications of Raman Spectroscopy in Art and Archaeology (RAA). The most
significant contributions were published in proceedings from consecutive meetings
in London 2001 [22], Ghent 2003 [23], Paris 2005 [24], Modena 2007 [25], Bilbao
2009 [26], Parma 2011 [27], Ljublana 2013 [28], and Wroclaw 2015 [29].
The Raman effect was discovered by Raman in 1928 at the Indian Association
for the Cultivation of Science (IACS) in Calcutta, and just after 2 years, in 1930 his
work was awarded by the Nobel Prize [30]. While Raman spectroscopy has shown
several technical advantages over applications of infrared spectroscopy, it was not
used in the analysis of the art objects until introduction, in 1975, of the Molecular
Optical Laser Examiner (MOLE) Raman microprobe [31]. Over the next 30 years,
several technical developments and innovations introduced into the Raman
instrumentation made this technique an essential tool for examinations of the wide
variety of artistic materials [32]. The Raman spectroscopy is characterized by
several features [9], which made this technique very attractive for the applications
in the analytical archaeometry [33]. The molecular spectra can be collected on very
small particles enabling the work with the microsamples due to its high spatial
resolution (typically ranging between 1 and 10 µm) both in terms of the lateral
resolution and depth profiling (with confocal instruments) [11]. The Raman spec-
troscopy as a molecular spectroscopic technique allows identification of inorganic
and organic compounds in admixtures without their separation, which is of par-
ticular importance in the analysis of paintings. Moreover, the molecular specificity
of the technique is enhanced by the availability of ad hoc-built databases of inor-
ganic and organic materials used in the cultural heritage (CH) field [11]. During
analysis, the laser power is kept at the sufficiently low level preventing the sample
or object from damage. To prevent the most sensitive artifacts from a light-induced
damage coming from the laser, during adjusting and positioning of the laser the
beam is usually de-focused [9]. Thus, the Raman spectroscopy preserves the ana-
lyzed specimen intact and enables the use of other, comprehensive elemental and
molecular analysis techniques (FTIR, UV-Vis/FORS, LIBS, XRF, AFM, and
others).
Apart from already mentioned advantages of the Raman spectroscopy, there are
also other specific requirements which have to be fulfilled during the analysis of
cultural heritage objects. It is a quite common situation that sampling is not allowed
and the direct study on the objects is performed in the laboratory or on site (in situ
investigations) with the application of a mobile instrumentation in the place where
the artwork is exhibited or stored. The measurements carried out in the museum
environment require careful risk assessment to minimize the potential hazard of an
accidental damage of studied objects. The first mobile Raman instrument used by
Vandenabelle et al. [34] in the museum facility was a noncommercial fiber optic
Raman spectrometer adjusted to the analysis of the vulnerable art objects. Over the
years, the increasing number of the CH studies made a significant impact on the
development of commercially available Raman spectrometers. Recently, the por-
table Raman spectrometers and some handheld spectrometers allow to record
excellent quality Raman spectra comparable with laboratory dispersive Raman
microspectrometers. Jehlicka et al. [35] compared seven portable Raman spec-
trometers successfully applied to the study of 17 beryl specimens including gem-
stones: emeralds (eight samples plus two synthetic emeralds) and aquamarines
(three samples).
Regardless of where the Raman analyzes are performed (museum or laboratory)
and the size and type of the device used, the laser power and its irradiance (mea-
sured in W m−2) have to be kept sufficiently low in order not to cause any damage.
Also, the time of the analysis should be as low as possible in order to reduce the
total absorbed dose of light. Moreover, very often the access to the objects is strictly
limited and analysis cannot be repeated so the quality of the spectra should be
efficient and has to deliver all the data required on the analyzed objects. During the
measurements carried out in the laboratory environment, it is possible to perform
analysis with the longer time of the acquisition and increase the spectral
signal-to-noise ratio. Also, the good wavelength stability and spectral calibration
are crucial. There are additional aspects that need to be considered in the analysis—
spectral range and spectral resolution used. Investigating inorganic materials, like
minerals (rock crystals, anatase, rutile) or mineral pigments (azurite, smalt, ocher),
it might be sufficient to record the bands at the low wave number in the spectral
range up to 2000 cm−1. But in analysis of paintings, the presence of m(CH)
stretching region at around 3000 cm−1 might reveal the presence of some organic
substances, so recording of the higher spectral region of Raman spectra would be
beneficial.
For many cultural heritage studies, the lasers with low excitation power are
needed, so the special attention should be given to the choice of the laser wave-
length adequate to the particular study. The Raman scattering intensity depends on
the laser wavelength (with the exponent 4), laser intensity and detector and optics
efficiency [13]. The most popular are green (514 nm, 532 nm) and red (785 nm)
lasers. It is important to remember that for the red laser usually the higher exciting
power is required (200–300 mW) and the green laser needs lower power (80–250
mW), thanks to the associated increase in the Raman effect and detector efficiency
[13]. The green laser usually is used in the investigations of inorganic materials like
minerals and gemstones, while the red laser is mostly applied in the studies of
organic materials like binding media, varnishes, or dyes.
The first reports presenting the application of the Raman microprobe in the
investigations of the painted manuscripts and colored archaeological materials [36]
and for the identification of azurite and malachite [37] were published by Guinea in
France. In the next year, the advantages of the application of the Raman spec-
troscopy in the analysis of the pigments and dyes were presented by the multi-
disciplinary group of the researches [38, 39]. In the beginning of the nineties,
several research groups [21] started various studies with the application of Raman
spectroscopy in the field of art and archaeology. Edwards et al. published the first
scientific study on the biodegradation of the Renaissance frescos identifying cal-
cium oxalate as the main compound in the incrustation with other organic com-
pounds of the lichen metabolism [40] and characterizing the mono- and dihydrate
forms of the oxalate compounds [41]. Year later, paper concerning the biodegraded
cave art [42] and the research of the biodegraded archaeological materials origi-
nated from the mummified skin of Otzi the Alpine Iceman dating from 5200 BC
[43] were published. Just a few months later, Clack et al. published in England the
study of the illuminations in the unique medieval manuscripts [44] and character-
ization of the historically important dyes [45]. He continued the research of art
objects with the use of the Raman spectroscopy, which resulted 3 years later in
publishing the first review highlighting the advantages of this technique in the
noninvasive investigations of manuscripts [45]. In the next years, he published
study of Persian [46], German [47], and Latin [48] illuminated manuscripts.
Simultaneously, French researcher carried out investigation of the blue pigments
used by French artist in nineteenth century [49] and systematically developed the
application of Raman spectroscopy in the study of works of art, forensic science
[50] and study on the past technology for manufacturing the archaeological blue
(Egyptian Blue) and green pigments [51]. The study of the marbles from Carrara in
the Certosa of Pavia revealed that the inorganic pigment Pb3O4 is responsible for
the presence of the red stains on the marble instead of previously assumed bacterial
strain identified as Micrococcus roseus [52]. The comprehensive study of works
assigned to the Italian Renaissance Master—Tintoretto—showed that the yellowish
shade over not protected areas of the realgar (As4S4) was wrongly assumed as an
orpiment (As2S3). The nonprotected realgar exposed to the sunlight developed into
a yellow material, identified by the Raman microspectroscopy as a pararealgar—the
light-induced polymorph of the realgar [53].
The great potential of the Raman spectroscopy in the analysis of the minerals,
liquid, and solid inclusion in minerals as well in the analysis of ceramics was
reported by Colomban in 1979. Investigations of the archaeological ceramics were
initiated by the Italian groups in the middle of nineties [54, 55]; further progress in
the study of the ancient and archaeological ceramic was made at the beginning of
the twentieth century [56]. Colomban et al. [57] with other researchers studied
selected porcelain representative for the Imperial Manufacture and now the
Manufacture Nationale de Sevres: French soft paste (‘pate tendre’ or ‘frit ware’),
English-like (bone China), the latest soft pastes made from glass (silica-rich soft
paste, Vogt soft paste), and ancient hard and new hard pastes. They revealed that
the presence of a strong beta-wollastonite and/or tricalcium phosphate peak allows
to differentiate a soft-paste body from a hard-paste body in Raman spectra.
Colomban et al. [49] continued the study of porcelain from Sevres Factory and
successfully identified and differentiated the ancient and modern European
porcelains by Raman macro- and microspectroscopy [58]. Simultaneously, several

other studies focused on the archeological ceramics were published. The Neolithic
pottery found on the bank of the River Danube was analyzed by Colomban et al.
[59], who identified the yellow-to-red decor present on the external faces of pottery
as made of ocher, a mixture of hematite, quartz, and phyllosilicates. Red deposit,
found on some internal surfaces of a pottery, consists of cinnabar (HgS) with some
quartz and phyllosilicates. Following, many studies of various archaeological
findings were published, for example, study of material collected from the wall of
an ancient Celtic vitrified fort [60], ceremonial polished ‘greenstone’ axe heads
from Mesoamerica [61], and the mineral species jadeite from Guatemala [62]. In the
following years, studies focused on the analysis of the Iberian ceramics [63],
archaeological ceramics from the fourteenth and seventeenth centuries from Parma
[64], ceramics from the period between the end of the mid-seventh century and the
first half of the fourth century BC were reported, as well as brief overview about
identification of enameled pottery, glass artifacts and associated pigments [65],
application and problems of the use of micro-Raman spectroscopy in the study of
ancient ceramics [66], pre-Roman ceramics excavated near the suburbs of Canosa
(Puglia, Italy) [67], and ancient ceramics from Ebla (Syria) [68].
One of the first scientific investigations focused on better understanding of the
techniques used in wall paintings, and their deterioration was published by Edwards
[69]. The next research focused on painting techniques used in wall paintings
elucidated the misinterpretation of ancient recipes. The term ‘minium’ was used for
both, cinnabar (HgS) and red lead (Pb3O4) pigments, in antiquity. In Roman times,
the term minium was reserved for mercury(II) sulfide (HgS) but was applied
increasingly to lead tetroxide (Pb3O4) by the Renaissance. The differentiation of the
components of this mixture was possible by application of Raman spectroscopy
[70, 71]. Another interesting mixture of pigments was found in the medieval
frescoes in Spain. The mixed iron red with cinnabar was applied to impart the red
color in the wall paintings in the church, but pure cinnabar was being used for only
the most important biblical figures [72]. Investigations of the Italian wall paintings
were started by the group from Parma (Italy) which applied the micro-Raman
spectroscopy combined with gas chromatography and ultraviolet fluorescence
photography to study some wall paintings in the S. Giovanni Evangelista Abbey in
Parma, Italy. The micro-Raman spectroscopy determined the palette used by two
painters of the sixteenth century: Parmigianino and Michelangelo Anselmi, leading
to the identification of different white, yellow, red, brown, green, blue, and black
pigments [73] and binding media [74]. The study of Italian historical wall paintings
was continued by Baraldi et al. who used the Raman spectroscopy combined with
other analytical methods to carry out the analytical characterization of Roman
painted plasters from the ‘Domus farini’ in Modena [75]. In 2007, the first results of
scientific examination focused on characterization of pigments present in paint
samples collected from the south and east wall decorations of the fifteenth-century
Thubchen Lakhang monastery located in Nepal were published. The first results of
the micro-Raman analyses identified that azurite was used alone and sometimes in
combination with lazurite, red ocher, malachite, brochantite, which may represent
an alteration product of malachite, orpiment, and vermilion, both alone and in

combination and finely a very interesting gilding technique that used pararealgar
with traces of orpiment and vermilion [76].
The group from Parma developed analysis of the palette of pigments used in the
mural painting in Pompeii and identified original pigments found in bowls dis-
covered there [77]. The influence of the different environmental conditions on the
state of wall painting in Pompeii became the object of the comprehensive study
carried out by Madariaga and co-workers. The group from the University in Bilbao
applied the Raman spectroscopy combined with other analytical techniques to the
characterization of the original composition and decay products in the mortars and
pigment layers [78], the blackening of red cinnabar, efflorescence [79], a severe
chemical attack of the acid gases on the murals after the eruption of the Vesuvius
volcano [80], and study of salts formed in protected wall paintings and the
mechanism of the decaying process [81].
Contemporarily to research of the wall paintings, one of the first applications of
the Raman spectroscopy in the study of the prehistoric rock art was published by
Smith et al. [82]. In the microsamples collected from the red and black parts of
Paleolithic paintings collected from the three caves situated in the Lot Department
in France mineral pigments: haematite, goethite, amorphous carbon, and manga-
nium oxide/hydroxide were identified. Then, in the same year, the chemical and
mineralogical study of ancient paints from the prehistoric rock paintings from the
Big Bend region of Texas was reported [83]. The application of Raman spec-
troscopy supported by the FTIR, XRD, and GC-MS analyses allowed characteri-
zation of the paint materials, including pigments and binders, and established the
relationship between the paints and the environmental matrix, which included
calcium sulfate (gypsum) and calcium oxalate monohydrate (whewellite). The
systematic study of the caves decorated by prehistoric paintings [83] and situated on
the Iberian Peninsula was started by Hermanz few years later. His investigations
encompassed the characterization of the chemical composition of pigments, iden-
tification of the components present in the crusts (whewellite, weddellite, and
amorphous carbon), and elucidation of the mechanism of sandstone weathering
process [84], revealing the nature and distribution of the components of the oxalate
layers: alpha-quartz, haematite, whewellite, gypsum, muscovite, microcline, ana-
tase, and rutile [85], identification of the rarely used white pigment (a-quartz,
anatase, muscovite and illite), identification of whewellite and weddellite as
products fungi or lichens activity (e.g., Verrucaria nigrescens) [86], characteriza-
tion of oxalate accretions of whewellite and weddellite on post-Paleolithic rock
paintings from open-air rock shelters of Triassic sandstone in the Sierra de las
Cuerdas (Cuenca, Spain) [87], and analysis of pigments from the caves from the
Tito Bustillo and El Buxu Caves, Asturias [88].
Molecular identification, by Raman spectroscopy, is particularly reliable in the
analysis of materials with crystalline structure [89], like minerals. In the CH field,
the study of minerals as raw structural materials [90] or pigments [91] played an
important role in the understanding of the nature and the internal structure of the
artifact. The synopsis covering the last decade of the applications of Raman
spectroscopy to issues in which raw minerals, including mineral pigments, are

