Modern Polymer Spect..

Modern Polymer Spectroscopy 

Edited by Giuseppe Zerbi 

@ WILEY-VCH

This Page Intentionally Left Blank

Modern 

Polymer Spectroscopy 

Edited by 

Giuseppe Zerbi 

CB WILEY-VCH 

Weinheim * New York - Chichester 

Brisbane * Singapore - Toronto

Prof. G. Zerbi 

Dipartimento di Ingegneria Chimica e Chimica Industriale 

Politecnico di Milano 

Piazza Leoiiardo da Vinci 32 

20132 Milano 

Italy 

This book was carefully produced. Nevertheless, authors, editor and publisher do not warrant 

the information contained therein to be free of errors. Readers are advised to keep in mind 

that statements, data, illustrations, procedural details or other items may inadvertently be 

inaccurate. 

Cover illustration: Dr. Erica Mannucci 

Library of Congress Card No. applied for 

A catalogue record for this book is available from the British Library 

Deutsche Bibliothek Cataloguing-in-Publication Data 

Modern polymer spectroscopy / ed. by Giuseppe Zerbi. - 1. AuA. - Weinheim; 

New York; Chichester; Brisbane; Singapore; Toronto: Wiley-VCH, 1999 

ISBN 3-527-29655-7 

8 WILEY-VCH Verlag GmbH. D-69469 Weinheim (Federal Republic of Germany), 1999 

Printed on acid-free and chlorine-free paper. 

All rights reserved (including those of translation in other languages). No part of this book may 

be reproduced in any fomi - by photoprinting, microfilm, or any other means - nor transmitted 

or translated into machine language without written permission from the publishers. Registered 

names, trademarks, etc. used in this book, even when not specifically marked as such. are not to be 

considered unprotected by law. 

Composition: Asco Typesetters, Hong Kong. 

Printing: Strauss Offsetdruck Gin bH., D-69509 Morlenbach. 

Bookbinding: Wilhelin Osswald 8 Co. D-67433 Neustaclt. 

Printed in the Federal Republic of Germany.

Preface 

For unfortunate reasons the success of vibrational [infrared and Raman spectroscopy) 

in industrial and university laboratories seems to fade quickly in favor of 

other physical techniques which aim at chemical or structural diagnosis of unknown 

samples. On the other hand, engineers and instrument manufacturers in the field 

of vibrational spectroscopy keep producing magnificent and very sophisticated 

new instruments and accessories which enable the recording of vibrational spectra 

of samples under the most awkward experimental conditioiis which can never be 

attained by the other very popular new physical techniques. 

At present, infrared spectra can be obtained with fast and very fast FTIR interferometers 

with microscopes, in reflection and microreflection, in diffusion, at very 

low or very high temperature, in dilute solutions, etc. Ranian scattering can span a 

very large energy range in the excitation lines, thus reaching off-resonance or resonance 

conditions; spectrometers and interferometers can be used, microsampling is 

common, a few scattered photons can be detected with very sensitive CCD detectors 

and optical fibers provide a variety of new sampling procedures. 

Parallel to the technological development we watch the invasion (even on a 

coinmercial level) of theoretical and computational techniques (both ah initio or 

semiempirical) which, by simply pushing a button, provide vibrational frequencies 

and intensities and even show the animated wiggling of molecules on the screen of 

any personal computer. 

In spite of the wealth of experimental and theoretical data which are easily 

available for the study of molecules and of their behavior, other fields of physics 

and chemistry seem to have the priority in the teaching at Universities. Vibrational 

spectroscopy is, in general, no longer taught in detail and, at most, students are 

quickly exposed to molecular vibrations in some courses of analytical chemistry 

or structural organic chemistry. It is curious to notice that molecular dynamics is 

considered a very complicated mathematical machinery which must be avoided 

by simply mentioning briefly that inolecules ‘wiggle’ in a curious way. Then, the 

traditional old-fashioned structure-group frequency spectral correlations (mostly in 

infrared) are considered the only useful tool for the interpretation of spectra. This 

very limited spectroscopic culture is obviously transferred to industrial laboratories 

who rely on the use of vibrational spectra only for very simple chemical diagnosis 

and routine analytical determinations. 

The conclusion of this analysis is that the ratio: (number of inforniation)/(capa-

vi 

Prefucr 

bility of the experimental and theoretical techniques) turns out to be very small, in 

spite of the great potential offered by vibrational spectroscopy. 

The above analysis, shared by many spectroscopists in the field of small molecules, 

can be further expanded when vibrational spectroscopy is considered in the field of 

polymers and macromolecules in general. The wiggling of polymers adds new flavor 

to physics and chemistry. The translational periodicity of infinite polymers with 

perfect structure generates phonons and collective vibrations which give rise to 

absorption or Raman scattering bands that escape the interpretation based on the 

traditional spectroscopic correlations. The concept of collective motions forins the 

basis for the understanding of the vibrations of finite chain molecules which form a 

nonnegligible part of industrially relevant materials. On the other hand, real polymer 

samples never show perfect chemical, strereochemical, and confoimational 

structure. Symmetry is broken and new bands appear which become characteristic 

of specific types of disorder. 

If a few simple theoretical concepts of the dynamics of ordered and disordered 

chain molecules are taken into account it can be easily perceived that the vibrational 

infrared and Raman spectra contain a wealth of information essential to 

analytical polymer chemistry, structural chemistry, and physics. 

The content of this book has been planned to rejuvenate the vibrational spectroscopy 

of polymers. At present, the classes of polymeric materials are very many 

in number and range from classical bulk polymers of great industrial and technological 

relevance to highly sophisticated functional polymers which reach even the 

interest of photonics and molecular electronics. Also, the whole world of biopolymen 

requires great attention from spectroscopy. 

This book touches on a very few classes of polymeric materials which we consider 

representatives for introducing problems, spectroscopic techniques, and solutions 

prototypical of many other classes of polymers and plastics. 

Chapters 1 and 2 introduce new experimental techniques that provide new sets of 

relevant data for the study of the local and overall mobility of polymer chains. In 

Chapter 1, the development of two-dimensional infrared spectroscopy is described 

with a discussion of the mathematical principles, the description of the instrumental 

technique, and a detailed analysis of a few cases. In Chapter 2 the success of Fourier 

transform infrared polarization spectroscopy is shown for the study of segmental 

mobility in a polymer or a liquid-crystalline polymer under the influence of an 

external directional perturbation such as electric, electromagnetic, or mechanical 

forces. Industrial research laboratories should pay much attention to the information 

which can be acquired with this technique for technologically relevant polymeric 

materials. 

Chapter 3 is a guided tour of molecular dynamics and infrared and Raman frequency 

and intensity spectroscopy of polymethylene chains, from oligomers to polymers, 

in their perfect and disordered states. The reader is exposed in the easiest 

possible way to the basic theoretical concepts and to the numerical techniques 

which can be applied to such studies. Many references are provided for the spectroscopist 

who wants to develop his or her own independent skill and judgment.

However, theory and calculations yield concepts and data to be used also in polymer 

characterization, even for routine analytical work. After discussing several 

oligoniers or polymers the reader is guided step-by-step in the practical exercise 

to derive the detailed scenario of the mechanism of phase transition and melting of 

n-alkanes. 

Some of the theoretical tools laid in Chapter 3 are fully exploited in Chapter 4, 

which exposes the reader to the actual problems (and proposed solutions) of the 

chemistry and physics of modern and technologically relevant polyconjugated polymers 

in their intact (insulating) and in the doped (electrically conducting) states. 

Chapter 5 is an up-to-date review of the spectroscopic and structural problems 

and solutions reached by a modern approach to the dynamics of polypeptides. A 

clear and comprehensive discussion is presented on the force fields necessary for 

a reliable structural analysis through the vibrational spectra. Tools are thus becoming 

available for a systematic study of the structure even of complex polypeptides. 

The message we wish to send to the reader of this book is that modern machines 

provide beautiful infrared and Raman spectra of polymers full of specific, unique 

and detailed information which can be extracted, not just merely and lazily using 

group frequency correlations. We will be very pleased to find that this book has 

provided the reader with enough motivation to overcome the potential barrier of 

some theoretical technicalities in order to enjoy fully the wealth of information 

inherently contained in the vibrational spectra of ordered or disordered chain 

molecules. 

Milan, June 1998 

Giuseppe Zerbi


Contents 

1 Two-Dimensional Infrared Spectroscopy 

I. Noda, A. E. Downrey and C. Marcott 

1.1 Introduction 

1.2 Background 

1.3 Basic Properties of 2D Correlation Spectra 

1.4 Instrumentation 

1.5 Applications 

2 Segmental Mobility of Liquid Crystals and Liquid-Crystalline 

Polymers Under External Fields: Characterization by 

Fourier-Transform Infrared Polarization Spectroscopy 

H. W. Siesler, I. Zebger, Ch. Kulinna, S. Okretic, S. Shilov and 

U. Hoflinann 


2.2 Measurement Techniques 

2.3 Theory 

2.4 Structure Dependent Alignment of Side-Chain Liquid-Crystalline 

Polyacrylates on Anisotropic Surfaces 

2.5 Electric-Field Induced Orientation and Relaxation of 

Liquid-Crystalline Systems 

2.6 Alignment of Side-Chain Liquid-Crystalline Polyesters Under 

Laser Irradiation 

2.7 Orientation of Liquid-Crystals Under Mechanical Force 

2.8 Conclusions 

1 

1 

2 

10 

12 

15 

33 

33 

34 

36 

37 

38 

60 

67 

81 

3 Vibrational Spectra as a Probe of Structural Order/Disorder in 

Chain Molecules and Polymers 87 

G. Zerbi and M. Del Zoppo 

3.1 Introduction 87 

3.2 The Dynamical Case of Small and Symmetric Molecules 88 

3.3 How to Describe the Vibrations of a Molecule 94 

3.4 Short and Long Range Vibrational Coupling in Molecules 95 

3.5 Towards Larger Molecules: From Oligoiners to Polymers 98 

3.6 From Dynamics to Vibrational Spectra of One-Dimensional 

Lattices 107

X 

Corztrnts 

3.7 The Case of Isotactic Polypropylene: A Textbook Case 

3.8 Density of Vibrational States and Neutron Scattering 

3.9 Moving Towards Reality: From Order to Disorder 

3.10 What Do We Learn from Calculations 

3.11 A Very Simple Case: Lattice Dynamics of HCI-DCl Mixed 

Crystals 

3.12 Cis-Tram Opening of the Double Bond in the Polymerisation 

of Ethylene 

3.13 Defect Modes as Structural Probes in Polymethylene Chains 

3.14 Case studies: 

Case 1 

Conformational Mapping of Fatty Acids Through 

Mass Defects 

Case 2 Liquid Crystalline Polymers: Polyesters 

Case 3 Chain Folding in Polyethylene Single Crystals 

Case 4 The Structure of the Skin and Core in Polyethylene 

Films (Normal and Ultradrawn) 

Case 5 Moving Towards More Complex Polymethylene 

Systems 

3.15 Simultaneous Configurational and Conformational Disorder. 

The case of Polyvynylchloride 

3.16 Structural Inhomogeneity and Raman Spectroscopy of 

LAM Modes 

3.17 Fermi Resonances 

3.18 Band Broadening and Conformational Flexibility 

3.19 A Worked Out Example: Froin N-Alkanes to Polyethylene. 

Structure and Dynamics 

4 Vibrational Spectroscopy of Intact and Doped Conjugated Polymers 

and Their Models 

Y Furukawa and M. Tasuiiii 


4.2 Materials 

4.3 Geometry of Intact Polymers 

4.4 Geometrical Changes Introduced by Doping 

4.5 Methodology of Raman Studies of Polarons, Bipolarons 

and Solitons 

4.6 Near Infrared Raman Spectroscopy 

4.7 Poly(p-phenylene) 

4.8 Other Polymers 

4.9 Electronic Absorption and ESR Spectroscopies and Theory 

4.10 Mechanism of Charge Transport 

4.1 1 Summary 

5 Vibrational Spectroscopy of Polypeptides 

S. Kvinim 


113 

1 I7 

122 

128 

131 

138 

141 

144 

146 

147 

150 

152 

156 

158 

159 

163 

172 

180 

207 

207 

208 

209 

210 

215 

216 

217 

228 

230 

23 1 

234 

239 

239

Con ten t~ 

xi 

Index 

5.2 Force Fields 

5.3 Amide Modes 

5.4 Polypeptides 

5.5 Summary 

240 

249 

257 

280 

287


Contributors 

M. Del Zoppo 

Dipartimento di Chimica Industriale 

Politecnico di Milaiio 

Piazza L. Da Vinci 32 

20133 Milano 

Italy 

A. E. Do?;rey 

The Procter and Gamble Company 

Miami Valley Laboratories 

P.O. Box 398707 

Cincinnati, OH 45239-8707 

USA 

Y. Furukawa 

Department of Chemistry 

School of Science and Engineering 

Waseda University 

Shinjuku-ku 

169 Tokyo 

Japan 

U. Hoffmann 

Bruins Instruments 

D 82178 Puchheim 

Germany 

S. Krimm 

Biophysics Research Division and 

Department of Physics 

University of Michigan 

Ann Arbor, MI 48109 

USA 

Ch. Kulinna 

Bayer AG 

D 40789, Monheim 

Germany 

C. Marcott 



P.O. Box 398707 

Cincinnati, OH 45239-8707 

USA 

I. Noda 



P.O. Box 398707 

Cincinnati? OH 45239-8707 

USA 

S. Okretic 

Fachhochschule Niederrhein 

Fachbereich 04 

D 7805 Krefeld 

Germany 

S. Shilov 

Institute of Macromolecular 

Compounds 

199004 St. Petersburg 

Russia 

H. W. Siesler 

Department of Physical Chemistry 

University of Essen 

D 45 1 17 Esseii 

Gerinaii y

xiv 

Coiitribirtovs 

M. Tasumi 

Department of Chemistry 

Faculty of Science 

Saitama University 

Urawa 

Saitama 338 

Japan 

I. Zebger 

Department of Physical Chemistry 

University of Essen 

D 451 17 Essen 

Germany 

G. Zerbi 

Dipartiniento di Chiinica Iiidustriale 

Politecnico di Milano 

Piazza L. Da Vinci 32 

20133 Milano 

Italy

1 Two-Dimensional Infrared (2D IR) 

Spectroscopy 


Two-dimensional infrared (2D IR) correlation spectroscopy [ 1-81 is a relatively new 

addition to the spectroscopic methods used for the characterization of polymers. In 

2D IR, infrared spectral intensity is obtained as a function of two independent 

wavenuinbers as shown in Figure 1-1. The well-recognized strength of IR spectroscopy 

arises from the specificity of an IR probe toward individual molecular vibrations 

which are strongly influenced by the local molecular structure and environment 

[9-121. Because of such specificity toward the submolecular state of a sample, 

surprisingly useful information about a complex polymeric system is provided by 

expanding IR analysis to the second spectral dimension. 

The initial idea of generating two-dimensional correlation spectra was introduced 

several decades ago in the field of NMR spectroscopy [13-161. Since then, numerous 

successful applications of multidimensional resonance spectroscopy techniques 

have been reported, including many different types of studies of polymeric materials 

by 2D NMR [17-191. However, until now the propagation of this powerful concept 

of inultidimensional spectroscopy in other areas of spectroscopy, especially vibrational 

spectroscopies such as IR and Rainan, has been surprisingly slow. 

One of the reasons why the two-dimensional correlation approach as applied in 

NMR spectroscopy was not readily incorporated into the field of IR spectroscopy 

was the relatively short characteristic times (on the order of picoseconds) associated 

with typical molecular vibrations probed by IR. Such a time scale is many orders 

of magnitude shorter than the relaxation times usually encountered in NMR. Consequently, 

the standard approach used so successfully in 2D NMR, i.e., multiplepulse 

excitations of a system followed by the detection and subsequent double 

Fourier transformation of a series of free induction decay signals, is not readily 

applicable to conventional IR experiments. A very different experimental approach, 

therefore, needed to be developed in order to produce 2D IR spectra useful for the 

characterization of polymers using an ordinary IR spectrometer.

Figure 1-1. A fishnet plot of a 

synchronous two-dimensional 

infrared (2D IR) correlation 

spectrum of atactic polystyrene 

in tlie CH-sti-etching vibration 

region at room temperature. 

1.2 Background 

1.2.1 Perturbation-Induced Dynamic Spectra 

The basic scheme adapted for generating 2D IR spectra 13, 81 is shown in Figure 

1-2. In a typical optical spectroscopy experiment, an electromagnetic probe (eg, 

IR, X-ray, UV or visible light) is applied to the system of interest, and physical 

or chemical information about the system is obtained in the form of a spectrum 

representing a characteristic transformation (eg, absorption, retardation, and 

scattering) of the electromagnetic probe by the system constituents. In a 2D IR 

experiment, an external physical perturbation is applied to the system [2, 31 with the 

incident IR beam used as a probe for spectroscopic observation. Such a perturbation 

often induces time-dependent fluctuations of the spectral intensity, known as 

the dyrzamic spectrum, superposed onto the normal static IR spectrum. 

There are, of course, many different types of physical stimuli which could induce 

such dynamic variations in the spectral intensities of polymeric samples. Possible 

sources of perturbations include electrical, thermal, magnetic, acoustic, chemical, 

optical, and mechanical stimuli. The waveform or specific time signature of the 

perturbation may also vary from a simple step function or short pulse, to more 

complex ones, including highly multiplexed signals and even random noises. In this 

chapter, dynamic spectra generated by a simple sinusoidal mechanical perturbation 

applied to polymers will be discussed. 

Mechanical. eleclncal. 

opllcal. lliermal. elc 

Eleclro.magneltc 

probe (eg IR. UV) 

D y n a m I c 

Figure 1-2. A generalized experimental scheme for 

2D correlation spectroscopy based on perturbationinduced 

dynamic spectral signals [S).

1.2 Brick~qrorrnrl 3 

. 

Figure 1-3. Schematic diagram of the dynamic 

infrared linear dichroism I DIRLDi experiinent 

1231. A small-amplitude sinusoidal strain is 

applied to a sample. and submolecular level 

reorientation responses of cheniical moieties are 

monitored with a polarized IR beam. 

DETECTOR 

1.2.2 Dynamic IR Linear Dichroism (DIRLD) 

Figure 1-3 shows a schematic diagram of a dynamic IR linear dichroism (DIRLD) 

experiment [20-251 which provided the foundation for the 2D IR analysis of polymers. 

In DIRLD spectroscopy, a small-amplitude oscillatory strain ica. 0.1% of the 

sample dimension) with an acoustic-range frequency is applied to a thin polymer 

film. The submolecular-level response of individual chemical constituents induced 

by the applied dynamic strain is then monitored by using a polarized IR probe as 

a function of deformation frequency and other variables such as temperature. 

The macroscopic stress response of the system may also be measured simultaneously. 

In short, a DIRLD experiment may be regarded as a combination of two 

well-established characterization techniques already used extensively for polymers: 

dyizanzic inechnnicul unalysis (DMA)[26, 271 and ijlfrured dichroisin (IRD) spectroscopy 

[ 10, 113. 

The optical anisotropy, as characterized by the difference between the absorption 

of IR light polarized in the directions parallel and perpendicular to the reference 

axis (i.e., the direction of applied strain), is known as the IR linear dichroism of the 

system. For a uniaxially oriented polymer system [ 10, 28-30], the dichroic dzflerence, 

AA(v) = A,,(Y) - Al(v), is proportional to the average orientation, i.e., the 

second moment of the orientation distribution function, of transition dipoles (or 

electric-dipole transition moments) associated with the molecular vibration occurring 

at frequency 1’. If the average orientation of the transition dipoles absorbing 

light at frequency if is in the direction parallel to the applied strain, the dichroic 

difference AA takes a positive value; on the other hand, the IR dichroism becomes 

negative if the transition dipoles are perpendicularly oriented. 

If a sinusoidally varying small-amplitude dynamic strain is applied to a polymeric 

system, a similar sinusoidal change in IR dichroic difference, as shown in Figure 

1-4, is usually observed [3, 231. The dynamic variation of IR dichroism arises from 

the time-dependent reorientation of transition dipoles induced by the applied strain. 

Interestingly, however, the dichroism signal often is not fully in phase with the 

strain. There is a finite phase difference between the two sinusoidal signals representing 

the externally applied macroscopic perturbation and the resulting dynamic 

molecular-level response of the system. This phase difference is due obviously to

I I D W l 

0 : 2n 4n 

+ p H- Phase Angle 

0 2n 4n 

Figure 1-4. A small-amplitude sinusoidal 

strain applied to a sample and resulting 

dynamic IR dichroism I DIRLDI response 

induced by the strain. The two siiiusoidal 

signals are not always in phase with each 

other. 

- 

c 

n 

L 

m 

58msec 

Figure 1-5. A time-resolved 

dynamic IR dichroism 

(DIRLD I spectrum of an 

atactic polystyrene film under 

a small-amplitude (ca. 0. I'X)) 

sinusoidal (23-Hzl dynamic 

strain at room temperature. 

the rate-dependent nature of the reorientation processes of various submolecular 

constituents. 

The advantage of using IR spectroscopy in dynamic studies of polymers is that 

such a measurement can be used to examine individual submolecular constituents 

or chemical functional groups by simply changing the wavelength of the IR probe. 

The variation of IR dichroism, depicted as a simple sinusoidal signal in Figure 1-4. 

can actually be measured as a function of not only time but also IR wavenumber. 

In other words, dynamic IR dichroism is obtained as a time-resolved spectrum 

representing the molecular level response of polymers, as shown in Figure 1-5. 

For a given sinusoidal dynamic strain i(t) with a small amplitude i and fre-

1.0 , 

Figure 1-6. The in-phtrse 

and cprrt/r.ct/irr.c components 

of the DIRLD spectrum 

shown in Figure 1-5. A 

normal static absorbance 

spectrum of the sample is 

also provided for reference. 

Wavenumber 

quency w, 

E( t) = i sin ux (1-1) 

a time-resolved DIRLD spectrum may be represented by 

where AA(v) and P(v) are. respectively, the riiagnitude and phase (loss) angle of 

the dynamic IR dichroism [23]. By using a simple trigonometric identity, the above 

expression can also be rewritten in the form 

Ak(i9, t) = AA’(v) sin cot + AA”(v) coswt. (1-3) 

The wavenumber-dependent terms, AA’( I!) and AA”(v), are known respectively as 

the in-phase spectrum and quadrature spectrunt of the dynamic dichroism of the 

system. They represent the WN/ (storage) and irnagirrury (loss) components of the 

time-dependent fluctuations of dichroism. Figure 1-6 shows an example of the inphase 

and quadrature spectral pair extracted from the continuous time-resolved 

spectrum shown in Figure 1-5. These two ways of representing a DIRLD spectrum 

contain equivalent information about the reorientation dynamics of transition 

dipoles. However, the orthogonal representation of the time-resolved spectrum 

using the in-phase and quadrature spectra is obviously more compact and easier to 

interpret than the stacked-trace plot of the time-resolved spectrum. 

1.2.3 IR Dichroism and Molecular Orientation 

In a traditional characterization study of polymeric materials, the IR dichroism 

technique is most often employed for the determination of the degree of orientation

Reference 

axis 

Molecular 

segment 

Transition 

dipole 

Figure 1-7. Uniaxial orientation of a polymei- chain segment with respect 

to a reference axis and corresponding alignment of a transition dipole. 

of molecular chain segments [lo, 28-30]. For a uniaxially oriented polymer system 

(Figure 1-7), the orimtatioii Jbctor P?(Q), i.e., the second moment of the orientation 

distribution function of polymer chain segments, is given by 

PZ(0) = (_?(COS? 6,) - 1)/2 (1-4) 

where 6, is the orientation angle between each polymer chain segment and the optical 

reference axis of the system. The notation (cos2 Q) indicates that the squared-cosine 

of orientation angles for individual segments are averaged over the entire space. 

The orientation factor Pz(B) is a convenient measure of the average degree of 

orientation of polymer chains in the system. Under an ideal orientation state where 

all polymer chains are aligned perfectly in the direction parallel to the reference 

axis, the value of P2(6,) becomes unity. On the other hand, if polymer chain segments 

are all perpendicularly aligned, P2(8) becomes - 1/2. An optically isotropic 

system gives the Pl(0) value of zero. 

According to the classical theory of the IR dichroisiii of polymers [lo, 28, 291, the 

diclzroic ratio, D(v) = AII(V)/A~(V), of the system is directly related to the orientation 

of polymer chain segments by 

D(V) - 1 Ow(.) + 2 

PI(@) = (1-5) 

D(v) +2 Dw(v) - 1 

The ultinzate dichroic ratio D, (v) is the value of dichroic ratio if the polymer chain 

segments are all perfectly aligned in the direction of the reference axis, i.e., 

Pz(S) = 1. Given the local orientation angle a between the polymer chain segment 

and transition dipole probed at the IR wavenumber v (Figure 1-7), the ultimate 

dichroic ratio is given by 

2 

DN,(V) = 2cot a. 

(1-6) 

It can be easily shown that Eq. (1-5) simplifies to the form

1.2 Bcrckgroinizd 7 

where AA (i~) is the iilti/ncrte dichroic d~juwrce which is related to D, ( v) by 

The ~ t r ~ ~ absorbcince, r ~ w d A,( if) = [A (1)) + 2AI (v)]/3, is believed to be independent 

of the degree of molecular orientation as long as the system remains symmetric 

with respect to the optical axis. 

If one accepts that the transition dipole orientation angle ci is a fixed molecular 

constant unaffected by the conformation of polymer chains, then D, and AAz 

should also be independent of the state of chain orientation. The classical theory 

of IR dichroisin for polymers thus makes an interesting assertion, according to Eq. 

(1-7), that IR dichroic difference AA(v) is alivays linearly proportional to the average 

orientation Pz(0) of polymer chains regardless of the IR wavenumber v. In 

other words, one should be able to determine the state of orientation of the entire 

polymer chain by simply observing the local orientation of a transition dipole 

associated with the molecular vibration of any arbitrarily chosen local functional 

group, as long as ci is known. 

1.2.4 Breakdown of the Classical Theory 

An interesting observation often made during the DIRLD measurement of polymers 

is that the phase angle p(~) between the applied strain and dynamic IR 

dichroic diff‘erence is strongly dependent on the IR wavenumber [3, 33-25]. Timedependent 

dichroism intensities measured at certain wavenumbers change much 

faster than others. Individual dynamic IR dichroisin signals thus become out of 

phase with each other as depicted in Figure 1-8. As already demonstrated in Figure 

1-6, the shape of the in-phase spectrum also becomes quite different from that of the 

quadrature spectrum. The relative amount of dynamic dichroism signal appearing 

in the in-phase and quadrature components varies considerably as a function of 

wavenumber. 

t 

Figure 1-8. Tinie-dependent sinusoidal u 

DIRLD signals from an atactic poly- 5 

(methyl methacrylate) sample detected at 0” 

different IR wavenumbers. Signals are out 

of phase with each other. 

a: 

(2952 cm-1) 

r 

Time, t

Quadrature 0 

+ 

3150 3 

Wavenumber 

Figure 1-9. DIRLD spectra of atactic polystyrene in the phenyl 

CH-stretching region. Transition dipoles for IR bands marked 

with + are reorienting totally in phase with the applied dynamic 

strain, while those with 0 are moving at rates substantially 

diffei-ent from the strain. 

This surprising discovery reveals that a inajor discrepancy exists between the 

classical theory of IR dichroisni and actual experimental observations made for 

reorientation dynamics of polymer chains. Figure 1-9, for example, indicates that 

the intensity of the quadrature spectrum is close to zero, i.e., signals are in phase 

with the applied strain for dichroism peaks marked with (t). If indeed IR dichroism 

signals at these wavenumbers reflect the time-dependent orieiitational state of 

the entire polymer chain, one would conclude that polymers are reorienting instantaneously 

under the dynamic deformation. On the other hand, if the dichroism signals 

are measured for other IR bands, for example, at the peaks inarked with (O), 

the polymer chain seems to reorient at a rate substantially out of phase with tlie 

applied strain. This second statement clearly contradicts the conclusion made as a 

result of the previous observation. Thus, for a certain polymer system undergoing a 

dynamic deformation, the well-accepted classical view of IR dichroisin (i.e., 

dichroic difference must always be linearly proportional to the average orientation 

of polymer chains, regardless of the wavenuinber of IR probe, as described in Eq. 

(1-7)) is no longer valid. 

The logical explanation for tlie experimentally observed, wavenumber-dependent 

behavior of the DIRLD phase angle is that the reorientation rates of individual 

traiisitioii dipoles in the system are not the same. For a given macroscopic perturbation 

such as dynamic strain, different submolecular constituents (e.g., backbone 

segments, side chains, and various functional groups comprising the polymer chain) 

may respond at different rates, more or less independently of each other. Consequently, 

transition dipoles associated with the molecular vibrations of different 

parts of a polymer chain have independent local reorieiitational responses not fully 

synchronized to tlie global motioiis of tlie entire polymer chain.

1.2 Background 9 

1.2.5 Two-Dimensional Correlation Analysis 

The wavenumber-dependent orientation rates of individual transition dipoles 

observed for dynamically stimulated polymer systems poses several intriguing 

questions: What makes these transition dipoles reorient at different rates? Why do 

some of the transition dipoles seem to reorient at a rate similar to each other? Is 

there any underlying mechanism responsible for synchronization, or lack thereof, in 

the local reorientation processes of submolecular structures‘? To answer such questions, 

one must introduce an effective way of representing a measure of the siniilarity 

oi- difference of the reorientation rates of transition dipoles. 

One convenient mathematical approach to comparing the behavior of a set of 

variables is a correlation analysis [2, 3, 81. For a pair of dynamic dichroisni signals, 

Aj(v1, f) and AA(v2, t + T ), which are observed at different instances separated by 

a fixed correlation time z at two arbitrarily selected wavenumbers, v1 and vz, for a 

certain length of observation period T, a two-dimerisionul c‘ross-correlation fiiriction 

X(z) is defined by 

For a pair of sinusoidally varying signals with a fixed frequency 01, the crosscorrelation 

function reduces to a simple form [3], 

X(z) = @(vl,v2)cosoz+ Y(vl,v2)sincoz. (1-10) 

The real and iimqinnry components, @(vl, 1’2) and Y(v1, v2), of the cross-correlation 

function X(z) are referred to, respectively, as the synclzroiioiis and asyiichronoLis 

correkition intensitji. These quantities are related to the in-phase and quadrature 

spectra of dynamic dichroism by 

(1-1 1) 

The synchronous correlation intensity @( VI ,v2) represents the siniilurity between the 

two dynamic dichroisni signals measured at dif‘fereiit wavenuinbers. This quantity 

becomes significant if the two dynamic IR dichroism sigiials are changing at a similar 

rate but vanishes if the time dependency of the signals is very different. The 

asynchronous correlation intensity ‘€‘(vl, vz), on the other hand, represents the dissiniilmity 

between signals. It becomes significant only if the two IR dichroism signals 

are changing out of phase with each other but reduces to zero if they are 

changing together. 

The two-dimensional nature of our correlation analysis arises from the fact 

that the correlation intensities are obtained by comparing the time dependence of

Wavenumber, vI 

Figure 1-10. A schematic contour diagram ofa synchronous 

2D IR correlation spectrum 131. Shaded areas represent 

negative correlation intensity. 

I 

n 

Wavenumber, v, 

Figure 1-11. A schematic contour diagram of an asynchronous 

2D IR correlation spectrum [3]. Shaded areas represent 


dynamic IR signals observed at two independently chosen wavenumbers. Plots of 

correlation intensities as functions of two wavenumber axes are referred to as 2D 

IR correlution spectra. Such spectra reveal important and useful information not 

readily accessible from conventional one-dimensional spectra. 

There are many different ways to plot a 2D IR correlation spectrum. A pseudothree-dimensional 

representation (so-called fishnet plot) as shown in Figure 1-1 is 

well suited for providing the overall features of a 2D IR spectrum. Usually, however, 

2D IR spectra are more conveniently displayed as contour maps to indicate 

clearly the location and intensity of peaks on a given spectral plane. In the following 

section, basic properties of and information extracted from 2D IR spectra are 

reviewed by using schematic contour maps (Figures 1-10 and 1-1 1). 

1.3 Basic Properties of 2D IR Correlation Spectra 

1.3.1 Synchronous 2D IR Spectrum 

The correlation intensity at the diagonal position of a synchronous 2D spectrum 

(Figure 1-10) corresponds to the autocorrelation function of perturbation-induced

1.3 Bnsic Properties of 20 IR Correlotioii Spectra 11 

dynamic fluctuations of the IR signals. Local intensity maxima along the diagonal 

are thus referred to as autopolis. Since the magnitude of dynamic variations of IR 

dichroism represents the susceptibility of transition dipoles to reorient under a given 

external perturbation, autopeaks indirectly reflect the local mobility of chemical 

groups contributing to the inolecular vibrations associated with the transition 

dipoles. The schematic example in Figure 1-10 indicates that functional groups 

contributing to the molecular vibrations with frequencies A, B, C, and D are undergoing 

reorientational motions induced by the applied dynamic strain. 

