Beruflich Dokumente
Kultur Dokumente
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.
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
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
vii
viii Contents
ix
x Contributors
Andrzej Koleżyński
1.1 Introduction
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
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);
1 Computational Methods in Spectroscopy 5
• 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.
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].
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
1 Computational Methods in Spectroscopy 7
@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
@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:
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
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:
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
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.
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
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).
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 Computational Methods in Spectroscopy 11
/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
P
n P
m
^hi where ^hi ¼ T^e þ V
^Ne ¼ h2 r2i Za e 2
ð1:21Þ
2me ria
i¼1 a¼1
X X e2
^ ee ¼
V ð1:22Þ
j
r
i [ j ij
X
n
^eff 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
X
^fi ¼ h r2
2 m
Za e2 ^
þ V eff ðiÞ ð1:29Þ
2me i a¼1
ria
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).
X Za e2
Vext ðrÞ ¼ ð1:32Þ
a
j r ra j
1 Computational Methods in Spectroscopy 15
And
X
^ext 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
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
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
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
1 Computational Methods in Spectroscopy 17
dF½nðrÞ
l¼ þ Vext ðrÞ ð1:44Þ
dnðrÞ
dTS ½nðrÞ
l¼ þ VKS ðrÞ ð1:45Þ
dnðrÞ
where
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
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
1 Computational Methods in Spectroscopy 19
where
sffiffiffiffiffiffiffiffiffiffiffiffiffiffi
3
rs ¼ ð1:54Þ
3
4pnðr Þ
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:
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):
1 Computational Methods in Spectroscopy 21
B3LYP
EXC ¼ a0 EXHF þ ð1 a0 ÞEXLDA þ ax EXBecke þ ð1 ac ÞECVVN þ ac ECLYP ð1:64Þ
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
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
1 Computational Methods in Spectroscopy 23
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
@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
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
ki ¼ 2pm2i ð1:78Þ
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
1 Computational Methods in Spectroscopy 27
^ 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:
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
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
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]:
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
0 dq½Vext ðr; tÞ
0
vðr; r ; t; t Þ ¼ ð1:90Þ
dVext ðr0 ; t0 Þ V0
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
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
1 Computational Methods in Spectroscopy 31
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.
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.
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
1 Computational Methods in Spectroscopy 35
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)
Fig. 1.3 Structure of zeolite A in a exemplary hydrated [125] and b dehydrated form [126]
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
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].
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, …);
1 Computational Methods in Spectroscopy 41
1.4 Conclusions
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
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
1 Computational Methods in Spectroscopy 45
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
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
2.1 Introduction
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
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
54 O. Bąk and P. Borowski
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
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
2 Scaling Procedures in Vibrational Spectroscopy 55
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
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
56 O. Bąk and P. Borowski
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
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
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 Scaling Procedures in Vibrational Spectroscopy 59
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
60 O. Bąk and P. Borowski
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
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
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.,
2 Scaling Procedures in Vibrational Spectroscopy 61
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
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.
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
1
U ¼ sT Fs: ð2:28Þ
2
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.
X
Nvib 2
F ðf Þ ¼ wp mscl
p ð f Þ m expt
p ð2:29Þ
p¼1
@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
2 Scaling Procedures in Vibrational Spectroscopy 65
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.
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
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
N mol
1X vib
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
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 .
68 O. Bąk and P. Borowski
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/
2 Scaling Procedures in Vibrational Spectroscopy 69
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.
70 O. Bąk and P. Borowski
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
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
72 O. Bąk and P. Borowski
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.
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
2 Scaling Procedures in Vibrational Spectroscopy 73
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
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
74 O. Bąk and P. Borowski
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.
2 Scaling Procedures in Vibrational Spectroscopy 75
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
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.
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
76 O. Bąk and P. Borowski
3
As mentioned in the introductory section in the following we will not cite papers, where only
applications are presented.
2 Scaling Procedures in Vibrational Spectroscopy 77
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
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
2 Scaling Procedures in Vibrational Spectroscopy 79
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
80 O. Bąk and P. Borowski
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
2 Scaling Procedures in Vibrational Spectroscopy 81
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.