involved were studied by Bersani and Lottici [92]. The special attention in this
review was devoted to studies where the Raman analysis of minerals was deeper
than an identification of the phases present in an archaeological or artistic object.
Apart from the analysis of the mineralogical composition of the paste and body of
ceramics, the identification of typomorphic [93, 94] minerals present in the body
might be applied in the provenance study of the archeological artifacts. The
provenance of archaeological vessels manufactured in late ninth and the turn of the
tenth century that were found in Gilów (southwestern Poland) was verified due to
the Raman identification of sillimanite [95]. The typomorphic mineral present in all
analyzed groups of ceramics suggested that the raw material used to manufacture
the ceramics originated from local sources. The Raman spectroscopy in the
gemology studies [96] was present since the beginning of the techniques and widely
evolved in the archeometry [97]. The possibility of the noninvasive and nonde-
structive investigation of the gemstones also carried out in situ without the necessity
of the transportation of very often unique and precious jewelry objects made the
Raman spectroscopy the leading tool in this field. In contrast to the present day, a
decade ago the possibility of investigations with the application of the mobile
spectrometer was quite rare. Analyses of the reliquary cross called Heinrich’s cross
dating back to the late Middle Ages (first quarter of the eleventh century) were
made directly in the museum, without the need to relocate the object [98]. The
determination of the nature of the gems generally verified the identification based
on previous optical observations, but most of the mounted materials were found to
be glass paste or quartz varieties. The mobile Raman microprobe allowed mea-
surements of mineral surfaces as well as inclusions in the interiors of transparent
minerals during the investigations, mineralogical studies of the large collection of
minerals and rocks of the Prussian kings in the Grotto Hall (Grottensaal) of the New
Palace (Neues Palais), Park Sanssouci in Potsdam [99]. In recent years, several
comprehensive papers emphasizing the main aspects of the applications of Raman
spectroscopy to gemology were published [100, 101], reflecting the increasing
number of studies focused on gemstone in the CH field.
In the last decade, the analysis of the organic pigments and colorants has also
developed. One of the first reports presenting application of the Raman spec-
troscopy to the analysis of the colorants and lithographic inks used in the group of
posters manufactured between 1890 and 1920 was published by Centeno [102].
However, the analysis of the organic materials used in the archeological artifacts
and art usually delivers more comprehensive results when the Raman analysis is
coupled with infrared techniques like ATR and noninvasive reflectance spec-
troscopy, as in successful studies of various resins [103] and old varnishes [104].
Recently, the potential of portable and benchtop Raman instruments complemented
by FTIR, NIR, and GC-MS has been shown in the study of plastic materials used in
the sculptures created by Kantor belonging to the contemporary art collections of
the National Museum in Krakow [105].
Surface-enhanced Raman scattering (SERS) spectroscopy was applied to the
identification of organic pigments and dyes in artworks and archaeological objects
because vibrational ‘fingerprints’ can be measured from microscopic samples [106,

107]. Several developments of SERS techniques applied in the studies of textiles,
paintings, and archaeological objects presenting the importance of the appropriate
parameters and substrate selection were reported by Casadio et al. [11].
Additionally, the wide range of applications of the SERS technique in the identi-
fication of natural and some synthetic dyes in microscopic samples from objects of
artistic, historical, and archaeological significance recently was reported by Leona
[108].
New frontiers of the Raman spectroscopy in the cultural field came from the
application of the portable instruments and from coupling Raman with other ana-
lytical techniques. The introduction of the Raman imaging permits noninvasive
investigations of the composition and structure of a material and delivers the more
representative characterization of the samples than the point-by-point micro-Raman
analysis [109]. For example, Raman spectral imaging was successfully applied in
the analysis of materials from middle ages, like pigments from the ‘Last Judgment’
by Hieronymus Bosh, inks from Glagolitic, Greek, and Latin manuscripts, and
pigments widely used in modern and contemporary paints like ultramarine blue
(Na6Al6Si6O24S4) [110]. Nowadays, in the studies of the artworks and archeolog-
ical objects the Raman spectroscopy is usually combined with several elemental
and molecular techniques, for example, SEM-EDS [111], XRF [112], LA-ICP-MS
[113], XRD [114], µFTIR [115], FT-NIR [116], ATR FTIR [117], micro-ATR,
ESI-MS [118], and FORS [119].
16.2.1 Case Studies
16.2.1.1 Medieval Manuscripts
The Raman spectroscopy, due to its efficiency, became one of the most widely
applied techniques in the noninvasive analysis of the unique and precious illumi-
nated manuscripts. Especially, the Medieval Hebrew manuscripts are rare objects
among European manuscripts collections, and due to religious reason, they are not
easily accessible for scientific studies. Very restrictive rules governing writing of
Hebrew manuscripts drastically limited materials that were available for the
preparation of inks and palette of pigments used in the illuminations. Due to reli-
gious constraints hampering access to the manuscripts, the available scientific data
on such materials is very limited [10, 120]. The Medieval Tanakh (Hebrew Bible)
was investigated [121, 122] at the Wroclaw University in the Cultural Heritage
Research Laboratory. The laboratory specializes in applications of Raman spec-
troscopy in the CH studies and was trusted with the access to Hebrew manuscript
from the collection of the Ossolinski National Institute in Wrocław (Poland). The
investigated manuscript consists of five volumes: Torah with the Aramaic transla-
tion, Book of Isaiah, Book of Judges, Book of Job, Book of Proverbs, Song of
Songs, Book of Ruth, Book of Lamentation, Ecclesiastes and Book of Esther. The
opportunity to investigate a valuable medieval codex created the unique possibility

to come one step closer to understanding how the Hebrew manuscripts were
prepared.
The major goal of the project was to confirm the provenance and dating of the
Bible. Additionally, the study was expected to establish the pallet of pigments used
in the preparation of illuminations and kind of inks. The art historians dated
manuscript roughly at the fifteenth century and suggested its western origin;
therefore, the influence of local manufacturing on preparation of religious manu-
scripts that were usually governed by eastern rules was highly probable. Additional
valuable information could be delivered by the comparison of the analyzed Hebrew
Bible to a western manuscript dated at the same time. The presence of the colorful
and high artistic quality illuminations, showing fantastic creatures, animals, and
numerous human figures distinguished this object from other Hebrew manuscripts.
The type of illumination was generally atypical for Hebrew texts, and during the
investigation, it has been established [123] that such illuminations are characteristic
to western, especially German style. At the same time, the image of a mermaid in
one of the illustrations suggested strong influence of eastern style. The text of the
invocation present in the manuscript listed a date (1492) when the Bible was sold,
and it could aid in placing this manuscript in time and region. Additionally, the
invocation contains also a name of the location where the transaction took place.
However, the place does not exist nowadays and the speculation was that the place
was located somewhere in northern France or in Germany. Besides, knowing the
time when the manuscript was sold or where the transaction took place did not
mean the Bible was made at the place of sale or that it was prepared directly before
the transaction. For more detailed study of the provenance, the comparative analysis
of a reference book made around the time included in the invocation (1492) and
closely to the expected location (France or Germany) was required. Comparison to
the reference manuscript Breviarium Romanum of French origin should aid dating
and might help in provenance determination of the investigated Hebrew Bible. The
richly illuminated reference medieval western manuscript also belonged to the
collection of The Ossolinski National Library in Wrocław.
The micro-Raman spectroscopy analysis was implemented in the investigation
of palette of pigments and inks applied in the Hebrew and the western manuscripts.
The measurements were carried out in situ (Fig. 16.1) with time of the analysis
strictly limited due to the high historical value of both manuscripts. The study was
performed using the Horiba Yvon Jobin T64000 spectrometer equipped with a
confocal microscope with 50 and 100 objectives. The Ar/Kr ion laser lines at
514 and 647 nm were employed for the Raman excitation, and the laser power at
the exit was below 6 mW for the green laser and below 40 mW for the red laser.
The spectra were dispersed using a diffraction grating (1800 line/mm) and detected
by a CCD camera (1024 256) with the spatial resolution of 200–1100 nm, and a
notch filter was used to separate the Rayleigh signal. The spectra were recorded at
room temperature in the range 100–3800 cm−1 with a spectral resolution of ca
2 cm−1. The integration time and accumulations (minimum 2) were chosen sepa-
rately for each analyzed spot. The analysis of inks and red and orange pigments was
Fig. 16.1 Micro-Raman

measurements of the Hebrew
Bible from The Ossolinski
National Library in Wrocław
carried out with the 647 nm laser; the 514 nm laser was used for other colors. There
were two reasons leading to utilization of the red laser: first, to overcome
the possible fluorescence, especially during analysis of inks; second, to avoid the
decomposition of red lead (minium). Clark et al. [124] have shown that the
application of the laser line 514 nm or the energy at the exit bigger than 2.3 mW for
the line 647 nm caused decomposition of the red lead. Degradation of the pigment
is displayed by the change of the Raman intensity of bands at 125, 485, 555 cm−1
and by the presence of the additional band at 1095 cm−1. Due to this fact, the
application of the adequate laser line and power was of the utmost importance.
The Raman spectra database created on-site for the historical pigments produced
according to the original recipes by Kremer manufacture [125] was used for the
comparative study of both investigated manuscripts. In order to perform the most
efficient comparative analysis, Raman spectra from the database and spectra of
analyzed pigments were collected under the same parameters and equipment.
Additionally, reference samples were prepared for all inks present in the Bible.
Carbon-based and iron gall inks were prepared according to old recipes.
The Hebrew Bible contains 20 extraordinary illuminations presenting the

impressive range of colors. In all of the investigated illuminations, the parts pre-
senting golden characters were prepared using pure lapis lazuli, the pigment made
from the mineral identified due to the presence of the Raman bands at 258, 548,
822, and 1096 cm−1. Various blue tonalities from light blue to navy blue were
achieved by mixing blue pigment with lead white or with carbon black. The Raman
spectra recorded on several white ornaments and on many other white details within
the decoration displayed the Raman band at 1055 cm−1, characteristic for the
presence of the lead white (2PbCO3Pb(OH)2). Lead white was quite commonly
used in carnations, while application of red lead (118, 149, 220, 390, 480, and
548 cm−1) instead of vermilion or red ocher (the clay containing the mineral
hematite Fe2O3) in order to create the warm tonality of the carnation as well in
creating other details like red lips or reddish cheeks was less common. Analysis of
the red orange elements revealed the presence of the pure red lead. The Raman
spectra recorded on the red details have shown that in many illuminations the red
color was achieved by mixing read lead with vermilion (Fig. 16.2a). There are two
explanations of this approach—one being that it helped artist to create the specific
tonality. The other explanation is economy related—for centuries, vermilion was
very expensive so mixing it with red lead lowered the cost substantially.
The bright, strong colors of the illumination gave the impression that many
pigments were utilized. In fact, the palette of pigments was rather limited. It was
expected that the brownish colors were painted using natural earth pigments. The
body color of the animal that looks like a lion (Fig. 16.2b) was created by
the mixing of three pigments—lapis lazuli, lead white, and a black pigment. The
Raman spectra with two characteristic bands at 1358 and 1585 cm−1 suggested the
Fig. 16.2 Pigments identified in the Medieval Hebrew Bible: a Raman spectra of the mixture red
lead and vermilion detected in the red area; b Raman spectra of the brownish color achieved by
mixing lapis lazuli, lead white, and carbon black
use of a carbon black. Also, the analysis of gray area indicated unusual approach by
the artist to the creation of various hues of colors. The specific gray tonality was
achieved by the combination of such pigments like lapis lazuli, lead white, and
carbon black. Unfortunately, strong fluorescence obscuring Raman spectra col-
lected on the greenish and reddish illuminated area of the Hebrew Bible suggested
the application of organic dye used quite frequently in manuscripts [107].
The comparative Raman analysis of the pigments used in the Breviarium
Romanum revealed the presence of the pigments popular in the medieval minerals,
like azurite (2CuCO3Cu(OH)2) present in the blue areas, lead white (found in a
carnation), yellow ocher (the yellow earth mineral composed of goethite (FeO(OH)
and clay) found in the background in the illumination presenting King David
praying in the garden (Fig. 16.3), and red lead detected in many reddish areas in
illuminations and decorative frames. Also, the red lead was found in the decoration
of initials and in the carnation where it was mixed with the lead white. Another red
pigment, vermilion, was found only in the initials in the decorations. Further
analysis of the palette of pigments resulted in the detection of a lead-tin-yellow,
type I (lead(II) stannate Pb2SnO4). It was the first synthetic pigment used since the
thirteenth century and replaced later by the Naples yellow. Interestingly, the pig-
ment discovered in the Breviarium Romanum was not used between 1750 and its
rediscovery in 1940 [126]. The lead-tin-yellow pigment was also detected in green
illuminations (trees, leaves, etc.). It was expected that a green mineral malachite
(CuCO3Cu(OH)2), one of the most widely used green pigments in the medieval
Fig. 16.3 Pigments identified in the Breviarium Romanum: a Raman spectrum of lead-tin-yellow
(type I) and the reference spectrum; b Raman spectrum of yellow ocher; c the illumination of King
David praying in a garden from the medieval French manuscript Breviarium Romanum
illuminations and paintings, was used in the green areas. However, the analysis
indicated that the green color was achieved by mixing lead-tin-yellow and azurite.
Further, Raman analysis included a study of inks present in both investigated
manuscripts. From the Hebrew Bible, the main text written in Hebrew and two
types of the text written in the Aramaic language (the text placed on the bottom of
the page called Masorah Magna and text written between columns of the main text
called Masorah Parva) were analyzed. The Raman analysis confirmed that the main
text and Masorah Parva were written using iron gall ink. In the spectra collected for
Masorah Magna (Fig. 16.4), the characteristic band for iron gall inks present at
1490 cm−1 was not present. Unfortunately, due to the fluorescence obscuring the
Raman spectra it was not possible to unambiguously ascertain the usage of the iron
gall ink or ink based on the carbon black in the analyzed manuscript. On the
contrary to the Hebrew manuscript, the Breviarium Romanum was written using
inks of various colors. The Raman analysis suggested the application of the iron
gall ink in the main text and revealed the presence of the vermilion used in the
richly decorated initials.
Traditional illuminations found in manuscripts utilize pure colors (like lead
white, lapis lazuli, vermilion, red led, led-tin-yellow) with an exception of lead
white that was mixed with other pigments in purpose of brightening and the mixture
of the blue and yellow pigments applied to create green color, as it was shown in
the Breviarium Romanum. Interestingly, some illuminations from the investigated
Fig. 16.4 Raman spectra of the inks analyzed in the Hebrew Bible (the page from the Bible is
shown on the right: The text placed on the bottom of the page is called Masorah Magna, and text
written between columns of the main text is called Masorah Parva)
Hebrew Bible were prepared by applying mixtures of various colors delivering