Peaks located at off-diagonal positions of a 2D correlation spectrum are called 

cyoss peaks. They appear when the dynamic variations of IR signals at two different 

wavenumbers corresponding to the spectral coordinates (111 , ~ 2 are ) correlated with 

each other. For a synchronous spectrum, this occurs when the two IR signals are 

fluctuating in phase (i.e., simultaneously) with each other. Synchronized variations 

of dichroism signals result from the simultaneous reorientation of transition dipoles 

associated with corresponding IR wavenumbers. Coordinated local motions of 

submolecular groups lead to such simultaneous reorientations. In turn, highly correlated 

local reorientations of chemical groups in response to a common external 

stimulus imply the possible existence of interactions or connectivity which restrict 

the independent motions of these submolecular structures. 

Functional groups which are not strongly interacting, on the other hand, can 

move independently of each other. The transition dipoles associated with molecular 

vibrations of these groups may then reorient at different rates (or somewhat out of 

phase with each other), resulting in a much weaker synchronous correlation. Thus, 

as long as the normal modes of vibrations correspond to reasonably pure group 

frequencies, one can use the cross peaks in a synchronous 2D IR spectrum to map 

out the degree of intra- and intermolecular interactions of various functional 

groups. In Figure 1-10, functional groups contributing to the molecular vibrations 

with frequencies A and C may be interacting. Likewise, the pair B and D could be 

connected. 

The signs of cross peaks indicate relative reorientation directions of transition 

dipoles and consequently their associated chemical groups. If the sign of a synchronous 

cross peak is positive, the corresponding pair of transition dipoles reorient 

in the same direction. If negative, on the other hand, the reorientation directions are 

perpendicular to each other. In Figure 1 - 10, mutually perpendicular reorientation 

of a pair of transition dipoles at wavenumbers A and C, as well as parallel reorientation 

for B and D, are observed. 

1.3.2 Asynchronous 2D IR Spectrum 

An asynchronous 2D IR spectrum (Figure 1-1 1) provides complementary information; 

cross peaks appear if IR signals are not synchronized to be completely in 

phase with each other. This feature is particularly useful. since IR bands arising 

from molecular vibrations of different functional groups, or even of similar group5 

in different local environments, may exhibit substantially different time-dependent 

intensity fluctuations. Thus, asynchronous 2D IR spectra can be used to differen-

tiate highly overlapped IR bands, since asynchronous cross peaks should develop 

among these bands. 

From the sign of asynchronous cross peaks, one can obtain temporal information 

about the perturbation-induced reorientation processes of transition dipoles and 

their corresponding chemical groups. A positive peak in an asynchronous spectrum 

indicates the transition dipole with the vibrational frequency 1’1 reorients hc;fbr.r that 

for 1’2. If the sign is negative, reorients ufi~r 

1’2. However. this temporal relationship 

is reversed for perpendicularly reorienting pairs of transition dipoles, i.e., the 

synchronous correlation intensity at the same spectral coordinate becomes negative. 

In Figure 1-1 1, the reorientational motions of functional groups contributing to the 

molecular vibrations with frequencies B and D occur before those for A and C. More 

detailed discussions on the properties of 2D 1R spectra are found elsewhere [3). 

1.4 Instrumentation 

Dynamic infrared spectra suitable for 2D correlation analysis can, in principle, be 

measured using any conventional IR Spectrometer [ 71. The spectrometer, however, 

must be equipped with the ability to stimulate samples by some physical means and 

measure the resulting time-dependent fluctuations of the IR signals. Both dispersive 

monochromators (231 and Fourier transform infrared (FT-IR) instrumentation 151 

have been successfully used to measure dynamic 2D 1R spectra. Although welldesigned 

dispersive spectrometers can often achieve better signal-to-noise ratios 

(S/N) over small spectral regions, FT-IR measurements cover much broader spectral 

regions in less time and are, more importantly, readily available commercially. 

Conventional rapid-scan FT-IR spectrometers are not well suited for these types 

of measurements when the dynamic strain frequencies of interest are in the range 

between 0.1 Hz and 10 kHz. Using a step-scanning interferometer substantially 

simplifies dynamic 2D FT-IR measurements. Figure 1-12 shows an example of a 

n 

Rolahng-Blade 

chopper. wc 

nred 

P*er 

Figure 1-12. A schematic view 

of a DIRLD spectrometer based 

on a small-amplitude mechanical 

deformation 1231.

2D IR spectrometer coupled with a dynamic rheometer capable of applying a smallamplitude 

mechanical perturbation [23]. 

In addition to the time-resolution constraints on dynamic infrared measurements. 

sensitive instrumentation is also required. The signals of interest, which are often 

lo4 times smaller than the normal IR absorbance of the sample, exist only as ii 

result of the small dynamic strain applied to the sample. 

The type of perturbation applied to the sample is not restricted to simple 

mechanical strain. Electrical perturbations, for example, can be effectively used to 

induce the necessary fluctuations of IR signals to produce 2D IR spectra. The use of 

electrical stimuli has been especially successful in studies of neniatic liquid crystalline 

systems [3 1-33], where selective orientation of liquid crystals under an alternating 

electric field was observed. Electrochemical experiments modulated with an 

alternating electrical current have also been used to generate dynamic 1R spectra 

suitable for 2D correlation analysis [34, 351. Recently, a two-dimensional photoacoustic 

spectroscopy (2D PAS) experiment was conducted with a step-scanning 

FT-IR spectrometer to obtain depth profiles of layered samples [36, 371. The characteristic 

time dependence of PAS signals representing sample components located 

at different depths inside samples were successfully differentiated by 2D correlation 

analysis. Each of these applications is similar in that a dynamic perturbation is 

applied to the sample and the time-resolved response is measured using phasesensitive 

detection [ 6 -81. 

1.4.1 Dispersive 2D IR Spectrometer 

A high-optical-throughput dispersive monochromator turns out to be an excellent 

choice for dynamic 2D IR spectroscopy [3, 231. As shown in Figure 1-12, the incident 

IR beam, originating from a high-intensity source, is modulated at three separate 

places by a chopper, photoelastic modulator (PEM), and dynamic mechanical 

analyzer (i.e., polymer stretcher). The chopper labels photons originating from the 

source so they can be distinguished from background IR emission. The PEM, which 

immediately follows a fixed linear polarizer in the beam path, enables the polarization 

direction of the beam to be switched back and forth rapidly (-100 kHz) between 

directions aligned parallel and perpendicular to the sample strain direction. 

Sample strain frequencies used are typically between 0.1 and 100 Hz. Each niodulation 

frequency is thus separated from the other two modulation frequencies by at 

least one order of magnitude. This makes analysis of the individual signal coniponents 

with lock-in amplifiers straightforward [23]. 

Figure 1-13 shows an example of a train of lock-in amplifiers used to demodulate 

DIRLD signals obtained with a spectrometer described in Figure 1-12. In order to 

obtain the time-dependent dynamic absorbance and DIRLD responses, quadrature 

lock-in amplifiers are used. These devices monitor signals both in phase and 90" out 

of phase (quadrature) with the sinusoidal strain reference signal. The monochromator 

is scanned one wavelength at a time through the spectrum. Data are collected 

on six separate channels (e.g., in-phase and quadrature dynamic dichroism, in-phase 

and quadrature dynamic absorbance, static dichroism, and normal IR absorbance)

Slalic dichroism 

In-phase dynamic dlchrolsm 

Quadrature dynamic dichrolsm 

Preamp 

- Lockin 

D 

- 

VD Slatic absorbance 

at each wavenumber position until an acceptable S/N is achieved. Multiple scans 

of the entire spectrum can be collected in order to further improve the S/N and to 

average out longer-tenii drift. 

1.4.2 Step-Scanning 2D FT-IR Spectrometer 

The dynamic 2D IR spectroscopy experiment performed on a step-scanning FT-IR 

spectrometer is in many ways similar to the way it is done on a dispersive instrument 

[5, 71. For example, a step-scanning interferometer measures an interferogram 

one mirror retardation position at a time in the same way a monochromator scans 

through a spectrum one wavelength at a time. The mirror remains in a fixed retardation 

position while data are accumulated from the output channels of lock-in 

amplifiers tuned to the signals of interest (e.g., in-phase and quadrature dynamic 

absorbance, normal 1R absorbance). As in the dispersive experiment, signals are 

time averaged long enough to achieve an acceptable S/N and multiple scans of the 

entire interferogram can be coadded to average out longer-term drift. When a scan 

is completed, the resulting interferogram does not depend on the moving mirror 

velocity, as in the case of an interferograni collected in rapid-scan mode. The 

Fourier frequencies in a step-scanning experiment are typically in the sub-Hz range 

where they are well separated from the polarization modulation and strain modulation 

frequencies. This is in contrast to a typical rapid-scan FT-IR experiment 

where each IR wavelength is modulated at a different acoustic-range frequency.

1.5 Applirtrtions 15 

Unlike a dispersive spectrometer which uses a mechanical light chopper, the stepscanning 

interferometer often performs better when a technique called phase niodu- 

Irr/ion is used to label the IR photons originating from the source. Phase modulation 

is achieved by dithering either the fixed or moving mirror of the interferometer 

a small amount (on the order of 1 pi) at a fixed frequency. This motion causes 

a modulation of the IR signal intensity due to a change in optical path difference 

between the two arms of the interferometer. The amplitude of the signal detected is 

thus the slope of the actual interferogram at the average optical path difference. 

Most of the early measurements of DIRLD spectra were made using a photoelastic 

modulator IPEM) to modulate the polarization rapidly between parallel and 

perpendicular to the dynamic strain direction. Although polarization modulation 

does improve the S/N of dynamic IR spectroscopy, by about 50% over experiments 

using only a fixed polarizer aligned parallel to the sample strain direction, a PEM 

adds significant complexity to the experiment. More lock-in amplifiers and careful 

optical alignment are required when using polarization modulation. Indeed the first 

2D IR spectra reported in the literature [2] were obtaining using unpolarized light. 

When a PEM is not in use, the step-scan FT-IR experiment will work equally well 

with either a room-temperature deuterated triglycine sulfate (DTGS) detector or 

a liquid-nitrogen-cooled mercury-cadmium-telluride (MCT) detector. Although 

MCT detectors are typically an order of magnitude more sensitive than DTGS 

detectors, the full throughput of a commercial FT-IR spectrometer is enough to 

saturate the most sensitive MCT detectors. Thus, the beam is normally attenuated 

by using smaller apertures, neutral-density filters, or optical filters when using an 

MCT detector. 

Although dispersive dynamic IR experiments are presently more sensitive than 

FT experiments over limited spectral ranges, recent introductions of commercial 

FT-IR spectrometers with step-scanning capability should make dynamic 2D IR 

spectroscopy accessible to many more laboratories. 


2D IR spectroscopy has been applied extensively to studies of polymeric materials. 

A recent review of 2D IR spectroscopy cites nunierous applications in the study of 

polymers by this technique [6]. In this section, some representative examples of 2D 

IR analysis of polymers are presented. We will start our discussion with a simple 

homogeneous amorphous polymer then move to more complex multiphase systems, 

such as seinicrystalline polymers. Alloys and blends consisting of more than one 

polymer components are of great scientific and technical importance. Both immiscible 

and miscible polymer blend systems may be studied by 2D IR spectroscopy. 

Analysis of microphase-separated block copolymers is also possible. Finally, the 

possible application of 2D IR spectroscopy to the studies of natural polymers of 

biological origin is explored.

- 

, 

i 

n 

-2950 

' 

E, Figure 

- > 

1-14. A synchronous 2D IR 

spectrum of a thin film of atactic 

polystyrene in the CH-stretching vibration 

region at room temperature. 

Regular one-dimensional spectra of 

the same system are provided at the 

top and left of the 2D spectrum for 

, I I l I I I I I 315" reference. Shaded areas represent 

I50 2950 2750 

Wavenumber. VI 


1.5.1 Amorphous Polymers 

A homogeneous one-phase system consisting of a single. noncrystallizing homopolymer, 

such as atactic polystyrene [3, 38-41 ] or poly(methy1 methacrylate) [42- 

441, provides one of the least complicated testing grounds for the applicability of 

the 2D IR technique for the study of local dynamics of polymers. By using 2D IR, 

surprisingly rich information can be extracted even from a simple polymeric system 

[25, 39-41 1. Figure 1-14 shows a contour-map representation of the synchronous 

2D IR correlation spectrum of an atactic polystyrene film in the CH-stretching 

region. A three-dimensional fish-net plot of the same 2D IR spectrum corresponding 

to Figure 1-14 has already been shown in Figure 1-1. 

DIRLD data used for the 2D correlation analysis was obtained by applying a 

23-Hz dynamic strain with an amplitude of 0.1% to a thin solution-cast film of atactic 

polystyrene (MI, 150,000) at room temperature 1391. The resulting dynamic dichroism 

spectra have already been shown in Figures 1-5 and 1-6. The time-dependent 

reorientations of transition dipoles associated with the molecular vibrations of 

backbone and side groups are all observable in the original dynamic dichroism 

spectra. However, much more detailed features are recognized in the 2D correlation 

spectrum. 

Autopeaks at the diagonal positions of Figure 1-14 represent transition dipoles 

that are susceptible to reorientational motions in the polystyrene sample. Peaks at 

2855 cm-I and 2925 cm-l are, respectively, assignable to the symmetric and antisymmetric 

CH2-stretching vibrations of the methylene groups in the backbone of 

the polystyrene chain; those peaks above 3000 cm-' arise from the phenyl ring CHstretching 

vibrations and indicate the reorientation of the side groups. The appearance 

of positive synchronous cross peaks at spectral coordinates corresponding to 

the two methylene bands indicates the transition dipoles associated with these two

\ 

'\ 

Melliyleiie 

1'7- 

-3000 

Posilivo cross peaks 

(phenyl 4 rnelhylena) 

, 

L 

a, 

n 

-3075 5 

C 

9 

- 2 

Figure 1-15. An asynchronous 2D 1R 

spectrum of a atactic polystyrene film 

in the CH-stretching vibration region 

I I I I I I I I I I 

3150 

3000 2900 2800 

wavenumbers reorient at a similar rate. Such synchronized responses are of course 

expected, since both 1R bands originate from vibrations of the same submolecular 

structural unit. 

The positive sign of the methylene cross peaks suggests the transition dipoles for 

the symmetric and antisymmetric methylene vibrations are both realigning in the 

same relative direction. i.e., perpendicular to the direction of applied strain. The 

molecular architecture dictates that these two transition dipoles are locally aligned 

more or less perpendicular to the backbone of the polymer segment. It is not difficult 

to conclude that the molecular chain segment of polystyrene must be realigning 

in the direction parallel to the applied dynamic tensile strain. 

Negative (shaded) synchronous cross peaks between the methylene and phenyl 

bands indicate that some of the phenyl transition dipoles are orienting parallel to 

the strain direction. It is known [40, 411 that the signs of these particular cross peaks 

for atactic polystyrene become positive at temperatures above the glass-to-rubber 

transition temperature (ca. 100 "C). The inversion of the cross-peak sign, and consequently, 

the relative reorientation direction of the side group with respect to the 

main chain, indicates a dramatic change in the local mobility of the polymer segments 

and functional groups is occurring around the glass transition temperature. 

Figure 1-15 shows the asynchronous 2D IR spectrum of the same sample comparing 

the reorientation dynamics of transition dipoles for backbone methylene and 

side-group phenyl. The spectral region corresponding to the asynchronous correlation 

of methylene stretching vibrations below 3000 cm-' is not shown since no 

asynchronous cross peaks are observed. The appearance of strong asynchronous 

cross peaks between phenyl and inethylene bands suggests the existence of rather 

complex reorientation dynamics involving the side groups. The rate of reorientational 

motion of the side groups under dynamic strain is apparently different from 

that of the backbone for this polymer. The signs of asynchronous cross peaks shown

in this 2D spectrum are all positive. Coupled with the fxt that the synchronous 

correlation intensities at corresponding coordinates i Figure 1 - 14) are all negative, 

one can conclude the side groups of this glassy polystyrene sample complete the 

realignment induced by the dynamic strain well before the main chain. 

This observation cannot be explained if the average local orientation angle of the 

phenyl groups with respect to the chain axis of polystyrene is fixed. There apparently 

is substantial freedom of the side groups to temporarily realign ahead of the 

main chain 141 1. This finding clearly demonstrates that it is impossible to monitor 

the main-chain dynamics of polystyrene by simply observing the reorientation of 

side groups. Furthermore, since phenyl groups contribute significantly toward the 

refractive index of polystyrene, the above observation also casts significant doubt 

over the validity of directly relating birefringence results of glassy polystyrene under 

a dynamic deformation to main-chain orientation dynamics. 

The observation of highly localized reorientational motions of functional groups 

in glassy amorphous polymers is not limited to atactic polystyrene. Similar independent 

motions of different side groups are observed in many other systems, including 

atactic poly(methy1 methacrylate) [42-441. Different side groups attached to 

the same polymer chain, such as ester methyl and a-methyl groups of polyjmethyl 

methacrylate). exhibited completely different reorientation rates (see Figure 1-8). 

2D IR spectroscopy has been especially useful in elucidating submolecular-level 

realignment of functional groups controlling the mechanical properties of glassy 

amorphous polymers [4, 43). 

The relationship between the local mobility of functional groups and glasstransition 

phenomenon has also been probed successfully. New insight into the 

dynamics of amorphous polymers going through glass-to-rubber transition processes 

was provided. Specifically, the local mobility of individual functional groups 

was found to play a very significant role in the glass-to-rubber transition of amorphous 

polymers [41]. In turn, it is now possible to determine if an amorphous 

polymer is in glassy or rubbery state by simply observing the characteristic local 

reorientation dynamics of functional groups using 2D IR spectroscopy. 

1.5.2 Semicrystalline Polymers 

Semicrystalline polymers, such as polyethylene 145-471 and polypropylene (5, 481, 

may also be studied by using the 2D IR technique. By taking advantage of the 

enhanced spectral resolution of 2D IR, overlapped IR bands assigned to the coexisting 

crystalline and amorphous phases of semicrystalline polymers can be easily 

differentiated. Such differentiation has become especially useful, for example, in the 

study of blends of high-density polyethylene and low-density polyethylene [47). 

Here it was found that blends of polyethylenes are mixed at the molecular scale 

only in the amorphous phase, while each component crystallizes separately. In this 

section. an example of a 2D IR analysis applied to a film of linear low-density 

polyethylene is discussed [46]. 

Figure 1-1 6 shows the asynchronous 2D IR spectrum of a thin film of linear lowdensity 

polyethylene consisting mainly of ethylene repeat units with a small amount

11 Amorphous Crystalline ,...." I 

Figure 1-16. An asynchronous ZD IR 

spectrum for the main chain syminetric 

CH2-stretching vibration of linear 

low-density polyethylene [46]. 

\ I...." 

I I , , 1 1 -2870 

2870 2855 2840 

Wavenumber, VI 

(ca. 3 mole'%,) of deuterium-substituted 1-octene comonomer units. The octene 

comonomer is incorporated into the polyethylene chain to produce short (sixcarbon) 

side branches. which reduce the melt temperature and crystallinity of the 

polymer. By selectively labeling the octene units with deuterium, one can unambiguously 

differentiate the dynamics of main chain and short side branches. The 

spectral region shown in Figure 1-16 represents IR signals arising exclusively from 

the polyethylene chain, with little contribution from the octene side branches. 

The presence of asynchronous peaks in the CHZ-stretching vibration of backbone 

methylene groups indicates there are two distinct bands in this region of the IR 

spectrum of the linear low density polyethylene sample. The band located near 

2855 cm-' can be assigned to the contribution from the crystalline phase of the 

sample. while the band near 2863 cm-' is due to the amorphous component of this 

semicrystalline polymer. The above assignment can be easily verified by heating the 

sample above its melt temperature to remove the IR spectral contribution from 

the crystalline component of the specimen. Because of the significant difference in 

the reorientational responses of polymer chains located in the crystalline and 

amorphous phase domains to a given dynamic deformation. 2D IR spectroscopy 

can easily differentiate IR bands associated with the two-phase domains, even 

though these bands are highly overlapped. 

Figure 1-17 shows an asynchronous 2D IR spectrum correlating the backbone of 

polyethylene segments located in either crystalline and amorphous domains and 

deuterium-substituted branches. Asynchronous cross peaks develop between the 

crystalline component of the symmetric CH2-stretching band of the polyethylene 

segments and bands for octene side branches. Such asynchronicity suggests short 

side branches can move independently of the polyethylene chain located in the 

crystalline phase of the sample. No noticeable asynchronicity. however, is observed

20 

I 

2840 

chain 

- 

Crystal 

line 

phous 

t .. 

Figure 1-17. An asynchronous 2D IR 

spectrum of linear low-density poly- 

~ 

ethylene comparing the strain-induced 

, , , I , , , , -2870 

50 2150 2050 local reorientational motions of the 

Wavenumber, VI main chain and side chain [46]. 

between the reorientation dynamics of side branch and amorphous components of 

the system. 

The above result reveals that the octene side branches are excluded from the 

crystalline lattice and are accumulating preferentially in the noncrystalline region of 

the polyethylene. This result agrees with the view that the depression of the melt 

temperature and crystallinity of linear low-density polyethylene is caused by the 

inability of the crystalline lattice to incorporate chain branches above a certain 

length. It is not possible, however, to determine conclusively from this 2D IR data 

if the short branches of linear low-density polyethylene are uniformly distributed 

within the amorphous phase domain, or preferentially accumulate near the interface 

between the amorphous and crystalline regions. 

The high-resolution capability of 2D IR spectroscopy to differentiate highly 

overlapped crystalline and amorphous IR bands has been successfully employed 

for the characterization of many other semicrystalline polymers. Biodegradable 

poly(hydroxyalkanoate)s, for example, were studied by this technique to show that 

molecular defects created by the incorporation of comonomer units tend to accumulate 

in the amorphous regions [49] in a manner similar to the case of linear lowdensity 

polyethylene. Additives, such as plasticizers, which are miscible in the 

molten state of semicrystalline polymers also tend to be excluded from the crystal 

lattice and preferentially accumulate in the amorphous phase once the polymer is 

brought down to a temperature below the melt temperature. 

1.5.3 Immiscible Polymer Blends 

Mixtures of polymers are often immiscible and spontaneously phase separate because 

of the much reduced entropy contribution to the free energy of mixing [50].


21 

/ I 

1430 

Figure 1-18. A synchronous 7D IR 

spectrum in the CH deformation and 

aromatic ring semicircle stretching 

vibration region of a blend of polystyrene 

and low-density polyethylene 

at room temperature [2]. 

I I I I I I I I 

1510 1470 14 

Wavenumber, V I 

c, 

> 

i 

n 

E 

1470 2 

a, 

2 

1510 

Such phase-separated polymer mixtures serve as an excellent model system where 

molecular-level interactions between components are small. The system studied 

here is a film made of an immiscible binary mixture of atactic polystyrene and lowdensity 

polyethylene (21. The dynamic IR measurement was carried out as before by 

mechanically stimulating the system at room temperature with a 23-Hz dynamic 

tensile strain with an amplitude of 0.1% The time-dependent fluctuations of IR 

absorbance induced by the strain were recorded at a spectral resolution of 4 cm-* . 

Figure 1-18 shows the synchronous 2D IR spectrum of the blend. Autopeaks 

observed on the diagonal positions of the spectrum near 1454 and 1495 cm-' 

represent the strain-induced local reorientation of polystyrene phenyl rings. The 

1454-cm-] band also contains a contribution from CH2 deformation of the backbone 

of polystyrene [lo, 11, 511. A pair of intense cross peaks appear at the offdiagonal 

positions of the spectral plane near 1454 and 1495 cm-', indicating the 

existence of a strong synchronicity between the reorientation of transition dipoles 

associated with these two IR bands of polystyrene phenyl side groups. 

Similarly, autopeaks corresponding to dynamic IR intensity variations for the 

CH, deformations of the polyethylene are observed near 1466 and 1475 cm-'. 

These autopeaks arise from the reorientation of molecular chains in the amorphous 

and crystalline domains of polyethylene. A pair of cross peaks clearly correlate the 

two IR bands originating from the polyethylene component of this blend. It is important 

to note that there is little development of synchronous cross peaks correlating 

polystyrene bands to polyethylene bands. 

The lack of synchronous cross peaks between polystyrene and polyethylene bands 

indicates these polymers are reorienting independently of each other. Cross peaks 

appearing in the asynchronous spectrum (Figure 1-19) also verify the above conclusion. 

For an immiscible blend of polyethylene and polystyrene, where molecularlevel 

interactions between the phase-separated components are absent, the timedependent 

behavior of IR intensity fluctuations of one component of the sample


spectrum of a blend of polystyrene and 

1510 1470 1430 low-density polyethylene at room tern- 

Wavenurnber, v( 

perature [Z]. 

may become substantially different from those of the other. Therefore, 2D IR correlation 

spectroscopy can easily differentiate individual spectral contributions from 

each of the immiscible components. 

Another notable feature in Figure 1-19 is the development of asynchronous 

cross peaks correlating the 1459-cm-’ band attributed to the CHz-deformation 

of the polystyrene backbone to the 1454-cm-’ and 1495-cm-’ bands arising from 

semicircle-stretching vibrations of the phenyl side groups. The asyncronicity among 

IR signals from these polystyrene bands again reflects the difference in mobilities of 

backbone and side-group functionalities, as already discussed in Section 1.5.1. The 

asynchronicity expected between the crystalline and amorphous bands of the polyethylene 

component are not apparent due to the intense cross peaks associated with 

polystyrene. 

The assignments of the 1454-cm-’ and 1459-cm-’ bands can be independently 

verified by selectively substituting the hydrogen atoms of polystyrene backbone 

with deuterium atoms. The elimination of spectral contributions from the backbone 

of polystyrene leaves only the pure phenyl vibration spectrum in this region, as 

shown in Figure 1-20. The asynchronous 2D IR spectrum (Figure 1-19) differentiated 

the two IR bands even without deuterium substitution. This result clearly 

demonstrates the truly unique feature of 2D IR spectroscopy to enhance substantially 

the resolution of highly overlapped spectral regions by effectively spreading 

IR bands along the second dimension according to the characteristic reorientational 

response times of individual functional groups. 

1.5.4 Miscible Polymer Blends 

Certain classes of polymer blends are known to be truly miscible and form homogeneous, 

one-phase mixtures. A specific interaction at the submolecular level. e.g..

1.5 Appliccitioiis 23 

Methylene 

Figure 1-20. IR spectra of normal atactic polystyrene and 

backbone-deuterated polystyrene films. 

I I I I 

I 475 14: 5 

Wavenumber 

hydrogen bonding, is often involved in such blends to overcome the tendency 

to phase segregate. In those cases, 2D IR spectroscopy is especially suited for the 

investigation of the molecular origin of such specific interactions. One of the wellknown 

miscible systems, a blend of atactic polystyrene and poly(viny1 methyl ether) 

1191. was investigated using 2D IR to elucidate the molecular origin of the miscibility 

of these polymers [52. 531. 

Interestingly. polystyrene and poly(viny1 methyl ether) are very different polymers. 

Polystyrene is a hard hydrophobic plastic resin. while polyivinyl methyl ether) 

is a soft water-soluble polymer. It is quite surprising that such a pair of polymers 

could produce a molecularly mixed homogeneous one-phase system. The specific 

origin of the miscibility of this polymer pair is not well understood, although some 

level of interaction is speculated between the methoxyl groups of poly(viny1 methyl 

ether) and the phenyl groups of polystyrene [54]. 

To simplify the spectral assignment, polystyrene which had been selectively 

deuterium substituted to eliminate the backbone contribution (Figure 1-21) was 

blended with poly(viny1 methyl ether). Special attention was paid to the symmetric 

CHZ-stretching peak of the methoxyl group of the poly(viny1 methyl ether) component 

near 2820 cm-'. Figure 1-22 shows the asynchronous 2D IR spectrum of 

poly(viny1 methyl ether) in this region. The presence of a pair of asynchronous cross 

peaks at the spectral coordinates at 2825 and 2813 cm-' indicates there are two 

distinct populations of methoxyl groups belonging to the poly(viny1 methyl ether) 

component of this blend, each having a different IR absorption band. 

Figure 1-23 shows that the methoxyl band of poly(viny1 methyl ether) at 

28 13 cm-' is synchronously correlated with the IR bands located at 3055 cm-' and 

3024 cm-' (assigned to the phenyl side groups of the polystyrene component of this 

blend). The presence of such synchronous correlation cross peaks strongly suggests 

the existence of a specific interaction synchronizing the local motions of the 

methoxyl groups of polyivinyl methyl ether) with the polystyrene phenyl groups.

3200 2950 2700 

Figure 1-21. IR spectra of backbone-deuterated 

' ' ' ' ' 2950 I ' ' ' ' 27do polystyrene and poly(viny1 methyl ether) in the 

Wavenumber 

CH-stretching vibration region. 

3200 

2825 

PVME melhoxyl 

L 

...' 

- Figure 1-22. An asynchronous ZD IR 

spectrum for the methoxyl vibration of 

a blend of backbone-deuterated poly- 

I 

2850 

, I O I 

2820 

, I 2850 

I 

2790 styrene and poly(viny1 methyl ether) 

Wavenumber, v, [571. 

Such a specific interaction most likely involves the lone-pair of electrons on the 

methoxyl oxygen atom of poly(viny1 methyl ether) and n-electrons of the polystyrene 

phenyl groups [54]. 

Figure 1-23 indicates only one methoxyl band is interacting with the polystyrene 

phenyl group, even though there are two distinct IR contributions of poly(viny1 

methyl ether) methoxyl groups, as already shown in Figure 1-22. The other 

methoxyl band of poly(viny1 methyl ether) at 2824 cm-' is asynchronously correlated 

with the phenyl bands of the polystyrene component (Figure 1-24). This result 

suggests that while one component of polyivinyl methyl ether) methoxyl groups

1.5 Ajqdicrrtions 25 

3024 

Figure 1-23. A synchronous ZD IR 

spectrum of a blend comparing the 

reorientational motions of transition 

dipoles associated with the polystyrene 

phenyl and poly(viny1 methyl ether) 

methoxyl groups [57]. 

.A 

t 

2850 

075 3045 3015 

Wavenumber, v, 

I 

U 

/I 

I 

d3-PS phenyl 

I ' 2790 


spectrum of a blend comparing the 

reorientational motions of transition 

dipoles associated with the polystyrene 

phenyl and poly(viny1 methyl ether) 

methoxyl groups 1571. 

I 

represented by the 2813-cm-' band is strongly interacting with the polystyrene 

phenyl groups, the other component represented by the 2824 cm-' band is not. 

There are many other examples of miscible polymer blends arising from specific 

submolecular interactions of the components. The miscible polymer pair of polystyrene 

and poly(pheny1ene oxide). for example, was recently studied by the 2D FT- 

IR technique in an effort to understand the molecular origin of the miscibility of 

this system [55]. Specific interactions between polymers and various small molecules, 

such as additives and plasticizers, were also studied by 2D IR. The importance 

of localized submolecular level interactions among functional groups of 

mixture components was demonstrated.

1.5.5 Block Copolymers 

Block copolymers made of dissimilar polymer segments joined by a covalent bond 

are interesting systems to study by 2D IR spectroscopy. Because of the repulsive 

interaction between the dissimilar segments, block copolymers often phase separate 

at the submolecular scale to form microdomains. It has been speculated for some 

time [56, 571 that the composition of block segments comprising such microdomains 

does not change sharply at the domain interface. Instead, a diffuse interlayer is 

thought to exist where substantial mixing of dissimilar block segments actually takes 

place. Analysis based on 2D 1R was used to probe the existence of such an iiitrrphirse 

region between microdomains [58, 591. 

A thin solution-cast film of a diblock copolymer consisting of a polystyrene block 

(a,, 50,000) and a deuterium-substituted polyisoprene block (a,, 100,000 J was 

stimulated at room temperature with a 0.1%)-amplitude, 23-Hz dynamic tensile 

strain. In this composition range, diblock copolymers of styrene and isoprene produce 

a hexagonally packed regular array of rod-like polystyrene microdomains 

embedded in a continuous matrix of the polyisoprene block. The diameter of each 

polystyrene microdomain is well below 0.1 pm. 

Figure 1-25 shows the synchronous 2D IR spectrum of this block copolymer 

in the CH-stretching region. Since the polyisoprene block of this polymer is fully 

deuterium substituted, the responses shown in Figure 1-25 represent exclusively the 

contribution from the polystyrene segment. One may expect the effect of the applied 

mechanical perturbation on the molecular reorientation of the block copolymer will 

be felt predominantly by the rubbery polyisoprene segments, since the dispersed 

polystyrene microdomains are in the glassy state at room temperature. It is, therefore, 

rather surprising to find autopeaks indicating substantial molecular mobility 

of polystyrene segments under dynamic strain at this temperature. 

2800 

2975 

7 1 

I I 1 1 1 ’ 

i0 2975 2E 

Wavenurnber, vI 

I 

3150 

Figure 1-25. A synchronous 2D IR 

spectrum in the CH-stretching vibration 

region of a microphase-separated diblock 

copolymer consisting of polystyrene and 

perdeuterated polyisoprene segments.

1.5 Applicutions 27 

Furthermore. the sign of the synchronous cross peaks correlating the motion 

of side-group phenyl and backbone methylene bands is positive. As discussed in 

Section 1.5.1, the positive cross peaks indicate the polystyrene segments giving rise 

to these dynamic TR signals are in the rubbery state, even though the measurement 

was made at room temperature. This result may be interpreted as the polystyrene 

segments showing rubber-like behavior being located in the diffuse interlayer region 

between the microdomains. The glass-transition temperature of the polystyrene 

segments in this region is substantially depressed due to mixing with the polyisoprene 

segments. 