1X K
U ðsÞ ¼ Flm sl sm : ð2:51Þ
2 l;m¼1
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;
2 Scaling Procedures in Vibrational Spectroscopy 83
• 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
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
84 O. Bąk and P. Borowski
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
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
2 Scaling Procedures in Vibrational Spectroscopy 85
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,
2 Scaling Procedures in Vibrational Spectroscopy 87
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.
88 O. Bąk and P. Borowski
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
2 Scaling Procedures in Vibrational Spectroscopy 89
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].
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.
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.
90 O. Bąk and P. Borowski
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.
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.
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,
2 Scaling Procedures in Vibrational Spectroscopy 91
References
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
94 O. Bąk and P. Borowski
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
59. Blom CE, Altona C (1977) Application of self-consistent-field ab-initio calculations to
organic-molecules. IV. Force constants of propene scaled on experimental frequencies. Mol
Phys 33:875–885
60. Blom CE, Altona C (1977) Application of self-consistent-field ab-initio calculations to
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
2 Scaling Procedures in Vibrational Spectroscopy 95
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
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
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
100 D. Wróbel and B. Barszcz
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.
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
102 D. Wróbel and B. Barszcz
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
104 D. Wróbel and B. Barszcz
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)
porphyrin-gold dyads was also studied with the use of advanced methods, and it
showed dynamics of the processes [41–45].
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.
3 Quantum Dot and Fullerene with Organic Chromophores as … 107
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
110 D. Wróbel and B. Barszcz
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
112 D. Wróbel and B. Barszcz
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:
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
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)
3 Quantum Dot and Fullerene with Organic Chromophores as … 115
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)
116 D. Wróbel and B. Barszcz
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
3 Quantum Dot and Fullerene with Organic Chromophores as … 117
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.
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
118 D. Wróbel and B. Barszcz
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
120 D. Wróbel and B. Barszcz
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
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
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
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
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.
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
128 A. K. Ojha and H. M. Heise
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
4 Material Analysis Using Raman Spectroscopy 129
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
130 A. K. Ojha and H. M. Heise
Fig. 4.3 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 785 nm
excitation wavelength
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
4 Material Analysis Using Raman Spectroscopy 131
Table 4.2 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 785 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
with a Lorentzian line shape function
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)
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.
132 A. K. Ojha and H. M. Heise
Table 4.3 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 1064 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
with a Lorentzian line shape function
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)
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
4 Material Analysis Using Raman Spectroscopy 133
Fig. 4.4 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 1064 nm
excitation wavelength
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.
134 A. K. Ojha and H. M. Heise
4.5 Conclusions
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
136 A. K. Ojha and H. M. Heise
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
4 Material Analysis Using Raman Spectroscopy 137
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
5.1 Introduction
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
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.
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
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
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]:
5 Ligand-Core NLO-Phores 145
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
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
5 Ligand-Core NLO-Phores 149
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
^a jkihkjl ^a jf i
^b jkihkjl
Tab ¼ þ ð5:5Þ
k
xk xf =2 xk xf =2
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.
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:
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
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
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)
5 Ligand-Core NLO-Phores 155
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.
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
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
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
5 Ligand-Core NLO-Phores 159
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
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.
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
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
166 K. Czamara et al.
Fig. 6.2 Raman and AIRROA spectra of H-aggregates formed by astaxanthin enantiomers.
Adapted with permission from [44], copyright (2016) American Chemical Society
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
6 Small and Large Molecules Investigated by Raman Spectroscopy 167
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)
168 K. Czamara et al.
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
6 Small and Large Molecules Investigated by Raman Spectroscopy 169
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
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.
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
6 Small and Large Molecules Investigated by Raman Spectroscopy 171
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.
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].
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
174 K. Czamara et al.
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
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].
176 K. Czamara et al.
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
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
178 K. Czamara et al.
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.
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].
6 Small and Large Molecules Investigated by Raman Spectroscopy 181
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
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
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.
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
6 Small and Large Molecules Investigated by Raman Spectroscopy 185
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
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].