valuable information on preparation rules of Hebrew codices that are believed to be
very inflexible. In conclusion, the reported study delivered priceless information on
the type of religious Hebrew manuscript extremely rarely investigated.
16.2.1.2 The Provenance Investigation of the Archeological Amber
Due to the ability to perform in situ measurements without any preparation of the
sample, the Raman spectroscopy is recognized as an effective technique used in the
provenance investigation of the amber. It is quite common that the weathered layer
is present on the surface of archeological amber objects causing strong fluorescence
obscuring Raman spectra. Unfortunately, in many cases this outer layer cannot be
removed without the damage to the object; hence, the implementation of supple-
mentary techniques is necessary in order to determine the provenance of such
artifacts. The applicability of the positron annihilation spectroscopy (PAS) in the
analysis of the raw material was proved during investigations of the amber objects
discovered in the archaeological excavations [127]. Moreover, other spectroscopic
methods like transmission infrared spectroscopy and ATR can be used for the
identification of amber [128, 129]; however, these techniques require preparation of
samples.
The successful investigation cannot be completed without the access to the
adequate database. The Cultural Heritage Research Laboratory at the Wrocław
University developed Raman spectra database of amber and copal which was
successfully applied in the provenance studies of archaeological amber objects
discovered in Poland [127]. The spectral database was created using reference
samples of succinite, fossil resins, subfossil resins, and imitations from a very
comprehensive collection of The Museum of the Earth’s Amber Department,
Warsaw. The database is systematically expanded. Recently, it was extended to
incorporate also spectra for amber and copal collected with other techniques: ATR,
PAS, near infrared (NIR), and SEM-EDS [129].
The applicability of this comprehensive scientific database was verified during
provenance investigation of the archeological amber objects from the collection of
amber deposits discovered in 1906 [130, 131] and 1936, nearby the horse racing
track in Wroclaw Partynice [132]. The total weight of the raw material accumulated
in pits amounted from 1240 to 1760 kg, according to various estimates [133],
which so far is the biggest find of this type in the world. Unfortunately, a large part
of the amber collection was sold by the conservatory service or destroyed during the
WWII. Until today, at the Museum of Archaeology in Wroclaw only its small
remainder weighing 4.5 kg has retained and is exposed at its permanent exhibition.
Many archaeological studies conducted on the site of the find have shown the
relationship of the findings with an extensive settlement of the Przeworsk and
perhaps the La Tène cultures (1st BC). The analysis of the findings indicates that
they had character of storage pits, not ‘treasures’ as they were often described. The
discovered pieces of raw material and semifinished products of amber in the
settlement suggest that its inhabitants practiced both the amber trade and process-
ing. The discovery of the commercial and craft amber center in Partynice plays
important role in history in Silesian (west-south Poland) archaeology and is still the
most convincing evidence concerning the course of the Amber Road and the role
that Wrocław area played in it.
As a method of choice, the FT-Raman spectroscopy was implemented in the
investigations. The Raman spectra of analyzed samples and reference materials
were collected using a Raman research system FT-Nicolet Magma 860 employing
for excitation In:Ga:Ar laser line at 1064 nm. The spectra were recorded at the
room temperature in the range 100–3800 cm−1 with the spectral resolution 4 cm−1
and with a constant number of scans (512/measurement). The laser power at the exit
has been set below 0.8 W and was chosen separately for each sample. The
archeological samples were analyzed without any sampling. Due to the huge
number of the objects varying in color, the analyses were carried out after the
selection of the reference samples.
A comparative analysis of excavated artifacts and reference materials originating
from amber sources within reasonable proximity to the excavations was performed.
Two deposits were considered: deposit of the Baltic amber (Poland and Russia) and
deposit from Moravia (Czech Republic). Reference materials including succinite
(Baltic amber) from tertiary deposit at the seashore from North Harbor in Gdansk
(Poland) and Sambian Peninsula, Yantarny near Kaliningrad (Russia), valchovite
from Valchov in Moravy (Czech Republic) dated back to Upper Cretaceous were
used.
Analysis of Raman spectra allows differentiating amber material of various
geological and geographical origins. There are three wave numbers regions
observed in amber Raman spectra that permit proper origin assignment. The first
region covers high wave number area from 3700 to 2000 cm−1; the second and the
third regions cover low wave number areas from 2000 to 1100 cm−1 and from 1100
to 100 cm−1, respectively. The main spectral feature that could be used as a guide
and help in distinguishing between various kinds of amber is the intensity ratio of
two bands at about 1449 and 1645 cm−1. The ratio can be used as an indicator of
the maturity level of the analyzed sample and its geological age [134].
Traditionally, Baltic amber has been characterized according to the band present
at 1150 cm−1 in the Raman spectra assigned to succinic acid and its esters and the
intensity ratio of peaks at 1645 cm−1 (m (C=C)) and 1450 cm−1 (d (CH2,CH3)). The
spectra of Baltic and Moravian amber reveal some differences in the normalized
intensity of the characteristic bands [127]. Samples of valchovite and succinate
amber show intensity ratios ranging between 0.48–0.63 and 0.62–0.86, respec-
tively, allowing differentiation of these two kinds of amber. The I1645/I1449 ratio is
higher for succinite (Baltic amber) what suggests much higher degree of maturity
and different geological ages [135]. High repeatability of Raman spectra collected
for analyzed amber samples discovered in the excavation was noticed. The position
of the characteristic bands and their relative Raman intensity observed in the spectra
of analyzed samples and reference material (succinite) from North Harbor in
Gdansk (Poland) and Sambian Peninsula, Yantarny near Kaliningrad (Russia)
Fig. 16.5 Raman spectra of amber samples: a archeological amber from the excavation in
Wrocław; b reference sample of succinite from Yantarny, Russia; c lumps of raw amber found in
1906 during the construction of the horse racing track in Wrocław; d the excavated specimens of
raw amber
overlapped almost completely in the regions of interests (Fig. 16.5). The compre-
hensive comparative analysis indicated the Baltic origin of the investigated amber
objects. However, due to the enormous number of discovered objects and the fact
that some part of the finds is still inaccessible, the possibility that other objects than
succinate were stored in the Wrocław-Partynice depots cannot be excluded.
Another amber study discussed below was also performed with the use of the
Raman technique but this time for the objects which underwent some conservation
treatments in the past which might obscure Raman spectra collected for them. The
studied artifacts were amber beads excavated in Jartypory (Poland) which origi-
nated from the Roman period (3–4th BC) and now belong to the collection of the
Archeological Museum in Warsaw [136]. The conservation treatment of the ana-
lyzed beads was carried out 10 years prior to the study with the application of a
mixture prepared from nitrocellulose lacquer, ethanol, and acetone. The aim of the
analysis was to establish the geological origin of the amber beads and the influence
of the conservation layer on the applicability of the Raman spectroscopy for such
multilayered objects. The study encompassed the analysis of archeological beads as
well as pretreated reference samples. In the first step, the provenance of the ref-
erence samples based on the spectra from previously developed Raman database
was established. Then, the reference samples were subjected to the same conser-
vation procedure that was used for the archeological beads and the newly applied
superficial layer was matured in the museum environment for 1 year. The analysis
of treated and untreated reference samples has shown that the applied lacquer does
not interfere with the Raman spectral bands used for amber identification. It was
then possible to determine the provenance of both sets of studied objects, beads
which underwent conservation treatment and reference samples. The recorded
Raman spectra suggested the use of succinite (Baltic amber) as a raw material in the
manufacturing process of the archeological beads.
Analyses of larger data sets collected with the Raman spectroscopy require new
approaches which can go beyond simple comparisons of the spectra. Successful
application of the chemometric methods in investigations of cultural heritage
objects has been documented in several recent papers [137–139]. In particular, the
Raman spectroscopy coupled with partial least squares (PLS) method was used for
the classification of amber samples according to their provenance and geological
age [135]. The Raman spectra of samples from Czech Republic, Baltic region, and
various objects dating back to the Upper Cretaceous and Cenozoic ages were
analyzed. In this work, an alternative chemometric approach was proposed to select
additional intervals on the spectra (Table 16.1). Several metrics (number of mis-
classifications and permutation test) which evaluated the capability of the alterna-
tive PLS models were used to predict future measurements. The achieved results
indicated the possibility to assess the geological age and provenance of amber
considering an interval of wave numbers instead of some single wave numbers.
Furthermore, differences in the Raman spectra were highlighted using the Variable
Importance Projection methodology over the PLS model. The applied model clearly
distinguished amber from Baltic, the Czech Republic and resins from Cenozoic and
Upper Cretaceous periods and supported classical results interpretations.
16.3 Vis
Color is one of the most important aspects of an artwork. Human eyes allow
resolving between approximately two million shades of colors even though our
vision capabilities are based only on three types of receptors sensitive to red, green,
and blue light. Analysis of the Vis spectra of an artifact, recorded with a spectral
resolution of many hundreds of points, could provide information on the chemical
nature of colorant used by an artist, their alteration caused by natural degradation
and/or human interventions. The analysis is usually done by comparing measured
spectra to a pigment/lake/dye database containing reflectance spectra collected for
different pure colorants dispersed in various binders (different oils, egg tempera,
varnishes). The spectra recorded in the reflectance mode contain also the scattered
light, where intensity and spectral characteristics depend on many factors (type of
Table 16.1 Results of the measures of variable importance in PLSDA for the best predictive
performance models assessed by variable importance projection (VIP) for geological age of amber
Interval 600–3100 (cm−1) Interval 1100–1550 (cm−1) Approximate assignment
of vibrational mode
654 m (C–C)
696 m (CC)
720 m (CC)
800 m (COC,CC)
886 m (COC)
938 q (CH2)
972 q (CH2,CH3)
1042 m (C–O)
1098 1106 m (CC)
1142 1142 m (CC)
1380 d (CH2,CH3)
1432 1432 d (CH2,CH3)
1502
1636 m (C=C)
1723 m (C=O)
2852 m (CH2,CH3)
2872 m (CH2,CH3)
2894 m (CH2)
2900 m (CH2,CH3)
2930 m (CH2)
2949 m (CH2)
the binder, pigment particle size, paint layer thickness and structure, state of
degradation) which in turn make conclusions based on a spectral resemblance
between artwork and database samples often difficult and sometimes misleading.
Other spectroscopic techniques (e.g., Raman, FTIR, XRF, LIBS, to name the most
often used ones) might provide additional information allowing for proper identi-
fication of the artists’ palette [140]. Efforts to build such spectral libraries were
undertaken by many researchers and organizations—see [141] for organic dyes in
colored papers [142]—for modern gouache paints [143–145]—for pigments and
pigment mixtures. Especially, Italian researchers have done a great deal of work in
this respect—compare the Web page [146] devoted to the topic which combines
reference spectra collected from 24 papers published by Bacci and co-workers
between 1992 and 2007.
The mode of operation often used in analysis of artworks for Vis reflectance (but
also UV, NIR, and some other spectroscopic techniques) is the fiber optic reflec-
tance spectroscopy (FORS), first introduced to the analysis of heritage objects in
1987 by M. Bacci and co-workers. FORS allows performing point measurements
for objects regardless of their size and location, as the equipment could be made
portable and easy to use even on high scaffolds required to reach mural paintings on
ceilings and high walls. The relatively low cost of equipment (depending on the
spectral range and resolution) also contributes to the widespread of the technique.
16.3.1 Microfading Tests
Measuring methods described so far and their interpretations were used for
chemical identification of components of the tested object. They could allow dis-
covering the secrets of the artist’s techniques, serve for its dating or authentication,
point to the presence of degradation products, and allow assessing the threats to the
durability of the artwork. Knowing the chemical composition of dyes and/or pig-
ments used by the artist, one can try to conclude about its resistance to environ-
mental factors, in particular—to light. Knowledge about the ease with which a
given compound photodegrades, obtained in tests in an isolated system, does not
translate directly into the durability of the objects in which it was applied. The same
dye on different substrates may show very different lightfastness as is shown and
discussed in the example below (Fig. 16.8).
The two main goals of museums, archives and libraries, are to share and protect
the collection. The simultaneous implementation of them creates an antinomy—
light, necessary for the viewer’s contact with the work, is also a factor that causes
its degradation. The compromise that museums use, based on observations and
research results, involves the use of the smallest possible doses of light, which are
believed to not cause degradation for a given category of the object. For objects
considered to be the most sensitive, it means the need to reduce the radiation power
of light sources to the lowest acceptable level and shorten the total time of illu-
mination (e.g., by periodically turning off the display or installing a proximity
switches) controlling in this way a total dose of light. Both these measures limit the
viewer’s contact with the work—for most people, the ability to recognize the colors
and details of the drawing in lighting at the 50 lx level recommended in such cases
is already significantly limited (especially for dark objects and objects with a low
contrast).
The solution that would allow for individual adjustment of illumination to a
given object could be to conduct photostability-accelerated aging tests with the use
of light source with spectral power distribution similar to the one used on exhibit.
Standard photostability tests lead to irreversible destruction of the sample, and in
the case of works of art and heritage materials, only a noninvasive (i.e., not
requiring a sample from the object) and non- or microdestructive methods could be
accepted. Determining here what microdestructive testing means in the context of
art analysis is not easy and will never be unambiguous. If the change in the object is
not perceptible to the human eye and does not affect its strength parameters, then
such a test is usually considered acceptable. In the case of microfading tests
described below, this condition is usually met.
In the world of art conservation, it is practiced to test the lightfastness of sac-

rificial materials or model samples (sometimes produced according to historical
recipes) as well as modern products used in conservation (paints, glues, papers,
cardboards, threads) based on the methods described in the relevant industrial
standards (ISO, etc.), using laboratory light aging chambers or direct sun exposure
methods [147]. In 1999, Paul Whitmore and co-authors published a paper in which
a new approach to photoaging studies was presented, opening exciting possibilities
to perform lightfastness measurements on originals [148].
16.3.1.1 Description of the Methodology
The device invented by Whitmore allows to perform aging experiments on a very

small area of the object (see insert in Fig. 16.6), so the induced color changes are
practically not visible to the naked eye. Whitmore’s idea is based on the use of the
FORS technique, except that the object is illuminated by a broadband light source
with so much light power that it causes photodegradation of the object during a
relatively short aging experiment. The Vis spectrum (% reflectance) recorded
during the measurement is converted on the fly into CIE L*a*b* color coordinates,
from which the DE*ab index (a metric of the color change) is calculated, informing
about the susceptibility of the tested point to light fading. The higher the value of
the DE*ab, the larger the color change induced by light.
Fig. 16.6 Microfading measurement for a watercolor drawing performed with a noncontact
instrument with 0° illumination −45° reflectance geometry. Insert: image of the aging light spot
A typical MFT instrument setup consists of two basic elements:

• A light source—a xenon lamp or a LED source connected by a fiber optic to a
lens which focuses the beam on the surface of the material being tested (distance
between the lens and the objects’ surface ranges from few centimeters to zero—
a contact MFT designs were recently developed [149]).
• The Vis fiber optic spectrometer, which collects the signal using its own lens,
most often used in 0°/45° geometry or sometimes in retrofocus arrangement,
with light paths separated by a beam splitter between illumination and spectra
collection.
For the proper operation of a microfader, a high stability of the light power of the
used light source is essential as its fluctuations would be interpreted as changes in
the tested objects’ color. In case of a xenon lamp, which typically has short-term
flicker and longtime drift of it’s light power, a closed-loop feedback system is
necessary to ensure proper operation of the light source.
When carrying out lightfastness measurements by MFT, in addition to calibra-
tion on white and black standards, necessary to convert spectrometer signal into %
reflectance and CIELab color values, an additional calibration for the system light
power is usually performed. For this calibration, a reference material for light-
fastness is used, which allows for a direct comparison with the results obtained for
examined historical objects, which immediately indicates the maximum light
exposure allowed for a given object to reach its intended preservation target (i.e.,
maximum color change acceptable for a given timeframe). More details on the most
commonly used lightfastness standards (the Blue Wool Standards) will be given in
the section below.
The MFT technique is currently applied at around 60 museums and dedicated
heritage research centers around the world. Most instrument setups are based on the
design described by Whitmore and have a character of DIY setups. Recently, a
ready-made instrument with software dedicated to conducting MFT tests with the
automated operation of many functions has been developed and became available
on the market.
16.3.1.2 Data Processing and Interpretation
An important criterion in the lightfastness analysis for the heritage materials is

whether the expected color change of the object exceeds the just-noticeable dif-
ference (JND) threshold. Art conservation literature recommendations often refer to
this value [150], although its precise determination encounters difficulties due to the
perceptual nonuniformities of the CIELab color space. The microfading tests allow
for many materials to achieve color change values that exceed the JND value as the
dose of illumination used in the tests could be very high—depending on the
instrument design and the light power of the source (usually being between 2 and
20 Mlx) and the time of testing (5–20 min), it could easily reach a dose of light
equivalent to more than 1 year of exposure under a typical museum illumination

(i.e., 100 lx per 10 h/day). The outcomes of MFT experiments allow pointing out
the most fragile objects or elements within a given object as well as compare
lightfastness of materials evaluated for their possible application in conservation
work. Analysis of test results can be carried out directly upon finishing a mea-
surement series, and conclusions are immediate and usually do not require any
interpretive measures and skills. Such a simplicity of transition from measurements
to practical conclusions makes the MFT method particularly useful for laboratories
and conservation workshops operating within cultural institutions, where the
microfader next to the pH meter, microscope, and digital camera might become a
standard piece of equipment used for everyday investigation of artifacts.
In order to classify art objects into light sensitivity categories, a method taken
from the textile industry was adopted in 1979 [151] where a set of standardized
materials—the ISO Blue Wool Standards (ISO 105-B08:1995)—is used to express
lightfastness of a given material as the Blue Wool Equivalent. In museum science
materials which colors fade faster than the Blue Wool Standard 3 (BWStd 3) are
considered as fugitive [152], and a strict exhibition policy is usually practiced,
following established maximum temporary light intensities and annual light doses
for objects belonging to BWStd 1–3 categories [153, 154] (Fig. 16.7).
Microfading test has a purely phenomenological character as it is a qualitative
evaluation of the objects’ property without any presumptions on its composition
and type of reaction(s) responsible for the observed color change.
In the analysis of MFT data, there is a temptation to recalculate it into objects’
exposure time under specified museum lighting conditions and thus to predict an
‘exact’ color change after a selected time of objects’ exhibit. Such conversions are
based on the reciprocity principle, postulated by Bunsen and Roscoe in 1865 [155],
which states that change induced by light is directly proportional to the total energy
Fig. 16.7 Color change (calculated according to CIE 1976) for a series of Blue Wool Standards
1–4, measured on MFT instrument with a LED light source
dose irrespective of the regime of its administration. The reciprocity principle holds
only in rare cases when multiple conditions are met:
• The material contains only one compound that is affected by light, and there is
only one pathway of its photodegradation.
• By-products of the photochemical reaction do not accumulate and interfere with
any subsequent photochemical reactions and do not participate in any secondary
reactions with other compounds present in the material.
• There are no reactions of any kind that are independent of light (especially
oxidation).
• There are no dark-time reactions.
As illustrated in [156], where microfading tests were performed for several
lightfastness reference materials and silk samples dyed with turmeric, pomegranate,
and cochineal, the reciprocity low cannot be used to obtain color change predictions
for samples highly susceptible to fading as consistent results were obtained only for
the most stable colors from the tested set. For the BWStd 1, authors established that
the light dose required to reach color change equal to 5 (DE*ab calculated with the
CIE 1976 formula) depends strongly on the illumination intensity used—for 1Mlx
illumination, the dose required was 1.7 Mlx h, for 4Mlx—only 0.4 Mlx h.
The significance of the ability to directly test the lightfastness of an object is
illustrated by the test results in which microfading measurements were performed
on the same dyes applied on papers of different compositions (dyeing technique and
concentration of dyes were identical; samples were taken from a sample book
produced by National Aniline Division, Allied Chemical & Dye Corporation, N.Y.,
1957). As can be seen in Fig. 16.8, changes in color, expressed here as DE*ab,
depend to a large extent on the type of paper used as the substrate, not on the
lightfastness of the colorant used.
Fig. 16.8 Color change measured in microfading tests (5 min of illumination by 4.2 Mlx white
LED) for four papers of different compositions dyed with two dyes, alizarine or methyl violet
Fig. 16.9 Fading of the BWStd 1: time series of reflectance (top) and reflectance difference R0–Rn
(bottom)
MFT measurements are based on FORS and carry more information than just
DE*ab, and information on the direction of color changes (e.g., yellowing, dark-
ening, fading), but the spectra interpretation leading to chemical identification of
components is not usually carried out by MFT practitioners. Some authors though
express the tested light resistance of faded samples by other measures than color
changes. Liang [157] and Bacci [158] proposed to use spectral change in the
absorption region (DR) as an alternative way to quantify light sensitivity of an
object. The time series of spectra for the faded sample presented in Fig. 16.9
illustrates this approach—for the tested BWStd 1 two regions of reflectance change
can be clearly seen at 440 and 705 nm. As the color change expressed in DE units
prevails in the conservation literature and because it gives direct guidance for the
expected changes of appearance for the tested artwork, this way of reporting
microfading results is of the greatest importance.
16.3.1.3 Applications of the MFT Method
The MFT method was up to now used exclusively for the study of heritage
materials although the technique could be implemented in various industries (i.e.,
paint and varnish, textile, printing media) to perform quality control and in testing
new products. The ability to quickly perform lightfastness tests and continuous
monitoring of color changes are strong arguments for the further dissemination of
the method, and if this happens, it will be a unique case of the transfer of mea-
surement technique from the CH studies domain to others, and not vice versa.
Bruce Ford, one of the pioneers of the MFT method, has compiled and updates a
comprehensive list of research papers (60+) in which authors implement the
technique for various types of materials (microfading.com/resources). The types of
materials on the list are as numerous as diverse are items in museums collections:
different media paintings (oil, watercolors) on different supports (including wall
paintings in grottoes), prints, photographs, inks in manuscripts (predominantly iron
gall ink), historical banknotes, textiles (including Maori and Alaskan natives). One
of the recurring research topics is the use of the MFT technique to evaluate
application of inert gases for a safe display of sensitive objects in anoxic frames or
showcases. Removal of oxygen was proved to slow down photodegradation for
many of the tested colorants, particularly a highly fugitive Prussian blue pigment
[159–161].
Also, modern art materials have been evaluated by means of the MFT.
Morales-Merino studied a selection of contemporary writing media (fineliners and
roller ball pens) and original artworks where type of a pen used by the artist was
known [162] and found out that it is not possible to relate directly results obtained
for model samples to the artwork with identical colorants used—authors strongly
recommend testing originals. In the same paper, it was shown also how big
influence on lightfastness has a colorant loading—the lower the concentration of the
ink on a tested spot, the higher the color change in aging tests. Del Hoyo-Melendez
and co-authors [163] surveyed textiles with various dye/mordant/silk combinations,
and one of their findings supported cited Morales-Merino conclusion: Color satu-
ration plays an important role in photostability of materials—textiles with higher
colorant load have shown also better lightfastness.
16.4 Conclusions
Scientific research on heritage and art objects began already in the eighteenth
century. The advent of modern analytical techniques in the mid-twentieth century
has brought enormous progress in this field. The introduction of Raman spec-
troscopy followed by an intensive development of its technical improvements and
following modifications greatly increasing its sensitivity (SERS) and flexibility in
applications requiring a nondestructive approach (FORS) has had probably the
greatest impact on heritage studies. Presented applications of Raman spectroscopy
in the study of cultural artifacts are just a small selection of the possibilities offered
by this technique in the field.
The second method presented in details in the chapter opens a possibility to
predict a future behavior of the studied sample upon its exposure to light, light—an
indispensable enemy of many art and heritage objects created with the use of
light-sensitive materials (organic dyes, some pigments, varnishes, etc.). Thanks to
the MFT method, preservation policies must no longer be based only on an edu-
cated guess and a general classification of objects to different light sensitivity
classes, but on a far more solid foundation of individual (and nondestructive) tests.
Although the technique is still at the initial stage of development, we believe that its
use will spread in the cultural constitutions and the MFT will become a widely used
tool in hands of not only CH researchers but also art conservators.
References
1. Ferchault de Réaumur RA (1719) Observations sur la matière qui colore des perles fausses et
sur quelques autres matières animales d’une semblable couleur, à l’occasion de quoi on
essaie d’expliquer la formation des écailles de poissons. In: Académie des Sciences
(ed) Memoires Académie des Sciences. Académie des Sciences. Paris
2. Ferchault de Réaumur RA. (1727) Idée générale des différentes manières dont on peut faire
la Porcelaine et quelles sont les véritablesmatières de celle de la Chine. In: Académie des
Sciences (ed) Memoires Académie des Sciences. Paris
3. Ferchault de Réaumur RA (1729) Second mémoire sur la porcelaine ou suite des principes
qui doivent conduire dans la composition des porcelaines de différents genres et qui
établissent les caractères des matières fondantes qu’on ne peut choisir pour tenir lieu de celle
qu’on employe à l. In: Académie des Sciences (ed) Memoires Académie des Sciences. Paris
4. Ferchault de Réaumur RA. (1736) Mémoire sur l’art de faire une nouvelle espèce de
Porcelaine par des moyens extrêmement simples et faciles ou de transformer le verre en
porcelaine. In: Académie des Sciences (ed) Memoires Académie des Sciences. Paris
5. Edwards HGM, Vandenabeele P (2016) Raman spectroscopy in art and archaeology. Phil
Trans R Soc A 374:1–3
6. Pollard, A. M., Batt, C. M., Stern, B. and Young SMM (2007) Analytical chemistry in
archaeology. Camb Manuals Archaeol
7. Eccles HRB (1922) Analysis of English porcelains in the V&A museum collections. Victoria
and Albert Museum Publishing, South Kensington
8. Vandenabeele P, Edwards HGM, Moens L (2007) A decade of Raman spectroscopy in art
and archeology. Chem Rev 107:675–686
9. Vandenabeele P, Edwards HGM, Jehlička J (2014) The role of mobile instrumentation in
novel applications of Raman spectroscopy: archaeometry, geosciences, and forensics. Chem
Soc Rev 43:2628
10. Centeno SA (2016) Identification of artistic materials in paintings and drawings by Raman
spectroscopy: some challenges and future outlook. J Raman Spectrosc 47:9–15
11. Casadio F, Daher C, Bellot-Gurlet L (2016) Raman Spectroscopy of cultural heritage
materials: overview of applications and new frontiers in instrumentation, sampling
modalities, and data processing. Top Curr Chem 5:374
12. Bersani D, Conti C, Matousek P, Pozzi F, Vandenabeele P (2016) Methodological
evolutions of Raman spectroscopy in art and archaeology. Anal Methods 8:8395–8409
13. Colomban P (2018) On-site Raman study of artwork: procedure and illustrative examples.
14. Burgio L, Clark RJH (2001) Library of FT-Raman spectra of pigments, minerals, pigment
media and varnishes, and supplement to existing library of Raman spectra of pigments with
visible excitation Spectrochim Acta PA 57:1491–1521
15. Bouchard M, Smith DC (2003) Catalogue of 45 reference Raman spectra of minerals