This hypothesis was later verified by selective deuterium substitution of part of 

the polystyrene segment [58, 591. Near the junction between the two blocks (most 

likely to be located at the interphase region), polystyrene segments exhibited rubbery 

behavior even at room temperature. This part of the polystyrene segment 

reoriented extensively for a given dynamic deformation and contributed strongly 

to the dynamic IR dichroism signals. On the other hand, the tail end of the polystyrene 

segment (distributed more uniformly within the polystyrene microdomain) 

remained predominantly glassy. The local submolecular mobility of the tail end of 

polystyrene was noticeably limited. 

1.5.6 Biopolymers 

2D IR spectroscopy has also been used extensively to study not only synthetic 

polymers but also biopolymers. The exceptionally high spectral resolution of 2D IR 

has provided a major advantage over conventional IR analyses plagued by heavily 

overlapped bands of biopolymers. For example, amide I and I1 regions of IR spectra 

of various proteins are thought to be composed of numerous overlapping bands, 

each representing different protein conformation, e.g., helix, sheet, and random 

coils. However, it is usually very difficult to separate individual spectral contributions. 

2D IR analysis of the protein component of human skin has been attempted 

in the past [3, 211. Asynchronous 2D IR spectra suggest IR bands assignable to 

different protein conformations can be readily differentiated by the development of 

asynchronous cross peaks. Synchronous 2D IR spectra, on the other hand, can 

correlate different IR bands belonging to similar protein conformations. Thus, 2D 

IR spectroscopy may be used systematically to map out the 1R band assignments of 

complex protein molecules. 

Similar 2D IR studies have been carried out for human hair keratin [44, 60, 61). 

Figure 1-26 shows the asynchronous 2D IR correlation spectrum of a protein film 

made from solubilized hair keratin [44]. In a normal one-dimensional 1R spectrum, 

individual band components are highly overlapped and completely obscured. The 

enhanced spectral resolution is quite apparent in the 2D IR spectrum. The position 

of many of the asynchronous 2D IR cross peaks are in good agreement with other 

band assignments and normal-mode analyses [62].

1' 

1656 

1661 

1669 

1679 

I I I I I I I I I 1 1788 

BE 1650 1608 

Wavenumber, u, 

Figure 1-26. An asynchronous ZD IR spectrum in the amide I vibration region of a film made of 

solubilized human hair keratin 1441. 

1.6 Further Extension of 2D Correlation Technique 

The basic concept of 2D IR spectroscopy based on the correlation analysis of 

perturbation-induced time-dependent fluctuations of IR intensities could be readily 

extended to other areas of polymers spectroscopy. The 2D correlation analysis has 

been successfully applied to the time-dependent variations of small angle X-ray 

scattering intensity measurement [4]. In this study, a small amplitude dynamic 

strain is applied a sheet of microphase-separated styrene-butadiene-styrene triblock 

copolymer sample. Intensity variation of scattered X-ray beam due to the 

strain-induced changes in the interdomain Bragg distances coupled with the reorientation 

of microdomain structures is analyzed by using the 2D correlation map. 

Similarly, the formal approach of 2D correlation analysis to time-dependent spectral 

intensity fluctuations has been extended to UV, Raman [63), and near-IR 

spectroscopy (641. There seems no intrinsic limitation to the application of this 

versatile technique in polymer spectroscopy. 

The introduction of the generalized 2D correlation method [8] has made it possible 

to further extend the utility of 2D correlation analysis beyond the study of 

simple sinusoidally varying dynamic spectral signals as originally described in Eqs. 

(1-2) and (1-3). By utilizing the time-domain Fourier transform of spectral intensity 

variations, a formal definition of 2D correlation spectra resembling Eq. (1-9) can be

1.7 Coizclirsions 29 

derived in a straightforward manner. Using this generalized method, the 2D correlation 

analysis of any time-dependent spectral signals can be carried out. Highly 

nonlinear time-dependent reactions of crosslinking polymer systems, for example, 

are followed by the time-resolved IR measurement [65, 661. The resulting sets of 

time-dependent IR spectra can be transformed into 2D correlation spectra to yield 

valuable information to probe details of curing reactions of polymers. 

One of the very promising new developments in the recent 2D correlation analysis 

of polymers is the application of this versatile technique to the study of spectral 

changes recorded as functions of a general physical variable, which is no longer 

limited to time. For example, it is possible to apply the 2D correlation analysis 

to the near-IR spectral changes corresponding to the development of disorder in 

a hydrogen-bonded polyamide system induced by the rising temperature [67]. Of 

course, spectral changes of polymers induced by other variables, such as pressure, 

age, composition, and the like, may be analyzed in the same way. 


2D IR spectroscopy based on the correlation analysis of the individual timedependent 

behavior of localized characteristic reorientational motions of various 

submolecular moieties comprising a system has been shown to be very useful for a 

broad range of applications in the study of complex polymeric materials. In 2D IR, 

the spectral resolution is substantially enhanced by spreading the overlapped IR 

bands along the second dimension. The presence or lack of chemical interactions or 

connectivity among functional groups located in various parts of the polymer system 

are detected. The relative reorientation directions and the order of realignment 

sequence of submolecular units are also provided. 

Notable discoveries made through the use of 2D IR spectroscopy include the fact 

that local dynamics of side groups can proceed independent of the polymer main 

chain, and there can be abrupt changes in the side group realignment mechanism 

above and below the glass transition temperature. It is possible to probe the niicroscopic 

spatial distribution of submolecular components of polymers, especially in 

phase-separated systems such as semicrystalline polymers and block copolymers, 

and the degree of interactions between various components can be estimated. 

Specific interactions between Components in polymer mixtures are quite effectively 

identified. The application of 2D IR spectroscopy certainly is not limited to the 

characterization of traditional synthetic polymers. Complex macromolecules of 

biological origin can also be readily studied by this technique. The recent extension 

of the 2D correlation concept to (1) spectroscopic study of polymers using probes 

other than IR, (2) nonsinusoidally varying spectral changes, and 13) dependence 

on physical variables other than time has greatly expanded the scope of possible 

spectroscopic applications of this technique to the study of polymeric materials.

30 I T~.vo-Dii7iensioiiLiI Infmr-ed (2D IR) Spectroscopji 

1.8 Symbols and Abbreviations 

A 

All 

AL 

A 0 

p2 

D 

Dm 

DIRLD 

T 

t 

AA 

AAW 

AA 

AA 

AA’ 

AA” 

0 

X 

Y 

CI 

P 

I 

& 

& 

V 

8 

T 

a) 

0 

Absorbance 

Parallel absorbance 

Perpendicular absorbance 

Structural absorbance =All + 2Al 

Orientation factor 

Dichroic ratio = Ai,/Al 

Ultimate dichroic ratio 

Dynamic infrared 1inea.r dichroism 

Observation period 

Time 

Dichroic difference 5 All - A1 

Ultimate dichroic difference 

Dynamic dichroic difference 

Amplitude of dynamic dichroic difference 

In-phase dynamic dichroic difference E AA cos P 

Quadrature dynamic dichroic difference = AA sin p 

Synchronous correlation intensity 

Two-dimensional cross-correlation function 

Asynchronous correlation intensity 

Transition dipole orientation angle 

Dynamic dichroism phase angle 

Dynamic strain 

Dynamic strain amplitude 

Wa ven urn ber 

Polymer chain segment orientation angle 

Correlation time 

Frequency of dynamic strain / dynamic dichroism 

Spatial average 

1.9 References 

[l] Noda, I. Bd. Am. Plzys. SOC. 1986, 31, 520. 

[2] Noda, I. .I. Am. Cliem. SOC. 1989, Ill, 8116. 

[3] Noda, I. Appl. Spectrox. 1990. 44, 550. 

[4] Noda, I. Clierrifracfs-l~fflcromof. Ed 1990, I , 89. 

151 Palmer, R. A.; Manning, C. J.; Chao. J. L.; Noda, 1.; Dowrey, A. E.; Marcott, C. Appl. 

SiJeCtrO,SC. 1991, 45, 12. 

161 Marcott, C.; Dowrey, A. E.; Noda, I. Aridj;t. C/zem. 1994, 66, 1065A. 

[7] Marcott, C.; Dowi-ey, A. E.: Noda. 1. A~JJ/. Spectrosc. 1993, 47, 1324. 

[8] Noda, I. Appl. Spectrosc,. 1993, 47, 1329.

1.9 References 31 

[9] Colthup, N. B.; Daly, L. H.; Wilberley, S. E. Introduction to Infrared urzd Ramon Spectrosc,oipv: 

Academic: New York, 1975. 

[ 101 Zbinden. R. Itfrored Spectroscopy of High Polymers; Academic Press: New York, 1964. 

[I 11 Painter, P. C.; Coleman, M. M.: Koenig, J. L. The Theory uf Vibrationnl Spectroscopj~ rind Its 

AjJjdircitiori to Po/j]rireric Materials: Wiley: New York, 1982. 

[ 121 Koenig, 1. L. Sj~ei-tro~co/i.y of Po/yners; American Chemical Society: Washington, D.C., 1992. 

[I31 Aue, W. P.: Bartholdi, E.; Ernst, R. R.; J. Cheni. Pl~ys. 1976, 64, 2229. 

1141 Nagayama. K.; Kumar. A.; Wiithrich, K.; Ernst, R. R. J. Muyn. Rcson. 1980, 40, 321. 

1151 Ernst, R. R.: Bodenhausen, G.; Wokaun, A. Princ@les of Nriclecil- Muyi7rtic Resonunce in One 

and Two Dimensions; University Press: Oxford, 1987. 

[16] Kessler, H.; Gehrke. M.; Griesinger, C. Angew. Clzem. Int. Ed. Engl. 1988, 37, 490. 

[17] Bovey, F. A. Po/j777. Eny. Sci. 1986, 26, 1419. 

[18] Bliimich, B.; Spiess. H. W. Angew. Chem. Znt. Ed. Engl. 1988, 27, 1655. 

[19] Mirau, P.; Tanaka, H.: Bovey, F. Marromolecules 1988, 21, 2929. 

[20] Noda, I.; Dowrey. A. E.: Marcolt, C. J. Polym. Sci., Polym. Lett. Ed. 1983, 31. 99. 

[21] Palmer, R. A,; Manning, C. J.; Chao, J. L.; Noda, I.; Dowrey, A. E.: Marcott, C. Appl. 

Spectrosc. 1991, 45, 12. 

[22] Meier, R. J. Macroniol. Syinp. 1997, 119, 25. 

[23] Noda, I.; Dowrey, A. E.: Marcott, C. Appl. Spectrosc. 1988, 42, 203. 

[24] Noda, I.; Dowrey, A. E.; Marcott, C. J. Molec. Struct. 1990, 224, 265. 

[25] Noda, I.; Dowrey, A. E.: Marcott, C. Polyrn. Neivs 1993, 18, 167. 

[26] Ferry, J. D. I/iscoela.rtic Properties of'Polwiers; 3rd ed.: Wiley: New York, 1980. 

[37] McCrum, N. G.; Read, B. E.; Williams. G. Anelastic and Dielectric Effects in Polymer Solids; 

Wiley: New York, 1967. 

[28] Fraser, R. D. B. J. Chi. Plrys. 1953, 31, 1511. 

[29] Fraser, R. D. B. J. Chern. Phys. 1958, 28, 1113. 

[30] Stein, R. S. J. Appl. Phys. 1961, 32, 1280. 

[31] Gregoriou, V. G.; Chao, J. L.; Toriumi, H.; Marcott, C.; Noda, I.; Palmer, R. A. SPIE 1991, 

1575, 209. 

[32] Gregoriou. V. G.; Chao, J. L.; Toriumi, H.; Palmer, R. A. Chi. Pl7ys. Lett. 1991, 175, 491, 

[33] Nakano, T.; Yokoyama, T.; Toriumi, H. Appl. Spectrosc. 1993, 47, 1354. 

[34] Chazalviel, J.-N.: Dubin, V. M.; Mandal, K. C.; Ozaman, F. Appl. Sj~ertrosr. 1993, 47, 1411. 

[35] Budevska, B. 0.; Griffiths, P. R.; Analyt. Chem. 1993, 65, 2963. 

[36] Dittmar, R. M.; Chao, J. L.; Palmer, R. A,, in: Photoacoustic and Pliototherni~il Phenomena 

III: Bicanic, 0. (Ed.). Springer: Berlin, 1992; pp, 492-496. 

[37] Marcott, C.; Story, G. M.; Dowrey, A. E.; Reeder, R. C.; Noda, I., Mikrochirn. Acta [Sicppl.] 

1997, 14, 157. 

[38] Noda, I.; Dowrey, A. E.: Marcott, C. Mikvochim. Acta [Wien] 1988, I, 101. 

[39] Noda, I.; Dowrey, A. E.; Marcott, C.; Polym. Prep. 1992, 33(1), 70. 

[40] Noda, I.; Dowrey, A. E.; Marcott, C.; Makromol. Chem., Macromol. Synp 1993, 72, 121. 

[41] Noda, I.; Dowrey, A. E.; Marcott, C.; J. Appl. Polyni Sci.: Appl. Polynz. Svnip. 1993, 52, 55. 

[42] I. Noda, Polym. Prep Jclpan. Engl. Ed. 1990, 39(2), 1620. 

[43] Noda, I.: Dowrey, A. E.: Marcott, C. Po[vm. Pvepr,. 1990, 31(1), 576. 

[44] Noda, I.; Dowrey, A. E.; Marcott, C. in Time-Resoloed Vihrational Spectroscopy V; Takahashi, 

H. Ed.; Springer-Verlag: 1992: pp. 331-334. 

[45] Noda, I.; Dowrey. A. E.; Marcott, C. J. Molec. Stl-rict. 1990, 234, 265. 

[46] Stein, R. S.; Satkowski, M. M.; Noda, 1. in Polyri7er Blends. Solrifiorrs, and Interfirces, Noda, I.: 

D. N. Rubingh, Eds.; Elsevier, New York: 1992; pp. 109-131. 

[47] Gregoriou, V. G.; Noda, I.; Dowrey, A. E.; Marcott,C.; Chao, J. L.; Palmer, R. A.; J. Polym. 

Sci: Part B: Polym. Phys., 1993, 31, 1769. 

[48] Palmer, R. A,: Manning, C. J.; Rzepicla, J. A.; Widder, J. M.; Thomas. P. J.; Chao, J. L.; 

Marcott, C.; Noda, I. SPIE 1989, 1145, 277. 

[49] Marcott, C.; Noda, I.: Dowrey A. E. Analytica Chim. Acta 1991, 250, 131. 

[SO] Flory, P. J. Princip1c.s of Polymer. Chemistry; Cornell University Press: Ithnca. 1953. 

[51] Snyder, R. W.; Painter. P. C. Polymer 1981, 22, 1633. 

[52] Noda, I.; Dowrey, A. E.; Marcott, C. Oyo Buturi 1991, 60, 1039.

32 1 T~iio-DinieiisioiiiIInfrared (20 IR) Spectroscopy 

[53] Satkowski, M. M.; Grothaus, J. T.; Smith, S. D.; Ashraf, A,; Marcott, C.; Dowrey, A. E.; 

Noda, I. in Polymer Sohitions, Blends. and Interfaces, Noda, I.; Rubingh, D. N., Eds.; Elsevier: 

Amsterdam, 1992; pp. 89-108. 

[54] Garcia, D. J. Polynz. Sci., Polym. P11y.r. Ed 1984, 22, 107. 

[SS] Palmer, R. A.; Gregoriou, V. G.; Chao, J. L. Po/ym. Prep.. 1992, 33111, 1222. 

[56] Learry, D. F.: Williams, M. C.: J. Polvin. Sci.. Po/p. Phys. Ed 1974, 12, 265. 

1571 Hashimoto, T.; Fujimura, M.; Kawai, H. M~~cronio/rcn/es 1980, 13, 1660. 

[58] Noda, I.; Smith, S. D.; Dowrey, A. E.; Grothaus, A. E.; Marcott, C. Mur. Res. Soc. Sjv7p. 

Proc. 1990, 171, 117. 

[59] Smith, S. D.; Noda, I.; Marcott, C.; Dowrey, A. E. in Polymer Solutions, Blrnd~s, and hiterfrees, 

Noda, I.; Rubingh! D. N., Eds.; Elsevier: Amsterdam, 1992; pp. 43-64. 

[60] Dowrey, A. E.; Hillebrand, G. G.; Noda, I. Marcott, C. SPIE 1989, 1145, 156. 

[61] Noda, I.; Dowrey, A. E.; Marcott, C. Polym Prep. 1991, 32(3). 671. 

[62] Byler, D. M.; Susi, H. Biopolynzen 1986, 25, 469. 

[63] Ebihara, K.; Takahashi, H.; Noda, I. AppL Spectrosc. 1993, 47, 1343. 

[64] Noda, I.; Story, G. M.; Dowrey, A. E.; Reeder, R. C.; Marcott, C. Mncroniol. Symp. 1997, 

119, 1. 

[65] Nakano, T.: Shimada, S.; Saitoh, R.; Noda, I. Appl. Spccrrosc. 1993, 47, 1337. 

[66] Shoda, K.; Katagiri, G.; Ishida, H.; Noda, I. SPIE 1994, 2089, 254. 

[67] Ozaki, Y.; Liu, Y.; Noda, I. Macroiizolec~iles 1997, 30, 2391.

2 Segmental Mobility of Liquid Crystals and 

Liquid-Crystalline Polymers under External 

Fields: Characterization by Fourier-tr ansform 

Infrared Polarization Spectroscopy 

H. W Siesler, I. Zebger, Ch. Kuliniia. S. Okretic, S. Shilov and U. Hoffnuinn 


Liquid crystals and maia-cliain/side-chain liquid-crystalline polymers have gained 

scientific and technical importance due to their applications as display materials, 

their exceptional mechanical and thermal properties, and their prospective applications 

for optical information storage and nonlinear optics, respectively [ 1-13]. 

Among other aspects, the study of segmental mobility as a function of an external 

(electric, electromagnetic, or mechanical) perturbation is of basic interest for the 

understanding of the dynamics of molecular processes involved in technical applications 

of liquid-crystalline materials. 

Fourier-transform infrared (FTIR) polarization spectroscopy has proved an 

extremely powerful technique to characterize the segmental mobility of specific 

functionalities of a polymer or a liquid-crystal(1ine polymer) under the influence of 

an external perturbation [ 14-23]. In a side-chain liquid-crystalline polymer 

(SCLCP) aligned due to the influence of an external field, for example, the orientation 

of the polymer backbone relative to the rigid mesogeii is strongly influenced by 

the coupling via a more or less flexible spacer (Figure 2-1). Thus, the availability 

of characteristic absorption bands for these individual functionalities will allow 

a quantitative characterization of their relative alignment and contribute towards 

a better understanding of their structure-property correlation. Based on this 

knowledge, tailor-made systems can then be synthesized by a systematic variation 

of partial structures. 

In what follows, different phenomena will be discussed for selected classes of 

liquid-crystalline compounds: 

structure-dependent alignment of side-chain liquid-crystalline polyacrylates on 

anisotropic surfaces 

electric field-induced orientation and relaxation of a liquid crystal, a side-chain 

liquid-crystalline polymer and a liquid-crystalline guest-host system 

reorientation dynamics of a ferroelectric side-chain liquid-crystalline polymer in 

a polarity-switched electric field 

alignment of side-chain liquid-crystalline polyesters under laser irradiation 

shear-induced orientation in chrial smectic liquid-crystalline copolysiloxanes

igid inesugen 

mail1 chain 

Figure 2-1. Schematic structure of a side-chain liquid-crystalline 

polymer. 

. orientation mechanism of a liquid single crystal elastomer during cyclic deformation 

and recovery. 

Static and time-resolved spectroscopic data acquired both, by the conventional 

rapid-scan and by the novel step-scan technique will be presented and interpreted in 

terms of their individual applications. 

2.2 Measurement Techniques and Instrumentation 

Since the mid-l970s, FTIR spectroscopy has been widely used for polymer analytical 

applications and for studies of time-dependent phenomena in chemical and 

physical processes [ 141. With the current iinproveinent in electronics and data 

acquisition procedures, the time resolution for such investigations with an acceptable 

signal/noise-ratio is presently about 20 111s. Time resolutions far beyond this 

value are not accessible by the conventional rapid-scan technique with continuous 

mirror moveinelit without additional interferogram data manipulation. Thus, for 

time-resolved, rapid-scan measurements, generally, the retardation time of the 

moving mirror has to be considerably shorter than the time duration of the investigated 

event (Figure 2-2a). In search of new approaches, stroboscopic techniques 

have been proposed, where interferograms taken in the rapid-scan mode are subsequently 

subjected to a sampling of reordered interferograin segments and have 

thus led to an improvement in time resolution by a factor of about 500 [24, 251. 

A further improvement in time resolution to the submicrosecond level has been 

achieved with the step-scan technique 116, 26, 271. Here, the mirror is moved stepwise 

and the optical retardation is held constant during sampling of the individual 

interferogram elements. Thus, the time resolution is only limited by the response of

SRMPLE/DETECTOR 

e3 

‘retardation 

2.3 Theory 

Throughout this chapter the general form of the order parameter S will be used 

[3, 3 1 j to describe the long-range orientational order of the investigated liquidcrystalline 

materials. In contrast to the theory of Maiei- and Saupe 132, 331, which 

considers the uniaxial distribution of the mesogens in a nematic phase relative to the 

macroscopic preferential direction (director), this formalism also allows negative 

values. The relation of S to the inclination angle 0 of the inesogenic/structural-unit 

axis and the reference direction (if not otherwise stated, the external field direction) 

is outlined for the case of an SCLCP in Figure 2-3. Also shown are the values of 

S for parallel and perpendicular orientation of the mesogens and their isotropic 

distribution. The derivation of S, which corresponds to the orientation factor f in 

general polynier orientation terminology [14, 17-20], is given in Figure 2-3, where 

R is the dichroic ratio determined from the experimental absorbance values measured 

with light polarized parallel and perpendicular, respectively, to the chosen 

reference direction. Ro is the theoretical dichroic ratio for perfect aligiiinent of the 

mesogen, and p is the angle between the transition moment of the absorbing group 

and the local chain axis (in the present case the long axis of the mesogenic group) 

[14j. The structural absorbance Ao has been chosen as parameter which represents 

the intensity of an absorption band exclusive of contributions from the orientation 

of the investigated polymer. It is given by [14]: 

Both, order parameter S and structural absorbance A0 are based on the assumption 

of a uniaxial orientation syniinetry [14, 18-20]. 

external field 

(referencc direction) 

s= I 

(e =or) 

parallel orientation 

a 

1 

S=-- (0=900) 

2 

perpendicular orientation 

s=o 

isotropic distribution 

Figure 2-3. Geometrical considerations for the 

derivation of the order parameter S of a side-chain 

liquid-crystalline polymer under the influence of an 

exterilal field (ff = inclination angle of the inesogen 

with the reference axis, v, = angle between the 

transition moment of the absorbing group and the 

local chain axis, Rn = theoretical dichroic ratio for 

a perfectly aligned structural unit).

2.4 Srvuc me-Dependenr Alignment 37 

In the case of the laser irradiation experiments (see Section 2.6), the polarization 

direction of the incident laser beam was chosen as reference direction. Contrary to 

the former publications we have changed the reference direction for the calculation 

of the photo-induced anisotropy due to a better understanding of the underlying 

orientation mechanism. We assume that the mesogens are aligned preferentially 

perpendicular to the polrivizniion iiectov [34, 351 and not to a polnvizntion plane [15, 

361 of the irradiating laser. Thus, we obtain negative S-values, whereas for the lastmentioned 

case the order parameter values would be positive, according to the 

Maier-Saupe theory. 

2.4 Structure-Dependent Alignment of Side-Chain 

Liquid-Crystalline Polyacrylates on Anisotropic 

Surfaces 

For the characterization of the anisotropic properties of liquid-crystalline (LC) 

materials it is necessary to induce a well-defined orientation of the mesogenic 

groups. Although spontaneous orientation phenomena can be regularly observed in 

limited domains, the necessary prerequisite for the application of LC-materials as 

optical construction elements, for example, is the preparation of uniform textures 

over larger areas. Such structures can be prepared by the alignment of LC-systems 

on anisotropic surfaces (e.g., glass, polymer films). The origin and theory of the 

interaction and the induced anisotropy has been treated in several books and 

articles [37-441. Here, the attention is primarily directed towards the influence of 

the spacer length in a homologous series of nematic side-chain liquid-crystalline 

polyacrylates on their alignment on an anisotropic polyimide surface. 

2.4.1 Materials and Experimental 

The polymers had a molecular weight M, (determined by gel permeation chromatography) 

of about 5000 [45] and their structural formula is shown alongside their 

nematic-isotropic transition temperatures (determined by polarization microscopy) 

[45] in Figure 2-4. 

For the preparation of the substrate surface, KBr-plates have been spin-coated 

with a thin layer of polyiinide (polyimide-kit ZLI 2650, Merck, Darinstadt, Germany). 

Upon drying and curing this coating to the final thickness of about 1 pm, the 

surface anisotropy was induced by unidirectional rubbing with a polymethacrylimide 

hardfoam-roll. The LC-polymers were sandwiched between the SUrfdCetreated 

KBr-plates above their clearing temperatures and finally annealed 3 "C 

below these temperatures for several hours. FTIR polarization spectra with a resolution 

of 4 cm-' were recorded by accumulating 20 scans.

main chain 

rt- 

YH2 

spacer rnesogeri 

+A 0 

H-C-C-O-CH2 ) -O-O-$-O-O-CN 

95 Figure 2-4. Structure and transition temperatures 

of the investigated side-chain liquid-crystalline 

polyacrylates. 

2.4.2 Results 

Despite identical sample pretreatment, gradually decreasing dichroic effects could 

be observed for the LC-polymers with decreasing spacer length. The FTIR polarization 

spectra of tlie P6CN- and P2CN-systen1, measured with radiation polarized 

parallel and perpendicular to the rubbing direction of the polyimide surface, 

respectively, are shown in Figure 2-5. From these data the coupling effect of the 

spacer between the main chain and the mesogen is clearly evident. Thus, the hesamethylene-spacer 

almost completely decouples the mesogen from the main chain 

and allows the induction of an extremely high orientation of the rigid mesogen 

(S(v(C3N)) = 0.62) (see also insert in Figure 2-5). The short diniethylene-spacer, 

on the other hand, strongly links the mesogen to the polymer backbone and no 

orientation of the mesogen on the anisotropic surface can be observed. The spectra 

of the system with the tetramethylene-spacer (not shown here) display intermediate 

effects (S(i(C-N) = 0.46). A detailed analysis of the dichroic effects in the polarization 

spectra of the whole series in terms of the relative orientation of the mesogen, 

spacer and main chain has been given in [36]. 

2.5 Electric Field-Induced Orientation and Relaxation 

of Liquid-Crystalline Systems 

The reorientation of liquid crystals in an external electric field is the basis of their 

technical application in liquid-crystal displays (LCDs) and has been treated in 

numerous books and reviews [I-3, 12, 38,46-501. This section is intended to denionstrate 

tlie application of time-resolved FTIR spectroscopy for a better understanding 

of tlie electro-optical performance of liquid-crystalline systems. Due to tlie 

reversible character of the orientational and relaxational motions of liquid crystals 

under tlie perturbation of an electric field, the time-resolution of the FTIR measurements 

can be extended down to the microsecond range by the implementation 

of tlie step-scan technique (see Section 2.2). Thur, an insight on ;I molecular level

2.5 Electric Field-Innductid Orirntcitioii 39 

f 1 

P6CN 

0 

I 

Figure 2-5. FTIR polarization spectra of the P6CN- and P2CN-system ithe insert schematically 

shows the aligniiieiit of the mesogens on the anisotropic surface; parallel and perpendicular refers to 

the polarization of the IR radiation relative to the rubbing direction). 

can be provided into the relevant mechanisms. In what follows. the application of 

vibrational spectroscopy to the investigation of different phenomena will be trcated 

separately. On the one hand we will discuss the electric field-induced changes of 

nematic LC-systems (low-molecular weight compounds, side-chain licluici-crystallirie

40 2 Segnwital Mobility of Liquid Crystuls aid Liquidcrystalline Polymevs 

6CPe 

0 

C H 3 - ( C H z ) e C - O D C = N 

5 

44.5°C 48.5OC 

K - N - I 

NLCP 

- 0 

2, H -&-O-(CHZ)~$$C O a C - N 

-. 

CH2 

-n 3S°C 130°C 

G - N - I 

Figure 2-6. Sti ucture aiid transition temperatures 

of the investigated low- aiid highmolecular 

weight LC systems I6CPB and 

NLCP). 

polymers and liquid-crystalline guest-host system) in switch-on/switch-off experiments, 

and on the other hand we will describe the structural consequences incurred 

by a ferroelectric side-chain polymer in a poled electric field. 

Nematic liquid-crystalline molecules, which consist of a rigid inesogenic unit with 

an attached flexible part (sequence of methylene units, for example) (Figure 2-6), 

are among the most characteristic thermotropic systems. FTIR spectroscopy has 

been applied to study the mechanisiii of orientation of the liquid crystals in an 

electric field; however, at present the experimental data are not always unequivocal, 

and conclusions obtained are not in complete agreement with each other. T~us, 

some authors did not detect any differences in orientational rates of the rigid and 

the flexible part of the LC molecule [16, 51-54], which is in accordance with the 

commonly accepted mechanism of the orientation of an LC as a cooperative 

motion of a molecular ensemble [l, 2, 381. Other authors reported, that the motion 

of the flexible part precedes that of the inesogen in both orientation and relaxation 

[29, 55-57], and some investigators suggested that the motion of the mesogen precedes 

the motion of the flexible part [58, 591. With respect to a more detailed exploitation 

of the spectroscopic data in terms of the synchronization of the segmental 

motions of different functionalities of the investigated liquid-crystalline systems, 

2D-correlation analysis [ 16, 30, 60, 611 has proved an extremely valuable tool. In 

view of the detailed discussion of this mathematical technique in Chapter 1 of this 

book; however, this topic will not be treated here, although several results extracted 

from the presented spectroscopic data are based on the application of the 2Dcorrelation 

formalism. 

2.5.1 Materials 

To study the addressed phenomena, a commercial sample of p-cyanophenyl p-nhexylbenzoate 

(hereafter abbreviated as 6CPB) (Figure 2-6) was obtained from 

Roche (Basel, Switzerland), and used without further purification [53, 541. The 

transition temperatures of this LC are also presented in Figure 2-6. It should be 

mentioned, that the nematic phase will be retained during cooling down to 7 "C 

below the crystal-nematic transition point. To demonstrate the reduction of

2.5 Electric Field-nclticetl Orientntioi? 41 

mobility in an external electric field when the mesogen is attached to a polymeric 

backbone, the nematic liquid-crystalline side-chain polymer (NLCP), whose sidechain 

is largely identical to the structure of the 6CPB (Figure 2-6), was investigated 

under analogous experimental conditions. 

2.5.2 Experimental 

The schematic construction of the measurement cell and the principles of operation 

are outlined in Figure 2-7. For the construction of this cell, two Ge-plates (approximately 

10 x 10 x 2 mm3) are used as IR-transparent window material and electrodes. 

The distance between the two Ge-plates is ensured by thin stripes of 

poly(ethyleneterephtha1ate) film (7 pm) acting as spacer. To avoid the occurrence of 

interference fringes in the spectra, the precise parallel aligninelit of the Ge-windows 

was intentionally disturbed by two supplementary spacers of polycarbonate film (2 

pnij on one side of this assembly (Figure 2-7a). To fill the assembled cell, the LCsample 

was heated above the nematic-isotropic transition temperature and then 

introduced by capillary forces. For the preparation of a prealigiied nionodoniain- 

LC, the inner surface of each Ge-plate was covered with a thin layer of polyimide 

and then rubbed in one direction with polyamide cloth (see also Section 2.4.1). This 

pretreatment induced a homogeneous orientation of the investigated LC-molecules 

or mesogenic functionalities (in the case of the polymer) along the rubbing direction 

of the surface (Figure 2-7b). Upon application of an electrical voltage across the 

Ge-plates, the liquid crystals rotate with the long axes of their niesogens into the 

direction of the applied field (Figure 2-7c). Switching off the voltage leads to a 

relaxational motion back to the original state. The accompanying intensity changes 

were monitored with IR-radiation polarized parallel to the rubbing direction of the 

Ge-windows (Figure 2-7b). A home-built heating cell has been used to keep the 

investigated sample at a constant temperature within -1- 0.1 "C during the experiment 

(Figure 3-8). 

The principle of the electronic set-up for the step-scan measurements, and the 

method of data acquisition relative to the experimental time-scale, are shown in 

Figures 2.9 and 2.10, respectively. When the scanning mirror has stepped to a definite 

retardation point, a control electronic unit coupled with a sine-wave generator 

sends a 6 Vpp (peak-to-peak height) pulse with a frequency of 10 kHz and a definite 

length (25 nis for the experiments presented here) to the sample cell to induce the 

reorientation of the LC. After this period the voltage is switched off during 225 ms 

to allow relaxation to the original state. Simultaneously to the application of the 

electric field, the same electronic unit sends a trigger pulse to the spectrometer to 

start the data collection. The data are sampled by the internal analog-to-digital 

converter (ADC) of the instrument during 50 ms with 0.05 ms time resolution. 

During this time interval 1000 data points are acquired for one fixed retardation 

(the applied data-acquisition scheme is imposed by the capacity of the computer). 