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)
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
6 Small and Large Molecules Investigated by Raman Spectroscopy 189
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
190 K. Czamara et al.
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
6 Small and Large Molecules Investigated by Raman Spectroscopy 191
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.
J Raman Spectrosc 47:1310–1317
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
192 K. Czamara et al.
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
6 Small and Large Molecules Investigated by Raman Spectroscopy 193
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
6 Small and Large Molecules Investigated by Raman Spectroscopy 195
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
196 K. Czamara et al.
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
Spectrosc 64:255–261
6 Small and Large Molecules Investigated by Raman Spectroscopy 197
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.
Phys Chem Chem Phys 11:7499–7504
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
198 K. Czamara et al.
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
7.1 Introduction
7.2.1 Mercaptoimidazoles
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.
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
204 R. Fausto et al.
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
7 Hydantoins and Mercaptoimidazoles: Vibrational … 205
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:
• 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
7 Hydantoins and Mercaptoimidazoles: Vibrational … 207
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
208 R. Fausto et al.
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].
7.3.1 Mercaptoimidazoles
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
210 R. Fausto et al.
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
7 Hydantoins and Mercaptoimidazoles: Vibrational … 211
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
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
7 Hydantoins and Mercaptoimidazoles: Vibrational … 215
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.
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
216 R. Fausto et al.
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
7 Hydantoins and Mercaptoimidazoles: Vibrational … 217
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
218 R. Fausto et al.
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).
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.
7 Hydantoins and Mercaptoimidazoles: Vibrational … 219
7.5 Conclusion
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 Hydantoins and Mercaptoimidazoles: Vibrational … 221
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
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.
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.
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
F Up
¼ ln Up þ Us ln Us þ vUs Up ð8:1Þ
kB T M
Z
v¼ ð8:2Þ
ð2kB T Þ 2esp ess epp
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
228 M. Kozanecki et al.
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
8 Vibrational Spectroscopy in Analysis of Stimuli-Responsive … 229
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
8 Vibrational Spectroscopy in Analysis of Stimuli-Responsive … 231
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
232 M. Kozanecki et al.
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
[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
8 Vibrational Spectroscopy in Analysis of Stimuli-Responsive … 235
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.
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
8 Vibrational Spectroscopy in Analysis of Stimuli-Responsive … 237
Also cations (having the same anion) could be ordered in similar series:
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
238 M. Kozanecki et al.
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 Vibrational Spectroscopy in Analysis of Stimuli-Responsive … 239
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
8 Vibrational Spectroscopy in Analysis of Stimuli-Responsive … 241
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].
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
244 M. Kozanecki et al.
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
8 Vibrational Spectroscopy in Analysis of Stimuli-Responsive … 245
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.
Fig. 8.12 A range of applicability of various computing methods compared with typical length
and timescales in polymer materials
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
250 M. Kozanecki et al.
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
8 Vibrational Spectroscopy in Analysis of Stimuli-Responsive … 251
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
252 M. Kozanecki et al.
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
8 Vibrational Spectroscopy in Analysis of Stimuli-Responsive … 253
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
254 M. Kozanecki et al.
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].
8 Vibrational Spectroscopy in Analysis of Stimuli-Responsive … 255
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
8 Vibrational Spectroscopy in Analysis of Stimuli-Responsive … 257
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
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
260 M. Kozanecki et al.
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.
8 Vibrational Spectroscopy in Analysis of Stimuli-Responsive … 261
References
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
8 Vibrational Spectroscopy in Analysis of Stimuli-Responsive … 263
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
264 M. Kozanecki et al.
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.
J Chem Phys 123:114502
94. Rezus YLA, Bakker HJ (2006) Orientational dynamics of isotopically diluted H2O and D2O.
J Chem Phys 125:144512
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
8 Vibrational Spectroscopy in Analysis of Stimuli-Responsive … 265
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).
Macromolecules 30:1111–1117
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
266 M. Kozanecki et al.
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
8 Vibrational Spectroscopy in Analysis of Stimuli-Responsive … 267
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
268 M. Kozanecki et al.
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
8 Vibrational Spectroscopy in Analysis of Stimuli-Responsive … 269
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
270 M. Kozanecki et al.
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
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
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].