concerning research in art history or archaeology, especially on corroded metals and
coloured glass. Spectrochim Acta A 59:2247–2266
16. Fremout W, Saverwyns S (2012) Identification of synthetic organic pigments: the role of a
comprehensive digital Raman spectral library. J Raman Spectrosc 43:1536–1544
17. Vandenabeele P, Moens L, Edwards HGM, Dams R (2000) Raman spectroscopic database
of azo pigments and application to modern art studies. J Raman Spectrosc 31:509–517
18. Edwards HGM, Farwell DW, Daffner L (1996) Fourier-transform Raman spectroscopic
study of natural waxes and resins. Spectrochim Acta A 52:1639–1648
19. Leona M, Stenger J, Ferloni E (2006) Application of surface-enhanced Raman scattering
techniques to the ultrasensitive identification of natural dyes in works of art. J Raman
20. Edwards HGM, Farwell DW, Holder JM, Lawson EE (1997) Fourier-transform Raman
spectroscopy of ivory: II. Spectroscopic analysis and assignments. J Mol Struct 435:49–58
21. Madariaga JM (2015) Analytical chemistry in the field of cultural heritage. Anal Methods
7:4848–4876
22. Edwards HGM, Chalmers JM (eds) (2005) Raman spectroscopy in archaeology and art
history. Royal Society of Chemistry, Cambridge
23. Vandenabeele P (2004) Raman spectroscopy in art and archaeology. J Raman Spectrosc
35:607–609
24. Bellot-Gurlet L, Pages-Camagna S, Coupry C (2006) Raman spectroscopy in art and
archaeology. J Raman Spectrosc 37:962–965
25. Baraldi P, Tinti A (2008) Raman spectroscopy in art and archaeology. J Raman Spectrosc
39:963–965
26. Madariaga JM (2010) Raman spectroscopy in art and archaeology. J Raman Spectrosc
41:1389–1393
27. Bersani D, Madariaga JM (2012) Applications of Raman spectroscopy in art and
archaeology. J Raman Spectrosc 43:1523–1528
28. Ropret P, Madariaga JM (2014) Applications of Raman spectroscopy in art and archaeology.
29. Łydżba-Kopczyńska B, Madariaga JM (2016) Applications of Raman spectroscopy in art
and archaeology. J Raman Spectrosc 47:1404–1407
30. Bohning JJ, Misra TN, Choudhury M (1998) The Raman effect. American Chemical
Society, Washington
31. Delhaye M, Dhamelincourt P (1975) Raman microprobe and microscope with laser
excitation. J Raman Spectrosc 3:33–43
32. Analytical Methods Committee AN 67 (2015) Raman spectroscopy in cultural heritage:
Background paper. Anal Methods 7:4844–4847
33. Edwards HGM, Vandenabeele P (2012) Analytical archaeometry. Selected Topics. The
Royal Society of Chemistry, Cambridge
34. Vandenabeele P, Weis TL, Grant ER, Moens LJ (2004) A new instrument adapted to in situ
Raman analysis of objects of art. Anal Bioanal Chem 379:137–142
35. Jehlička J, Culka A, Bersani D, Vandenabeele P (2017) Comparison of seven portable
Raman spectrometers: beryl as a case study. J Raman Spectrosc 48:1289–1299
36. Guineau B (1984) Microanalysis of painted manuscripts and of colored archaeological
materials by Raman laser microprobe. J Forensic Sci 29:471–485
37. Guineau BGV (1984) Analyse non destructive des pigments par microsonde Raman laser:
exemples de l’azurite et de la malachite. Stud Conserv 29:35–41
38. Delhaye M, Guineau B, Vezin J, Coupry C (1985) La microsonde Raman au secours des
oeuvres d’art. Mesures 11:119–124
39. Guineau B (1987) Non-destructive analysis of organic pigments and dyes using raman
microprobe, microfluorimeter rand absorption microspectrophotometer. Stud Conserv
34:38–40
40. Edwards HGM, Farwell DW, Seaward MRD (1991) Raman spectra of oxalates in lichen
encrustations on Renaissance frescoes. Spectrochim Acta A 47:1531–1539
41. Edwards HGM, Farwell DW, Jenkins R, Seaward MRD (1992) Vibrational Raman
spectroscopic studies of calcium oxalate monohydrate and dihydrate in lichen encrustations
on renaissance frescoes. J Raman Spectrosc 23:185–189
42. Russ J, Palma RL, Loyd DH, Farwell DWEH (1995) Analysis of the rock accretions in the
Lower Pecos Region of SW Texas. Geoarchaeology 10:43–63
43. Williams AC, Edwards HGM, Barry BW (1995) The ‘Iceman’: molecular structure of
5200-year-old skin characterised by raman spectroscopy and electron microscopy.
BBA-Protein Struct Mol 1246:98–105
44. Best SP, Clark RJH, Withnall R (1992) Non-destructive pigment analysis of artefacts by
Raman microscopy. Endeavour 16:66–73
45. Clark RJH, Cooksey CJ, Daniels MAMWR (1993) Indigo, woad, and Tyrian purple:
important vat dyes from antiquity to the present. Endeavour 17:191–199
46. Ciomartan DA, Clark RJH (1996) Raman microscopy applied to the analysis of the pigments
used in two persian manuscripts. J Braz Chem Soc 7:395–402
47. Burgio L, Ciomartan DA, Clark RJH (1997) Pigment identification on medieval
manuscripts, paintings and other artefacts by Raman microscopy: applications to the study
of three German manuscripts. J Mol Struct 405:1–11
48. Burgio L, Ciomartan DA, Clark RJH (1997) Raman microscopy study of the pigments on
three illuminated mediaeval latin manuscripts. J Raman Spectrosc 28(7):9–83
49. Coupry C, Lautié A, Revault M, Dufilho J (1994) Contribution of Raman spectroscopy to art
and history. J Raman Spectrosc 25:89–94
50. Turrell G (1996) Raman microscopy developments and applications. Elsevier, Academic
Press, Amsterdam
51. Pagès-Camagna S, Colinart S, Coupry C (1999) Fabrication processes of archaeological
Egyptian blue and green pigments enlightened by Raman microscopy and scanning electron
microscopy. J Raman Spectrosc 30:313–317
52. Bruni S, Cariati F, Cl Bianchi, Zanardini E, Sorlini C (1995) Spectroscopic investigation of red
stains affecting the Carrara marblefaçade of the Certosa of Pavia. Archaeometry 37:249–255
53. Trentelman K, Stodulski L, Pavlosky M (1996) Characterization of pararealgar and other
light-induced transformation products from realgar by Raman microspectroscopy. Anal
Chem 68:1755–1761
54. Barone G, Crupi V, Galli S, Majolino D, Migliardo P, Venuti V (2003) Spectroscopic
investigation of Greek ceramic artefacts. J Mol Struct 651–653:449–458
55. Bruno P, Caselli M, Curri ML, Favia P, Laganara C, Traini A (1997) Surface examination of
red painting on medieval pottery from the South of Italy. Anal Chim 87:539–553
56. Lofrumento C, Zoppi A, Castellucci EM (2004) Micro-Raman spectroscopy of ancient
ceramics: a study of Frenchsigillata wares. J Raman Spectrosc 35:650–655
57. Colomban P, Treppoz F (2001) Identification and differentiation of ancient and modern
European porcelains by Raman macro- and micro-spectroscopy. J Raman Spectrosc 32:93–102
58. Colomban P, Sagon G, Faurel X (2001) Differentiation of antique ceramics from the Raman
spectra of their coloured glazes and paintings. J Raman Spectrosc 32:351–360
59. Mioč UB, Colomban P, Sagon G, Stojanović M, Rosić A (2004) Ochre decor and cinnabar
residues in Neolithic pottery from Vinča, Serbia. J Raman Spectrosc 35:843–846
60. Smith DC, Vernioles JD (1997) The temperature of fusion of a celtic vitrified fort: a
feasibility study of the application of the Raman microprobe to the non-destructive
characterization of unprepared archaeological objects. J Raman Spectrosc 28:195–197
61. Smith DC, Gendron F (1997) Archaeometric application of the Raman microprobe to the
non-destructive identification of two pre-columbian ceremonial polished ‘greenstone’
axe-heads from Mesoamerica. J Raman Spectrosc 28:731–738
62. Gendron F, Smith DC, Gendron-Badou A (2002) Discovery of jadeite-jade in Guatemala
confirmed by non-destructive Raman microscopy. J Archaeol Sci 29:837–851
63. Parras D, Vandenabeele P, Sánchez A, Montejo M, Moens L, Ramos N (2010)

Micro-Raman spectroscopy of decorated pottery from the Iberian archaeological site of
puente tablas. J Raman Spectrosc 41:68–73
64. Bersani D, Lottici PP, Virgenti S, Sodo A, Malvestuto G, Botti A, Salvioli-Mariani E,
Tribaudino M, Ospitali F, Catarsi M (2010) Multi-technique investigation of archaeological
pottery from Parma (Italy). J Raman Spectrosc 41:1556–1561
65. Colomban P (2013) The destructive/non-destructive identification of enameled pottery, glass
artifacts and associated pigments—a brief overview. Arts 2:77–110
66. Medeghini L, Lottici PP, De Vito C, Mignardi S, Bersani D (2014) Micro-Raman spectroscopy
and ancient ceramics: applications and problems. J Raman Spectrosc 45:1244–1250
67. De Benedetto GE, Nicolì S, Pennetta A, Rizzo D, Sabbatini L, Mangone A (2011) An
integrated spectroscopic approach to investigate pigments and engobes on pre-Roman
pottery. J Raman Spectrosc 42:1317–1323
68. De Vito C, Medeghini L, Mignardi S, Ballirano P, Peyronel L (2015) Technological
fingerprints of the Early Bronze Age clay figurines from Tell Mardikh-Ebla (Syria). J Eur
Ceram Soc 35:3743–3754
69. Edwards HGM (1998) Raman spectroscopy offresco fragment substrates. Asian J Phys 7:383–389
70. Edwards HGM, Farwell DW, Newton EM, Perez FR, Edwards HGM, Farwell DW,
Newton EM, Perez FR (1999) Minium; FT-Raman non-destructive analysis applied to an
historical controversy. Analyst 124:1323–1326
71. Edwards HGM, Farwell DW, Rull Perez F, Jorge Villar S (1999) Spanish mediaeval frescoes
at Basconcillos del Tozo: a fourier transform Raman spectroscopic study. J Raman Spectrosc
30:307–311
72. Edwards HGM, Farwell DW, Rull Perez F, Jorge Villar S (1999) Spanish mediaeval frescoes
at Basconcillos del Tozo: a fourier transform Raman spectroscopic study. J Raman Spectrosc
30:307–311
73. Bersani D, Antonioli G, Lottici PP, Casoli A (2003) Raman microspectrometric investi-
gation of wall paintings in S. Giovanni Evangelista Abbey in Parma: a comparison between
two artists of the 16th century. Spectrochimica Acta A 59:2409–2417
74. Bersani D, Paolo Lottici P, Antonioli G, Campani E, Casoli A, Violante C (2004) Pigments
and binders in the wall paintings of Santa Maria della Steccata in Parma(Italy): the ultimate
technique of Parmigianino. J Raman Spectrosc 35:694–703
75. Baraldi P, Bonazzi A, Giordani N, Paccagnella F, Zannini P (2006) Analytical character-
ization of Roman plasters of the ‘Domus farini’ in Modena. Archaeometry 48:481–499
76. Mazzeo R, Baraldi P, Lujàn R, Fagnano C (2004) Characterization of mural painting
pigments from the Thubchen Lakhang temple in Lo Manthang, Nepal. J Raman Spectrosc
35:678–685
77. Aliatis I, Bersani D, Campani E, Casoli A, Lottici PP, Mantovan S, Marino IG (2010) Pigments
used in Roman wall paintings in the Vesuvian area. J Raman Spectrosc 41:1537–1542
78. Maguregui M, Knuutinen U, Castro K, Madariaga JM (2010) Raman spectroscopy as a tool
to diagnose the impact and conservation stateof Pompeian second and fourth style wall
paintings exposed to diverseenvironments (House of Marcus Lucretius). J Raman Spectrosc
41:1400–1409
79. Madariaga JM, Maguregui M, De Vallejuelo SFO, Knuutinen U, Castro K, Martinez-Arkarazo
I, Giakoumaki A, Pitarch A (2014) In situ analysis with portable Raman and ED-XRF
spectrometers for the diagnosis of the formation of efflorescence on walls and wall paintings of
the Insula IX 3 (Pompeii, Italy). J Raman Spectrosc 45:1059–1067
80. Madariaga JM, Maguregui M, Castro K, Knuutinen U, Martínez-Arkarazo I (2016) Portable
Raman, DRIFTS, and XRF analysis to diagnose the conservation state of two wall painting
panels from pompeii deposited in the Naples National Archaeological Museum (Italy). Appl
81. Prieto-Taboada N, Fdez-Ortiz De Vallejuelo S, Veneranda M, Marcaida I, Morillas H,
Maguregui M, Castro K, De Carolis E, Osanna M, Madariaga JM (2018) Study of the
soluble salts formation in a recently restored house of Pompeii by in-situ Raman