This cycle is repeated several times (five times for the presented results on 6CPB) 

and the data points corresponding to the same time in the period of the event are 

averaged to improve the signal-to-noise ratio. When the data collection at a given

42 2 Segnien frrl Mobility of Liquid Cyystrils crrzd Liqziirl-Cr~~stallii~e Polyrnel-s 

GERMAHIUM WINDOU (ELECTRODE) 

I \ 

A 

a 

I \ 

I N I 

bIWFRARED BEAH 

DIRECTION 

' 

,- 

b 

POLARIZED INFRARED BEAM 

A 

C 

h 

I NFRRRED' BERM 

AND ELECTRIC FIELD DIRECTION 

ED 

Figure 2-7. Electric field-induced orientation 

of a liquid crystal (a) Construction 

of the meaauiement cell (b) Hornogeiieous 

al~gninent of the LC between the 

dnisotromc electrode surfaces before 

application of the electric field. 

(c) Hoineotropic alignment of the LC 

during application of the electric field. 

retardation is finished, the mirror is moved to the next retardation point (Figure 

2-2b). 

After a settling time of 40 ms, the experiment is repeated as described above. For 

the presented experiments all data were collected symmetrically around the centerburst 

of the interferogram. A total number of 2368 mirror positions were scanned, 

leading to a spectral resolution of 8 c1n-l. After completing the data collection at 

each retardation point, the data were resorted to interferograiiis on the time scale 

(Figure 2-2b) and transformed to the corresponding spectra. According to this

a 

-@ 

COWER 

SRHPLE 

._ 

2.5 Elec tric Fielrl-hiclirc.td Orieri tci tiori 43 

CORXIRL HERIIHG UIRE 

INPRRRED BERM 

STEEL CORE UITH SRMPLE COHPRRTMENI 

CELL / 

TEFLOH IHSULRTION 

'i 

GERMANIUM 

I , 

INFRARED BEAM 

b -. , ,,__ TEFLON JACKET 

Figure 2-8. Variable-temperature cell utilized for the spectroscopic characterization of the electric 

field-induced alignment of liquid crystals. (a) Complete assembly of the accessory. (b) Expanded 

ciew of the sample cell. 

procedure, 1000 spectra with a time resolution of 0.05 ms were processed. A further 

improvement of the signal-to-noise-ratio was obtained by averaging 10 spectra 

recorded in 0.5-nis intervals, starting with the time corresponding to the onset of 

applying the electric field. Thus, finally 100 spectra with a time resolution of 0.5 ms 

were obtained. The data collection for the whole experiment required approximately 

50 minutes. 

Due to the increase of viscosity in the polymeric NLCP system, the reorientation 

times varied between minutes and seconds, depending on the experimental temperature. 

Thus, the conventional rapid-scan technique (Figure 2-2a) was applied to 

follow the orientation and relaxation dynamics of the NLCP. In order to synchronize 

the application of the electric field with the data collection. a common switch 

was used to control the voltage [53].

44 2 Srgnientcrl Mobility of Liquid Crystcr1.Y mid Lipid-Cryctallirzc Polymer.\ 

TRIGGER PULSE 

SIHEWRUE 

GENERATOR 

SPEC TROMETER 

ELECTRONIC CONTROL ~ - - - l f l i , b 

___ SAMPLE CELL 

/I 

I 

Figure 2-9. Block diagram of the electronic measurement set-up. 

ORIENTATION 

PERIODICRL 

EVENT 

RELAXRTION 

TRIGGER 

PULSE (TTL) 

E 25 50 258 TIME/ms 

Figure 2-10. Time scale of the periodic event, control signal and data collection sequence for the 

step-scan measurements. 

2.5.3 Results and Discussion 

2.5.3.1 p-Cyanophenyl p-n-Hexylbenzoate (6CPB) in a Switching 

Experiment 

The FTIR polarization spectra of a prealigned 6CPB in the nematic phase are presented 

in Figure 2-11. The band assignment of the most prominent absorptions 

based on the spectral analysis of similar compounds [59, 551 is given in Table 2-1. 

The v(C-N) band (2230 cin-') and the two v(C=C)~~ stretching bands (1605 and 

1503 cm-') are polarized along the long axis of the mesogenic group and show 

parallel dichroisni with All > A1 (Ail and Al are the absorbance values measured

2.5 Electric Field-Itzduced Orientation 45 

Figure 2-11. FTIR polarization 

spectra of 6CPB (41 "C) taken 

with radiation polarized parallel 

(upper spectrum) and perpendicular 

(lower spectrum) to the 

rubbing direction of the polyimide 

surface. 

3300 3000 2750 2500 2250 2000 1750 1500 1250 1003 600 

Wavenuiber cm-' 

Table 2-1. Band assignment and waveiiuiiiber position of selected absorption bands for 6CPB (as = 

antisymmetric; s = symmetric; ar = aromatic; oop = out-of-plane). 

Band assignment Wavenumber (cm- ' ) 

2929 

2558 

2230 

1742 

1605 

1503 

1415 

1464 

1264 

1207- 101 7 

843 

757 

with light polarized parallel and perpendicular to the rubbing direction of the polyimide 

layer, respectively). This means, that the long axis of the mesogenic unit is 

oriented predominantly parallel to the rubbing direction (Figure 2-7b). The 757 

cni-' band (ring CH out-of-plane defoimation) exhibits a perpendicular dichroism 

with Al > A,,, since the transition moment for this band is perpendicular to the 

ring plane and hence, perpendicular to the long axis of the mesogen. These four 

bands are characteristic of the mesogenic units. For the characterization of the 

flexible part of the LC-molecule the 2929 cm-' and 2858 c1n-l bands (antisymmetric 

and symmetric v( CH2)-stretching vibrations) were evaluated. All the selected 

bands had absorbance values < 1 for the given cell thickness. 

When a voltage higher than the threshold value is applied to the electrodes, the 

LC molecules undergo a transition from the homogeneous (parallel to the rubbing 

direction) to the honieotropic (parallel to the electric field) orientation. According 

to our data for 6CPB, the threshold voltage to induce such a reorientation is

Figure 2-12. Stack-plot of spectra measured during the electric field-induced orirnration and duriiip 

part of the relaxation of thc 6CPB molecules (41 -C'. 

Figure 2-13. Relative absorbance/time-plot during orientation and relaxation of 6CPB (41 C! 

for selected absorption bands: (0) 2929 cni-'. (7) 2858 cm-', (012230 cin-l. I + ) 1605 cm-'. I x 

1503 cm-'. (+) 757 cni.~'. 

approximately 1.5 Vpp. In the homeotropic type of orientation, the long axis of the 

LC molecule is normal to the electrode surface and parallel to the IR beam propagation. 

According to geometrical considerations [62] a reorientation to this honieotropic 

state thus leads to intensity changes of the Ail-component only. This component 

asymptotically approaches the Al intensity value at sufficiently high fields. 

For this reason, all time-resolved measurements have been performed with light 

polarized parallel to the polyimide-layer rubbing direction. 

A stack-plot of time-resolved spectra recorded during the honiogeneous-homeotropic 

transition and part of the relaxation process is presented in Figure 2-12. A 

good signal-to-noise-ratio is observed for all spectral regions (except when the 

absorbance exceeds a value of 2). Relative absorbance changes A, for selected 

absorption bands were calculated from: 

where A, is the absorbance at time t and A0 is the absorbance just before reorientation. 

These relative intensity changes of selected absorption bands are presented 

in Figure 2-13 and remarkable changes are observed only after a definite induction

2.5 Electric Field-IwrElrced Orientation 47 

period of several milliseconds after application of the electric field. This phenomenon 

is well described in the literature [46]. The intensities of the 2230, 1605 and 

1503 cni-' bands decrease, whereas that foi- the 757 cm-' band increases M,hen the 

field is switched on. This indicates that the mesogens 1-eorient with their long axes in 

the direction of the applied field (Figure 3-7c). The process of orientation at the 

given experimental conditions takes about 25 ms. After switching off the field, thz 

relaxation to the initial homogeneous state (Figure 2-7b) takes place. The intensity 

changes of the 2929 and 2858 cnir' bands reflect the orientation and relaxation 

process of the flexible part of the LC molecule. For an exact quantitative comparison 

of the orientational and relaxational rates of the rigid and flexible segments of 

the LC molecule, the order parameters of the different bands had to be calculated 

as a function of time. However, this is possible only for the case of a pure homogeneous 

orientation (it., the director of the LC monodomain and, hence) the axis of 

uniaxial orientation is perpendicular to the IR beam propagation) or for the case of 

a pure homeotropic orientation (where the axis of uniaxial orientation is parallel 

to the beam propagation). In intermediate states the orientation of the director is 

not constant. Thus, near the surface, due to strong interactions, the molecules are 

oriented perpendicular to the electric field whereas at the center of the cell they are 

oriented parallel to the electric-field direction [38]. For a qualitative compai-ison 

of the orientational rates of the rigid and the flexible parts of the investigated LC 

molecule a normalized intensity A,, was utilized which has been calculated from: 

where At is the band peak absorbance at time t, A0 is the value before the npplication 

of the electric field, and A,,, is the corresponding absorbance value at t = 25 

ms. The result for the 2929 and 2230 cm-* bands are presented for a temperature of 

41 "C in Figure 2-14a. According to these data there are no detectable differences in 

A, both, during the orientation and relaxation process, for the rigid and the flexible 

segments of the LC molecule. This means that under the given experimental conditions 

and with the evaluatioii by the usual one-dimensional technique, the different 

segments of the LC molecule orient and relax at comparable rates. The same 

conclusion can be derived from Figure 2-14b where the corresponding A,,-values 

are presented for 45°C. Due to the small temperature interval, the diRerence in 

orientational behavior is not significant. However, it is not possible to perform 

the experiment in a wider temperature range because of the narrow temperature 

interval of the investigated mesophase. 

The results of similar rates for the reorientation of the flexible and rigid segments 

of the 6CPB molecule in an electric field are in accordance with the coninioiily 

accepted mechaiiisni of orientation by a cooperative motion of the LC molecules [ l~ 

2, 381. However, in recent investigations it has been reported, that the response of 

the mesogens and the flexible part on the applied field may be different [29, 55-59]. 

If these observations are correct they could be the basis of a principally new 

description of LC orientation on a molecular level. Two explanations for thc Jisagreement 

of o~iresults with these views can be given:

48 2 Segliientril Mobility qf Liquid Crystals arid Liquid-Cr.~~stiillirie Polymers 

ORIENTATION c+ 

RELAXATION 

---------- 

I0 20 30 40 5n 

TIME lms 

n 10 213 310 40 50 

TlMElms 

Figure 2-14. Normalized absorbaiice/time-plot during orientation and relaxation of 6CPB at two 

different temperatures for selected absoiptioii bands: (0) 2929 cm-', (0) 2230 cin-'. 

1. The effect of different response of the flexible and the rigid part of the LC 

molecule towards application of an electric field is not general. This is supported 

by results [56], where this phenomenon is detected under certain experimental 

conditions only. Moreover, up to now most experiments have been performed 

on one LC (4-n-pentyl-4'-cyanobiphenyl, 5CB) only, which has a different 

structure compared to the LC molecule investigated here. 

2. The effect is too small to be detected under the given experimental conditions 

with conventional data analysis. To detect this effect, Urano et al. [55, 561 have 

applied tiine-resolved spectroscopy by using a modified dispersive instrument. 

With this instrument, absorbance changes down to lop6 a.u. have been detected. 

However, the time requirement for the experiment in the step-scan inode 

depends on the period of the event (0.25 s in our case). To reach such a high 

signal-to-noise-ratio we had to average the data of many more events at the 

individual retardation points. For example, the data collection for 2368 retardation 

points over 100 events in each point (compared with our five events) would 

require approximately 17 h of experimental time. Due to uncontrolled changes 

in the environmental conditions and drifts of electronics during such a time 

period, the main requirement of the step-scan technique, i.e., the reproducibility 

of the periodical event, could be violated. 

To extract more information, the data were processed with the newly developed 

method of two-dimensional spectroscopy [60, 6 11. This technique is especially useful 

for the detection of small differences in the orientational rates of different molecular

2.5 Electric Field-Im/irced Orientutioii 49 

I 

- 1 ' I I I I 

32OU 3000 2750 2500 2250 2000 1750 1500 1250 I000 750 500 

WAVENUMBER / cm' 

Figure 2-15. FTIR polarization spectra of NLCP (25 "C) taken with radiation polarized parallel 

and perpendicular to the rubbing direction. 

groups. With the development of this formalism it is now possible to construct 2D 

correlation plots based on conventional time-resolved spectral data. The 3D analysis 

reveals [ 161, that there is a strong correlation in the synchronous spectrum and 

no appreciable asynchronicity in the asynchronous spectrum between the niesogen 

bands and the absorptions of the hydrocarbon chain (see Table 2-1). This means, 

that the mesogen and the flexible segment of the liquid crystal reorient synchronously. 

The strong correlation between the inesogen bands themselves indicates 

that the mesogen reorients as a rigid core. No differences in 2D correlation spectra 

between the orientation and relaxation processes were found. 

2.5.3.2 Nematic Side-Chain Liquid-Crystalline Polymer in a 

Switching Experiment 

The FTIR polarization spectra of the prealigned NLCP are presented in Figure 

2-15. Due to the almost identical structure of the side-chain compared with the lowmolecular 

weight liquid crystal (6CPB), most of the functionality-specific band 

assignments of Table 2-1 are also valid for the NLCP. The 1603 c1n-l and 1503 

cm-I (phenyl C=C-stretching vibrations) bands and the 2230 cm-' (v(C_N)) band 

are polarized along the long axis of the mesogenic functionality and show parallel 

dichroisni with All > AL. This means, that the long axis of the mesogenic unit is 

predominantly oriented parallel to the rubbing direction. The 757 cni- ' bund (ring 

CH out-of-plane deformation) exhibits a perpendicular dichroisni with Al > All 

because the transition moment for this band is perpendicular to the ring plane and 

hence, perpendicular to the long axis of the mesogen. Since the transition moment

50 

w 

CJ 1.50 

z 

3 1.00 

2 

s 

0. so 

w 

5 0.00 

4 

g -0.50 

ORIENTATION 

RELAXATION 

0 25 SO 75 100 0 100 200 300 400 

TIME / s 

TIME / s 

$ 1.50 

z 

4 2.00 

f 0. 50 

* 

W 

2 0.00 

k- 

4 

Fl -O. 50 

0.0 I. 5 3.0 4. 5 0 

TIME / s 

5 10 1s 20 

TIME / s 

Figure 2-16. Relative absorbance/time-plot during orientation and relaxation of NLCP at different 

temperatures for selected absorption bands: (0) 2926 cm-', (0) 2229 cm-', (*) 757 cm-I. 

for this band is also nearly prependicular to the traiis-polymethylene chain, this 

indicates that the spacer is oriented preferably parallel to the rubbing direction of 

the polyimide support and the mesogen. We used the 2130 and the 757 an-' bands 

for the characterization of the movement of the mesogen and the 2916 cm-' band 

for that of the spacer. Unfortunately, the bands which characterize the backbone 

are heavily overlapped with absorptions of the spacer and they are much less 

intense than the corresponding spacer bands (a monomeric unit contains one 

methylene group in the iiiaiii chain und six in the spacer). Therefore, in the case 

of this polymer, we cannot follow the inoveiiient of the backbone. Owing to the 

experimental geometry, the reorientation of NLCP led only to changes in intensity 

of light polarized parallel to the rubbing direction. For this reason, as before, we 

used only polarized radiation to monitor the molecular motion of the NLCP. 

The relative intensity changes during orientation and relaxation at different temperatures 

for selected absorptions bands are shown in Figure 2-16. As can be seen 

from this figure, the orientation and relaxation rates incrcased more than 20-fold 

when the temperature increased from 80 to 120 "C. At temperatures close to the 

glass transition, the orientation times increase to a few tens of minutes. This is 

apparently related to the decrease of viscosity with increasing temperature. At each 

temperature the relaxation takes more time than the reoreination. The A, values 

change less at higher temperatures than at lower temperatures (Figure 3-16). This is 

the consequense of the decrease in the order parameter of the NLCP with increasing 

temperature. The values of A, for different absorption bands depend on the angle

2.5 Electric Field-Induced Orientlition 51 

Figure 2-17. Normalized 

absorbaace/time-plot during 

orientation and relaxation of 

NLCP at dif'rerent temperatures 

for selected absorption 

bands: (x) 2926 ciii-l, i+) 

2229 cm-' . 

0 1 2 3 4 5 0 5 1 0 

m / r m / r 

between their transition moment and the molecular axis of the LC molecule. In 

order to coinpare the rates of molecular motion for the differents segments, the 

normalized absorbance A, was used. The A, values for bands characterizing the 

niesogen and the flexible part are shown in Figure 2-17. As is evident from this 

figure, no significant differences of their intensities can be detected during the orientation 

and relaxation process. Thus, it must be assumed, that the mesogen and 

the flexible part reorient as a rigid unit and the application of the electric field does 

not lead to conformational changes in the flexible spacer. This conclusion, based on 

the direct monitoring of the movement of different molecular segments, is in agreement 

with the results of [63] derived from the measurement of rotational viscosity of 

the same NLCP in a magnetic field and from theoretical modeling. However, it has 

been reported that the application of the electric field led to a decrease of trans 

conformers in a flexible spacer of an NLCP similar to that studied here. This conclusion 

was based on the analysis of the 1490-1410 cni-' region which is sensitive 

to the conformation of the polymethylene sequence. As has been shown recently 

for a set of side-chain liquid crystalline polymers with different length of the polymethylene 

spacer, this region also contains absorption bands of the aromatic ring 

[64]. In our opinion, the conformational analysis based on the absorbance in this 

region is thus very complex and ambiguous. 

If the main chain takes part in the reorientation, it is expected that its motion will 

be, at least partially, shifted in phase with respect to the side chain (spacer and 

rnesogen). A detailed 2D analysis [30] in the region of the in(CH2) bands (2936 and 

2863 cni-' ) does not reveal any significant asynchronicity. Hence, we may conclude 

that the main chain does not take part in the reorientation, or it moves only slightly.

Figure 2-18. Scheinatic representation 

of the segmental 

motions of the NLCP during 

the switching process. 

There is a strong correlation in the synchronous spectrum and no appreciable 

asynchronicity in the asynchronous spectrum between the mesogen band (2230 

cm-') and the v(CH2) bands (2926 and 2863 cm-I). The main part of the intensity 

changes in the region of 2850-2950 cm-' is due to the spacer; thus, on the basis of 

the 2D results we may draw the conclusion that the spacer and the mesogen reorient 

simultaneously. 

From a further detailed 2D analysis of selected absorption band intensities of the 

corresponding power spectrum (16, 60, 611 it was finally concluded that only part of 

the spacer of the NLCP takes part in the reorientation. This result is summarized 

graphically in Figure 2- 18, where the orientational behavior of a NLCP-mesogen 

during the switching process is symbolized schematically relative to the entire 

polymeric structure, including the spacer and the main chain. 

2.5.4 Nematic Liquid-Crystalline Guest-Host System in a 

Switching Experiment 

Generally, the phenomenon of dissolving and aligning a molecule or a group of 

molecules, such as dyes or probes, for example, by a liquid crystal can be called the 

guest-host phenomenon [65, 661. The host liquid crystal can be a single compound 

or a multicomponent mixture. Depending on its structural geometry (e.g., elongated 

or spherical) the guest molecule couples more or less significantly to the anisotropic 

intermolecular interaction field of the liquid-crystalline host. The liquid-crystalline 

solutions can be easily oriented by electric, magnetic, surface, or mechanical forces, 

resulting in highly oriented solvent and solute molecules. This phenomenon provides 

the basis for the application of liquid crystals as anisotropic solvents for 

spectroscopic investigations of anisotropic molecular properties (67, 681. Actually, 

in a broad sense. most of the commercially available liquid-crystal mixtures may be 

regarded as being related to the guest-host effect, as they incorporate some nonmesomorphic 

molecules to generate the desired electro-optical effects in the mixtures 

(69-711. In the present section we present an example of how a guest molecule 

is oriented anisotropically in a liquid-crystalline solution which is exposed to an

2.5 Electric Field-Induced Orientatioii 53 

GUEST-HOST SYSTEM 

ZLI - 1695 

Figure 2-19. Chemical structures and 

transition temperatures of the investigated 

nematic liquid-crystalline 

guest-host system. 

H 

I 

N 

2-NAPHTHALDEHYDE 

GCIEST 

external electric field in a switching experiment. Time-resolved step-scan FTIRspectroscopy 

has been applied to study how a solute follows the reorientation 

motion of the nematic solvent [72]. 

2.5.4.1 Materials 

A 10% solution of 2-naphthaldehyde in the nematic liquid-crystal mixture 4-cyano- 

4'-alkylbicyclohexyls (ZLI-1695, Merck) (Figure 2- 19) was prepared for the measurements. 

This nematic phase is very convenient as an anisotropic solvent because 

of its weak IR spectrum, allowing one to obtain polarization spectra in practically 

the whole mid-IR region. 

2.5.4.2 Experimental 

For the switching experiment, the cell described in Figures 2-7 and 2-8 was utilized 

(here the path length was 10 pin). The spectral resolution was 16 cni-' and a function 

generator supplied the electrical signal (Figures 2-9 and 2-10) needed to orient 

the liquid-crystalline sample (16V,,, 1 kHz, 20 ms on, 260 ms off). The positive 

edge of this signal also synchronized the start of the data acqusition at each step 

during the complete step-scanning experiment. The time resolution of the experiment 

was 100 ps; each spectrum contains coadded data of two orientation/relaxation 

cycles. 

The reorientation behavior of the nematic solvent was determined by the v(C-N) 

and the v,(CH?) absorption bands at 2239 and 2857 cm-', respectively, whereas the 

v(C=O) absorption at 1698 cm-' was utilized to characterize the segmental motion 

of the solute (Figure 2-20). 

2.5.4.3 Results and Discussion 

The step-scan spectra of the studied solution, sampled at 0.0 and 19.5 ms, are pi-esented 

in Figure 2-20. The observed intensity changes correspond to the reorientation 

of the solute and solvent molecules from their preferential alignment in the 

rubbing direction (0.0 ms) into the applied electric field axis i 19.5 nisi. The detailed 

relative intensity changes of the three references bands versus time after the electric 

field was switched off are shown in Figure 2-21. Here, the lower values for the rela-

54 2 Segmental Mobility qf Liquid Cyvstrrls cmd Licliiid-Cvystalliiie Po1jwer.r 

I 

I 

3000 2400 is00 1200 600 

WAVENUMBER /[ cm-'] 

Figure 2-20. Step-scan FTIR spectra of the investigated nematic liquid-crystalline guest-host 

solution before (0.0 ms) and after (19.5 ms) reorientation in the electric field. 

0. 

L 

2 0. 

4 

B 

I 

0 10 20 30 40 so 

Figure 2-21. Relative absorbance changes versus time of the v,(CH?) (0) and v(C=N) fV) nbsorption 

bands of the solvent and the solute-specific v(C=O) (0) band during orientation and relaxation 

of the individual segments. 

tive absorbance of the v(C-N) relative to the v(C=O) band are due to two effects: 

(1) better initial orienation in the rubbing direction; and (7) the more eficient 

reorientation of the C-N functionality into the electric field direction. 

The positive values for the relative absorbance of the vs(CH2) absorption reflect 

the preferential perpendicular alignment of the transition moiiieiit direction of this 

band relative to the long axis of the nematic solvent molecule. A recent study 1291

2.5 Electric Field-Induced Orienfation 55 

n 10 rn 30 40 50 

TIME / ms 

Figure 2-22. Normalized absorbance versus time during orientation and relaxation of the solventand 

solute-specific functionalities (see Figure 2-21). 

on the nematic liquid-crystal 4-cyano-4'-pentylbiphenyl (often denoted as 5CB) 

by means of time-resolved FTIR spectroscopy under the conditions of dynamic 

reorientation give evidence for different reorientation times of its non-ridgidly 

bonded fragments. 

Although Figure 2-2 1 shows that neither the orientation nor the relaxation of the 

solute and solvent molecules is complete under the experimental conditions, we 

expected that the delay phenomena between the solvent and solute molecules would 

be clearly expressed in a normalized absorbance versus time plot. This plot, however, 

as shown in Figure 2-22, does not provide any indications of such a behavior. 

In othei- words, the reorientation rates are equal for the solvent and the solute. 

Without drawing any general conclusions from a single example, the present results 

show that one may expect similar behavior in other nematic solutions. 

2.5.5 Reorientation Dynamics of a Ferroelectric Side-Chain 

Liquid-Crystalline Polymer in a Polarity-Switched 

Electric Field 

The surface-stabilized ferroelectric liquid crystals in the smectic C* (SmC*) phase 

are among the most interesting types of liquid-crystalline systems because of their 

potential applications in high-resolution flat panel displays and fast electro-optical 

devices [73-761. Within this class of compounds, ferroelectric liquid-crystalline polymers 

(FLCPs) have gained theoretical and practical interest as systems which combine 

the properties of polymers and ferroelectric liquid crystals. This combination is 

achieved by attaching the ferroelectric mesogen to a main chain via a flexible spacer

56 2 Segrnental Mohilitj~ of Liquid Crystals arid Lirjuid-C~.vstallirie Polymers 

(mole ratio 0.95 : 0.05 : 2.70) 

C=O N 

I II 

? v 

o y 0 I 0 

8 "C 55 'C 73 'C 

G-S:e---,S,c---*I 

Figure 2-23. Chemical structure, 

coniposition and transition 

temperatures of the side-chain 

FLCP under investigation. 

[77, 781. Since the spacer is effectively decoupling the movement of the mesogen 

froni the polymer backbone, a different behavior of mesogen, spacer, and main 

chain can be expected in a reorientation process. So far, most of the research on the 

mobility of FLCPs has been limited to the determination of relaxation times by 

using dielectric spectroscopy [79] or response times by using electro-optical switching 

studies [80] without differentiation of the motions of individual functionalities 

in the macromolecule and only few time-resolved vibrational spectroscopic measurements 

are available on this subject [81-831. However, the movements of the 

individual segments control the reorientation process and hence, the study of their 

dynamics is a key to understand the behavior of an FLCP in an electric field. Based 

on the recent advances in data-acquisition techniques [ 161 we have attempted in 

the present work to follow the reorientation dynamics of the different segments of 

an FLCP from one stable state to another on the submillisecond time scale during 

poling of an electric field [83]. 


The chemical structure, composition and the transition temperatures of the investigated 

side-chain FLCP are shown in Figure 2-23. The molecular weight of this 

polymer is approximately 18800 and the synthesis and characterization have been 

described in detail elsewhere [84]. 


The variable-temperature cell was the same as described in Section 2.5.2, the only 

difference being that a spacer with a uniform thickness of 5 pm was used between 

the electrodes. The empty cell and the FLCP were heated to 90 "C to introduce the

2.5 Elrc tric Field-Iiidircwl Orim tri tiori 57 

INFRARED BEAM 

Figure 2-24. Polai-ization geometry and 

mesogen alignment of the investigated 

FLCP in the polarity-switched electric field 

IA: polarity +. B: polarity -J. 

POLARIZATION 

polymer into the cell through the gap between the electrodes by capillary forces. 

After filling the cell, it was slowly cooled (O.S"C/min) and an AC-voltage of 30 

Vpp, IOHz, was applied during cooling with the aim to reduce defects in orientation. 

The FLCP was investigated in the SmC*-phase where two surface-stabilized 

orientational states exist, depending on whether the mesogens incline to the right or 

to the left relative to the rubbing direction of the polyimide support (Figure 2-24). 

The switching between these two states is possible by changing the polarity of the 

applied field. The propagation direction of the polarized 1R beam was perpendicular 

to the plane of the Ge-windows. The largest absorption intensity changes during 

the switching process were obtained by fixing the polarizer at 45" relative to the 

rubbing direction of the polyimide coating. All spectra were collected with a spectral 

resolution of 8 cm-' and a time resolution of 0.1 ms. For a more detailed 

characterization of the reorientation mechanism, the experiments were carried out 

at different temperatures (28 "C, 42 "C and 47 "C ( f 0.1 "C)). At the lower temperature 

only rapid-scan measurements were required during orientation at 20 V (DC), 

whereas at the two higher temperatures the step-scan technique had to be applied 

due to the much faster segmental movement. Here, for the switching between the 

two orientational states a bipolar rectangular signal with a voltage of 20 Vpp and a 

frequency of 20 Hz and 100 Hz was applied at 42 "C and 47 "C, respectively (Figure 

2-25). The spectral data collection was synchronized with the negative edge of the 

electric signal. 


Polarization spectra of the preoriented FLCP are shown in Figure 2-26. For the 

characterization of the mesogen and the flexible spacer, the 1607 cm-' ( Y(C=C),,) 

and 2926 cm-' ( vas(CH?)) absorptions, respectively, were chosen. Unfortunately, 

no bands free of overlap can be attributed to the backbone. The band at 1020 cm-' 

(if( Si-0-Si)) is strongly overlapped by the i!,(C-O-C) absorption of the ester group 

and the band near 807 cm-' is also superimposed in a rather complex manner. 

For the characterization of the methyl substituents of the siloxane-backbone, the 

h'(Si-CH3) vibration at 1260 cm-' was chosen, although it contains a contribution 

from the I*~,,(C-O-C) vibration related to the mesogen. The 1607 and 1260 cm-' 

bands have parallel dichroism and the 2926 cm-' band has perpendicular dichro-

58 2 Segnientrrl Mobility of Liqriid Crystals aid LiCIuid-Ci.vJtulliiie Po1jwiei.s 

PERIODICRL EUEHT .............. 

................ 

TRIGGER PULSE ................. 

DRTR RCOUISITION .1([1[([(/(1/1/ i- 

........-........ RESIILUTIoH z5us 

.................. 

COHTRDL SIGHRL ..............m 

ZOU,~, ioonz 

, "------, 

I. 0 

1 ' 1 . 1 1 . 1 I 

> 

-_---- __- -____ -.- 

Figure 2-25. Time scale of the 

periodical event. control signal and 

data collection sequence for the poling 

experiment of the FLCP (2OHz at 

0 10 28 I0 48 50 TIME/ns 42 "C, 100 Hz at 47°C). 

~ 

- - ~ 

POLARITY: @ t 47OC 

u 

3 1 2 

LL 

0 

!nn 

I 

n. 4 

1.6 

'3200 3flK 2800 2600 2400 2200 2000 1000 1600 1400 1200 1000 800 600 

WAVENUMBER / =in ' 

Figure 2-26. FTIR-polarization spectra of the investigated FLCP prealigned at 47 "C and different 

polarity of the electric field (radiation polarized at 45" to the rubbing direction according to the 

geometry of Figure 2-74). 

ism. Due to the perpendicular transition moment of the 1260 cm-' band relative to 

the local chain axis it can be concluded that in the FLCP the mesogen and the 

spacer are aligned preferentially parallel to each other but are perpendicular to the 

backbone [16, 53, 831. The sinusoidal changes observed in the baseline of the spectra 

are a consequence of the interference of the IR-radiation between the Ge-windows, 

but do not disturb the evaluation of the absorption bands in the present case. A 

stack-plot of time-resolved spectra recorded during several cycles of polarity 

changes is shown in Figure 2-27. A good signal-to-noise-ratio is observed for all 

spectral regions, except where the intensity exceeds an absorbance of 1.7. 

Normalized absorbance/time-plots of the selected absorption bands are plotted 

for experiments at different temperatures in Figure 2-28. As can be seen from this 

figure, the reorientation at 47 'C takes only approximately 5 ms compared with

2500 2008 I s00 

WRVtNdMRtR CM-I 

Figure 2-27. Spectroscopic response of the FLCP upon changing the polarity of the applied electric 

field (1: 2926 cm-I, 2: 2854 cni-', 3: 1607 cm-'. 4: 1507 cm-I, 5: 1260 cm-I). 

I. 25 

1.00 

0. i5 

0. 50 

0. 25 

n. 00 

-0. 25' 

0. 0 

w u 1.2s 

z 

c 1.00 

m 

Q 0.75 

2 0.50 

0. 25 

1.0 2. 0 3.0 4.0 5. 0 6 0 

Tinds 

i 

0.00 

-0.2s 

a -0.50 

p ! o 5 10 15 20 25 30 35 40 45 

0 Time/ms 

z 1 . z 

!. 00 

0. i s 

0. 50 

0.25 

n. 00 

0. 25 

01 

"%, 

5 ID 15 20 iE 

Timelms 

Figure 2-28. Normalized absorbance/time-plots for poling experiments of Ihe FLCP at different 

temperatures. (a) 28 "C, rapid-scan measurement, 20 V (DC): (b) 42 "C, step-scan measurement, 

20 V,,, 20Hz; (c) 47"C, step-scan measurement, 20 V,,, 100 Ha, data acquisition for 2.5 periods. 

(1260 cm-' (o), 1607 cm-' (+), 2926 cni-' (0)).

about 1.5 s for 28 "C. The shape of the response curves are similar and can be 

described as exponential. The reorientation starts without time delay when the 

polarity of the applied field is changed. This means, that there is no retardation 

in the response of different structural segments. No distinct differences have been 

found in the orientation rates of the flexible and the rigid part of the molecules 

of ferroelectric nonpolynieric LCs [53, 581. However, it has been shown that the 

molecular motion cannot be described by a single exponential process but consists 

primarily of a fast and a subsequent slow process [52]. This motion also includes 

confornia tioiial changes in the flexible part of the molecule. In the polymeric ferroelectric 

system under investigation, such conforniational changes may take place 

in the spacer and the main chain. Due to the lack of conformationally sensitive 

absorptions in the FLCP, we were not able to detect or differentiate such changes. 