276 P. Stoch and A. Stoch
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.
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,
9 Mössbauer Spectroscopy of Magnetoelectric Perovskite Oxides 277
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.
278 P. Stoch and A. Stoch
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].
9 Mössbauer Spectroscopy of Magnetoelectric Perovskite Oxides 279
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].
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]:
280 P. Stoch and A. Stoch
1=2
1 1
EQ ¼ eQVzz 1 þ g2 ð9:5Þ
4 3
Vxx Vyy
g¼ ð9:6Þ
Vzz
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
9 Mössbauer Spectroscopy of Magnetoelectric Perovskite Oxides 281
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
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
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
9 Mössbauer Spectroscopy of Magnetoelectric Perovskite Oxides 283
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 is rather low because these two orders originate from different
kinds of ions [27, 29].
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
286 P. Stoch and A. Stoch
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,
9 Mössbauer Spectroscopy of Magnetoelectric Perovskite Oxides 287
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]
288 P. Stoch and A. Stoch
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
9 Mössbauer Spectroscopy of Magnetoelectric Perovskite Oxides 289
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
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
9 Mössbauer Spectroscopy of Magnetoelectric Perovskite Oxides 291
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
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
57
Fig. 9.6 a An exemplary high-temperature Fe Mössbauer effect spectra; b dependence of
Mössbauer line width on temperature
9 Mössbauer Spectroscopy of Magnetoelectric Perovskite Oxides 295
9.4 Conclusions
References
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
9 Mössbauer Spectroscopy of Magnetoelectric Perovskite Oxides 299
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
300 P. Stoch and A. Stoch
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
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
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).
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
10 Vibrational Spectroscopy of Zeolites … 305
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.
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
10 Vibrational Spectroscopy of Zeolites … 309
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.
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
10 Vibrational Spectroscopy of Zeolites … 311
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.
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.
10 Vibrational Spectroscopy of Zeolites … 313
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.
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.
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. 10.10 Experimental and theoretical IR spectra of LTA, FAU, CHA, MOR, FER, and
HEU-type structures
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.
10 Vibrational Spectroscopy of Zeolites … 319
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
10 Vibrational Spectroscopy of Zeolites … 321
Fig. 10.11 IR spectra of LTA model structure calculated for various extra-framework cations [63]
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.
10 Vibrational Spectroscopy of Zeolites … 323
Fig. 10.12 Extreme atomic displacements during restrained tetrahedra rotations (a); cation–
oxygen vibrations (b); and vibrations of the cationic sublattice (c)
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).
10 Vibrational Spectroscopy of Zeolites … 325
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.
10.8 Conclusions
Acknowledgements This work was partially supported by the National Science Centre, Poland,
under grant No. 2015/17/B/ST8/01200.
10 Vibrational Spectroscopy of Zeolites … 329
References
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.
Spectrochim Acta A 195:62–67
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
10 Vibrational Spectroscopy of Zeolites … 331
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
Spectrosc 9:225–228
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.
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
11.1 Introduction
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.
336 P. Jodłowski and J. Łojewska
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].
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
338 P. Jodłowski and J. Łojewska
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).
11 In Situ and Operando Techniques in Catalyst Characterisation … 339
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
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
342 P. Jodłowski and J. Łojewska
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
11 In Situ and Operando Techniques in Catalyst Characterisation … 343
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],
copyright (2018) Elsevier
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.
344 P. Jodłowski and J. Łojewska
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.
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,
348 P. Jodłowski and J. Łojewska
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
350 P. Jodłowski and J. Łojewska
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.
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.
11 In Situ and Operando Techniques in Catalyst Characterisation … 351
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
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
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
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.
Phys Chem Chem Phys 4:2667–2672
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
11 In Situ and Operando Techniques in Catalyst Characterisation … 357
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
Spectrosc 47:1376–1381
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
11 In Situ and Operando Techniques in Catalyst Characterisation … 359
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-
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
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].