spectroscopy. Sci Rep 8:1613
82. Smith DC, Bouchard M, Lorblanchet M (1999) An initial Raman microscopic investigation
of prehistoric rock art in caves of the Quercy District, S.W., France. J Raman Spectrosc
30:347–354
83. Edwards HGM, Drummond L, Russ J (1999) Fourier transform Raman spectroscopic study
of prehistoric rock paintings from the Big Bend region, Texas. J Raman Spectrosc 30:421–
428
84. Hernanz A, M. Gavira-Vallejo J don J, Ruiz-Lopez F (2006) Introduction to Raman
microscopy of prehistoric rockpaintings from the Sierra de las Cuerdas, Cuenca, Spain.
85. Hernanz A, Gavira-Vallejo JM, Ruiz-López JF (2007) Calcium oxalates and prehistoric
paintings. the usefulness of these biomaterials. J Optoelectron Adv Mater 9:512–521
86. Hernanz A, Gavira-Vallejo JM, Ruiz-Lopez JF, Edwards HGM (2008) A comprehensive
micro-Raman spectroscopic study of prehistoric rock paintings from the Sierra de las
Cuerdas, Cuenca, Spain. J Raman Spectrosc 972–984
87. Hernanz A, Ruiz-López JF, Gavira-Vallejo JM, Martin S, Gavrilenko E (2010) Raman
microscopy of prehistoric rock paintings from the Hoz de Vicente, Minglanilla, Cuenca,
Spain. J Raman Spectrosc 41:1394–1399
88. Hernanz A, Gavira-Vallejo JM, Ruiz-Lõpez JF, Martin S, Maroto-Valiente Á, De
Balbín-Behrmann R, Menéndez M, Alcolea-González JJ (2012) Spectroscopy of
Palaeolithic rock paintings from the Tito Bustillo and El Buxu Caves, Asturias, Spain.
89. Colomban P, Slodczyk A (2009) Raman intensity: an important tool to study the structure
and phase transitions of amorphous/crystalline materials. Opt Mater (Amst) 31:1759–1763
90. Katsaros T, Ganetsos T (2012) Raman characterization of gemstones from the collection of
the Byzantine & Christian Museum. Archaeology 1:7–14
91. Košařová V, Hradil D, Hradilová J, Čermáková Z, Němec I, Schreiner M (2016) The
efficiency of micro-Raman spectroscopy in the analysis of complicated mixtures in modern
paints: Munch’s and Kupka’s paintings under study. Spectrochim Acta A 156:36–46
92. Bersani D, Lottici PP (2010) Applications of Raman spectroscopy to gemology. Anal
Bioanal Chem 397:2631–2646
93. Ermakov NP (1965) Research on the nature of mineral-forming solutions: with special
reference to data from fluid inclusions. Oxford, Pergamon Press Ltd., Oxford
94. Roedder E (1984) Fluid inclusions. In: Ribbe PH (ed) Reviews in mineralogy and
geochemistry. Mineralogical Society of America, p 644
95. Łydzba-Kopczyńska B, Zych E, August C, Rusek G, Pankiewicz A (2008) Analytical
techniques in provenance determination of archaeological objects from Lower Silesia. J Mol
Struct 887:41–47
96. Kolesov BA, Geiger CA (1998) Raman spectra of silicate garnets. Phys Chem Miner
25:142–151
97. Barone G, Bersani D, Lottici PP, Mazzoleni P, Raneri S, Longobardo U (2016) Red
gemstone characterization by micro-Raman spectroscopy: the case of rubies and their
imitations. J Raman Spectrosc 47:1534–1539
98. Reiche I, Pages-Camagna S, Lambacher L (2004) In situ Raman spectroscopic investigations
of the adorning gemstones on the reliquary Heinrich’s Cross from the treasury of Basel
Cathedral. J Raman Spectrosc 35:719–725
99. Ziemann MA (2006) In situ micro-Raman spectroscopy on minerals on-site in the Grotto
Hall of the New Palace, Park Sanssouci, in Potsdam. J Raman Spectrosc 37:1019–1025
100. Coccato A, Bersani D, Coudray A, Sanyova J, Moens L, Vandenabeele P (2016) Raman
spectroscopy of green minerals and reaction products with an application in Cultural
Heritage research. J Raman Spectrosc 47:1429–1443
101. Bersani D, Lottici PP (2016) Raman spectroscopy of minerals and mineral pigments in
archaeometry. J Raman Spectrosc 47:499–530
102. Centeno SA, Llado Buisan V, Ropret P (2007) Raman study of synthetic organic pigments
and dyes in early lithographic inks (1890–1920). J Raman Spectrosc 37:111–1118
103. Daher C, Bellot-Gurlet L, Le Hô AS, Paris C, Regert M (2013) Advanced discriminating
criteria for natural organic substances of cultural heritage interest: spectral decomposition
and multivariate analyses of FT-Raman and FT-IR signatures. Talanta 115:540–547
104. Daher C, Pimenta V, Bellot-Gurlet L (2014) Towards a non-invasive quantitative analysis of
the organic components in museum objects varnishes by vibrational spectroscopies:
methodological approach. Talanta 129:336–345
105. Klisinska A, Łydżba-Kopczyńska B, Czarnecka M, Koźlecki T, del Hoyo Mélendez J,
Mendys A, Kłosowska‐Klechowska, Anna Obarzanowski M, Frączek P (2018) Raman
spectroscopy as a powerful technique for the identification of polymers used in cast
sculptures from museum collections. J Raman Spectrosc. https://doi.org/10.1002/jrs.5407
(First published: 08 June 2018)
106. Van Elslande E, Lecomte S, Le Ho A-S (2008) Micro-Raman spectroscopy (MRS) and
surface-enhanced Raman scattering (SERS) on organic colourants in archaeological
pigments. J Raman Spectrosc 39:10001–11006
107. Melo MJ, Nabais P, Guimarães M, Araújo R, Castro R, Oliveira MC, Whitworth I (2016)
Organic dyes in illuminated manuscripts: a unique cultural and historic record. Philos
Trans R Soc A Math Phys Eng Sci 374:20160050
108. Pozzi F, Leona M (2016) Surface-enhanced Raman spectroscopy in art and archaeology.
109. Kiefer W, Mazzolini AP, Stoddart PR (2007) Recent advances in linear and nonlinear
Raman spectroscopy I. J Raman Spectrosc 38:1538–1553
110. Cappa F, Pintus V, Ofner J, Schreiner M, Lendl B (2015) Raman imaging for cultural
heritage investigations. In: 8th international conference on advanced vibrational spec-
troscopy, Wien, 7–17th July 2015, pp 246–247
111. Antunes V, Candeias A, Mirão J, Carvalho ML, Dias CB, Manhita A, Cardoso A,
Francisco MJ, Lauw A, Manso M (2018) Analytical characterization of the palette and
painting techniques of Jorge Afonso, the great 16th century Master of Lisbon painting
workshop. Spectrochim Acta A 193:264–275
112. Christiansen MB, Sørensen MA, Sanyova J, Bendix J, Simonsen KP (2017) Characterisation
of the rare cadmium chromate pigment in a 19th century tube colour by Raman, FTIR, X-ray
and EPR. Spectrochim Acta A 208–214
113. Basso E, Invernizzi C, Malagodi M, La Russa MF, Bersani D, Lottici PP (2014)
Characterization of colorants and opacifiers in roman glass mosaic tesserae through
spectroscopic and spectrometric techniques. J Raman Spectrosc 45:238–245
114. Casadio F, Bezúr A, Fiedler I, Muir K, Trad T, MacCagnola S (2012) Pablo Picasso to
Jasper Johns: a Raman study of cobalt-based synthetic inorganic pigments. J Raman
115. Otero V, Sanches D, Montagner C, Vilarigues M, Carlyle L, Lopes JA, Melo MJ (2014)
Characterisation of metal carboxylates by Raman and infrared spectroscopy in works of art.
116. Carlesi S, Ricci M, Cucci C, Lofrumento C, Picollo M, Becucci M (2016) Multivariate
analysis of combined reflectance FT-NIR and micro-Raman spectra on oil-paint models.
Microchem J 124:703–711
117. Cheilakou E, Troullinos M, Koui M (2014) Identification of pigments on Byzantine wall
paintings from Crete (14th century AD) using non-invasive fiber optics diffuse reflectance
spectroscopy (FORS). J Archaeol Sci 41:541–555
118. Bronzato M, Zoleo A, Biondi B, Centeno SA (2016) An insight into the metal coordination
and spectroscopic properties of artistic Fe and Fe/Cu logwood inks. Spectrochim Acta A
153:522–529
119. Cucci C, Bartolozzi G, De Vita M, Marchiafava V, Picollo M, Casadio F (2016) The colors
of Keith Haring: a spectroscopic study on the materials of the mural painting Tuttomondo
and on reference contemporary outdoor paints. Appl Spectrosc 70:186–196
120. Sirat C (2002) Hebrew Manuscripts of the Middle Ages. Cambridge University Press,
Cambridge
121. Gruchalska A, Rogulska A, Rusek G, Łydzba-Kopczyńska BI (2010) Spectroscopic studies
of atypically illuminated medieval Hebrew bible in comparison to a XV century western
manuscript. AIP Conf Proc 1267:238–239
122. Łydżba-Kopczyńska B, Rogulska A (2016) Decoration techniques in mediaeval Hebrew
Bibles: non-invasive XRF and l-Raman analysis. In: 6th meeting X-ray and other
techniques in investigations of the objects of cultural heritage, Kraków, 19–21 May 2016,
pp 238–239
123. Ray E (1999) Sofer: the story of a Torah Scroll. Torah Aura Productions, Los Angeles
124. Burgio L, Clark RJH, Firth S (2001) Raman spectroscopy as a means for the identification of
plattnerite (PbO2), of lead pigments and of their degradation products. Analyst 126:222–227
125. Conservation KP offers RM for A. http://shop.kremerpigments.com/en/
126. Eastaugh N, Walsh V, Siddall R, Chaplin T (2004) Optical microscopy of historical
pigments. Taylor & Francis, London
127. Łydzba-Kopczyńska BI, Gediga B, Chojcan J, Sachanbiński M (2012) Provenance
investigations of amber jewelry excavated in Lower Silesia (Poland) and dated back to
Early Iron Age. J Raman Spectrosc 43:1839–1844
128. Angelini I, Bellintani P (2005) Archaeological ambers from northern Italy: an FTIR-DRIFT
study of provenance by comparison with the geological amber database. Archaeometry
47:441–454
129. Guiliano M, Asia L, Onoratini G, Mille G (2007) Applications of diamond crystal
ATR FTIR spectroscopy to the characterization of ambers. Spectrochim Acta A 67:1407–
1411
130. Seger H (1930) Der Bernsteinfund von Hartlieb bei Breslau. Altschlesien 3:171–184
131. Nowothnig W (1937) Der Bernsteinhandelsplatz von Breslau-Hartlieb. Altschlesische
Blätter. Altschlesische, Blätter, pp 48–51
132. Łydżba-Kopczyńska B, Krzywiecka M, Chojcan J, Madera PS (2015) Wrocław-Partynice
amber depots—the application of the comprehensive spectral database of succinate and
fossil and subfossil resins. In: 8th Congress on Application of Raman Spectroscopy in Art
and Archaeology Wroclaw, 1–5 Sept 2015, pp 146–147
133. Niedźwiedzki R (2014) Gigantyczny bursztynowy “skarb” partynicki z Wrocławia. In:
Amberif 2014. Bursztyn. Gemmologia – Muzealnictwo – Archeologia. Gdańsk-Warszawa,
23–26
134. Brody RH, Edwards HGM, Pollard AM (2001) A study of amber and copal samples using
FT-Raman. Spectrochim Acta A 57:1325–1338
135. Peris-Díaz MD, Łydżba-Kopczyńska B, Sentandreu E (2018) Raman spectroscopy coupled
to chemometrics to discriminate provenance and geological age of amber. J Raman
136. Łydżba-Kopczyńska B, Białek E, Rusek G, Weker W (2011) Badania pochodzenia
obiektów bursztynowych poddanych zabiegom konserwatorskim. In: XI konferencja
Analiza chemiczna w ochronie zabytków, Warszawa 1–2 Dec 2011, p. 33
137. Navas N, Romero-Pastor J, Manzano E, Cardell C (2010) Raman spectroscopic discrim-
ination of pigments and tempera paint model samples by principal component analysis on
first-derivative spectra. J Raman Spectrosc 41:1486–1493
138. Doherty B, Vagnini M, Dufourmantelle K, Sgamellotti A, Brunetti B, Miliani C (2014) A
vibrational spectroscopic and principal component analysis of triarylmethane dyes by
comparative laboratory and portable instrumentation. Spectrochim Acta A 121:292–305
139. Otero V, Sanches D, Montagner C, Vilarigues M, Carlyle L, Lopes JA, Melo MJ (2014)
Characterisation of metal carboxylates by Raman and infrared spectroscopy in works of art.
140. Delaney JK, Ricciardi P, Deming Glinsman L, Facin M, Thoury M, Palmet M, René de la Rie E
(2014) Use of imaging spectroscopy, fiber optic reflectance spectroscopy, and X-ray
fluores-cence to map and identify pigments in illuminated manuscripts. Stud Conserv 59:91–101
141. Montagner C, Bacci M, Bracci S, Freeman R, Picollo M (2011) Library of UV-Vis-NIR