Taking into account the assignment of the absorption bands under consideration 

and their dichroic changes during poling the electric field, we can conclude from 

Figure 2-28 [83] that not only the mesogen but also the spacer and the main chain 

(at least some molecular groups which are in the neighborhood of the spacer attachment 

to the backbone) take part in the reorientation process during poling the 

electric field. This picture, however, is not supported by the 2D-correlation analysis, 

which shows, that there is a strong correlation between the ring-stretching bands 

(1506, 1607 cm-') and the two carbonyl bands around 1750 cni-' [16]. This is evidence 

that the mesogen reorients as a rigid unit. The strong correlation in the synchronous 

spectrum and no appreciable asynchronicity in the asynchronous spectrum 

between the mesogen bands and the v(CH2) bands (2926 and 2855 cin-'), 

which are mainly due to the spacer, indicates that the mesogen aiid the spacer 

reorient synchronously. Figure 2-29 compares the power spectrum of the FLCP to 

the reference spectrum (average spectrum of a reorientation process). One can easily 

notice, that the intensities of the spacer bands in the power spectrum are considerably 

reduced relative to the mesogen bands. Based on these data, we have to coiiclude 

that only a small part of the spacer, directly attached to the mesogen, takes 

part in the reorientation. The 2D analysis does not supply any evidence that the 

main chain of the FLCP participates in the reorientation. Based on these data, the 

picture shown in Figure 2-30 can be derived for the response of the FLCP to a 

poled electric field. 

2.6 Alignment of Side-Chain Liquid-Crystalline 

Polyesters Under Laser Irradiation 

It has been demonstrated that the photo-induced orientation of dye-containing 

SCLCPs can be used for reversible optical data storage [85-901. FTIR polarization 

spectroscopy has been applied to monitor the orientation of the main chain and the 

mesogenic side-chains which are of critical importance in these storage processes 

[15, 91-94]. Here, some results obtained by laser irradiation of a homologous series

-3.6 Aligmient OJ' Side- Clinin Liyuicl- Cv!~stnlline Poljwstevs 61 

161 

WAVENUMBER I cmd 

14 - 

12 - 

10 - 

c 

c 

zi 

(0 8- 

6- 

Figure 2-29. Comparison of the reference (top) 

and power (bottom) spectra intensities of the 

FLCP for the reorientation process during 

poling the external electric field. 

4r 

1500 2000 2500 3000 

WAVENUMBER /at-' 

Figure 2-30. Schematic representation of the 

segmental motions of the FLCP during poling 

the electric field. 

POLARITY SWITCH 

RESPONSE 

of novel SCLC-polyesters [ 151 are presented. In unoriented polyester films with a 

dodecamethylene spacing of the ester groups in the main chain, holograms with a 

resolution of 6000 linesjmm can be recorded by the interference of an Argon-ion 

laser beam [95]. The thermal erasure and rewriting of the laser-induced alignment in 

the SCLC polyesters is the basis for such storage processes. In order to investigate 

the mechanism of the involved processes in more detail, a variable-temperature, 

computer-controlled sampling system for on-line polarization FTIR nieasurements 

during laser irradiation as well as thermal erasure has been designed and 

constructed.

62 2 Segniental Mobility of Lipid Crystals and Liquicl-Cvystalliiie Po1jwiei.s 

SIDE-CHAIN 

p 2 ' m = main-chain spacer length 

n = side-chain spacer length 

P 

"=: ESTERLINKAGE 

R = azobenzene substituent 

Design parameter 

a) variation of the length of the methylene units in 

main-chain rn = 4, 10, 11, 12 ,I4 

and 

side-chain n = 6, 8, 10 

b) variation of the azobenzene substituent R 

R=CN(a), NO, (b), OCH, (c), H(d) 

Figure 2-31. General structure 

and nomenclature (PnRm) of the 

investigated SCLC polyesters used 

for reversible optical data storage 

(e.g., P6a12: polyester with 12 

CH? groups in the acid part of the 

main chain and six CH2-groups 

in the side-chain spacer and a 

CN-suhstitucnt on the mesogen). 

Table 2-2. Phase behavior of P8a12 [96]. 

Structure 4 3 2 1 

Transition temperature range 30-40 "C 46-54 C 59-63 "C 65-71 "C 

2.6.1 Materials 

The variations in chemical structure of the investigated SCLC polyesters and their 

nomenclature is outlined in Figure 2-3 1. Differential-scaiining-calorimetry (DSC 1, 

polarization-optical-microscopy (POM) and X-ray investigations [95-971 reveal a 

rather complex phase behavior of this homologous polyester series, which strongly 

depends on the length (m) of the main-chain. For the polyadipates (m = 4), for 

example, an enantiotropic liquid-crystalline behavior is observed, involving disordered 

smectic mesophases in the range from T, at about 10°C to the sniectic tf 

isotropic transition (from about 55-70 "C, varying with the side chain spacer length 

n = 6,8, 10). The longer homologs, with a main chain length of 10-14 niethylene 

units could be classified as inonotropic LCPs, with polyinorphic character. Some 

of them exhibit up to four first-order transitions, which partially overlap, involving 

inany different organized phases. An exact phase assigninelit is very difficult. Candidates 

with the highest potential for optical data storage are the polytetradecanedioates 

(ni = 12) with 6 and 8 methylene groups in the spacer [95, 981. DSC and 

X-ray investigations reveal for both polyesters two low-order mesomorphic phases 

(structures 3 and 4) and two high-order mesomorphic or crystalline phases jstruclures 

I and 2). The glass-transition temperature of both systems is about 25°C. 

Representative of the discussed SCLC polyesters, the phase behavior of P8a12 is 

detailed in Table 2-2. 

The influence of azobenzene substituents with varying dipole moments has also

~ polyester 

2.6 Alignnient qf Side-Chaiii Liquid- Crystalline Polyrsters 63 

film 

polarization direction 

Figure 2-32. Simplified picture of the experimental 

set-up for laser-irradiation and symbolic 

representation of the motivation for the 

FTIR-spectroscopic detemiinatioii of the order argon.ion laser ,R beam 

parameters of the polyester functionalities. 

(488 nin) 

been studied, but no obvious iinproveinent of the data storage capability 1991 could 

be observed. 

Thin film of the polyesters were obtained by casting a chloroform solution 

(3mg/200 pl) of the polymer onto KBr plates with a 1.5 x 1Oinm' area. After 

evaporation of the solvent, the films were dried at room temperature under vacuum 

for 2 h. The film thickness obtained by this preparation procedure was approximately 

5 pni. 

2.6.2 Experimental 

Figure 2-32 provides a simplified picture of a laser irradiation/FTIR measurement 

experiment symbolizing the inherent problem to determine the laser-induced orientation 

of the different polymer segments. The detailed instrumental design of the 

apparatus used for these on-line ineasurements is given in Figure 2-33 and shows 

that the following experimental parameters can be varied: 

0 Based on an internally calibrated photo-diode, the laser intensity can be reduced 

from 800 mW/cin2 in a controlled fashion with a neutral-glass filter wheel. 

The irradiation time is adjusted by the shutter. 

The temperature of the sample is controlled by a Peltier element. 

The acquisition of the FTIR polarization spectra is performed by using the 

dedicated computer of the spectrometer, which is synchronized with the different 

devices. The laser beam passes through the shutter and the filter wheel and is 

directed via two mirrors onto the film sample, which is mounted on the Peltier 

heater/cooler arrangement. This set-up was used to follow the photo-induced 

anisotropy during and after irradiation as a function of time and temperature, as 

well as its thermal stability during erasing cycles. Time-resolved spectra with a 

resolution of 4 cin-' were taken with radiation polarized alternately parallel and

Figure 2-33. Instrumental details of the apparatus used for on-line FTIR measurements during 

laser irradiation. 

perpendicular to the argon-ion laser beam polarization direction, respectively. The 

irradiated sample area was about 3 mm’. 

2.6.3 Results and Discussion 

In the FTIR polarization spectra of the films prepared by casting from solution 

no dichroism could be detected before laser irradiation. Figure 2-34 shows the 

polarization spectra of a P8a12 polyester after irradiation with the 488 nm argonion 

laser line (laser power 800 mW/cm’, irradiation time 1000 s). A detailed assignment 

of the individual absorptions bands to specific structural units is given in 

Table 2-3. The observable dichroic effects clearly demonstrate the significant orientation 

induced in the different functionalities of the polymer. Thus, the mesogenic 

side groups are preferentially oriented perpendicular to the polarization direction 

of the laser beam. This picture is based on the 0-dichroism of the v(C-N) band and 

the aromatic ring modes (see also Table 2-3). Additionally, the 7c-dichroism of the 

v(CH2) absorptions indicates a perpendicular orientation of the side and main 

chains relative to the laser beam polarization direction, whereas the smaller K- 

dichroism of the ester linkage indicates a somewhat lower orientation. As the electromagnetic 

radiation influences only the orientation of the mesogenic groups with

I." 

0.9 

0.8 

Figure 2-34. FTlR 0.7 

polarization spectra of 

o, 0.6 

an irradiated P8a 12 0 

polyester (after 1000 s of 5 0.5 

irradiation with the 488 

0.4 

nm line of an at-gon-ion p 

laser at an intensity of a 0.3 

800 mW/cm' I. The IR 

0.2 

beam was polarized 

parallel and perpendicu- 0.1 

......... parallel 

- perpendicular 

lar to the polarization 0.0 

direction of the laser 3300 3000 2700 2400 2100 1800 1500 1200 900 600 

beam. 

Wavenumber I cm-1 

Table 2-3. Band assignment. wavenumber position and structural origin of selected absorptions of 

the polyester P8a12 (as = antisymmetric: s = symmetric; ar = aromatic: oop = out-of-plane). 

Band assignment Wavenumber Structural origin 

(cm-I) 

c+~(CH?) 

I's(CH2) 

v(C-N) 

I'(c=o) 

1*(C=C),, 

l'( C=C),, 

t ~ CH2) ( 

is,,( C-0- Ar) 

Various modes of coupled 

COOR group vibrations 

I*( N-Ar) 

11, (C-0- Ar) 

jjooP (CH I,, 

;',(CH?) 

2920 

2851 

1379 

1733 

1603 / 1583 

1501 

1403- 1355 

1257 

1201 . . . 1 I05 

1141 

1022 

839 

723 

main chain / spacer 


mesogen 

main chain / (ester linkage) 

mesogen 

mesogen 


mesogen / spacer 

main chain / (ester linkage) 

mesogen 

mesogen / spacer 

mesogen 


the azobenzene moieties. the alignment of the spacer and the main chain can only 

be induced by coupling effects. Investigations of polyesters with specifically deuterated 

aliphatic sites in the spacer and main chain demonstrate, that the perpendicular 

alignment of the acid-specific main chain relative to the laser polarization 

direction is even better than the alignment of the side-chain spacer units 1001. Thus, 

Figure 2-3% shows the region of the v(CH2) and v(CD2) polarization spectra of a 

P6a12 polyester with a completely deuterated main chain (except the CH-group

66 

i v(CHZ) 

fi 

0) 

m C 

* 

B 

P 

4 

a 

3000 2800 2600 2400 2200 2 

Wavenumbers (cm-1) 

' n 

v(C=N) 

I 

2300 2250 2200 2150 2100 2050 2000 

Wavenumbers (cm-q) 

30 

Figure 2-35. (a) FTIR polarization 

spectra of'the v(CH?), \s(C-N) and 

v(CD2) absorptions of a specifically 

denterated P6a13 (see text) after laser 

irradiation. (b) FTIR polarization spectra 

of the v(CD7) atid i(C=N) absorptions 

of a specifically deuterated 

P6a12 (see text) after laser ii-radiation. 

(IR radiation polarized parallel f. . .) 

and perpendicular (-i to the laser 

polarization). 

Laser polarization 

direction 

CN 

I 

Figure 2-36. Schematic model of the irradiation-induced segmental 

alignment in the polyesters relative to the polarization direction of the 

argon-ion laser beam. 

where the spacer merges into the main chain) and Figure 2-35b demonstrates the 

almost negligible dichroism of the v(CD*)-region of a polyester which has been 

deuterated in the alcoholic part of the main chain [loo]. On the basis of these 

experimental results, the model shown in Figure 2-36 can be derived for the laser-

induced alignment of the investigated polyester and the following scenario is suggested 

for optical storage: the film obtained by casting from solution consists of' 

macroscopically isotropic niicro-domains. The irradiation with linearly polarized 

laser light leads to a melting of these domains, caused by direct heating due to 

absorption or a deposition of heat through traris-cis-trms isonierization cycles of 

the azobenzeiie moieties [95]. The origin of the orientation effect is attributed to 

tivm,r-cis photoisomerization of the azo-group, followed by a c-is-trans thermal or 

photoisomerization. In the cis-form, this functionality is able to undergo some 

reorientation with respect to its original alignment. The absorption of the polarized 

'writing' beam will continue as long as a component of the electric dipole vector of 

the azo-group lies in the direction of the laser polarization. The trurzs-cis reorientation 

process will therefore continue until all the azo-groups have been oriented 

perpendicular to the polarization vector of the laser beam [86. 1011. In the polyesters 

investigated here, this orientation remains after the 'writing' beam is turned 

off and a long-tenn stable, macroscopic orientation is established. 

Apart from this long-term stability, the possibility to erase and reinduce this 

anisotropy in the polymer is a prerequisite for its application as a reversible optical 

data-storage medium. To demonstrate the loss of mesogen-alignment during heating 

and the perfect reversibility of a 'write-erase-write' cycle, Figure 2-37a shows a 

stack plot of the i(C-N)-band region in the polarization spectra of a P10a12 film 

during heating from 32 "C to 100 "C (isotropization temperature, 68 "C) and in 

Figure 2-37b the dichroism of the v(C-N)-band of the same polymer is shown 

(from left to right) after irradiation, upon heating the irradiated sample to 100°C 

and after reirradiation. The aliiiost perfect coincidence of repeated writing cycles 

is demonstrated in more detail in Figure 2-38a, which displays the simultaneous 

photo-induced alignment of the different segments during laser irradiation in an 

order parameter/tinie-plot. Figure 2-38b exhibits the intermediate thermal erasure 

of the first orientation as a function of temperature. The loss of orientation for the 

different polymeric segments is also coupled and takes place in a narrow teniperature 

interval. A comparison of Figure 2-38b with the transition temperature ranges 

of the individual phases in Table 2-2 shows that, during erasure, the anisotropy 

is mainly retained in the less ordered mesophases 3 and 4. Beyond 54 "C, only a 

small, residual anisotropy is observed due to recrystallization effects [96], which 

completely disappears at the clearing point of phase 2. 

2.7 Orientation of Liquid Crystals Under 

Mechanical Forces 

Apart from the orientation mechanisms discussed in the previous sections, a preferential 

alignment of liquid crystals or liquid-crystalline polymers may also bc 

induced by the application of mechanical forces. Depending on the investigated 

system, this can be either done by shearing between plates or in stress-strain niea-

100 “C 

2300 2200 2100 

walenumbers [cm-’] 

I00 “C 

2300 2200 2100 

wavenumbers [cm-11 

1 

e 

2 

0 0.20- 

z - 

4 

80 

wavenumbers [cm-l J 

Figure 2-37. Thermal erasure of a laser-induced alignment in a PlOal2 polyester demonstrated by 

the disappearance of the v(C=N) dichroism at the isotropization point (68°C) (a) and the effect of 

irradiation, thermal erasure and reirradiation on the v(C-N) dichroism of the same polymer (b). 

surements of bulk samples [36, 102-1041. Spectroscopic data on the orientation 

induced by these pretreatments may then be utilized for a better understanding of 

the different responses of specific functionalities to the external force field, or they 

can establish the basis for a correlation of the mechanical properties and the structure 

of the liquid-crystalline system.

2.7 Orientation oj Liquid Ci:i,stals Under Mechanicul Forces 69 

a 

b 

solid symbol 1st writing cycle, open symbol 6th wntlng cycle 

1 ' 1 9 I , I I I ' I 

0 200 400 600 800 1000 30 35 40 45 50 55 60 65 70 75 80 

Irradiation time Is Temperature I "C 

Figure 2-38. Reversibility of repeated write/erase-cycles demonstrated in time-resolved FTIR 

polarization measurements by the order parameter for the different segments of PSal?: (a) As a 

function of irradiation time (488 nm line of an argon-ion laser, at 30"C, SOOmW/cm') in the 1st 

and 6th cycle; and (b) Versus temperature during thermal erasure with a heating rate of 30 "C/niin. 

2.7.1 Shear-Induced Orientation in Side-Chain Liquid-Crystalline 

Copolysiloxanes 

As pointed out already in Section 2.5.5, low-molecular weight ferroelectric liquid 

crystals (FLCs) and FLCPs are attracting a lot of interest because of their potential 

for electro-optical applications. The polymers offer new possibilities, e.g., as elastomers 

for piezoelectric elements or by copolymerization [77, 78, 1051 due to the 

formation of intrinsic mixtures between SniCY' mesogenic units and other coinonomers. 

This leads to FLCPs combining several material properties which might be 

utilized for colored displays in the case of comonomers containing chi-omophores. 

For the differentiated evaluation of such copolymers with reference to the possible 

exploitation of nonlinear optical (NLO) properties, the interplay of the different orientation 

tendencies of the side-chain functionalities is of crucial importance [36, 1061. 


The copolymers presented here are copolysiloxanes whose liquid-crystalline phase 

behavior is characterized by the formation of a SmC* phase. By the additional

70 2 Segrnentrrl Mobility of Liquid Crystals and Liquirl-Ci.vstalliiit1 Po1jwiei.r 

r- 

-45 % -47 70 -8 % 

- [ .I-[ 7- f - .} 

4H2 41 

C=O 

c=o 

I 

?. 

H 3C-FH 

C=O 

I 

9 

CH2 -CH3 

chiral mesogen 

phase transition 

main chain + mesogen 

copolymer 

98 

(- 8”h) Figure 2-39. Chemical structure, coiiiposition 

and phase-transition temperatures of the 

different investigated copolymers. 

+ ChreNphore n 

introduction of chromophores with extended n-conjugation it was attempted to 

combine the ferroelectric properties of the mesomorphic phase with the NLOproperties 

of the chromophore. The synthesis has been described in detail 11061, and 

Figure 2-39 shows the structure and composition of the investigated homopolymer 

(containing no chromophore) and the modified copolymers alongside their corresponding 

phase-transition temperatures. 


Oriented polymeric films were obtained in a mechanical force field by applying 

a shear stress. The NaCl plates used as sample supports were coated with a thin 

polyiinide layer to improve the surface properties. The shearing process along a 

defined direction was first performed in the isotropic state above the clearing temperature 

and then at the isotropic/SmC*-transition. Afterwards, the temperature 

was gradually decreased to room temperature. The mechanical process was simultaneously 

analyzed using polarization optical microscopy. The experimental set-up 

is shown in Figure 2-40. For the subsequent dichroic measurements the samples 

were mounted in the spectrometer so that the shearing direction and the polarization 

direction of the incoming radiation were initially parallel. By rotating the

2.7 Orieritation of Liquid Ciysttrls Under Mechanictrl Forces 71 

Figure 2-40. Experimental set-up for the shearing experiment 

combined with POM iiieasureinents (A: microscope, 

B: goniometer, C: fixed NaCl crystal, D: shearing NaCl 

crystal, E: temperature control). 

polarizer for 90" the spectrum with perpendicular polarization relative to the 

shearing direction was recorded. 


The question to be clarified by the FTIR spectroscopic polarization studies was 

whether: 

the interaction between the polymer chain and the side-on linked mesogenic dye 

functionalities dominates and the chroniophores orient parallel to the polyi~ier 

chain [ 107, 1081 and thus perpendicular to the non-chroniophoric mesogens 

(Figure 2-41a); or 

the interaction between the different mesogens (non-chromophoric and chromophoric) 

is dominating and leads to their parallel alignment (perpendicular to the 

chain direction) (Figure 2-41b). 

In order to determine the orientational behavior of the main chain, the spacer and 

the mesogenic and chroniophoric side groups relative to the shearing direction, 

detailed band assignments of the monomers, homopolymer and copolymers preceded 

the dichroic measurements of the sheared polymers [36, 1061. Based on these 

investigations, the polarization spectra of the sheared copolymers I and I1 (Figure 

2-42) yielded the surprising qualitative result, that the orientation of the side-on 

fixed chroniophores is doiniiiated by the interaction with the mesogenic groups and 

leads to the structure represented in Figure 2-41b. Thus, the dichroisiii of specific 

absorptions indicates only low orientation effects for the main chain parallel to the 

shearing direction, whereas the mesogens and the spacers clearly orient preferentially 

perpendicular to the direction of shear. The o-dichroisni of the v,(NO?)- 

absorption (Figure 2-42) supports the perpendicular alignment of the chromophore 

long axis relative to the shearing direction, and thus the validity of the model shown 

in Figure 2-4 1 b [36]. These dye-containing copolymers luay therefore not be LIS~~LI~ 

for nonlinear optics [ 1061.

. I 

polymer chain 

0 0 

“I 

chromophore with 

a donor I acceptor System 

(“sideon” - Ilnk) 

k 

cn, 

/\ 

H,C 

chiral mesogen 

(“endon“ - Ilnk) 

chromophore 

3‘ 

..-. I 

. . 

’. 

, polymer chain 

’. . ... .. . 

.. ’.. 

spacer 

oriented 

mesogen 

Figure 2-41. Possible models of mesogen 

and chromophore alignment in the sheared 

FLC-copolymers. (a) The chromophore 

orients parallel to the local polymer chain 

axis and the shearing direction, whereas the 

mesogen aligns perpendicular to the chromophore. 

(b) The interaction between the 

chromophores and the mesogens dominates 

and induces a parallel alignment to each 

other and a perpendicular orientation to the 

shearing direction. 

2.7.2 Orientation Mechanism of Liquid Single Crystal Elastomers 

During Cyclic Deformation and Recovery 

The application of the previously discussed techniques to induce monodomain 

structures in side-chain liquid-crystalline polymers by the application of electric or 

electromagnetic fields, by shearing or on anisotropic surfaces, frequently leads to 

comparatively low, macroscopically uniform orientation. Additionally, the methods 

are limited to a sample thickness of about 100 pm. Liquid-crystalline side-chain 

elastomers do not have this restriction, because a high macroscopic orientation can 

be induced in polymeric networks by mechanical deformation up to a sample 

thickness of about a centimeter [103, 1091. The synthesis of such systems can be 

performed by crosslinking linear, side-chain liquid-crystalline polymers to networks 

[ 1 lo]. The inherent combination of rubber elasticity and liquid-crystalline phase 

behavior, may then be exploited for the induction of a macroscopic mesogen orientation 

by mechanical deformation. 

Liquid single crystal elastomers (LSCEs) are characterized by a high macroscopically 

uniform orientation in the liquid-crystalline phase without the require-

0 8 

18 

16 

1 

a 

W 

u 

z a 

m 

a 

0 

Ln 

m 

Q 

0 6 

0 4 

0 2 

0 01 

1 4 

1 2 

10' I 

0 0. 

a 6 

0 4 

0 2 

0 0 

, 

I 

WAVEPJUMBER / [ cm-'I 

b 

WAVENUMBER / I cm?I 

Figure 2-42. FTIR polarization spectra of the copolymers with chromophore I (a) and I1 (b) after 

shearing (IR-radiation polarized parallel (- - -) and perpendicular (-) to the shearing direction) (the 

arrow indicates the \),(NO?) absorption of the chromophore). 

ment of an external stabilizing (mechanical) field. A characteristic feature of this 

state of order is the optical transparency of the material. The optical behavior of 

such elastomers is comparable with organic or inorganic single crystals. Kiipfer and 

Finkelinam [ 1 1 11 have described the synthetic route for LSCEs with a permanent 

mesogen orientation. 


The LSCE system investigated in the present work is based 017 a methylsiloxane 

elastomer which has been crosslinked in a two-step procedure (Figure 2-43 j. The 

mesogenic side groups have been attached via a vinyl end-group (compound 2) and 

the crosslinking agents are a hydrochinonedialkylether with vinyl end-groups

74 2 Seyrneiitiil Mobility qf' Liquid Cvj~stul~ iiizd Liquid-Cryrtulliiie Polyriievs 

L)FFORRl.AI~ION 

-%I {E".l 

CROSSLlhKING UNDER - 

STRESS 

.Y LSCE 

\ 

CROSSLINKING WITHOUT 

DEFORMATION 

I>F.I'ORMATION 

- = 

1 REFERENCE ELASTOMER 

CH3 4 

g,n 3°C n,i 83OC LSCE 

g,n 2 OC n.i 81 'C REFERENCE ELASTOMER 

Figure 2-43. Cheinical structure, transition temperatures and schematic presentation of the crosslinking 

processes involved in the synthesis of the investigated LSCE and reference elastomer [l 1 11. 

(compound 3) and a vinyl- and niethacryloyl-substituted derivative (compound 4). 

which has also liquid-crystalline properties, The key point of the crosslinking reactions 

is the fact that, under the experimental conditions, the addition of the vinyl 

group is faster by approximately two orders of magnitude [ll 11. Thus, in the first 

step, a network with a low crosslink-density is synthesized because predominantly 

compound 3 becomes active. In the second reaction step, the mesogens of this 

network are macroscopically oriented by application of a stress at elevated temperature 

and the induced anisotropy is fixed in the system by crosslinking via the 

niethacryloyl groups. The crosslinking reactions were performed in the nematic 

phase of the system. Part of the methoxy functionalities in compound 2 have been 

replaced by nitrile groups to provide an IR-active probe for the Orientation measurements. 

Additionally to the LSCE system, a structurally identical reference 

elastomer was investigated, which had not been subjected to a mechanical deformation 

during the second crosslinking step (Figure 2-43). In the reference polymer 

the mesogens do not exhibit a macroscopic orientation and, contrary to the LSCE, 

this material is opaque. 

2.7.2.2 Rheo-Optical FTIR Spectroscopy 

The increased need for more detailed data and a better understanding of the mechanisms 

involved in polymer deformation and relaxation has led to the search

2.7 Orientation oj Liquid Crystals Uiider Meclimiicul Forces 75 

Figure 2-44. (a) Principle of rheo-optical FTIR spectroscopy of polymer films. (b) Variabletemperature 

stretching machine for rheo-optical FTIR and FT-Raman spectroscopy. (A) stretching 

machine; (B) sample transfer mechanism; (1) stress transducer; (71 strain transducer: (3) film sample; 

k4'1 sample clamps [17, 181. 

for new experimental techniques to characterize transient structural changes 

during mechanical processes. With the advent of rapid-scanning FTIR spectroscopy, 

rheo-optical measurements have been applied to obtain data on the orientation, 

conformation, and crystallization during mechanical treatment of a wide 

variety of polymers [14, 17-20, 1121. The experimental principle of rheo-optical 

FTIR spectroscopy is based on the simultaneous acquisition of polarization spectra 

and stress-strain diagrams during deformation, recovery or stress-relaxation of 

the polymer under investigation in a miniaturized, computer-controlled stretching 

niachine which fits into the sample compartment of the spectrometer (Figure 2-44). 

The impleinentatioii of a heating cell as a closed, nitrogen-purged system offers the 

additional possibility to study the above mechanical phenomena under controlled 

temperature conditions ( i 0.5 "C) up to 200 "C. The electromechanical apparatus, 

the experimental procedure and the evaluation of the relevant parameters from 

the polarization spectra series have been described in detail in several publications 

[17-20, 1121. 

2.7.2.3 Near-Infrared Spectroscopy 

For samples exceeding the thickness accessible for ti-ansmission measurements 

by mid-infrared spectroscopy, near-infrared (NIR) spectroscopy may be used as 

an alternative technique [ 1131. Adjacent to the conventional mid-infrared region 

14000-400 cni-l) the NIR region covers the interval from about 12 500-4000 cm ' .

76 2 Segnrental Mobilitv of Liquid Crystals cmd Liquid-Crystalline Poljvners 

The absorption bands observed in this region can be assigned to overtones and 

combinations of characteristic fundamental frequencies. Apart from analytical 

applications, NIR spectroscopy has been demonstrated to be of value also to rheooptical 

applications [17, 114, 11 51. Thus, although NIR spectra are generally less 

informative than their mid-infrared analogs due to overlap of absorption bands, the 

details of the changes in crystallinity, conformation, and orientation are also contained 

in the overtone and combination vibrations [I 7, 11 31. 


The rheo-optical investigations were performed in the stretching machine shown in 

Figure 2-44b. For this purpose, film samples of the LSCE and the reference polymer 

(sample thickness about 300-350 pni, original length 8 mm, width 5 mm) were 

subjected to an elongation-recovery cycle up to about 45% strain, with a strain 

rate of ll'l/o/min and 20-scan interferograms with a spectral resolution of 5 c1n-l 

were collected siinultaneously with radiation alternatingly polarized parallel and 

perpendicular to the drawing direction. After completion of the experiment the 

interferograms were transformed to the corresponding spectra and evaluated in 

terms of the order parameter (orientation function) [14, 17-20, 1121 of selected 

absorption bands. The experiments discussed here were performed at 27 "C. In 

order to study the reorientation mechanism of the preoriented LSCE [ 1 16- 12 I] 

and the isotropic reference elastomer during the mechanical treatment of the rheooptical 

experiment, the samples were mounted in the clamps of the stretching 

machine after a rotation of 90" from the deformation direction in the second 

crosslinking step (Figure 2-45). Thus, in the LSCE the mesogens are originally oriented 

perpendicular to the subsequent drawing direction. This is reflected in the 

direction of deforni;ition 

during crosslinking 

Q 

0%) deformation 

direction of mechanical 

force during rheo-optical 

experiment 

maximum strain 45% 

n 

change in director 

orientation during deformation 

8 

polarization direciian 

of NIR-radiation 

transparent opaque transparent 

Figure 2-45. Experimental sequence for the observation of the director reorientation during the 

rheo-optical elongation-recovery cycle.

2.7 Orientation of Liquid Crjvtals Under Mechuniccil Forces 77 

2 00 

f- 

I50 

0 50 

0 25 

7000 6500 6000 5500 5000 4500 4000 3500 3100 

wavenumlxrs [cin-l] 

I 00 

0 50 

0 00 

2300 2200 2100 

NIR MIR 

Figure 2-46. FTIR/FTNIR polarization spectra of the undrawn LSCE system (the radiation 

was polarized parallel and perpendicular to the subsequent drawing direction of the rheo-optical 

experiment). 

dichroic effects of the polarization spectra taken before the elongation procedure of 

(Figure 2-46). In contrast, no dichroisin can be observed in the undrawn reference 

polymer (see below). 

Due to the large thickness of the samples, the spectrometer had been modified to 

collect NIR spectra extending into the mid-infrared region ( 10 000-2000 cm-’ ) 

(tungsten-halogen source, CaF1-beamsplitter, InSb-detector). Despite an extensive 

band-overlap in the NIR-region, several absorptions in the spectrum of Figure 2- 

46 could be assigned by spectral comparison with low-molecular weight reference 

compounds [36]. Here, however, only the data derived from the v(C=N) absorption 

will be discussed in some detail. 


In Figure 2-47 the stress-strain diagram of an elongation-recovery cycle up 

to 43% strain at 27°C is shown for the reference elastomer. Further elongationrecovery 

cycles do not reflect significant differences. The change in order parameter 

of the v(C-N) absorption band during the mechanical treatment is shown in Figure 

2-48. Throughout the whole experiment the drawing direction was the polarization 

reference direction. Due to the crosslinking conditions, no orientation (S = 0) can 

be observed for the original sample and uniaxial elongation leads to a linear increase 

of the order parameter up to S = 0.45 at the maximum strain of 43%. During 

recovery, an almost linear decrease of S can be observed with a small residual 

orientation of the mesogens upon return to the original starting position. This is in 

agreement with the hysteresis effect and the permanent elongation obserL ed in 

the stress-strain diagram (Figure 2-47). When the structural absorbance A0 of 

the v(C-N) absorption is plotted as a function of elongation 1361, an almost linear 

decrease is observed which is in good agreement with the theoretically calculated

78 2 Segnientril Mobility of Lipid Crystals uizd Licliri~-Cr)istulliiie Polvrwrs 

0.07 - 

0.06 .. 

- 

N 005 .* 

Figure 2-47. Stress-strain diagram of 

0 10 20 30 40 an elongation-recovery cycle of' the 

STRAIN I"/] 

reference elastomer at 27 "C. 

0. 60 

S 

0. 40 

0. 20 

Figure 2-48. Order parameter/ 

strain-plot for the v(C-N) 

I . 

I I I I I I 1 I absorption corresponding to the 

10 20 30 10 ' 40 30 20 10 elongation-recovery cycle shown 

STRAIN [Yo] in Figure 2-47. 

decrease in sample thickness based on the assumption of a constant sample volume 

during elongation [30, 361. A more complex behavior is observed during the elongatioii-recovery 

cycle of the LSCE system to a maximum strain of 45% at 27°C. 

The corresponding stress-strain curve is shown in Figure 2-49 and has been subdivided 

into different strain intervals. The order parameter of the v(C-N) absorption 

derived from the spectra acquired during this mechanical treatment is presented 

in Figure 2-50. Here too, the analogous strain intervals are indicated to assist 

the correlation and interpretation of the mechanical and spectroscopic changes. 