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]).
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.
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
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
(continued)
12 Application of Spectroscopic Methods in the Studies … 369
S28 S29
S33 S34
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
370 A. Chechelska-Noworyta et al.
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
12 Application of Spectroscopic Methods in the Studies … 373
(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
12 Application of Spectroscopic Methods in the Studies … 375
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
376 A. Chechelska-Noworyta et al.
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].
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
380 A. Chechelska-Noworyta et al.
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
382 A. Chechelska-Noworyta et al.
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
12 Application of Spectroscopic Methods in the Studies … 383
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.
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)
12 Application of Spectroscopic Methods in the Studies … 385
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
S22 S23
(continued)
386 A. Chechelska-Noworyta et al.
S24 S25
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
390 A. Chechelska-Noworyta et al.
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].
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.
394 A. Chechelska-Noworyta et al.
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
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
398 A. Chechelska-Noworyta et al.
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
12 Application of Spectroscopic Methods in the Studies … 399
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
400 A. Chechelska-Noworyta et al.
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.
Macromolecules 45:9020–9028
Chapter 13
Spectroscopic Aspects
of Polydimethylsiloxane (PDMS)
Used for Optical Waveguides
13.1 Introduction
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
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
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.
For waveguide fabrication using the PDMS materials, casting is adopted as the
suitable technology [5, 10, 11]. The first basic prerequisite for the replication
404 D. Cai and H. M. Heise
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
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.
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
13 Spectroscopic Aspects of Polydimethylsiloxane (PDMS) Used for … 407
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
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)
408 D. Cai and H. M. Heise
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
13 Spectroscopic Aspects of Polydimethylsiloxane (PDMS) Used for … 409
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
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.
410 D. Cai and H. M. Heise
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
13 Spectroscopic Aspects of Polydimethylsiloxane (PDMS) Used for … 411
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.
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
(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
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
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
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.
13 Spectroscopic Aspects of Polydimethylsiloxane (PDMS) Used for … 415
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
416 D. Cai and H. M. Heise
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.
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
13 Spectroscopic Aspects of Polydimethylsiloxane (PDMS) Used for … 417
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
418 D. Cai and H. M. Heise
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 Spectroscopic Aspects of Polydimethylsiloxane (PDMS) Used for … 419
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
420 D. Cai and H. M. Heise
Fig. 13.9 Relationship between sidewall roughness and optical waveguide loss at 850 nm.
Adapted with permission from [21], copyright (2010) IEEE
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
13 Spectroscopic Aspects of Polydimethylsiloxane (PDMS) Used for … 421
Fig. 13.10 Doctor blading by hand (a) and by controlled machine (b). Reprinted with permission
from [21], copyright (2010) IEEE
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.
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
13 Spectroscopic Aspects of Polydimethylsiloxane (PDMS) Used for … 423
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.
424 D. Cai and H. M. Heise
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.
Spectrochim Acta A 16:87–105
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
13 Spectroscopic Aspects of Polydimethylsiloxane (PDMS) Used for … 425
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
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
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:
430 J. Zmojda et al.
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
432 J. Zmojda et al.
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
14 The Luminescent Properties of Photonic Glasses … 433
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.
434 J. Zmojda et al.
Fig. 14.5 SEM images of SGS glasses morphology with visible (a) nano- and (b) micro-crystals
of EuPO4
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),
436 J. Zmojda et al.
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
14 The Luminescent Properties of Photonic Glasses … 437
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
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
14 The Luminescent Properties of Photonic Glasses … 439
Fig. 14.12 Luminescence spectra of antimony-germanate glasses doped with Eu3+ ions (a) and
modified with high-intensity glass-forming elements
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
14 The Luminescent Properties of Photonic Glasses … 441
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
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
14 The Luminescent Properties of Photonic Glasses … 443
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
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
444 J. Zmojda et al.
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)
446 J. Zmojda et al.
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)
14 The Luminescent Properties of Photonic Glasses … 447
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
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.
450 J. Zmojda et al.
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].