reflectance spectra of modern organic dyes from historic pat-tern-card coloured papers.
142. Depuis G, Menu M (2006) Quantitative characterisation of pigment mixtures used in art by
fiber-optics diffuse-reflectance spectroscopy. Appl Phys A 83:469–474
143. FORS (2016) Fiber optics reflectance spectra (FORS) of pictorial materials in the 270–
1700 nm range. http://fors.ifac.cnr.it. Accessed 16 June 2018
144. Bacci M, Picollo M, Trumpy G, Tsukada M, Kunzelman D (2007) Non-invasive
identification of white pigments on 20th century oil paintings by using fiber optic
reflectance. Spectrosc JAIC 46:27–37
145. Cosentino A (2014) FORS spectral database of historical pigments in different binders.
e-Conservation J 2: 57–68
146. Cavaleria T, Giovagnolia A, Nervoa M (2013) Pigments and mixtures identification by
Visible Reflectance Spectroscopy. Procedia Chem 8:45–54
147. Some examples for Textiles (2014) ISO 105-B01:2014 Textiles—tests for colour fastness—
part B01: colour fastness to light: daylight; ISO 105-B02:2014, Textiles—tests for colour
fastness—part B02: colour fastness to artificial light: Xenon arc fading lamp
148. Whitmore PM, Pan X, Bailie C (1999) Predicting the fading of objects: identification of
fugitive colorants through direct nondestructive lightfastness measurements. JAIC 38:395–
409
149. Pesme C, Lerwill A, Beltran V, Druzik J (2016) Development of contact portable microfade
tester to assess light sensitivity of col-lection items. J Am Inst Conserv 55:117–137
150. Ashley-Smith J, Derbyshire A, Pretzel B (2002) The continuing development of a practical
lighting policy for works of art on paper and other object types at the Victoria and Albert
Museum. In: ICOM Committee for Conservation, preprints, 13th triennial meeting, Rio De
Janeiro, Brazil. ICOM, Paris, pp 3–8
151. Feller RL (1979) Use of the international standards organization’s blue wool standards for
exposure to light. In: AIC preprints of papers presented at the 7th annual meeting, Toronto,
Canada
152. Michalski S (1987) Damage to museum objects by visible radiation (light) and ultraviolet
radiation (UV). In: Lighting in museums, galleries and historic houses. Museums
Association, UKIC, and Group of Designers and Interpreters for Museums, London
153. Thompson G (2013) The museum environment, 6th edn. Elsevier, Amsterdam
154. Colby KM (1992) A suggested exhibition/exposure policy for works of art on paper. J Int
Inst Conserv Can Gr 17:3–11
155. Roscoe HE (1865) The Bakerian Lecture: on a method of meteorological registration of the
chemical action of total daylight. Philos Trans R Soc L 155:605–631
156. del Hoyo-Meléndez JM, Mecklenburg MF (2011) An investigation of the reciprocity
principle of light exposures using micro-fading spectrometry. Spectrosc Lett 44:52–62
157. Liang H, Lange R, Lucian A, Hyndes P, Townsend J., Hackney S (2011) Development of
portable microfading spectrometers for measurement of light sensitivity of materials. In:
International Council of Museums, Committee for Conservation (ICOM-CC) triennial
conference, Lisbon
158. Bacci M, Cucci C, Mencaglia AA MA& PS (2004) Calibration and use of photosensitive
materials for light monitoring in museums: blue wool standard 1 as a case study. Stud
Conserv 49:85–98
159. Lerwill A, Townsend JH, Thomas J, Hackney S, Caspers C, Liang H (2014) Photochemical
colour change for traditional watercolour pigments in low oxygen levels. Stud Conserv
60:15–32
160. del Hoyo-Meléndez JM, Mecklenburg MF (2011) The use of micro-fading spectrometry to
evaluate the light fastness of materials in oxygen-free environments. Spectrosc Lett 44:113–121
161. Łojewski T, Thomas J, Gołąb R, Kawałko J, Łojewska J (2011) Light ageing with
simultaneous colorimetry via fiber optics reflection spectrometry. Rev Sci Instrum
82:076102
162. Morales Merino C, Röhrs S, Meyer F, Marten S, Reiche I (2016) Micro-fading testing on
modern ink based pens and contemporary drawings from the Kupferstichkabinett Berlin.
Berliner Beiträge zur Archäometrie, Kunsttechnologie und Konserv 24:89–102
163. del Hoyo-Meléndez JM, Mecklenburg M (2012) Micro-fading spectrometry: a tool for
real-time assessment of the light-fastness of dye/textile systems. Fibers Polym 13:1079–1085
Index
Symbols B
1-methyl-2-thiobenzimidazole, 202 Basic sites, 336
2-thiobenzimidazole, 202 Binomial distribution, 292, 293
a-cage, 302 Bio-imaging applications, 139, 150, 156
b-cage, 302 Biomolecules, 162, 177, 183
e-cage, 302 B-matrix, 61, 77, 78
p-A isotherm, 103, 104 Bond critical point, 207, 320
p bond polarization, 206 Born-Oppenheimer approximation (BOA),
p charges, 206 8–10, 14
p-electron, 97, 99, 100, 106, 109, 116 Breit-Wigner-Fano (BWF) line shape function,
p mesomerism, 206 123, 126, 127, 129–132, 135
R charges, 206 Brønsted sites, 337, 353
Bulk water, 230, 232, 233, 242, 251, 254, 255
A Bulky counterions, 154
Absorption loss, 401, 402, 410–414, 417, 423
Absorbance spectra, 410, 415 C
Acceptor, 97, 98, 105, 110, 113, 114, 116 Calibration constant, 281, 282
Acidic sites, 337, 353 Calsilite, 475–478
Active sites, 334–341, 353 Cancrinite, 302, 319, 326
Alcohol probe molecules, 353 CAN structure type, 319
Aluminosilicate glass, 475–478 Carbon monoxide probe molecule, 335, 338,
Amber, 483, 497–500 339
Ammonia probe molecules, 335, 338 Carbon Nanotubes (CNTs), 123–126, 129, 132,
Amorphous state, 457 134, 135
Anharmonicity, 49, 52, 67, 68, 73, 76, 85 Cartesian displacements, 60, 61, 84
Anti-Stokes effect, 346 Catalyst, 334–342, 344, 346–353
Archeometry, 490 Centrosymmetric, 284, 285, 288, 290
Artworks, 484, 490, 491, 508 Chabazite, 317, 318
Asymmetry ratio, 433, 438 Chain dynamics, 224
Asynchronous spectra, 345 Charge transfer excited state, 146
Atomic Force Microscopy (AFM), 334, 335, CHA structure type, 317
350–353 Chemisorption, 167, 325, 339, 340, 348
Atoms in Molecules (AIM), 206 Chemometrics, 500
Attenuated Total Reflectance (ATR), 340 Chirality, 162, 163, 165, 169, 170
Attenuation, 445, 447, 448, 451 Chromaticity diagram, 453

and Physics 26, https://doi.org/10.1007/978-3-030-01355-4
520 Index
Clinoptilolite, 312, 317, 318, 325 E

Coherent Potential Approximation (CPA), 37 Effective Scaling Factor (ESF), 50, 83, 86
Coil-to-globule transition, 226, 239, 243–245, Effective Scaling Frequency Factor method
252, 253, 256, 260 (ESFF), 49, 80–82, 84–87, 89–91
Compression module, 103 Electrical-Optical-Circuit-Boards (EOCB),
COnductor-like Screening MOdel (COSMO), 406, 416
33 Electron density, 276, 277, 281, 282
Configuration Interaction approximation, 14, Electron transfer, 99, 101–103, 105, 110, 111,
40 113, 116
Confocal microscopy, 155 Endogenous non-linear optics, 139
Conformational cooling, 208 Endothelium, 161, 171, 172, 184
Continuous random network theory, 459, 462 Energy level diagram, 431, 436–438, 444
Corrole, 97, 100, 106–112, 114–117 Energy migration, 448
Corrole-fullerene dyad, 97, 108, 110–112, 115, Energy transfer, 98, 102–105, 110, 111, 115,
116 116, 427, 428, 434, 437, 438, 440, 443,
Coupled Cluster Theory, 14, 40 444, 446, 450, 453
Cristobalite, 458, 460–463, 468, 470, 471 Energy Transfer Upconversion (ETU), 443
Cross-relaxation, 432 Ergothioneine, 201
Crystallite theory, 458–460, 462, 479 Exchange-correlation, 41
Crystallization temperature, 429, 433 Exchange interaction, 295, 296
C2v structure, 106 Excitation spectra, 440
Cultural heritage, 483–486, 497 Excited State Absorption (ESA), 443, 447
Curie temperature, 284–286, 289
Cyclosilicates, 307, 312 F
Factors’ optimization, 80, 90
D FAU structure type, 315, 317, 319
D-band, 123, 125 Fermi contact field, 278
Decomposition, 464–466, 470–477 Ferrierite, 317, 318
Deconvolution, 465, 466 FER structure type, 316–318
Degree of polymerization, 472, 479 Fiber optics reflectance spectroscopy, 483, 490,
Density Functional Theory (DFT), 1, 6, 14–16, 501
20–22, 26–29, 40–42, 97, 107–109, Flory’s theory, 225
113, 114, 304 Fluorescein, 151
Deuteration, 412, 416, 417 Fluorescence kinetics, 107, 111
Devitrificate, 478 Fluorescence quenching, 99, 105
Dexter, 111 Fluorescent Intensity Ratio (FIR), 433
Differential Scanning Calorimetry (DSC), 200, Fock operator, 12, 13
442 Force constants, 49
Diffuse Reflectance Infrared Spectroscopy Force Field (FF), 49, 52, 60, 64
(DRIFT), 340–346 Förster long-range resonance excitation energy
Disorder, 280, 286, 290, 291, 296 transfer (FRET), 110
Disordered structure, 467, 468 Fourier Transform Infrared Spectroscopy
Dissipative Particle Dynamics (DPD), 248 (FT-IR spectroscopy), 103, 107, 335,
Disulfide, 202, 212 337, 338, 340–342, 344–346, 353, 364,
DM interaction, 288 365, 370, 380, 393
Donor, 102, 103, 105, 110, 113–115 Free-base, 100, 101, 107, 108, 114
Doublet, 290, 294, 296 Fullerene, 97, 98, 106, 109–111, 114–117
Drug screening, 176
D2h structure, 100 G
D4h structure, 100, 106 Gaussian distribution, 291, 292
Dynamic Lattice Liquid (DLL) model, 249, G-band, 123–125, 127–134
250 Gel electrophoresis, 144
Dynamics of volume phase transition, 239, Generalized Gradient Approximation (GGA),
250, 252 18–21, 41
Index 521
Glass-ceramic, 431, 432, 453 In situ spectroscopy, 336

Glasses, 427–431, 434, 436–441, 443, 444, Intermolecular interactions, 200, 216, 218
453, 454 Internal Coordinates (ICs), 53, 62, 81
Glass-fiber, 435, 438, 452 Internal coordinates types, 74
Glass structure, 457–464, 467, 472, 473, Intramolecular hydrogen bond, 203, 214
477–479 Ion exchange processes, 324
Glutathione, 144, 154, 156 IR spectroscopy, 327, 464, 467, 469, 472, 475,
Gouterman, 107 477–479
G-type antiferromagnet, 287–289 Isomer shift (IS), 276, 277, 281, 282, 289, 291,
296
H Isothiocyanic acid, 211
Halogenation, 401, 405, 416, 417, 423
Harmonic approximation, 22, 23, 25, 26, 49, K
52, 56, 58, 60, 63, 68, 81 Kasha′s rule, 153
Harmonic oscillator, 54–59 Kinetic energy, 59, 60
Hartree-Fock equations, 12, 13 Kohn-Sham equations, 17, 28
Hartree-Fock method, 10, 20, 304, 305
Heavy metal cations, 325 L
Hessian matrix, 25, 35 Lanthanides ions, 453
Heterogenous structure, 459–461 Laplacian of electron density, 320, 322
HEU structure type, 316, 318 Lattice vibrations, 304, 314, 319, 323, 328
High-vacuum (HV), 335, 341 Least-Squares Merit Function (LSMF), 64–67,
Hofmeister series, 237, 243 75, 83
Hohenberg-Kohn theorem, 14–16 Lebiedev theory, 457
Hole transfer, 113, 115 Lewis sites, 337, 338
HOMO, 97, 99, 107, 115, 116 Lightfastness, 502–504, 506, 508
Hybrid xc functionals, 20 Linde Type A zeolite (LTA), 34–37
Hydration, 199, 201, 205–207, 211, 219, 220, Linear silicone polymers, 403
228, 231, 235, 237, 240, 242, 243, 245, Local Density Approximation (LDA), 18, 19,
250, 252, 253 21, 41
Hydrogel, 228, 231–233, 240, 243, 244, 251, Local Field Enhancement (LFE), 434, 454
254, 255, 257 Local Scaling (scale) Factor (LSF), 82
Hydrogen bond, 228–230, 234, 236, 239, 250, Local Surface Plasmon Resonance sensors
252–254 (LSPR), 435
Hydrogen bonding, 200, 216 Long-range order, 457, 458, 460, 463, 469,
Hydrophilic centre, 228, 231, 239, 240, 242, 470, 479
243, 251, 252, 256 Lorentzian line shape function, 123, 131–133,
Hydrophilic group, 228, 232, 250, 254 135
Hydrophobic interaction, 228, 239, 253 Lower Critical Solution Temperature (LCST),
Hydrophobic part, 232, 239 226
Hydrophobic site, 231 LTA structure type, 302, 303, 316, 318–324
Hydrosilylation, 361–367, 370–378, 380–383, Luminescence decay, 431, 432, 435, 446
385–393 Luminescence spectra, 431, 438–440, 443,
Hydrosilylation reaction, 403, 418 445, 447, 449, 450, 452, 453
Hyperfine interaction, 273–276, 278, 280, 281,
289, 291–293, 296 M
Macroring, 106, 108, 109
I Magnetic hyperfine field, 279, 281, 283,
Independent Electrons Approximation (IEA), 288–290, 294–296
6, 10, 13 Magnetic moment, 279, 284, 286–291,
Infrared Spectroscopy, 199–201, 219, 334, 340 294–296
In silico experiment, 1, 32, 41, 42, 328 Magnetoelectric, 273, 274, 283–286, 288–290,
In situ cell, 336 295, 296
522 Index
Marcus, 111, 116 Normal vibrations, 61, 86