Due to the anisotropy of the starting material and the reorientation of the mesogens 

during the mechanical treatment, specific assumptions in terms of the reference 

direction were necessary. Thus, for the derivation of the order parameter during 

elongation between 0 and 32% strain (intervals I and I1 in Figure 2-50), the deformation 

direction during the second crosslinking step (which is perpendicular to the 

applied deformation in the rheo-optical experiment) has been chosen as polarization 

reference direction. Beyond the inversion point of S = 0, the machine direction of 

the rheo-optical experiment is applied as polarization reference (interval 111). During 

recovery this procedure is reversed. Therefore, the order parameter always has 

a positive value throughout the whole experiment. From Figure 2-50 the following 

picture can be derived for the orientational changes taking place during the 

elongatioii-recovery cycle:

2.7 Orientatioiz of Liquid Crystals Under Mechanical Forces 79 

Figure 2-49. Stress-strain diagram 

of an elongation-recovery cycle of 

the LSCE at 27 “C. 

0 10 20 30 40 

STRAIN [“/.I 

I). so- 

Figure 2-SO. Order parameter/ 

strain-plot for the vjC_N) 

absorption corresponding to 

the elongation-recovery cycle 

shown in Figure 2-49 (see 

text). 

s 0.40- 

0.30. 

0. 20- 

0.10~ 

S’I’RALN [%.I 

During interval I the stress increases linearly with strain and no significant 

changes in the mesogen orientation can be observed. 

When the external stress (a,) reaches the internal stress (0,) of the LSCE systein 

(induced by the second crosslinking procedure) [lll] at about 15% strain, the 

stress-strain curve levels off and siinultaneously the order parameter decreases 

rapidly towards zero (interval 11) (Figures 2-49 and 2-50). 

During interval 111, the reorientation of the niesogens in the direction of the 

external field takes place until maximum elongation. This procedure is accompanied 

by a slight decrease in stress due to the cooperativity of the reorientational 

motion after exceeding the internal stress of the LSCE. 

Depending on the experimental conditions (temperature, niaxiinuni strain), the 

described phenomena are completely reversible upon recovery (Figure 2-51). 

Due to the intrinsic anisotropy in the original LSCE network, the spectroscopically 

monitored inversion point of the order parameter (S = 0) could principally 

be interpreted by the three different mesogen alignments shown schematically in 

Figure 2-52 [36]. For the elucidation of this question the structural absorbance A,,

Figure 2-51. Reversible dichroic behavior of 

the v(C-N) absorption of the LSCE during an 

elongation-recovery cycle to 45'%1 strain at 

'7 'C. 

C 

H b 

es~ = merogenlongsxls 

e 

Figure 2-52. Possible mesogen orientations 

during the reorientation process at S=O (ul: 

direction of internal stress induced during the 

second crosslinking step. 0,: direction of external 

stress applied during the rheo-optical 

experimentj. (a I 

Homeotropic orientation 

perpendicular to the film plane: (b) isotropic 

orientation; (c) planar biaxial orientation. 

3.40 0. 80 

3. 20 

S 

I 

0. so 

0.40 s 

2.60 

2.40 

I I I I 

10 LO 30 40 

S'IKAIN IYuI 

n. w) 

0. 20 

0. 10 

Figure 2-53. Structural absorbance 

and order parameter/strainplots 

for the v(C3N) absorption 

corresponding to the elongation of 

the LSCE to 45% strain at 27°C. 

of the v(CsN) absorption has been plotted alongside the order parameter for the 

LSCE elongation step in Figure 2-53. The reference direction for the evaluation of 

the structural absorbance has been chosen in accordance with the assumption made 

for the order parameter (see Figure 2-50). As can be seen from Figure 2-53 the 

structural absorbance decreases linearly up to about 25% strain. This is in agreement 

with the reference polymer and can be attributed to the decrease in sample 

thickness. Between 25'%, and 35% strain. however, the structural absorbance shows 

an abrupt increase followed by only a slight decrease up to maximum elongational

2.8 Conclusions 81 

strain. As A0 is an experimentally determined parameter based on a uniaxial orientation 

model whose value should represent the intensity of an absorption band 

exclusive of orientation effects, the anomalous increase of the structural absorbance 

during elongation-where the sample thickness actually decreases-points towards an 

invalidity of the uniaxial syinmetry. T~us, the geometries a and b of Figure 2-52 can 

be excluded on the basis that both would lead to a decrease of A" during elongation. 

The occurrence of the structural model c at the inversion point of the order 

parameter has first been proposed on the basis of small-angle X-ray measurements 

by Finkelinann and Kiipfer [117]. Thus, the appearance of four diffraction maxima 

have been explained in terms of a biaxial, planar orientation of the mesogens in the 

film plane (Figure 2-52c). Depending on the inclination angle of the mesogens with 

the stretching direction, such a transition state would lead to an order parameter 

close to zero and could also explain the anomalous behavior of the structural 

absorbance [36]. In conclusion, it could be shown that, during elongation of the 

anisotropic LSCE network, the network structure changes from a uniaxial symmetry 

perpendicular to the applied mechanical field (parallel to the orientation induced 

in the second crosslinking step) via a planar-biaxial model (inversion of the order 

parameter) to a system aligned in the stretching direction of the rheo-optical experiment 

[36]. Simultaneously, the originally transparent polymer becomes opaque 

at the inversion point of the order parameter and turns transparent again upon 

reorientation of the mesogens into the machine direction. 


The aim of this contribution was to demonstrate the potential of FTIR/FTNIR 

polarization spectroscopy at variable temperatures for the characterization of segmental 

mobility in liquid crystals and liquid-crystalline polymers under the influence 

of external fields. Selected examples have been discussed, where either timeresolved 

measurements or static experiments have been utilized to monitor the 

effects of electric, electroniagnetic, or mechanical perturbations on the liquidcrystalline 

system under investigation. For the extremely fast, reversible switching 

processes of low-molecular weight liquid crystals and side-chain liquid-crystalline 

polymers in electric fields, the novel step-scan FTIR technique has been applied. 

The selective character of vibrational spectroscopy generally allowed us to separate 

the orientational effects in different functionalities of the liquid-crystalline compound. 

Thus, side-chain liquid-crystalline polymers, for example, have been characterized 

under the influence of different external fields in terms of the relative 

alignment of their main chain, spacer, and mesogenic groups, respectively. Although 

the presented data are necessarily limited to selected examples, the variety of 

information and its detailed quantitative Character demonstrate, that FTIR/FTNIR 

polarization spectroscopy has certainly become an extremely valuable tool lo elucidate 

the dynamics of segmental motion in liquid-crystalline systems.

82 3 Segmental Mobility of Liquid CrVstLiIs und Liqziicl-Cr!).stalline Poljjmer-s 

2.9 Acknowledgments 

The authors would like to thank Prof. Dr. H. Finkelmann and Dr. J. Kiipfer 

(University of Freiburg, Germany), Prof. Dr. R. Zentel and Dr. T. Oge (University 

of Wuppertal and University of Mainz, Germany), Prof. Dr. B. Jordanov (University 

of Sofia, Bulgaria), Dr. N. C. R. Holm, Dr. S. Hvilsted, Dr. M. Pedersen and 

Dr. P. S. Rainanujam (Riso National Laboratory, Denmark), Dr. F. Andruzzi 

(CNR, Pisa, Italy), Dr. M. Paci, Dr. E. L. Tassi and Prof. Dr. P.-L. Magagnini 

(University of Pisa, Italy), and Dr. E. Happ (University of Ulm, Germany), Dr. C. 

Hendaiin (Deutsche Post, Darmstadt), Dr. M. Czarnecki (University of WroclaLy, 

Poland) for helpful discussions, individual experimental and theoretical data, and 

the supply of polymer samples. Financial and instrumental assistance by the Deutsche 

Forschungsgenieinschaft, Bonn, Germany, the Ministerium fur Wisseiischaft 

und Forschung, Dusseldorf, Germany, Fonds der Cheinischen Industrie, Frankfurt, 

Germany, and the European Community via contract BRE2.CT93.0449 is also 

gratefully acknowledged. 


[I] Vertogen G., de Jeu, W. H., T/iennotro/iic Liquid Crystcils, F~rthmerzt~d~, Berlin: Springer 

Verlag, 1987. 

[2] Koswig, H. D., Fli e Kri:r.istalki~, Berlin: VEB Verlag, 1984. 

[3] Donald, A. M., Windle. A. H., Lipid Crystallirze Polytners, Cambridge: Cambridge University 

Press, 1992, pp. 38-41. 

[4] Finkelmann, H., in: Lipid Crj~~rallinirv 

Pol,yniers: Prirrcrjhs mid Atntkiriierird Pwprrirs: 

Ciferri, A. (Ed.), Weinheim: VCH Verlagsgesellschaft, 1991; pp. 315. 

[5] McArdle, C. B., in: Side Chain Liquid Crystal Polymers: McArdle, C. B. (Ed.), Glasgow: 

Blackie & Son Ltd., 1989; pp. 357. 

[6] Wendorff, J. H., Eich, M.. Mol. Cryst. Liq. Cryst. 1989, 169, 133-166. 

[7] Ndkamura, T., Ueno, T., Tani, C., Mol. Cryst. Liq. Cryst. 1989, 169, 167-192. 

[8] Demus, D., Goodby, J., Gray, G. W., Spiess, H.-W., Vil, V. (Eds.), H~~rirlDooX o/ Liqid 

Crystals, Weinheim: Wiley-VCH, 1998. 

[9] Carfagna, C. (Ed.), Lipid Crystalline Polymers, Oxford: Pergamon, 1994. 

[lo] Chilton, J. A., Gossey, M. T. (Eds.), Special Po/j3rners ,/ijr Electronic cititl O/~roe/~,c~rro~iic.s, 

London: Chapman & Hall, 1995. 

[ll] Kelly, S. M., Liq. Cryst. Today 1996, 6(4), 1. 

[I?] Schadt, M. in: Lipid Crystal.y: Baunigartel, H., Frank, E. U., Grunsbein, W. (Eds.), Stegenieyer, 

H. (Guest Ed), New York: Springer, 1994, Chapter 8, p. 195. 

[ 131 Chilton, J. A,, Gossey, M. T. (Eds.), Special Pol~~mers ,for Elecfronics crritl Opto-electrotiics, 

London: Chapman & Hall, 1995. 

[ 141 Siesler, H. W., Holland-Moritz., K., Zr?/iar.ed crnd Rtrriinrz S{)t~c.fr.u,scopjJ of Poljwiets, New 

York: Marcel Dekker, 1980. 

[15] Hvilsted, S., Andruzzi, F., Kulinna, C., Siesler H. W., Ramanujam, P. S.? i~~icroriiole~~~rlc..r. 

1995,-?8, 2172. 

1161 Shilov, S.. Okretic, S., Siesler. H. W., Czarnecki, M. A., Appl. Spectrosc. Reus., 1996, 

31 (1&2), 125-165.

3. I0 Refcreiice.r 83 

[I 71 Hoffinann, LJ., Pfeifer, F., Olirelic, S., Vdlld, N.. Zahedi, M., Siesler. H. W., Appl. Spci~tro.~~~. 

1993: 47, 1531-1539. 

[l8] Siesler. H. W., in: Oriented Po(iwirr Muterinl.7: Fakirov F. (Ed.), Zug: Hiithig & Wepf, 996. 

pp. 138-166. 

[ 191 Siesler, H. W., in: Adi1aizce.s in Applied Fourier Transform In/kured Spectroscopy: McKenzie, 

M. W. (Ed.j, Chichester: Wiley 8: Sons Ltd., 1988; pp. 189-246. 

1201 Siesler, H. W.. in: Atloances in Pollvner Science. Berlin: Springer Verlag, 1984; 64. pp. 1- 77. 

1211 BrBuchler, M., Boeffel, C., Spiess, H. W., Makroniol. Clic7n. 1991, 192, 1153. 

[22] Menzel, H., Hallensleben, M. L., Schmidt, A., Knoll, W., Fischer, T., Stumpe, J. ilfiicrum~leculrs, 

1993, 26, 3644-3349. 

[231 Buffeteau, T., Natansohn, A,, Rochon, P., Pezolet, M., i~~ucroi7zolecitls~ 1996, 29, 8783. 

1241 Graham. J. A.. Hammaker. R. M., Fateley, W. G., in: Cl~errzical. Biological urrd Inchstrial 

Appli~~ufims nf’ h~firrr.c~r/ S~~CC/~O.SUJ~,I~: Dnrig, J. R. (Ed.), Chichester: Wiley Interscience, 

1985; pp. 301-334. 

[25] Souvignier, G., Gerwert, K.. Biopl7~s. J. 1992, 63, 1393. 

[26j Uhmann, W., Becker, A,, Taran, C., Siebert, F., Ajyl. Speetro.~c~. 1991, 45, 390. 

[27] Palmer. R. A., Spectroscopy 1993. 8, 26. 

[28] Palmer, R. A,, Chao, J. L., Dittmar R. M., Gregoriou, V. G.? Plunkett, S. E., &p/. Spectrosc. 

1993, 47, 1297-1310. 

[29] Nakano, T., Yokoyama, T., Toriunii, H.. Appl. Spectrosc. 1993, 47, 1354-1366. 

[301 Gregoriou, V. G., Noda, I., Dowrey, A. E., Marcott, C., Chao, J. J., Palmer, R. A., J. Polym. 

Sci. B, 1993, 31, 1769-1777. 

[31] Lafrance, C. P.: Nabet, A,, Prud’homnie R. E., Pezolet, M., Can. J. Clwni. 1995, 73, 1497. 

[32] Maier, W., Saupe, A.? Z. Nr/turfor.tc/ztiny 1959, Ida, 882. 

[33] Maier, W., Saupe, A.. Z. Nriturforschung 1961, 160, 816. 

[34] Anderle, K., Birenheide, R., Eich, M., WendorK, J. H., Ilfulcroniol. C//em. Rqid CO~ZIIILII/. 

1989, 10, 471. 

[35] Natansohn, A., Rochon. P., Gosselin, J., Xie, S., Mclcrori~olec,ri/e,~ 1992, 25, 2268. 

[36] Kulinna, C., Dissertotion, Universitat Essen, 1994. 

[37] Chatelain, P., Bull. Soc. Fr. Miner. 1943, 66, 105. 

[38] Blinov, L. M., Electro-Optical mid ~1~iyrieto-O~)ticcd Proprrtics of Liquid Crysfrtl.r, Chichester: 

Wiley-Interscience, 1983. 

[391 Frank, F. C., Disctrss. Faradq) Soc. 1958, 25, 19. 

[40] Geary, J. M., Goodby, J. W., J. Appl. Phys. 1987, 62, 4100. 

[41] Castellano, J. A., Mol. Cryst. Liq. Cryst. 1983, 94, 33. 

1421 Jerdme, B. in: Hcindbook of Liquid Crystuls: Demus, D., Goodby, J., Gray, G. W., Spiess. 

H.-W., Vil, V. (Eds.), Weinheim: Wiley-VCH, 1998, Vol. 1, p. 535. 

[43J Geary, J. M., Goodby, J. M.: Kmetz, A. P., Patel, J. S., J. Appl. Phjis. 1987, 62, 4100. 

[44] Toney, M. F., Russel, T. P., Logan, J. A,, Kikuchi, H., Sands, J. M., Kumar, S. K. Nature 

1995, 374, 709. 

[45] Happ, E., University of Ulm, Sektion Polymere, personal communication, 1994. 

[46] Kaneko, E., Principles and Applications of Liquid CrJtstul Displaj~s, Tokio: KTK Scientific 

Publ., 1987. 

[47] Finkenzellner, U., Spektruni deer. PL’issensclzuft, 1990, 8. 54. 

[48] Schadt, M., Helfrich, J., App/, Plzys. Lett. 1971, 18, 127. 

[49] Gelhaar, T., Pauluth, D., Ndir. Chern. Tech. Lnh. 1997, 4S( l), 9. 

[50] Schadt. M., Phys. Bl. 1996, 7/8, 695. 

[51] Kaito, A,, Wang. J. I(.. Hsu, S. L., Anal. Cheni. Acta 1986, 189, 27. 

[52] Masutani, K., Yokota, Y., Furukawa, Y., Tasumi, M., Yoshizawa, A., App/. Siiectrosc. 1993, 

47, 1370-1375. 

[53] Okretic, S., Dissertutiorz, Universitat Essen, 1995. 

[54] Shilov, S. V., Okretic. S., Siesler, H. W.. Vib. Spectrosc. 1995, 9, 57. 

[55] Urano, T., Hamaguchi, H., Clwm Phys. Lett. 1992, 195, 287. 

[S6] Urano, T., Hamaguchi, H., Appl. Spectrosc. 1993, 47, 2108. 

[57] Gregoriou, V. G., Chao, J. L., Toriumi, H., Palmer, R. A.. Chenz. Phys. Lett. 1991, 179, 

491.

84 2 Segniental Mobility oj Liquid Crystals and Liqui~~-Cr),st~illine Pol?,nzeus 

[58J Czarnecki, M. A,, Katayanm, N., Ozaki, Y., Satoh, M., Yoshio, K., Watanabe, T., Yanagl, 

T., Apyl. Spectrosc. 1993, 47, 1382-1385. 

(591 Powell, J. R.. Krishnan. K., Yokoyamn, T., Nakano, T., in: Proc. 9th 177t. Conference on 

Foiwier Trunsjbnn Spectroscopy: Bertie, J. E., Wieser, H. (Eds.), Proc. SPIE 2089. 1993; 428- 

429. 

[60J Noda, I., Appl. Spectrosc. 1993, 47, 1329-1336. 

[61] Noda, I., -4ypl. Spectrosc. 1990, 44, 550-561. 

[62] Michl, J., Thulstrup, E. W., Spectroscoyj, with Polu d Light. Solute Alignment by Photoselection 

in Liquid Crystals, Polymers and Membranes, Weinheim: VCH, 1986. 

1631 Gotz, S., Stille, W., Strobl, G., Scheuermann, H., Macronzoleczrles, 1993, 26, 1520. 

[64] Hendann, C.. Dissertation, Universitat Essen, Germany, 1997. 

[65] Heilnieier, G. H., Zanoni, L. A., Appl. Phys. Lett., 1968, 13, 91. 

[66] Bahadur, B., Mol. Cryst. Liq. Cryst. 1983, 99, 345. 

1671 Sackmann, E. in Applicntions of' Liqziid Cryytnls, Meier, G.. Sackniann, E., Grabmaier, G. 

(eds.j, New York: Springer-Verlag, 1975. 

[68] Schumann, C., in Harzdbook of Liquid Crystals, Kelker, H.. Hatz, R. (eds.), Weinheim: 

Verlag Chemie, 1980. 

[69] Scheffer, T. J., Nehring, J., in P/?.vsics and Chernisfry ofLiquid Crystal Devices, Sprokel, J. G., 

(ed.), New York: Plenum Press, 1979, pp. 173. 

[70] Uchida, T., Wada, M., Mol. Cryst. Liq. Cr,yst. 1981, 63, 19. 

[71] Bahadur, B. in Liquid Crystuls-Applicutioris and Uses, Bahadur B. (ed.), Singapore: World 

Scientific, 1992. 

[72] Jordanov, B., Okretic, S. Siesler, H. W., Appl. Spectrosc. 1997, 51(3), 447-449. 

[73] de Barny, P.. Dubois, J. C., in: Side CI7airt Liquid Crystal Polymers: McArdle, C. B. (Ed.), 

Glasgow: Blackie & Son Ltd., 1989; chap. 5. 

[74] Lagerwall, S. T. in Handbook of' Liquid Crystals, Demus, D., Goodby, J., Gray, G. W.. 

Spiess, H.-W., Vill, V. ieds.), Weinheim: Wiley-VCH, Vol. 2B, 1998, 515-664. 

[75] Dijon, J., in Liquid Crystals-Applications and Uses, Bahadur B. (ed.), Singapore: World Scientific, 

1992. 

[76] Goodby, J. W., Blinc, R., Clark, N. A,, Lagerwall, S. T., Osipov, M. A.; Pikiii, S. A,, 

Sakurai, T., Yoshino, K., Zeks, B., Ferroelectric Liquid Crj'stuls: Priizc@le,r; Properties und 

Applicutions, Philadelphia: Gordon and Breach, 1991. 

[77] Kapitza, H.. Poths, H., Zentel, R., Mulironiol. Cheni. Mucron7o1. Syinp., 1991, 44, 117. 

[78] Brehmer, M., Zentel, R., Wagenblast, G., Sieniensmeyer, K., Mukroinol. Cheni., 1994, 195, 

1891. 

[79] Schonfeld, A,, Kremer, F., Hofniann, A,, Kiihnpast, K., Springer, J., Scherowsky, G., 

Makronzol. Clwii. 1993, 194, 1149. 

[80] Poths, H., Zentel, R., Liq. Cryst. 1994, 16, 749. 

[Sl] Shilov, S. V., Skupin, H., Kremer, F., Gebhard, E., Zentel, R., Liq. Cryt. 1997,22(2), 203-210. 

[82] Verma, A. L., Zhao, B., Jiang, S. M., Sheng, J. C., Ozaki, Y., Phys. Rev. E, 1997, 56(3), 342. 

(831 Shilov. S. V., Okretic, S.. Siesler, H. W., Zentel, R., Oge, T., Mucron7ol. Rapid Connnun. 

1995, 16. 

[84] Og, T., Zentel, R., Mabornol. Chem. Phys. 1996, 197, 1805-1813. 

[85] Eich, M.! Wendorff, J. H., Reck, B., Ringsdorf, H., Makrornol. Chem. Rupid. Cornnzun. 1987, 

8, 59. 

[86] Fischer, Th., Lasker, L., Stumpe. J., Kostromin, J. Photocheni. Photobiol. A: Chem. 1994, 80, 

453-459. 

[87] Wiesner, U., Antonietti, M., Boeffel, C., Spiess, H. W., Macromolecules 1990, 8, 2133. 

[88] Stumpe, J., Miiller, L., Kreysig, D., Hauck, G., Koswig, H. D., Ruhmann, R., Rubner, J., 

hfuiirornol. Chew. Rapid. Cornnzun. 1991, 12. 81. 

[89] Haitjema, H. J., von Morgen. G. L., Tan, Y. Y.. Challa, G., Macromolecules 1994,27, 6201- 

6206. 

[90] Fischer, T., Lasker, L.. Czapla, S., Rubner, J., Stumpe, J., ilfol. Cryst. Liq. Cryst. 1997, 298, 

2 13-220. 

[9 I] Wiesner, U., Reynolds, N., Boeffel, C., Spiess, H. W., hfakrornol. Chern. Rupid. Cornnwn. 

1991, 12, 457.


[92] Wiesner, U., Reynolds, N., Boefel, C., Spiess, H. W., Liq. Crysr. 1992, 11, 251. 

[93] Natansohn, A., Rochon, P., Pezolet, M.. Audet, P., Brown, D., To, S., il/lacromo/ecu/es 1994, 

27, 2580. 

[94) Buffeteau, T. and PCzolet, M., ,4pp/. S~wctrosc. 1996, 50. 948. 

1951 Ramanujam, P. S., Holme, C., Hvilsted, S., Pedersen, M., Andruzzi, F., Paci, M., Tassi, E. 

L., Magagnini, P., Hoffmann. U., Zebger, I., and Siesler, H. W., Polyni. Ah. Teclmol. 1996, 

7, 768. 

[96j Tassi E. L., Paci, M., Magagnini, P. L.. Mol. Cryst. L;q. CFJJS~., 1995, ,736, 135. 

[97J Tassi E. L., Paci, M., Magagnini, P. L., Yang, B., Francescangeli, O., Rustichelli, F., Liq. 

Cryst. 1998, 24(3), 457-465. 

[SS] Holnie, N. C. R., Ramanujam, P. S., Hvilsted, S., Appl. Opt., 1996, 35, 4622. 

[99] Zebger, I., Kulinna, C., Siesler, H. W., Andruzzi, F., Pedersen, M., Ramanujam, P. S., 

Hvilsted, S.; Mrrlironiol. Cliern. Murrorid. Symp., 1995, 194, 159. 

[IOO] Kulinna, C., Hvilsted, S., Hendann, C.. Siesler, H. W., Ramanujam, P. S., Macr.orno/ecu/es 

1998,31,2141-2151. 

[I011 Rochon, P., Gosselin, J., Natausohn, A,, Xie, S., App/. Php. Lett. 1992, 60, 4. 

[lo?} Kneppe, H., Schnejder F. in Hunrllmok ofLiqitid Crysrds, Demus, D., Goodby, J., Gray, G. 

W., Spiess, H.-W., Vill, V. (eds.), Weinheim: Wiley-VCH, Vol. 2A, 1998, 142-169. 

[I031 Brand, H. R., Finkelmann, H., in Handbook oj. Liquid Crysluls, Demus, D., Goodby, J., 

Gray, G. W., Spies, H.-W., Vill, V. (eds.), Weinlieim: Wiley-VCH, Vol. 3: 1998, 277-302. 

[I041 Onogi, S., Asada, T., in YIZIhlerniitionnl Congwss of Rlieolog.y, Astarita, G., Marrucci, G., 

Nicolais, L. (Eds.), New York: Plenum Press, 1980, Vol. I, 127. 

[I051 Schmidt, H. W.? Angeir. Cfiem. Int. Ed. Adti Mateu., 1989, 28, 940. 

[lo61 Dunion, M., Zentel, R., Kulinna, Ch., Siesler H. W.; Liquid Crystals, 1995, 18(6j, 903. 

[lo71 Hessel, F., Finkelmann, H., Po(vrw7 Bztlletin, 1985, 14, 375. 

[IOS] Keller, P., Hardouin, F., Mauzac, M., Achard, M., Mol. Cryst. Lig. Cryst., IZS, 171. 

[I091 Finkelmann, H., Kock, H. J., Gleim, W., Rehage, H., Mukromol. Clleni. Rapid Commun. 

1984, 5, 287. 

[I 101 Finkelniann. H., Kock, H. J., Rehage, H., hfCikIw170/. C/wn. Rapid Cornmiin. 1981, 2, 317. 

[I 111 Kupfer, J., Finkelmann, H., Mdmmol. Cliem. Rupid Coninurn. 1991, I?? 717-726. 

[I 121 Siesler, H. W., Makroniol. Cllern. Macromol. Svmp. 1992, 53, 89-103. 

I1131 Siesler, H. W., n/Iakromot. Chern. Maruomol. Syrr2y. 1991, 52, 113. 

[114] Ameri, A., Siesler, H. W., J. Appl. Poljvn. Sci., 1998 (in press). 

[115] Wu, P., Siesler, H. W., Dal Maso, F., Zanier, N., Anahr.sis hfagazZi7c~. 1998, 26, 45-48. 

[I 161 Knndler, I., Finkelmann, H., Il/lrrcuonzo/. Cheni. Phys., 1998, 199, 677-656. 

[1l7] Finkelmann, H., Kupfer, J., Macromol. Chem. Phys. 1994, 1Y5, 1353-1367. 

[I 181 Kundler, I., Nishikawa, E., Finkelmann, H., Mucvoniul. S~.nzy., 1997, 117, 11. 

[I 191 Bladon, P., Warner, M., Terentjev, E. M., MmronmIearles, 1994, 27, 7067. 

[I201 Weilepp, J., Brand, H. R., Etirophvs. Lett., 1996, 34(7), 495. 

[121] Kiipfer, J., Finkelmann, H., Po/jwt Adu. Techtzol., 1994, 5, 110-115.


t 

Order/Disorder in Chain Molecules and 

Polymers 

G Zerhi mid M. Del Zoppo 


At present the theoretical calculation of vibrational properties of polyatomic molecules 

can be carried out with reasonable success with the help of large, fast computers 

using filly automatic computing programs. Normal mode calculations have 

developed along three main lines of thought. On the one hand, a major effort has 

been made by the pioneers of vibro-rotational spectroscopy with the aim of determining 

accurate ro-vibrational constants to determine harmonic and/or anharinonic 

vibrational force fields and the derived physical properties [ 11. Once norinal 

modes are known accurately (both fundamental frequencies and relative vibrational 

displacements [2, 3]), they become the essential ingredients for the treatment of 

vibrational intensities in infrared and scattering cross-sections in Raman [4-61 or 

neutron-scattering experiments [7]. A huge effort along these lines has been made 

by many authors in the past 50 years, with the pioneering works froin the group of 

B.L.C. Crawford at Minnesota and of T. Shimanouchi at Tokyo having been followed 

by many authors. The basic mathematics and the computing programs first 

derived by Curtiss [S] and then by Schachtschneider and Snyder [9] and freely distributed 

worldwide [ 101 have been reproduced with different names and are used 

ubiquitously and on most occasions without due mention and thanks to the original 

authors for their great efforts. 

A second school of thought has actively pursued the idea of calculating vibrational 

properties froin ‘ah initio’ quantum chemical calculations [ 1 I]. The underlying 

concept is that if optimal atomic wavefunctions (and if large, fast computers) 

are available all physical observables can be calculated crb initio, thus making 

nonnal-mode calculations from empirical force constants obsolete. A major effort is 

currently being made to assess critically the reliability of Lib initio calculations in the 

prediction of spectroscopic observables [ 11-14]. 

A third group of workers has introduced the idea that vibrational force constants 

derived from critically determined semiempirical potentials provide an easy way 

to calculate vibrational properties, as well as many other structural and dynaniical 

molecular properties. These semiempirical potentials are part of fully automated 

coinputer packages offered commercially and used widely. For expressing judgment 

on the reliability of these semiempirical potentials related to spectroscopic properties, 

one matter is to provide a qualitative description of molecular dynamics

88 3 Vibvationul Spectru us a Pvobe of Structurcil 0rdt.vlDisordt.r 

(molecules can always be made to ‘wiggle’ on a screen), another matter is to account 

quantitatively for frequencies, amplitudes, infrared (dipole derivatives), and Ranian 

or hyper-Raman intensities (polarizability derivatives at various orders). 

The latter attempt to obtain vibrational force constants is a combination of 

the two last methods, i.e., ab initio methods are used to obtain the parameters of 

two-body or many-body interactions generally adopted in semiempirical molecular 

mechanics calculations [15]. 

The size of the molecules that can be treated using the above-mentioned methods 

is quite large (20-30 atoms) and is not a problem if periodicity [ 16- 181 or symmetry 

[3] are intrinsic properties of the molecule. However, the problem becomes more 

complex when very large and irregular molecular objects need to be studied. 

In this chapter we do not wish to dwell on the problem of the usefulness and 

limitations of the various methods for calculating vibrational force constants. 

Rather, we will deal with the problems of carrying out normal-mode calculations on 

very large molecular systems which have no symmetry are of irregular shape, and 

chemically are highly disordered [19]. It appears obvious that none of the abovementioned 

methods of calculation can be applied to such huge and structurally 

complex systems. 

In nature, chemical systems with chemical and/or structural irregularities or disorder 

are very common and their relevance in natural phenomena is of fundamental 

importance. Biological as well as synthetic oligoiners and polymers form a large 

class of compounds whose structure and properties need to be characterized qualitatively 

and quantitatively. Vibrational infrared and Rarnan spectroscopy is one of 

the few physical tools available to study the microstructure of these systems. Many 

empirical assignment and correlative and analytical studies have been carried out, 

but the possibility of a theoretical understanding of the vibrational properties (and 

the derived infrared and Raman spectra) have not been yet thoroughly explored. 

In this chapter we present and discuss the dynamics and vibrational spectra of 

long-chain molecules with disordered structure, and discuss the current possibilities 

of calculating the vibrational infrared, Raman and neutron-scattering spectra of 

these systems. 

3.2 The Dynamical Case of Small and Symmetric 

Molecules 

Let us briefly define the elementary parameters and review the basic equations for 

the treatment of molecular and lattice dynamics. The concepts presented here form 

the background for a further discussion later in the chapter for the case of large 

molecular systems. 

In classical molecular spectroscopy, a vibrating molecule is always represented by 

a set of balls connected by weightless springs. The springs represent the chemical

onds which hold atoms together through directional forces generated by the interactions 

of valence electrons. Since the vibrational molecular potential is unknown it 

is approximated by a Taylor expansion about the equilibrium geometry in terms of 

a set of suitably chosen coordinates. Let x,, y, and z, be the instantaneous Cartesian 

coordinates of the a-th atom of the vibrating molecule and let x,", y: and z," be the 

corresponding equilibrium coordinates when the molecule is at rest. For simplicity 

let x, (i = 1-3N) label the vibrational displacements of the type xu - x:, yr - y: and 

z, - z," for the N atoms. The generally unknown vibrational potential can be 

approximated by the following Taylor expansion about the equilibrium structure 

2V = 2Vo + 2 C (CJV/~X~)~X~ 

1 'J 

+ C (8sV/8~,d~J)o~,~, 

+ . . . (3-1) 

The zero-th order term is removed by a suitable shift of the origin, the first-order 

term is zero because the forces (dV/ax,),, vanish at equilibrium, the second-order 

term defines a quadratic harmonic potential. In most cases of chemical interest the 

expansion is truncated at the second order. The analysis of the higher-order terms is 

outside the scope of our discussion. The truncation at the second order implies that 

the restoring forces are assumed to be linear with the infinitessimal displacements 

from the equilibrium position. 

The quadratic potential can be written in matrix notation as 

2V = x'Fxx (3-2) 

where x is the vector (and x' its transpose) of the 3N Cartesian displacement coordinates 

and F, is the matrix of the quadratic force constants: 

Correspondingly, the kinetic energy can be written as: 

2T = X'MX (3-4) 

where M is the diagonal matrix of the atomic masses. Only 3N - 6 of the xi are 

independent because the Cartesian coordinates are related by the Eckart-Sayvetz 

conditions [3]. 