14 The Luminescent Properties of Photonic Glasses … 451
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
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)
14.6 Summary
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
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
456 J. Zmojda et al.
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.
Fig. 15.1 Schematic representation of m-SiO2 structure according to crystallite theory (a), and
continuous random network theory (b)
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
15 Spectroscopic Characterization of Silicate Amorphous Materials 461
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
462 W. Mozgawa et al.
Fig. 15.3 Ways to connect ordered regions: in one type of arrangement sequence (a), and in
different arrangement sequences (b)
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
464 W. Mozgawa et al.
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.
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:
466 W. Mozgawa et al.
exp½2pðki c0 Þx
F 0 ð xÞ ¼ F ð xÞ expð2pc0 xÞ ¼ 0:5A0i ci cos 2pm0 x ð15:2Þ
Rci
X
n
v2 ¼ ½wr f ðxr =qÞ ð15:3Þ
r¼1
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
468 W. Mozgawa et al.
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:
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
470 W. Mozgawa et al.
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
15 Spectroscopic Characterization of Silicate Amorphous Materials 473
Fig. 15.9 MIR spectra of glasses from the K2O–SiO2 system. Adapted with permission from [19],
copyright (1993) Elsevier
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
474 W. Mozgawa et al.
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
15 Spectroscopic Characterization of Silicate Amorphous Materials 475
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:
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
15 Spectroscopic Characterization of Silicate Amorphous Materials 477
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.
478 W. Mozgawa et al.
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
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
480 W. Mozgawa et al.
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
15 Spectroscopic Characterization of Silicate Amorphous Materials 481
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
44. Dowty E (1987) Vibrational interactions of tetrahedra in silicate glasses and crystals. Phys
Chem Miner 14:122–138
45. Dowty E (1987) Vibrational interactions of tetrahedra in silicate glasses and crystals. Phys
Chem Miner 14:542–552
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
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 (&)
Faculty of Materials Science and Ceramics, AGH University of Science
and Technology, 30 Mickiewicza Av., 30-059 Krakow, Poland
e-mail: lojewski@agh.edu.pl
B. Łydżba-Kopczyńska
Faculty of Chemistry, University of Wroclaw, F. Joliot-Curie 14,
50-383 Wroclaw, Poland
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
486 T. Łojewski and B. Łydżba-Kopczyńska
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
16 Spectroscopy in the Analysis of Artworks 487
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
488 T. Łojewski and B. Łydżba-Kopczyńska
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
492 T. Łojewski and B. Łydżba-Kopczyńska
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.
494 T. Łojewski and B. Łydżba-Kopczyńska
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
16 Spectroscopy in the Analysis of Artworks 495
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
496 T. Łojewski and B. Łydżba-Kopczyńska
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)
16 Spectroscopy in the Analysis of Artworks 497
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
498 T. Łojewski and B. Łydżba-Kopczyńska
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)
16 Spectroscopy in the Analysis of Artworks 499
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
500 T. Łojewski and B. Łydżba-Kopczyńska
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
16 Spectroscopy in the Analysis of Artworks 501
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
502 T. Łojewski and B. Łydżba-Kopczyńska
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.
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.
16 Spectroscopy in the Analysis of Artworks 503
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
504 T. Łojewski and B. Łydżba-Kopczyńska
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
506 T. Łojewski and B. Łydżba-Kopczyńska
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
16 Spectroscopy in the Analysis of Artworks 507
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.
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.,
508 T. Łojewski and B. Łydżba-Kopczyńska
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
16 Spectroscopy in the Analysis of Artworks 509
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.
J Raman Spectrosc 49:921–934
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
510 T. Łojewski and B. Łydżba-Kopczyńska
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
512 T. Łojewski and B. Łydżba-Kopczyńska
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.
J Raman Spectrosc 47:67–77
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
Spectrosc 43:1761–1771
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.
J Raman Spectrosc 45:1197–1206
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
16 Spectroscopy in the Analysis of Artworks 515
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
Spectrosc 49:842–851
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.
J Raman Spectrosc 45:1197–1206
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
516 T. Łojewski and B. Łydżba-Kopczyńska
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