Mass spectrometry (MS), 144, 145, 345 N-vinylidenemethanamine, 211
Mass tensor, 62, 84 Nyquist rule, 409
Matrix isolation, 199–201, 216, 219
Medieval manuscripts, 487, 491 O
Medium polarization, 140 Operando Spectroscopy, 336
Mercaptoimidazoles, 199, 201, 208, 209, 219 Optical fiber, 427–429, 432, 435, 440, 441,
Merocyanine, 102 443–445, 447–453
Mesomeric structures, 206, 207 Optical glasses, 428
Meta-Generalized Gradient Approximation Optical insertion loss, 402, 405, 408, 409,
(MGGA), 18, 20 420–423
Metal nanoclusters, 140, 142, 144 Optical loss, 409, 413, 414, 416–419, 421–423
Metal oxide catalysts, 338–340, 345 Optical propagation loss, 420
Microfading tests, 484, 502, 506 Optical siloxane-based polymers, 402
Molecular aggregate, 104, 105 Organofunctional octahedral oligomeric
Molecular Dynamics (MD), 232, 247 silsesquioxnes, 383
Molecular switches, 209 Organofunctional spherosiloxanes, 361
Molecular vibrations, 108, 109 Ovothiol, 201
Møller-Plesset Perturbation Theory, 14, 40 Oxide glasses, 429
Monopole interaction, 276
Monte Carlo (MC) methods, 249 P
Moore–Penrose inverse, 78, 84 PED coefficients, 82, 91
Mordenite, 317, 318, 325 Pentasils, 317, 318
MOR structure type, 316–318 Periodic structure model, 325
Mössbauer effect, 275, 276, 278, 290–294 Perovskite, 38, 273, 274, 284–290, 292
Multicomponent glasses, 477, 479 Perturbation approach, 56, 58
Multiferroic, 283, 284, 289, 294, 296 Phonon energy, 428–430, 444, 450, 453
Multimode waveguide materials, 402 Photoacoustic (PAS), 340
Multi-photon optics, 140 Photochemistry, 199, 201, 219
Multivariate Curve Resolution Alternating PhotoInduced Detachment Association
Least Squares (MCR-ALS), 341, 342 mechanism (PIDA), 209
Multi-Walled Carbon Nanotube (MWCNT), Photonic glasses, 427–429, 444
123–135 pH-responsive systems, 228, 236, 238, 259
Phthalocyanine, 98–107, 116, 117
N Pigments, 484, 485, 487–490, 492–496, 502
Nanoparticles, 427, 432, 434, 436–438, 440, Plasmon resonance, 434, 438, 440, 454
443, 454 Polarizable Continuum Model (PCM), 33
Nanophotonics, 434 Polyhedral Oligomeric Silsesquioxanes
Natural Bond Orbital (NBO), 206 (POSS), 362, 387–389
Natural Internal Coordinates (NICs), 53, 73, Poly(2-(2-methoxyethoxy) ethyl methacrylate)
77–79, 81, 82, 85 PMEO2MA, 240, 253
Near-infrared (NIR) spectroscopy, 402, 405, Poly(N-isopropylacrylamide)—PNIPAM, 240,
414, 416 250
Near-infrared loss spectrum, 401, 414 Polydimethylsiloxane (PDMS) polymers,
Néel temperature, 286, 288, 291, 296 401–412, 414, 416–423
Nitrogen oxide probe molecule, 338, 339 Polymer chain, 225–227, 230, 231, 235–237,
NMR spectroscopy 239, 242, 244, 245, 250, 251, 254, 255,
1
H NMR spectroscopy, 367, 374, 379 259
29
Si NMR spectroscopy, 365, 366, 370, Polymer gel, 232, 242, 252
373, 377, 378, 381, 383, 386, 388, 390, Polymer hydrogel, 243, 253
392, 394 Polymer network, 224, 243, 244, 250
Non-local xc functionals, 21 Polymer optical fibres (POF), 402
Normal modes of vibrations, 3, 5, 23, 24 Polymer-solvent interactions, 226, 228
Index 523
Polymethylmethacrylate (PMMA), 402, 404 Rod-in-tube method, 445, 447, 451

Polymorphic forms of SiO2, 467 Rozenberg and coworkers′ empirical
Polymorphism, 215, 220 correlations, 216
Polymorphs, 215, 216, 218, 220 Runge-Gross theorem, 27
Polyoxymethylene (POM), 404
Polysiloxane, 361–366, 369–372, 374–377, S
379, 380, 382, 383, 393, 394, 402, 417 Sayvetz conditions, 60, 61
Polyvinylmethylether (PVME), 239 Scaffold, 142–144
Pore opening (PO) vibrations, 307 Scaled Quantum Mechanical (Force Field)
Porphyrin, 97–102, 106, 107, 116 method (SQM (FF)), 74
Potential energy, 49, 52, 53, 60, 62, 81, 82 Scaling (scale) Factors (SFs), 49, 64, 65, 68,
Potential Energy Curve (PEC), 52–54, 56, 58 69–79, 81, 84–87, 89, 90
Potential Energy Distribution (PED), 82 Scaling procedures, 49, 52, 61, 63–65, 84, 87,
Potential Energy Hypersurface, 8 89, 90
Potential Energy Surface (PES), 53, 54, 63 Schrödinger equation, 6–10, 12, 16, 17, 27
Primary bound water, 231, 233 Scanning Electron Microscopy (SEM), 432,
Primary Building Unit (PBU), 302 442
Primitive internal coordinates, 50 Secondary bound water, 231–233
Probe molecules, 335–339, 341, 347, 353 Secondary Building Unit (SBU), 302–304,
Proteins, 140, 156 307–309, 311, 312, 315, 318–320, 324,
Provenance studies, 497 325, 328
Pseudolattice vibrations, 314 Second Harmonic Generation (SHG), 140, 151,
Push-pull dyes, 141 156
Pyridine probe molecule, 333, 336–338, 340, Self Consistent Field method (SCF), 12
353 Sellmeier formula, 406
Sextet, 278, 279, 290, 291, 293, 294
Q Short-range order, 457, 458, 460, 463, 464
Quadrupole moment, 279, 283 Side group modified polysiloxanes, 379
Quadrupole splitting, 276, 282, 283, 290, 291 Silicate framework, 320
Quantum chemistry methods, 142 Silicate glass, 458–462, 464, 466–468,
Quantum dot, 97, 98, 103, 105 470–473, 476
Quartz, 470, 471 Silver doping, 156
Quasi-molecular elementary unit, 469 Silver nanoparticles, 427, 436, 438, 443
Quinone, 102 Single cell analysis, 170
Singlet, 277
R Single-Walled Carbon Nanotubes (SWCNT),
Raman imaging, 161, 172, 181, 182, 184, 187 123, 124, 126, 127, 131, 132, 134, 135
Raman Optical Activity (ROA), 161, 162, 165 Slater determinant, 10, 11, 13, 26
Raman spectroscopy, 124, 134, 199, 201, 216, Sodalite cage, 302
219, 334, 335, 345–348, 350, 483–485, Solid-solid transition, 215
487–492, 498, 500, 508 Sorption, 312, 320, 325, 327, 328
Rapid-scan, 344 Spherosilicates, 361–365, 386, 390, 392, 394
Rare-earth ions, 434, 444 Step-scan, 344
Recoilless emission, 274 Stimuli-Responsive Polymer System (SRPS),
Recoil process, 274 223
Redox sites, 336, 337 Stimuli sensitive, 224, 225, 250, 257, 259
Reflectance spectroscopy, 343, 345 Stokes effect, 346
Refractive index, 403, 405–407, 409, 418, 419 Structural properties, 435
Reorganization energy, 97, 99, 107, 110, 113, Superexchange, 286, 288–290, 296
114, 116, 117 Surface-Enhanced Raman Scattering (SERS),
Resonance Raman effect, 123, 125, 132 435
Resonant absorption, 274, 288 Surface-Enhanced Raman Spectroscopy
Ring opening (RO) vibrations, 307 (SERS), 161, 166, 169
Ring vibrations, 475, 477 Susceptibility, 140
524 Index
T Upconversion process, 427, 443, 450, 452

Tautomer, 106, 107, 202, 203, 205, 208, 209, Upper Critical Solution Temperature (UCST),
219 226
Tautomerism, 205, 219
Taylor expansion, 54 V
Tectosilicates, 301, 312 Variation approach, 56
Thermal contribution to enthalpy and entropy, Vibrational spectroscopy, 199, 200, 219, 302,
66 303, 312, 457, 460, 464
Thermal curing, 402–404, 418 Virtual Crystal Aproximation (VCA), 37
Thermal stability, 428–430, 440, 442, 444, 453 Visible light spectroscopy, 484
Thermochemical reduction, 430, 436, 442 Visible range, 428, 436, 443, 444, 446, 449,
Thermo-optic coefficients (TOC), 407 450
Thermo-responsive polymer, 229, 235, 239, Vitreous SiO2, 458–463, 467–469, 471–473,
250, 259 476, 479
Thermo-responsive system, 226, 251, 256 Volume Phase Transition (VPT), 239
Thiocyanic acid, 211
Thiols, 144 W
Third Harmonic Generation (THG), 140 Waveguide, 401–409, 412, 413, 419–423
Thiyl radical, 209, 211 Wavenumber linear scaling (WLS), 49, 52, 72
Time-Dependent Density Functional Theory Wilson-Decius-Cross method, 49, 52, 59
(TD-DFT), 1, 26, 27, 29, 31, 40, 97,
116, 139, 145, 209 X
Time Resolved Spectroscopy, 259 X-ray Absorption Spectroscopy, 334
Tip-enhanced Raman Spectroscopy, 350 X-ray diffraction (XRD), 200
Topological analysis of electron density, 320 X-ray Photoelectron Spectroscopy (XPS), 145,
TPA cross section, 147–149 378
TPE cross sections, 151
Transformation temperature, 436, 437, 441 Y
Transition dipole moments, 147, 148, 150 Young’s Modulus, 123, 125, 126, 134, 135
Transition energies, 147
Transmission Electron Microscopy (TEM), 436 Z
Transmission spectroscopy, 341 Zachariasen theory, 458
Tridymite, 461 Zeeman effect, 276, 278
Two-dimensional Correlation Spectroscopy Zeeman splitting, 278
(2D CoS), 345 Zeolite A, 303, 314, 323, 328
Two-Photon Excitation Fluorescence (TPEF), Zeolite framework, 303, 304, 308, 319, 320,
140, 141, 151–153, 155, 156 324
Zeolite Y, 303, 317, 325
U Zero-point vibrational energy correction
Ultra-High Vacuum (UHV), 335, 341 (ZPVE), 66, 67, 69–72
Ultraviolet-Visible Spectroscopy (UV-Vis), 98, Zero-point vibrational level, 203
103, 104, 107, 334
Uniform scaling (US), 49, 52, 65

(Challenges and Advances in Computational Chemistry and Physics 26) Andrzej Koleżyński, Magdalena Król - Molecular Spectroscopy—Experiment and Theory_ From Molecules to Functional Materials-Springer I.pdf

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

(Challenges and Advances in Computational Chemistry and Physics 26) Andrzej Koleżyński, Magdalena Król - Molecular Spectroscopy—Experiment and Theory_ From Molecules to Functional Materials-Springer I.pdf

Hochgeladen von

Copyright:

Verfügbare Formate

Challenges and Advances

in Computational Chemistry and Physics 26

More information about this series at http://www.springer.com/series/6918

ISSN 2542-4491 ISSN 2542-4483 (electronic)

Library of Congress Control Number: 2018955928

© Springer Nature Switzerland AG 2019

foundations and description of fundamentals of computational methods commonly

Kraków, Poland Andrzej Koleżyński

1 Computational Methods in Spectroscopy . . . . . . . . . . . . . . . . . . . . 1

10 Vibrational Spectroscopy of Zeolites . . . . . . . . . . . . . . . . . . . . . . . . 301

Rodolphe Antoine Institut Lumière Matière, UMR5306 Université Claude

Dominik Dorosz Faculty of Materials Science and Ceramics, AGH University of

Kamilla Malek Faculty of Chemistry, Jagiellonian University, Krakow, Poland;

Abstract Spectroscopy investigates the interaction of electromagnetic radiation

A broadly understood spectroscopy is devoted to the studies of the interactions of

© Springer Nature Switzerland AG 2019 1

the form of electromagnetic radiation. The spectrum of electromagnetic radiation

1.2 Theoretical Foundations for Modeling of Real Systems

The physical reality that surrounds us is extremely complex and impossible to be

1.2.1 Ab Initio Methods

1.2.1.1 Schrödinger Equation

possible results (measurement problem, collapse of the wave function), which

and the time-dependent Schrödinger equation takes its ﬁnal form:

Wðr; tÞ ¼ wðrÞTðtÞ ð1:4Þ

The Schrödinger equation then takes the form:

which after the transformation gives the equation:

and second on time:

Wðx; tÞ ¼ wðxÞTðtÞ ¼ wðxÞeiEt=h ð1:9Þ

where ri and Ra—the coordinates of electrons and atomic nuclei, respectively, n—

1.2.1.2 Born–Oppenheimer Approximation, Potential Energy

The ﬁrst simpliﬁcation, bringing us closer to the solution of the time-independent

The Born–Oppenheimer approximation leads to a very important concept—the

1.2.1.3 Hartree–Fock Method and Post-HF Extensions

The solution of the nuclear time-independent Schrödinger equation allows deter-

wel ðr1 ; r2 ; . . .; rN Þ ¼ /1 ðr1 Þ/2 ðr2 Þ    /N ðrN Þ ð1:18Þ

^ el of the form (1.14)

• Two-electron contribution (depending on the coordinates of two electrons):

The former can be calculated accurately, whereas the two-electron contributions

where Jij is a Coulomb integral deﬁned as:

describing Coulomb repulsion between electron 1 on orbital i and electron 2 on

with no classical interpretation—this integral is entirely a quantum mechanical

In order to be able to solve this Schrödinger equation, we must replace the

and as a result, the so-called one-electron Fock operator is obtained:

Using variational principle:

to the problem of energy minimization, we get a set of appropriate one-electron

^fi /ðiÞ ¼ ei /ðiÞ; i ¼ 1; 2; . . .; N ð1:31Þ

One-electron Fock operator ^fi depends, through effective potential V

• Next, the system of Hartree–Fock equations is solved;

1.2.1.4 Density Functional Theory

HF method with post-HF extensions is still commonly used (especially by chemists

Hence, the system Hamiltonian is deﬁned as the sum of two operators:

According to the ﬁrst Hohenberg–Kohn theorem, the ground state properties of

being a solution to the following Schrödinger equations:

and satisfying the orthonormality condition:

The second is Hartree energy, classic electrostatic energy:

energy EXC[n(r)], containing non-classical electrostatic interactions and the differ-

where k is undetermined Lagrange multiplier associated with constraints on the

In Kohn–Sham’s formulation, respective Euler’s equation transforms into:

VKS ðrÞ ¼ Vext ðrÞ þ VH ðrÞ þ Vxc ðrÞ ð1:46Þ

where eXC ðr Þ is exchange–correlation energy density. Based on the way in which

where e0XC ½nðr Þ is the exchange–correlation energy density of a homogeneous

e0XC ½nðrÞ ¼ e0X ½nðrÞ þ e0C ½nðrÞ ð1:51Þ

Often, in order to increase the flexibility of the approximate functional, the

nðrÞ ¼ na ðrÞ þ nb ðrÞ ð1:55Þ

wel ðr1 ; r2 ; . . .; rN Þ ¼ /1 ðr1 Þ/2 ðr2 Þ /N ðrN Þ ð1:18Þ

where M is a diagonal matrix with atomic masses on the main diagonal