With such a model, atoms are allowed to perform very small harmonic oscillations 

about their equilibrium positions and the dynamical treatment describes the 

normal modes Q, of a set of coupled harmonic oscillators. 

When the kinetic and potential energies are introduced into the Lagrange equation 

the solutions are of the form: 

where A, are called frequency parameters and define the frequency of oscillation of

90 3 Vihrntimzal Spectra as n Probe qf Strzrcturol Or-cler 

the atoms during the j-tli normal mode Q,. During each Q, all atoms move in phase 

(i.e., they go through their equilibrium positions at the same time) with frequency v, 

(in cm-') = (Eb,/4n-c-) 

7 7 112 and with amplitudes L,. 

Frequency parameters and vibrational amplitudes can be calculated by solving 

the secular equation 

M-'F,L, = L,A (3-6) 

Six of the 2, vanish because of the Eckart-Sayvetz conditions and describe three 

rigid translations and three rigid rotations of the molecule; each of the remaining 

3N - 6 nonvanishing A, is associated to the normal mode Q,. Depending on the 

shape (symmetry) of the molecule, degenerate symmetry species may occur and 

some of the nonvanishing may turn out to be equal (doublets, triplets, and even 

multiplets with very high symmetrical molecules, e.g., fullerene). Accidental degeneracy 

may occur for large and asymmetric molecules when intramolecular coupling 

is small or zero. When all 3N solutions are considered together, the relation between 

Cartesian displacements and normal modes is as follows: 

x = L,Q (3-7) 

The matrix L, is nonorthogonal because M-'F, is not symmetric. The normalization 

conditions are 

L,L', = M-' (3-5) 

which ensures that 

2T = Q'Q (3-921,) 

2V = Q'AQ 

(3-9b) 

i.e., normal modes are mutually independent and can be described as isolated 

harmonic oscillators. The shape of the vibrational motions sought in any work of 

vibrational assignment is described precisely by the solution of Eq. (3-6), once the 

intramolecular potential and the geometry of the molecule is suitably chosen. 

A few observations on the treatment of molecular and lattice dynamics in cartesian 

coordinates are necessary. 

1. When molecular dynamics is applied to the understanding of chemical intramolecular 

phenomena the use of chemical internal displacement coordinates as 

initially defined by Wilson et al. [3] are much more useful and may have a direct 

chemical meaning (see below). 

2. When molecular vibrations are studied by quantum chemical or molecular mechanics 

methods, the problem is first treated in Cartesian coordinates and possibly 

later transformed into internal coordinates. These programs generally 

provide the numerical values of each element of the F, and L, matrices; the

3.2 The Dyriamical Case qf Simll arid Symnietric Molecziles 91 

elements of Lx are simply the numbers needed to directly pro-iect on a screen the 

atomic vibrations for each Q,,. 

3. The treatment of vibrational infrared and Raman intensities provides quantities 

such as (7p/dx, and d((a))/ax; (,where p is the total molecular dipole electric 

dipole moment and ((a)) is the total niokcular polarizability tensor). These 

quantities have a direct physical meaning and are directly calculated ah iizitio or 

sometimes by extended programs of molecular mechanics calculations [4-61. 

4. When lattice dynamic calculations for one- or three-dimensional systems need 

to be carried out, the Cartesian space is more suitable in order to describe both 

inter- and intra-molecular vibrations [20]. 

When the chemical reality of an isolated molecule needs to be considered through 

its vibrational spectrum, new sets of ‘chemical’ coordinates can be defined [3]. Let 

R = BX (3-10) 

define a set of internal displacement coordinates where B is the matrix of geometrical 

coefficients and 

where 

xvib 1 AR (3-1 1) 

For a discussion of the A matrix, see [21]. For simplicity in this discussion 

we assume that a nonredundant set of 3N - 6 Rs has been defined [22]. The Rs 

describe bond stretching, valence angle bendings, and torsions. The vibrational 

potential can be rewritten in terms of Rs as 

2V = R’FRR (3-13) 

and the kinetic energy matrix as 

where 

2T = R’(GR)-’R (3-14) 

GR = BM-~B’ (3-15) 

G contains all information on equilibrium geometry and masses. The potential 

energy matrix is written as 

FR = A‘F,A (3-16) 

Equation (3-16) is the one to be used when F, is calculated by (16 iizitio methods and

92 3 J,’ibrationrrl Spectra as a Probe of Structural Order 

the problem requires a treatment in terms of internal displacement coordinates R. 

Care must be taken in the construction of A [21]. 

The eigenvalue equation in internal coordinates becomes 

and the relation to normal coordinates is given by 

(3-17) 

R = LRQ 

(3-18) 

with the iiornialization condition 

LL’ = GR (3-19) 

As discussed in Section 3.4, many norinal modes are characterized by the fact 

that only a few atoms belonging to a specific chemical functional group are moving, 

while the other do not feel the intramolecular geometrical and electronic changes 

which occur at a specific site of the molecule. Such group uihrutiorzs are then characteristic 

of a given functional group and allow its identification in a chemically 

unknown material. The existence of ‘localized’ group vibrations showing specific 

and characteristic group frequencies have fueled the development of spectroscopic 

correlations which have made infrared (and recently also Raman) spectra a valuable 

physical tool for chemical and structural diagnosis (23-261. These correlations 

form the basis of many automated computing programs for the chemical identification 

of unknown compounds. 

For the above reason, researchers felt the need to describe normal modes in terms 

of ‘group coordinates’ (or local symmetry coordinates) which can be defined by an 

orthogonal transformation 

X=CR (3-20) 

The expressions for Gx and Fx and for the corresponding secular equations are 

the results of a straightforward similarity transformation such as in Eqs. (3-13) and 

(3-14). The Japanese School of Shimanouchi has provided many force fields and 

vibrational assignment in terms of group coordinates [27]. 

When all redundancies between internal coordinates are removed, the size of the 

eigenvalue equation to be solved is equal to the number of normal modes (3N- 6) 

expected for the n~olecule under study. If the molecule has some symmetry it 

belongs to a given symmetry point group g; group theory provides the structure 

of the irreducible representation of g, i.e., the number of normal modes in each 

symmetry species r,. By a suitable linear and orthogonal transformation 

S=UR 01 (S=UC‘) (3-21) 

(with U’ = U-’) it is possible to define a new set of symmetry coordinates S which 

foiin the basis of an irreducible representation of the point group g 131.

3.2 The Dyrinniical Case of Sinnll and Sjwimetric Molecules 93 

The scalar quantities V and T in Eqs. (3-13) and (3-14) can be re-expressed in 

terms of symmetry coordinates 

2V = S’UFRU’S = S’FsS (3-22) 

2T = S’U(G,)p’U’S = S’(Gs)p’S (3-23) 

The new matrices Fs and Gs are factored into blocks with dimensions identical 

to the structure of the irreducible representation [3]. The corresponding secular 

equation becomes separated into blocks of smaller size. Numerical calculations 

carried out separately for each symmetry block enable to identify the normal modes 

which belong to a given symmetry species. This is essential when the vibrational 

assignment needs to be carried out [2]. Band shape analysis (in IR and Raman) 

of gaseous samples, Raman depolarization ratios (for gaseous or liquid/solution 

samples) and dichroic ratios in IR or Raman experiments in polarized light on 

single crystals or stretch-oriented solid samples [28] provide the experimental data 

supporting group theoretical predictions. 

The numerical solution of the eigenvalue Eq. (3-17) is obviously greatly simplified 

when symmetry factoring can be carried out for molecules which possess some 

symmetry elements. Symmetry factoring was essential in the past (even for small 

molecules) when numerical processes were necessarily cumbersome and time consuming 

because of primitive computing technologies. In spite of the explosive development 

of computational tools this problem cannot yet be fully overlooked. The 

molecular systems to be treated have become large and the size of the secular determinant 

can still pose some problems. 

A typical example for a modern textbook is the case of the molecule of fullerene 

(C~O) which shows a very high symmetry (point group Ih): its 174 normal vibrations 

are separated into smaller symmetry blocks. Since icosahedral symmetry gives rise 

to a large number of degenerate modes, only 46 distinct mode frequencies are 

expected: 

The application of group theory for the determination of the structure of the 

irreducible representation (for the prediction of optical selection rules for infrared 

and Ranian spectra) is generally easy and straightforward [2, 31. On the contrary 

the construction of symmetry coordinates (i.e., the construction of the U matrix, 

Eq. (3-21)) is sometimes diflicult and cumbersome for large systems especially when 

degenerate species occur (e.g., adaniantane and fdlerene). An automatic method 

for the construction of symmetry coordinates using computers based on the diagonalization 

of the GR matrix has been proposed [29]. The method is based on the 

fact that GR contains in itself all the information on the symmetry of the problem 

and on the redundancies, if some or all have not been previously removed. The 

diagonalization of GR provides directly the elements ofthe U matrix (Eq. 13-21)) as

94 3 Vibrational Spectro us a Probe oj Strzrctural Order 

follows. Let GR be diagonalized by the unitary transformation D 

GRD = Dl- (3-23) 

where r is the diagonal matrix of the eigenvalues. The new set of coordinates 

C = D’R (3-25) 

forms an irreducible representation of the symmetry point group of the molecule, 

thus the C, form a set of symmetry coordinates and D’ can replace U in Eq. (3-71). 

The method is particularly useful for the automatic removal of all redundancies 

in structurally complex systems. Indeed if in > 3N - 6 is the size of the eigenvalue 

equation 

r=m-(3N-6) (3-26) 

is the number of zero eigenvalues in Eq. (3-34) which provide the required r 

redundancy relations 

D’~R = o (3-27) 

between the coordinates used in setting up the starting B matrix (Eq. (3-10)). 

A fully automated computer method for the handling of group theory has been 

proposed and is recommended for extremely large and highly symmetrical systems 

with many local and cyclic redundancies [30]. The numerical methods in treating 

group theory and symmetry coordinates have been generally neglected, as few large 

and highly syininetrical molecules have required the attention of spectroscopists. At 

present, the need to understand the vibrations of fullerene, fulleroid, tubular molecules, 

dendrites, etc., may revive the use of such numerical methods. 

3.3 How to Describe the Vibrations of a Molecule 

The main purpose of a dynaniical analysis of a molecule is to assign the vibrational 

transitions observed in infrared and Rainan in terms of molecular motions. As 

mentioned above, chemical group frequency correlations have been the justification 

of most of the vibrational assignment. The description of the normal modes has 

posed some problems throughout the years. 

It would seem obvious that the only way to describe the atomic motion would be 

that of plotting the Cartesian displacements of each atom during normal mode Q, 

with frequency vJ. At present, with the availability of computer techniques and fully 

automated computing programs, the Cartesian atomic displacements are a norinal 

part of the output, while some programs even provide animated descriptions of the 

molecular vibrations which can be viewed in all directions by appropriate use of

3.4 S11or.f- and Loizg-Rcrizye Vibrational Coupliiig in Molmrles 95 

the software. Using Cartesian displacement coordinates, the identification of group 

frequencies is not always straightforward and unambiguous. Moreover, the publication 

of many drawings or of many tables of numbers becomes often graphical and 

editorial problems. 

In the past, in order to describe the normal modes much use has been made of the 

so-called potential energy distribution (PEDI. Torkington [3 I] has first proposed 

and Moriiio aiid Kuchitsu [32J have later generalized the concept of PED defined 

as: 

PED = APJF (3-28) 

where J,,,, = Li, and J,,ml = 2LmlLll. 

The advantage of PED is that it can be expressed either in internal or group 

coordinates, thus allowing a more direct description of the main characteristic of 

the nomal modes. 

In simple words, PED provides quantitatively (generally PED is expressed as 'XI 

contributions) the extent of involvement of one or several diagonal and off-diagonal 

force constants in a given normal mode. The classical textbook case is that of the 

well-known group frequency modes of a CH2 group. Using group coordinates (Eq. 

3-20)), the modes commonly described as CH2 antisymmetric stretch (d-), CH2 

symmetric stretch (d+), CH2 scissoring (a), wagging (w), twisting (t), and rocking 

(P) can be easily identified with reference to the corresponding Fx group force 

constants [27]. If the same motions were expressed in tenns of classical internal 

coordinates of CCH bending the PED would indicate unselectively that in all the 

four motions CCH bendings are involved without allowing any distinction of the 

characteristic group modes. The same situation is found in the case of the characteristic 

group modes of the -CH3 group which are recognized by their characteristic 

frequencies of bending, umbrella (U), in-plane, and out-of-plane rocking. 

A typical case, famous in organic vibrational spectroscopy, is represented by the 

work on trans-planar n-alkanes by Schachtschneier and Snyder [33] aiid by Snyder 

et a/. [34]; the latter has extended his study to branched paraffins [35], to a-alkanes 

in the liquid phase [36] and to oligo and poly-ethers [37]. The use of PED has 

allowed to distinguish clearly between band progressions and the evolution of the 

normal modes by changing the phase-coupling (see later in this chapter). 

3.4 Short- and Long-Range Vibrational Coupling in 

Molecules 

As in this chapter the aim is to understand tlie vibrations of large and highly disordered 

and structurally irregular molecules, tlie basic problem to be faced is 

whether and to what extent normal vibrations are able to probe the molecular 

intrainolecular environment, i.e., whether normal vibrations are mostly localized or 

extended over a large portion of the molecular system.

96 3 Vihratioiial Spectru as a Probe qf Structural Order 

This issue has rarely been faced by molecular spectroscopists for several reasons. 

All those who searched for chemically useful characteristic frequencies aimed at 

modes strongly localized within the functional group [23-261. In contrast, those 

who dealt with large systems generally considered oligomeric or polymeric molecules 

as structurally perfect systems, often with translational periodicity, and 

studied 'phonons' which, by definition, are collective phenomena [ 17-1 91. The 

problem cannot be overlooked when the vibrations of large and structurally complex 

molecules must be understood. 

The first theoretical approach seeded by the explosive development of chemical 

spectroscopic group frequency correlations has been presented by King and Crawford 

who gave the dynamical conditions under which a normal mode can be defined 

as 'localized' [38]. 

111 correlative vibrational spectroscopy, a group of chemically similar molecules 

shows characteristic group frequencies [23-261 vi (say, the stretching of the carbonyl 

group >C=O, very strong in infrared) which exhibit systematic changes that chemical 

spectroscopists would like to ascribe to inductive and/or mesomeric effects by 

the various substituents placed in the molecules either at short or large distances 

from the functional group of interest. On the other hand, the observed wavenumber 

shifts may also originate from changes of mass and/or geometry. 

Use is made here [38] of the dynamical quantities presented in Chapter 2. Let 

Go, Fo and La-' be the kinetic, potential, and normal coordinate transformation 

matrices of one molecule of the series taken as reference [39]. In going from one 

molecule to a chemically similar one within the same class, one may expect small 

changes in the geometry (AG) or in the force constants (AF), or in both, which may 

be the cause of the observed changes of vi. The corresponding matrices of the 

molecule so modified can be written: 

G=Go+AG 

F = Fo + AF 

L-I = L~-' + A(L-') 

(3-29) 

(3-30) 

(3-31) 

If the perturbations of G and F matrices are small it can be assumed that the 

eigenvectors do not change and the zeroth-order values in Eq. (3-31) are used. 

When the eigenvalue Eq. (3-17) is expanded in terms of the quantities in equations 

(3.29) through (3-31) in terms of the first-order changes AG and AF and the zerothorder 

eigenvectors, the final expression for the (group) frequency parameter turns 

out to be: 

in which (lo), is the frequency parameter of the i-th group-frequency mode of the 

reference unperturbed molecule, AG,1 and AFml are the changes of those elements 

of the kinetic and potential energy matrix whose contributions to the changes of A, 

are scaled by the related elements of the Lo and L,' matrices.

3.4 Slzort- and Loiig-Range Vibrational Couyliiig in Molecules 97 

According to King and Crawford [38], if li has to be insensitive to changes of the 

molecular framework (ix., if the motion is to be localized) the sufficient conditions 

are that the coupling terms in Eq. (3-32) be small, namely either 

or 

(3-33) 

(3-34) 

AGml z 0 

AFm, E 0 

(3-35) 

(3-36) 

It is interesting to note that the problem of dynamical localization versus collectivity 

has never been a key issue in molecular spectroscopy and dynamics in the past 

40 years as the molecules considered were of limited size, thus obviously implying 

the existence of collective motions throughout the whole molecule. Moreover, most 

of the molecules considered contained either only covalent CJ bonds or, if n bonds 

existed, they were either isolated or with limited conjugation (e.g., aromatic molecules). 

The recent explosive interest in highly conjugated polyene and polyaromatic 

systems [40] has asked molecular dynamics to reconsider the problem of the extent 

of conjugation (delocalization) of n: bonds aloiig the whole molecular chain 141, 421. 

At the present stage of the spectroscopic studies for organic systems, a few 

observations on Eqs. (3-32) to (3-36) must be made. 

1. From the definition of the GR matrix (Eq. (3-15)) or, more directly, from the 

analytical expressions of the GR matrix elements provided by Wilson et al. [3], it 

transpires that the element (GR)~, connecting internal coordinates i and j may be 

different from zero when they possess at least one atom in common. It follows 

that atoms away from the functional group under study certainly will not affect 

the group frequency; Eq. (3-35) holds, i.e., the G matrix favours localization. 

2. For molecules made up exclusively by CJ bonds or containing also isolated and 

noncoiijugated K bonds inductive effects are generally considered not to extend 

at large distances, thus limiting the distances of the possible electronic effects 

described by AFllli to a few bonds around the ftinctional group (hence to a few 

internal coordinates). Equation (3-36) holds and again localization seems to be 

favored when strong electronic delocalization over large distances is not likely to 

occur within the molecule. 

3. When intramolecular mechanical coupling is very large (i.e,, Eqs. (3-33) and (3- 

34) do not hold) mechanical delocalization over a certain molecular domain 

takes place and the concept of group frequency is lost. This generally happens 

in the energy range between z 1500-400 cm-' where skeletal stretching and 

bending motions are strongly coupled. In this case small mass, geometry. or 

electronic perturbations may induce large changes in the spectrum, as indeed 

observed experimentally.

98 3 I i’brational Spectra as a Probe of Structural Ordrv 

4. The situation represented by Eq. (3-36) is at present of particular interest when 

polyconjugated molecules (polyenes and polyaromatic systems) are considered. 

The role played by vibrational infrared and Ranian spectroscopy in this new 

field of material science has been essential for the understanding of the new 

physical and chemical phenomena associated with the existence of extended n 

electron delocalization. The reader is referred to specialized references for a 

thorough discussion on the spectroscopy of these systems [4143]. For these 

systems recent studies have pointed out the occurrence of characteristic normal 

modes which show ‘frequency dispersion’ with conjugation length, i.e., the 

characteristic fi-equencies lower by adding repeating units which tale part in the 

conjugation [41-431. Dealing with a chain molecule, the elements (Lo),m, and 

entering Eqs. (3-33) and (3-34) are necessarily large. The physically 

relevant problem is to discover whether, to what extent, and at which distance 

the elements of AF change by adding conjugated units. The solution of this 

problem would shine some light on the distance and the extent of electronic 

interactions in polyconjugated systems 1441. This is a challenge to spectroscopists, 

theoretical quantum chemists, and computational chemists. 

5. It is obvious that the addition of strong electron donors or acceptors at a site 

away from the functional group of interest will modify the electronic environment 

and possibly the geometry at the site of the substitution. Certainly for 

particular internal coordinates, large AF and/or AG must occur, but their contribution 

to the frequency shift of the group frequency of interest is reduced to 

zero since the corresponding elements of L or L-‘ which relate the motion of 

the functional group and the motions at the site of the substituent are all zero 

(Eq. (3-33)). 

3.5 Towards Larger Molecules: From Oligomers to 

Polymers 

Molecular spectroscopy and lattice dynamics of oligonieric and polymeric materials 

have been treated thoroughly in several textbooks or articles [16-181. We wish to 

point out here a few fundamental concepts which form the basis of the understanding 

of the spectra of disordered materials which will follow in this discussion. 

Let LIS first consider the vibrations of polymers considered as one-dimensional 

perfect and infinite lattices. The usual basic assumptions are the following: 

1. The polymer is obtained by forming a one-dimensional chain of chemical units 

linked with a pre-assigned cherwicnl seqzirnce. For sake of example, we consider 

propylene (CH~-CH=CHZ) as the prototype of a monomer unit capable of 

producing a polypropylene chain. The first step is to make a chain of polypropylene 

in which all monomers are chemically linked head-to-tail (cheiizicul

eyularity), e.g., 

2. During the polymerization process the catalysts is chosen in such a way that 

.stereo-rrgzilur.itJi is assured [45], e.g., 

isotactic: CH3 CH3 

I 

-C-CH?-C-CH2 

I 

H 

I 

I 

H 

3. Upon solidification from solution (or from the melt), inacromolecules coil in an 

energetically preferred minimum energy conformation and form long chains 

along which roriforniatioiznl periodicity is assured [46-48). Chains are considered 

long enough to be considered ‘infinite’. For isotactic polypropylene, the conforniational 

sequence is TGTGTG. The overall shape of the chain can be described 

as a regular helix with a threefold screw axis [45]. 

4. The process of solification proceeds with the formation (when possible) of crystalline 

material in which chains pack in a lattice with tri-diinensional periodicity. 

The conditions for perfect packing are that chains possess chemical regularity, 

configurational periodicity, and conforinational regularity. 

Since interniolecular Van der Waals-type forces which hold molecules in a crystalline 

lattice are very weak, when compared with intramolecular covalent forces, 

lattice dynamics of polymers is made easy by considering firstly the vibrations of 

infinite isolated polymer chains [ 16- 181. Three-dimensional order is considered 

later as a sinall perturbation of the normal modes of the one-dimensional chain 

by the weak intermolecular forces [49]. In other words, we are considering the 

dynamics of one chain ‘in ziacuo’. 

The structural characteristics of the polymer chain constructed as described 

above are that translational or roto-translational symmetry can be found along the 

chain in addition to the traditional point group symmetry operations (rotation axes, 

symmetry planes, inversion center, and identity) [18]. 

In a dynamical treatment the labeling of the internal coordinates starts froin the 

reference repeat unit and proceeds along the chain by applying the one-dimensional 

translational operator which identifies the translationally equivalent internal coordinates. 

The structural situation is clarified if one considers that a polymer chain 

can be generated by application to a chemical repeat unit of a certain rototranslational 

operator $(Q, 1) which corresponds to a rotation of an angle 8 about the 

chain axis and a translation of a fraction 1 = d/n of the crystallographic unit length 

(repeat distance) d. In general, the crystallographic repeat unit can contain L 

chemical units coiled in a helix with t turns [47].

100 3 T%xitionul Spectin as a Pvohe of Straictuval Order. 

Let p be the number of atoms in the chemical repeat units and let the indices n 

and n’ label the sites of the chemical repeat units along the polymer chain and i and 

1 label the 3p internal co-ordinates within the repeat unit. The potential energy can 

be formally written in a way similar to Eq. (3-13) [SO]. 

11; n‘ 

i. 1 

(3-37) 

where 

(FR)llll’i, = (FR)ni”li (3-39) 

let s : In - 11‘1 define the distance of interaction, which (see Section 3.6’1 is very 

important in determining the vibrations of the chain; it follows that because of 

translational symmetry 

Substituting Eq. (3-40) into Eq. (3-37), one obtains 

In Eq. (3-41), the first term gives the contribution of the intramolecular potential 

by the reference unit at the n-th site, and the second and third terms collect the 

contribution of intramolecular coupling of the n-th unit with the neighboring units 

at distances s and -s respectively. 

A similar equation can be written for the kinetic energy in a way similar to what 

has been done for the finite molecule (Eq. (3-14)). Since the number of oscillators 

is taken to be infinite (the chain is considered infinite) the dynamical problem results 

in the writing of an infinite number of second-order differential equations, the 

solutions of which are of the type 

R”+S. 

, - A, exp 1-i (At + sv,)] (3-42) 

It should be noted that A, is independent of n, v, is the phase shift between two 

adjacent roto-translationally equivalent internal co-ordinates, and A refers to the 

vibrational frequency as in the case of small molecules (see Eq. (3-5)). Physically, 

Eq. (3-42) tells us that the j-th coordinate at the n + s site oscillates with frequency /? 

with an amplitude equal to the amplitude of the j-tli coordinate in the reference cell 

at site n multiplied by a phase factor exp (-isq). Attention should be paid to the

3.5 Towards Lnrger Molecules: From Oligoiiiers to Polymers 101 

fact that y is the phase shift of two equivalent oscillators belonging to two adjacent 

chemical units within the crystallographic repeat unit (see above). 

Following the same algebraic procedures as for the finite molecule the final 

secular equation to be solved is [50] 

where 

GR(~) = (GR)’ + c [(GR)S eisq + (GR’)S ePisV‘] (3-44) 

S 

FR(~) = (FR)’ + c [(FR)~ eisv + (FR’)~ eCiSv] 

S 

(3-45) 

Equation (3-43) is of 3pth degree in A. There are 3p characteristic roots (i = 1 . . .3p) 

for each value of the phase y. The r = 3p functions i~,(v,) can be interpreted as 

branches of a multiple valued function which is the dispersion relation for a onedimensional 

polymer chain. The solution reached in Eq. (3-43) is analogous to that 

previously attained by many authors in solid-state physics for simple tridimensional 

lattices [51]. The important digerelice is that Eq. (3-43) treats the problem in internal 

chemical coordinates, while, generally, simple lattices are treated in Cartesian 

coordinates (Eq. (3-6)). The fhction is periodic with period ~ Tand C calculations are 

limited to values of q~ within the first Brillouin zone [51, 521 (BZ) (-TC I y i TC). In 

most cases, the results for only half of the BZ are reported, the second half being 

symmetrical through y z 0. 

In the spectroscopy of polymers as one-dimensional ID lattices it is generally 

easier to deal with the phase shift y at odds with solid-state physics, dealing with 

three-dimensional 3D crystal, which treats the whole dynamics in terms of the 

vector k. For ID lattices we can describe the wave-motion (phonon) propagating 

along the 1D chain either by the phase shift v, or by the vector Ik/ = y/d where k 

has only one component along the chain axis k,. In contrast, in the spectroscopy 

and dynamics of 3D lattices (where intermolecular forces are active in all directions) 

phonons are labeled with k vectors with three components (kk, k,, k,) [51]. When k 

is used in polymer dynamics Eq. (3-43) can be easily rewritten as 

It should be remembered that in a single and isolated polymer, chain atoms 

are allowed to move in the tridimensional space even if phonons are considered 

to propagate in one direction along the chain axis. This means that no neighboring 

intermolecular interactions are taken into account, i.e., the dynamics is that of a 

chain ‘in uacuo’.

102 3 l,i'brationcil Spectra as n Probe of Structural Order 

300G i 

- 

2Got 

3000 

r 

7 

250{ - 

1500 

l 5 O 0 L = 

loook 

Figure 3-1. Dispersion curves of single-chain trans planar polyethylene. (a) v/p; (b) i$/k. 

As an example, in Figure 3-1 we report the dispersion curves of a single chain of 

polyethylene in v, and k space. The chain of trims-planar polyethylene -(CW?),,- is 

generated by application to a single CH2 group of the operator $(n. d/3). It follows 

that the crystallographic repeat unit contains two CH? groups. The physics and the 

numbers are not changed. The dispersion curves in Figure 3-lb are obtained by just 

halving Figure 3-la at .n/2 [53]. 

As an example of a more complex polymer chain we quote the case of polytetrafluoi-oethylene 

-(CFZ)n- for which one can identify the rototranslational 

operator $(2n/15, d/15) which describes a polymer chain made up by z = 15 CF2 

units coiled in a helix with t = 7 turns. In cases such as polytetrafluoroetliylene the

3.5 TOIIYW~ Laiyer Molertiles: Fvoiii Oligomevs to Polimers 103 

- 

Figure 3-2. Dispersion curves \./a, for single chain helical 

(z = IS, t = 7) polytetrafluoroethylene. 

0' 60' 120" lBi 

plot of v(q) becomes usable while v(k) becomes unreadable (117 frequency phonon 

branches in k space) (Figure 3-2) [54]. 

Dealing with 1D chains iii vuczio carries an important consequence on the shape 

of the dispersion curves if compared with the same systems considered as 3D 

lattices. From lattice dynamics in 3D it is known that if q is the number of atoms 

in the crystallographic repeat unit the dispersion relation provides 3q phonon 

branches. three of which are labeled as 'acoustical' and 3q - 3 as 'optical' branches. 

The former three have zero frequencies at the so-called r symmetry point in k space 

(i.e., with k, = k, = k, = 0; see Figure 3-3) and correspond to the three nongenuine 

vibrations with zero frequencies which describe the three rigid translations of the 

whole lattice along the three Cartesian coordinates [49. 551. The 3q - 3 optical 

branches may interact with the electromagnetic waves according to specific selection 

rules and provide the system with particular optical properties (absorption and/ 

or Raman scattering), as discussed below. 

In ID systems iii u~~cuo the rigid rotation of the whole chain about its axis is not 

hindered. It follows that an additional zero eigenvalue is expected from Eq. (3.43) 

when applied to 1D systems at k = 0. The lack of intermolecular forces for the 

polymer chain in uact[o carries an important consequence on the shape of the two 

acoustic branches near k = 0. While the three acoustic branches for 3D crystals 

have a discontinuity at k = 0, for ID crystals two of the acoustical branches may 

approach ~(0) almost asymptotically [56]. It follows that the dynamics of 1D systems 

based on the model of an isolated chain at small k values and at very low 

frequencies is totally wrong and cannot be used for the interpretation of related 

physical properties of real systems (neutron-scattering experiments, thermodynamic 

properties, Debye Waller factors, mean square amplitudes, and other physical 

properties where thermal population is involved which is strongly affected by lowenergy 

vibrations). 

For the benefit of the discussion which follows we give here a calculated schematic 

example of the procedures followed in setting up the matrices F~(rp) and

104 3 ViLmrtional Spectra as u Probe oj'StrLrctzrra1 Order 

cm 

12oc 

1100 

1000 

900 

800 

700 

600 

500 

400 

300 

200 

100 

" 

n 

0 

I 

Figure 3-3. Dispersion curves of crystalhe orthorhoiiibic polyethylene with two molecules per unit 

cell (from [49]). Comparison with Figure 3-1 shows the splitting of the frequency branches and the 

shape of the acoustical branches at q~ - 0 (see text). Attention should be paid to the fact that the 

frequencies and the shape of the dispersion curves shown in Figure 3-1 and in this figure may differ 

because they have been calculated with two different force fields.

3.5 Towmu" Larger Molecules: F~om Oligoi?iers to Polyriers 105 

a 

-1 0 +I 

Figure 3-4. (a) Chemical 

structure of trarzs-polyacetylene 

and ib) model 

assumed in the calculation 

reported in the text. The 

model gives the numbering 

of the one-dimensional 

repeat unit cell and the 

labeling of the internal 

coordinates. 

GR (y) for a simplified chain of planar trans-polyacetylene, (-CH=CH-), where 

each CH group is taken as a point mass. Only in-plane motions are considered. 

From the viewpoint of chemistry and materials science, polyacetylene is a prototypical 

relevant material in molecular electronics and shows a low (HOMO- 

LUMO) band gap (= 1.4 eV) because pz orbitals of the carbon atoms in the sp2 

hybridization are strongly delocalized along the molecular chain (large conjugation). 

The distance of interaction between 7c bonds is yet unknown and is the subject 

of active research [40-421. 

Figure 3-4 shows the drawing of the actual molecule as well as of the simplified 

model with the labeling (1) of the sites of the translational repeating unit and the 

labeling of the internal coordinates of the 1 = 0 unit and the equivalent ones at sites 

+/- s. The unit cell is made up by two inasses (p = 2) and the in-plane internal 

coordinates are 2 x 2 = 4 which is also the size of the secular equation. We expect 

to calculate four branches in the dispersion relation, two of which are optical and 

two acoustical. 

We calculate explicitly the algebraic expressions of the elements of the GR(~) 

matrix which enters Eq. (3-43). The following simplifications have been accepted in 

such a simple calculation: (i) the groups C-H are taken as point masses of in = 13 

daltons; (ii) C=C and C-C bond lengths are equal; and (iii) bond angles are taken 

equal to 120".

106 3 Vibrutioricil Spectra as a Probe of Structural Order 

(3-48) 

where p is the reciprocal of the C-C distance and p is the reciprocal of the mass of 

the CH unit taken as point mass. 

It is immediately seen that nonzero elements occur in the diagonal block and only 

in the two neighboring blocks which describe the interactions of the central unit 

with the nearest neighbors in position +1 and -1, i.e., the distance of interaction is 

s = 1, very short indeed. The construction of the GR(w) matrix through Eq. (3-44) 

yields a 4 x 4 matrix where the generic element has the following form: 

(3-49) 

The opposite is the case of the potential energy matrix whose structure is synibolically 

given in Table 16, p. 294 of [41]. 

The most relevant fact is that while the Gp, matrix is certainly truncated at the +I 

and -1 units in the FR matrix the distance of interaction of the zero-th unit with its 

s-th neighbour is yet unknown since the so-called 'delocalization length' needs to be 

determined [41-431. The truncation at the site s of the long ribbon of interaction 

submatrices is then arbitrary. Many ab iiiitio calculations are at present dealing with 

this essential problem which is common to all poly-conjugated chains presently 

available. 

The truncation of the sum in Eq. (3-45) is then arbitrary, thus affecting the 

validity of the physical conclusions which may be derived. It follows that any 

addition or variation of terms or extension of the interactions (increase of delocalization 

length) will affect the numerical values of the elements of the FR(~) 

matrix, 

thus changing the shape of the dispersion curves [41, 421. It is apparent that the 

shape of the dispersion curves is a direct consequence of the electronic properties of 

the poly-conjugated chains [58] (Figure 3-5). 

An opposite case is found for trans-planar polyethylene taken as a prototype of 

all exclusively o bonded chains for which conjugation is not predicted (the existence 

of o-conjugated chains like polysilanes needs further studies). In the case of polyethylene, 

experiments or calculations indicate that the electronic interactions between 

o C-C bonds fall OR very quickly along the chain with s = 2 [59]. Indeed, as 

shown by Schachtschneider and Snyder, the normal modes of finite n-alkanes lie 

almost perfectly on the dispersion curves calculated for polyethylene [33-351. No 

chain length effect in the FK(~) matrix is revealed for these classes of systems. 

Analogously for o bonded tridimensional lattices like cubic diamond and silicon, a 

valence force field limited to a few first-neighbor interactions nicely accounts for the 

optical and neutron-scattering data available to date [60] (Figure 3-6).

1700 

- CIO 

--- ca 

c 22 

_- c 12 

l l l l l l l l l 

Figure 3-5. Effect of the long-range interactions (s in equation 3.45) on the dispersion curves of the 

frequency branch v3 (associated with the in-phase skeletal stretching) of trms polyacetylene calculated 

for the unsimplified and realistic equilibrium geometry. The force field has been derived by ‘LI~J 

initio’ calculations on polyene oligoinei-s with n = 2,4,5,6, and 11 double bonds [%I. Attention 

should be paid to the fact that the ‘softening’ of the ~ ( = 90) mode iiiodulated by the distance of 

iriteraction generates a ‘dispersion’ with chain length of a very strong and very characteristic 

Ranian line ([41, 421). 

These observations are essential when the dynamics of disordered chains will be 

discussed later in this chapter. 

3.6 From Dynamics to Vibrational Spectra of 

One-Dimensional Lattices 

The experiments which provide direct information on the vibrations of one- and 

tridimensional lattices are infrared absorption (reflection) spectra, Ranian scattering 

and neutron scattering. For a general discussion on the theoretical principles and 

experimental methods the reader is referred to the abundant literature on these

108 3 Vibrational Spectra as a Probe of Structurtil Order 

K 

DIAMOND r-x 

Figure 3-6. Example of lattice dyiiamical calculations on 0-bonded tridimenional crystals with 

short range interactions. Dispersion curves for cubic diamond along the r + (K) i X symmetry 

direction. Experimental points from neutron-scattering experiments; dispersion curves from least 

squares frequency fitting of a six parameters short range valence force field (from [60]). The Raman 

active phonon is the triply- degenerate state indicated with r;, near 1300 cm-'. Notice that at k # 0 

the degeneracy at is removed because of the lowering of the symmetry throughout the whole BZ. 

Notice also the three acoustic branches for which 1' - 0 at k i 

r. 

techniques (e.g., refs. [16-19]). Our aim is to present here methods for extracting 

from the experimental data the maximum amount of information on structure and 

properties of these relevant materials. 

The next step in our analysis is to define the experimental observables predicted 

by lattice dynamics. The spectroscopic selection rules are very restrictive and, of the 

very many phonon frequencies calculated, very few can be directly observed experimentally 

[ 16- 191 in the optical vibrational spectra. 

First, only phonons with k = 0 can interact with the electromagnetic wave; thus, 

only the fundamental frequencies at the centre of the Brillouin zone in k space are 

potentially infrared or Raman active. Second, further restrictions on the spectroscopic 

activity are introduced by the symmetry of the polymer chain. Once the

3.6 From Dynamics to Vibrational Spectra of One-Diniensiorzal Lattices 109 

translational (or rototranslational) symmetry element is properly taken into account 

(in combination with point-group symmetry operations) a factor group isomorphous 

to a symmetry point group can be defined and k = 0 phonons can be separated 

into symmetry species, just as in the case of finite molecules. Traditional 

infrared and/or Raman activities can be predicted as well as the optical transitions 

associated to specific dipole allowed transitions with their directional properties for 

oriented samples. 

The physical meaning of the selection rule which states that only k = 0 phonons 

can be observed spectroscopically is that the atomic displacements in translationally 

equivalent crystallographic units must be all in-phase. When the geometry of the 

polymer chain (i.e. $(8, 1) and t = no. of turns of the helix within the rototranslational 

unit cell) is taken into account, phonons are infrared active for y = 0 and 

y z 0 (doubly degenerate); Raman activity is achieved for phonons with y = 0 

(nondegenerate), 0 and 20 (both doubly degenerate) [15, 61, 621. Degeneracy is 

clearly understood when folding of the dispersion curves is carried out at the zone 

boundary in going from y to k space. If p is the number of atoms in the chemical 

repeat unit for a helical polymer, one expects in the infrared 3p - 2 totally symmetric 

11 modes with q = 0 and 3p - 1 doubly degenerated I modes with y = 0. 

Coincidences with the infrared spectrum are expected in the Raman with the modes 

at y = 0 and y = 8, while extra modes are expected at y = 28 (practically never 

observed because their Rainan intensity is very weak). 

Let us further consider the direction of the transition dipole moments in infrared 

for a stretch-oriented polymer sample in which it is assumed that polymer chains 

are aligned along the stretching direction. Phonons with 9 = 0 have their transition 

dipole moment parallel to the chain axis which coincides with the direction of 

stretching (11 modes); modes with y = 8 have transitions moments perpendicular to 

the chain axis (1 modes). They can easily be identified in an infrared experiment on 

stretch-oriented samples carried out in polarized light. Analogously, the elements of 

the polarizability tensor can be selectively excited by Ranian scattering experiments 

on oriented polymer samples examined in polarized light in diflerent scattering 

geometries. For a study of the Raman spectrum of a stretch-oriented polyethylene 

rod in various geometries see [28] and Figure 3-7 [63]. 

A few comments of general relevance need to be made. 

The shape of the dispersion curves and the corresponding eigenvectors which 

describe the vibrational displacements (phonons) depend necessarily on the vibrational 

force field adopted in the calculations. The fitting of the experiniental frequencies 

with the calculated ones, although being a good indication that the force 

field is reasonably acceptable, can by no means be taken as proof of its general 

reliability throughout the whole BZ. Indeed, fitting has been obtained only at two 

values of the phase coupling y. The intramolecular interactions may be different at 

y # 0 and y # 8 or 20, thus in principle changing the shape of the theoretically 

predicted dispersion curves (see. for example, [64]). Other independent experiments 

are needed before claiming the success and generality of the intramolecular force 

field. 

So far, we have discussed k = 0 phonons of the one-dimensional lattice and have

1 10 3 Vibrritionnl SIJectrcr as CI Pvohe of Stvuctrirnl Order 

< 

7 

T- 

I 

0 

0 Ln 

P- 

N 

0 

0 

co 

N 

0 

m 

co 

N 

0 

0 

m 

N 

0 

m 

Di 

0 

0 

0 

C’? 

I 

__- 

0 Lo 

0 

c3

3.6 Froin Dynaiiiics to T’ihratioiial Sjlectra of Onr-DiriimsioriLII Lattices 1 1 1 

neglected the possible existence of a tridimensional arrangement of the polymer 

chains in the solid. Experimentally, it transpires that for most of the polymers the 

entire observed spectrum can be accounted for in terms of the k = 0 modes of the 

single infinite chain [65]. When the material is melted, oi- is dissolved in some suitable 

solvents, k = 0 modes disappear and the spectrum becomes typical of a conformationally 

disordered ‘liquid-like’ molecule with the classical ‘group frequency’ 

modes which can be interpreted using the classical spectroscopic correlations (see 

Section 3.2). 

The fact that, in going from the solid to the liquid state, maiiy bands disappear 

has been taken by authors as an indication that these bands must be associated with 

the material in the Crystalline state. In spite of the existence of clear and easy 

dynamical theories on polymer vibrations, a whole body of literature has accepted 

the direct correlation between k = 0 modes of the single chain and content of 

material in the crystalline (3D) state. 

This view is wrong both in principle and in practice. In spite of strong warning 

and extensive discussions (theoretical and experimental [65, 661) the general definition 

of ‘crystallinity bands’ has been widely and uncritically accepted by the chemical 

polymer community and even analytical determinations on the concentration 

of crystalline material have been carried out on polymers which show no direct 

spectroscopic indications of crystallinity [65, 661. The disappearance of such bands 

upon melting is simply due to the fact that the conformational regularity of the 

chain collapses and no 1 D-translational periodicity can be found anymore over a 

reasonable chain length. Chemical and stereo regularities are not modified in the 

melt or in solution, but all optical selection rules are removed because of the lack 

of phase coupling between adjacent units. From the above discussion it becomes 

apparent that the k = 0 bands previously discussed (generally and reasonably 

called ’regularity bands’ [65, 661) arise from a polymer molecule organized as a onedimensional 

crystal iiz uacuo, as if there were no intermolecular lattice forces. It 

becomes apparent that their labeling as crystallinity bands is a conceptual error. 

Certainly, the dynamical treatment can be and has been carried out for a few 

cases by taking into account the tridimensional arrangement of polymer chains [49]. 

A model of suitable intermolecular nonbonded atom-atom potential has to be 

chosen critically [49, 671 and calculations can be carried out using the same principles 

discussed previously in this chapter. The number (q) of atoms per tridiniensional 

unit cell increases, the complexity of the BZ increases, many more phonon 

branches are calculated for different directions of the wave-vectors, and special 

symmetry directions and symmetry point in the BZ can be found depending on the 

syiiinietry of the whole lattice [60, 641. Typical examples are given in Figures 3-3 

and 3-6. 

Since intermolecular forces in polymers are weak, their effects on the phonons 

of the whole lattice are relatively small, thus originating small splitting of few of 

the regularity bands [68]. Rigorously speaking, the limited splitting of the regularity 

bands observed for a few polymers originate from phonons at the point ik, = 

k, = k, = 0) of the tridimensional BZ. Such splitting can indeed be considered as 

crystallinity band and certainly originate from material organized in a tridimensional 

lattice. This occurs when intermolecular forces are strong enough, the nuin-

112 3 Vibrutionnl Spectru as ci Probe of Strirctzirul Order 

do0 2600 2600 2300 zboo 1foo 1400 1ioo Boo 500 

WAVENUMBERS 

Figure 3-8. Infrared absorption spectrum of highly crystalline polyethylene where crystal field 

splitting is shown in the 1450 and 720 cm-' range due to crystal field splitting (crystallinity bands). 

ber of chains in the unit cell is appropriate, and space-group selection rules are 

favorable [67-691. Among the true crystallinity bands we have also to consider the 

low-frequency lattice modes associated with the translation and librations of the 

molecules which behave almost as rigid objects moving within the 3D lattice. 

The classical case of crystalline polyethylene (Figure 3-8) may be mentioned in 

reference also to the vibrational spectra of its model compounds consisting of a 

series of n-alkanes. Relatively short alkanes with an odd number of C atoms crystallize 

in the orthorhombic lattice with two molecules in the unit cell. Relatively 

short n-alkanes with an even number of C atoms crystallize in monoclinic or triclinic 

unit cells with one molecule per unit cell. Very few of the normal modes of 

orthorhombic n-alkanes show in infrared and/or Raman doublets due to phonons 

at r (i.e., true doublets due to weak intermolecular forces; such splitting is coinmonly 

known and studied as correlation field splitting 1701); the spectra of monoclinic 

(or triclinic) n-alkane do not show any splitting as if the molecules, in a first 

approximation, were trans-planar, but in vuczio. Indeed, there exists only one molecule 

per unit cell and no correlation field splitting can occur. Analogously, when 

orthorhombic n-alkanes molecules are heated just before melting they transform 

in a probably monoclinic unit cell (there may be different interpretation of the 

so-called monoclinic modification). The classical doublets (phonons at r) observed 

in infrared and Raman disappear and singlets (phonons at k = 0) are observed as 

if the molecules were iiz ouc~io. This problem will be matter of extensive discussion 

later in this chapter.

3.7 The Case of Isotactic Polypropylene - A Tesbook Cme 113 

The fact which needs to be justified is that in some analytical determinations of 

the concentration of crystalline material in a polymer bulk, correlations are found 

between the intensity of the regularity bands and the data from X-ray diffraction 

experiments (which measures scattering from 3D periodicity). The measure of the 

intensity in infrared or Raman provides the amount of material organized as 1 D 

straight chains. If the ‘rIinylzrtti’ model is accepted, the measure of the intensity 

of the regultrr-ity bands [65, 661 gives an estimate of the concentration of straight 

spaghetti in an otherwise disordered environment of ‘boiled’ spaghetti (conformationally 

disordered chains). Whether the chains are packed in a crystal (3D order) 

or exist in a sort of ‘liquid crystalline’ arrangement (1D order) cannot be revealed 

from k = 0 regularity bands. 

True crystallinity bands have been observed for only a few polymers and their 

origin experimentally verified by isotopic dilution studies [70]. We may quote the 

classical prototype case of orthorhombic polyethylene [68, 691, orthorhombic polyoxymethylene 

[7 I], aiid-with caution-possibly a few others [72, 731. 

3.7 The Case of Isotactic Polypropylene - 

A Texbook Case 

It is beneficial here to discuss the case of isotactic polypropylene (IPP) as a worked 

example of the analysis of the vibrational spectrum of a polymer molecule. Let us 

assume that polymerization of propene with Ziegler-Natta catalysts has produced a 

fully head-to-tail isotactic polymer. 

Let us first consider the vibrational spectra (infrared) of IPP in the melt (Figure 

3-9a) [74]. In this physical state, polymer chains possess a conformationally irregular 

structure like any liquid branched n-alkane. We expect the infrared spectrum to 

consist of relatively few bands easily identified as the group frequencies of CH3 and 

CH2 and CH groups. Other modes may be identified with caution, but their location 

is irrelevant to the present discussion. The experimental infrared spectrum of 

Figure 3-9a is in full agreement with the expectations. 

It must be pointed out that the infrared spectrum of liquid IPP shows a few bands 

(especially the band at 973 cmr’) which are neither observed in the spectrum of 

atactic PP nor in the spectrum of liquid methyl-branched hydrocarbons. Moreover, 

in the Rainan spectra of molten IPP, three characteristic tacticity bands have been 

located at 1002, 973 and 398 cnir’ which allow IPP to be distinguished from the 

syndiotactic stereosionier (993, 966 and 310 cnr’). It follows that the ‘isotactic’ 

chiral units generate vibrations mostly localized on the asymmetric carbon atom, 

but with some coupling with the neighboring units. This allows the chirality of the 

neighbors to be probed. These bands become characteristic of the tacticity and 

allow us to distinguish between isotactic and syndiotactic sequences [75, 761. 

Next, let the sample of molten IPP solidify. Polymer chains are allowed through 

their conformational flexibility to reach the minimum energy conformational struc-

114 3 Vil,rcitionciI S)ectra as ci Probe of Strzictitrnl Order 

li 

I)

3.7 The Cuse of Isotactic Polypvop.vleiie - A Tevbook Case 1 15 

ture which is known from calculations and experiments [77] to form a 1D periodical 

lattice with the coiiformational sequence -GTGTGT- which defines a threefold 

helix with $ = (3rr/3. d/3) and one turn (t = 1). The isolated chain belongs to the 

point group C3. Each chemical unit contains nine atoms and three chemical units 

(z = 3) are contained in the crystallographic repeat unit [77]. Thus, we expect 

3 x 9 = 27 phonon branches in q space or 3 x 9 x 3 = 81 in k space. 

Following the previous discussion we expect: 

0 27-2 totally symmetric q = 0 nonzero frequency modes of A species, infrared 

active with 1) dichroisni and Raman active in zz polarization. 

0 54-2 doubly degenerate modes (thus 27-1 frequencies) at y = B = 2n/3 of E 

species infrared and Ranian active (1). The phonons for q~ = 28 coincide with 

the phonons at q = 0 because of the periodicity of the functions ~ (y) (see Figure 

3-10a). 

The observed infrared spectrum of an isotropic sample of IPP is shown in Figure 

3-9a, and clearly identifies the regularity bands associated with 1 D periodicity. The 

assignment of the observed infrared bands to vibrations of species A, B = 0" (11) and 

E, B = 120" (1) is carried out on a stretch-oriented sample (Figure 3-9b). 

Dispersion curves have been calculated (Figure 3-10a) [56] from which spectroscopically 

active phonons have been derived for q = 0 and 27c/3. One-phonon density 

of states have been calculated and compared with the few data experimentally 

available from neuron-scattering spectroscopy (Figure 3-lob) (see Section 3.8). In 

Figure 3- 1 Oa, the dispersion curves of isotactic polypropylene are reported in terms 

of the phase coupling y; in the same figure the six high-energy branches (associated 

with the C-H-stretching modes in the 3000 cnir' range are omitted. The C-H 

branches are flat (i.e., C-H-stretching phonons practically do not show dispersion 

with p). 

Next, let IPP chains crystallize in a cell with space group isomorphous with the 

point group C, with four molecules per unit cell in three dimensions. In principle, 

one should expect (9 x 3 x 4 x 3) - 3 = 321 vibrations in the spectrum for the 

spectroscopically active phonons at the symmetry point r (i.e. k, = k, = k, = 0). 

This would mean that, in principle, each of the 25A + (2 x 26)E k = 0 bands of the 

single chain should split into four lines if intermolecular Van der Waals-type forces 

between atoms were strong enough to generate an observable splitting. Generally, 

such forces are very weak, especially since interchain distances are relatively large. 

If such splitting could be observed, true infrared crystallinity bands of crystalline 

IPP would be identified. The search for such splitting has not been easy and we 

made careful experiments at very low temperature aiming at shrinking the crystal, 

thus increasing the interchain interactions which should increase the correlation 

field splitting (i.e. k, = 0 line-group modes in 1D should show splitting into niultiplets 

due to phonons with k, = k, = k, = 0 in 3D (phonons at the syinnietry point 

r in the BZ). Moreover, bands should become sharper since 'multiphonon states' 

are depopulated. After much work the experiments were finally successful and the 

Ranian spectrum of IPP at liquid nitrogen temperature shows indeed multiplet 

splittings due to crystallinity (Figure 3-1 1) [78].

116 3 Vihtionul Syectru us n Probe oj Stritcturul Order 

Q

3.8 Density of Vibrational States and Neutuopi Scattering 117 

1360 

I/‘\ 

i r 

/ A 

1040 

‘. 

81 0 

I 

/ 

I 

I 

Figure 3-11. Raman spectrum of isotactic polypropylene from ‘smectic’ (- - -) to crystalline (-) 

(tridimensional order) obtained after long annealing processes at low temperatures. Raman regularity 

bands split into doublets (crystallinity bands) due to phonons at r (k, = k, = k, = 0) (from 

[781). 

The case of isotactic polypropylene becomes an interesting texbook case which 

shows clearly and separately the many steps in the formation of a polymeric 

material. 

3.8 Density of Vibrational States and Neutron Scattering 

The next step in the development of vibrational spectroscopy of large organic 

molecules is to realize that other physical measurements, such as neutron-scattering 

spectroscopy, can provide essential information on the normal vibrations of organic 

molecules. This information is complementary to that provided by infrared and 

Raman spectroscopy which is intrinsically limited by the optical selection rules. 

Indeed, the limitations to the spectroscopic activity imposed by symmetry selection 

rules are very strict and select only a few phonons. Moreover, when some sort of 

molecular disordering is introduced into the system (and this is always the case), 

symmetry is removed, selection rules are relaxed, and the spectra show sharp 

absorptions (or Raman scattering) originating only from characteristic localized 

vibrations floating on top of a broad background absorption whose origin needs to 

be explored and understood.

11 8 3 Vihrationul Spectra ns a Probe of Structural 0rdek.v 

a b C

3.8 Dtvz.sit,v of L’ihrntioizal Strrtes aid Nezrtroiz Scutteriny 1 19 

We will introduce the basic principle and use of neutron-scattering spectroscopy 

by first introducing an important dynamical quantity, namely the density of vibrational 

states g(i1). 

Let g(v) be the fraction of normal modes with frequencies in the interval v and 

v + dv with dv 0. g(v) can be calculated analytically only for very simple atomic 

chains which are never encountered when realistic polymeric materials must be 

studied. 

Let us take a very large molecule and let 3N z C be the number of vibrations of 

the whole system; if c is the number of frequencies in the interval AV = v + dv then 

go)) = c(Av)/C (3-50) 

The total density of states is evaluated as the sum over all the r branches of the 

dispersion relation (Eq. (3-43)). Numerical methods such as the ‘root sampling 

method’ have been proposed for simple one- or tridimensional lattices [511. When 

the systems become more complex, as in the cases discussed in this chapter, the 

easiest numerical method is that proposed by Dean based on the application of the 

Negative Eigenvalues Theorem (NET)[79]. The method is extremely useful and has 

been extensively used in the laboratory of the writer for many polymeric cases. The 

details of the method are discussed in the next section. 

NET is a way to calculate g(v), i.e., it provides the number (or the fraction) of 

vibrational energy states or normal modes in a chosen y (or k) interval. g(v) can be 

plotted as an histogram whose accuracy depends on the frequency interval Ait chosen 

in the numerical calculation (numerical resolution of the computer experiment). 

This will be used later in this chapter in the study of disordered systems. 

Let us consider g(v) of an ideal infinite polymer and examine the case of singlechain 

polyvinylchloride in its three possible structures (threefold helix isotactic, 

trans-planar syndiotactic, folded syndiotactic) [go]. Foi- easier understanding, Figure 

3-12 shows both dispersion curves and g(v) for this polymer in its three possible 

structures. It is apparent that the shape of g(v) is determined by the shape of the 

dispersion curves which in turn depend on the vibrational force field used in the 

calculation of ~(y). Flat sections of v(q) give rise to strong ‘singularities’ in g(v). 

The shape of the dispersion curves and of electronic and/or vibrational g(v) for 

both electronic bands and vibrational bands have been the subjects of extensive 

theoretical treatments [51]. For infinite and perfect systems, the shape of g(i1) is 

characterized by different types of singularities which are related to the dimensionality 

of the systems and to many other physical factors. We restrict our analysis to 

the case of polymeric materials which can be considered as one-dimensional lattices. 

Our discussion can only be sketchy and qualitative, but it is aimed at pointing out 

the features most relevant to the problems of disorder treated in this chapter. 

T 

Figure 3-12. Phonon dispersion curves v(p) (a, c). i(k) (b) and density of vibrational states gii,) from 

0 to 1500 cn-’ of three possible models of a chain of polyvinylchloride showing the lR-Rainan 

active and the inactive ‘singularities’ corresponding to flat sections of the disperion CUI-vcs (from 

[801).

120 3 Vibratiorial Spectra us a Probe of Structural Order 

The following observations should be kept in mind. 

1. Because of the shape of the calculated v(q), which may show many maxima and 

minima (due to the vibrational coupling intrinsic to the system and to the force 

field used), g(v) will also show corresponding strong or weak singularities. Some 

of these singularities necessarily coincide with k = 0 spectroscopically active 

phonons, while other will not be seen in the optical spectrum of a perfect system, 

but will (in principle) be observed in the neutron-scattering spectra and possibly 

in the optical spectra of a partially disordered polymer. In a disordered material, 

no symmetry selection rules are active and all modes may gain some activity. 

The vibrational spectra of a disordered material can thus be considered the 

mapping of g(v) dipole (or polarizability) weighted. The g(v) calculated with 

NET or any other numerical or analytical method is, instead, dipole (polarizability) 

unweighted. Any comparison with the experimental g(v) must be made 

with great caution. 

2. As discussed in Section 3.5 for tridimensional lattices, v i 0 for k 40 for the 

three acoustical branches with a positive slope and with a discontinuity at k = 0 

(r). The shape of g(v) i 0 has been matter of strong interest in physics for the 

calculation of specific heat and related thermodynamic quantities. As already 

mentioned for one-dimensional lattices in LIUCUO, some of the acoustical branches 

tend asymptotically to zero. The derived calculated g(v) shows a very strong 

singularity at v = 0. Such a singularity is meaningless and only due to the limitation 

of the molecular modes adopted in the calculations (1D lattice) which is 

unable to account for intermolecular forces. Also, for real polymer samples the 

experimental g(v) i 0 for k i 0 as expected for classical crystals since they do 

interact with the neighboring chains with very weak intermolecular forces. 

3. Additional singularities both in calculated and experimental g(v) are expected at 

very low energies (~0-30 cm-’) originating from very low-frequency 3D-lattice 

phonons. These singularities may coincide with the optical phonons in infrared 

and/or Raman spectra at very low frequencies. 

As already anticipated, a complementary experimental technique for deriving 

information on the dynamics (frequencies and vibrational amplitudes) of polymers 

or of materials in general is the use of inelastic neutron-scattering techniques (INS). 

After a long development time, during which experiments were difficult and provided 

limited information, the instruments in a few specialized centers recently 

began to provide detailed data covering the whole spectrum. Thus, we predict a 

‘renaissance’ of INS techniques for the studies of molecular and lattice dynamics. 

For a thorough discussion of INS spectroscopy we refer the reader to specialized 

publications [8 13 (see also other references quoted below); here, we restrict ourselves 

to the basic principles of the technique connected especially with the dynamic 

quantities which are of interest in optical spectroscopy. 

Let a beam of cold neutrons made mono-energetic with wave-vector kll be shone 

onto a sample, and let the beam be scattered inelastically and incoherently. The 

outgoing neutrons have wave-vector kf and are collected and detected by suitable 

devices. Since the energy of the thermal neutrons is of the same order as that of

3.8 Driisity of Vibratioiid StLitrs arid Neutron Scuttering 121 

lattice and molecular vibrations, neutrons impinging on the sample may gain or lose 

energy in a way proportional to g(v). 

It has been shown [S2] that the scattering cross-section of a polycrystalline material 

is approximately the same as that of cubic crystals. We are, at present, concerned 

with one-phonon processes and neglect the possibility that multiphonon 

processes may occur (also in vibrational spectra we have neglected anharmonic 

effect with the consequent appearance of overtones, combinations, and hot bands). 

The one-phonon scattering cross-section for a cubic lattice can be written as [83, 

S4]: 

d2a,,,,h/d0 do = A1 kf 1 /I ko lep2w (hK2/2M)[g(co)/oi] 

x [h'(of - wo + o)/(eP" - I)] (3-51) 

In Eq. (3-51), and in a few equations which will follow the vibrational frequencies 

are not expressed as v in wavenumbers, but as w in Hz. In Eq. (3-51), A is the 

scattering cross-section of the scattering nucleus (since the H atom has a very large 

cross section, namely AH = 82.5 barn, Ac = 5.5 barn, molecules containing hydrogen 

are very suitable for neutron experiments), lkfl and lkol are the moduli of the 

wave-vectors of the scattered and incident neutron respectively, e-2w the Debye- 

Waller factor or temperature factor, K = kf - ko, M the mass of the unit cell, g(w) 

the desired density of vibrational states, and p = h/?kBT (where kB and T are the 

Boltzman factor and the temperature respectively). 

It is clear that thermal population is strongly contributing to the scattering crosssection 

with the Debye-Waller factor and the term p. Measurements must then be 

carried out at very low temperature. 

The experimental g(co) (or g(v)), while showing qualitatively all the features of 

the calculated g(w), cannot yet be quantitatively compared because the vibrational 

amplitudes (generally called polarization vectors in neutron-scattering circles) of the 

various normal modes (phonons) are not yet taken into account. The situation is 

similar to the features of the vibrational spectra of an isotropic sample compared 

with the features of the spectra in polarized light of an oriented anisotropic sample. 

Let us take a stretch-oriented sample of a polymer subject to a neutron beam with 

incident wave-vector ko. A more complete expression of the scattering cross-section 

is the following [S2, 841: 

Some of the terms appearing in Eq. (3-52) are already defined for Eq. (3-51). The 

new terms are the following: A, is the scattering cross-section of the n-th nucleus of 

the unit cell, e-2W11 the Debye-Waller factor for the n-th nucleus of mass MI,, coi the 

frequency of the j-tli normal mode of the r-th branch in the dispersion relation. The 

most important quantity is ek. C:' which gives the projection along ko of the atomic

122 3 Vibrational Spectra as N Probe of Structural Order 

displacement C of atom n for the j-th normal mode of the r branch with frequency 

wi. The sum is extended to all the atoms n in the unit cell and to all branches r in 

the dispersion relation. The Debye-Waller factor W, of each atom n has an explicit 

expression [83, 841 which again contains quantities defined in Eqs. (3-51) and (3-52). 

In spite of the seemingly complex expressions given in Eqs. (3-51) and (3-52) for 

the scattering cross-sections, all terms appearing in these equations are known from 

dynamics and from experiments? and have been defined in the previous sections of 

this chapter. Automated computer programs have been added to the classical programs 

for normal coordinate calculations of polymers for the calculations of the 

INS spectra. 

The reason of requiring the experiments to be carried out on stretch-oriented 

samples mounted with known geometry with respect to the incoming neutron beam 

is the existence of the ek . Cy,r term which indicates that the scattering cross-section 

is determined by the directional interaction of the neutrons with wave vector ko and 

the direction of the atomic displacements. This situation is very similar in the experiments 

of infrared spectroscopy in polarized light with a stretch-oriented material. 

In dichroism experiments the absorption intensity is proportional to /pi. El’, i.e., 

to the square of the projection of the molecular dipole moment change p during the 

i-th normal mode onto the electric field of the light beam. 

The g(w) measured with INS experiments on stretch-oriented materials is then 

‘amplitude-weighted’ and is ‘directional’. The directionality of the displacements 

associated with a given normal mode can be determined by placing ko 11 and I to 

the chain direction. 

The history of polymer neutron spectroscopy can be divided in two different time 

domains. The enthusiasm for INS experiments in polymers reached its highest peak 

in the 1970s, but the experiments have been restricted to a few polymers with a 

limited number of data [84, 851 which could hardly solve particular problems in 

polymer dynamics. The availability of modern instruments with higher resolution 

and with the possibility of removing most of multiphonon scattering has presently 

revived the field and vibrational spectroscopists await with great interest for the 

revival of this important field of research. 

In Figure 3-13 we show the most recent INS results reported by Parker [86] on 

oriented polyethylene at 30 K obtained with the high-resolution, broad-band spectrometer 

(TFXA) at ISIS pulsed spallation neutron source at the Rutherford Appleton 

Laboratory, Chilton, UK. Notice that the whole spectral range 0-4000 cnc‘ 

range has been explored with considerable resolution, especially in the lower-energy 

region. 

3.9 Moving Towards Reality: From Order to Disorder 

3.9.1 Basic Concepts 

In reality, polymeric materials never possess the perfect chemical stereochemical 

and conforinational regularity which were assumed at the beginning of Section 3.5.

3 9 Mouiiiq ToIvmds Rerrlitv: From Order to Disordrr 123 

I " ~ ' " " l " " l ' ' ' ' 

I 1 

Figure 3-13. Inelastic neutron scat- 1 , , , , I , , , , I , , , I , , , , 

teriiig (INS) spectrum of polyethyl- 0 1000 2000 3000 4000 

ene at 30 K (from [86]). 

wavenumbedcm -' 

The ideal infinite polymer chain iiz ziucuo has allowed to consti-uct dynamical 

theories which describe ideal spectroscopic properties of the polymer determined 

through strict optical selection rules and obtained by ideally clean experiments. In 

reality, polymerization reactions never yield perfect chemical linking and perfect 

stereo-regularity; these chemical and stereochemical defects necessarily introduce 

conformational distortions as to avoid steric encumbrance. Moreover, the complex 

processes of crystallization from melt or solution of a very long and flexible object 

implies the generation of various kinds of morphologies which reflect particular 

packing of 'conformationally' ordered chains. New concepts such as chain folding, 

looping, tie molecules, fringed micelles, etc., are necessarily introduced for the description 

of the building up of the supermolecular structure of a polymer chain. It 

follows that any polymer material which has been allowed to crystallize consists of 

a fraction of matter crystallized in a 3D-lattice and a fraction of 'conformationally 

irregular' material [46, 871. 

In polymer material science it is essential to know the detailed independent 

structural characteristics (and the relative concentrations) of the crystalline and 

'amorphous' (or irregular) fractions of a polymer sample. All mechanical, viscoelastic 

or, in general, the physical properties of real polymeric materials depend on 

the relative fraction of crystalline versus non crystalline material. 

Since the spectroscopic manifestations of molecular disorder are many and very 

characteristic, the vibrational spectrum has become a usefyl probe for molecular 

order-disorder which can be described at the level of a few Angstroms. 

Spectroscopists are asked to find ways to account for the new spectral features 

associated to disorder, thus making vibrational spectra useful for the practical 

structural characterization of a real polymer material. Obviously, the theories presented 

in the previous sections of this chapter must be adapted or re-worked in 

order to understand the spectra of disordered organic polymers. 

The prototypical case of an experimental observation to be accounted for is the 

spectrum of a sample of orthorhombic polyethylene [88] (Figure 3-14). Let us con-

124 3 Vibvcrtinntrl Spectvci as ci Prohe qf Stvzicturrrl Order 

100, II I I I 

Figure 3-14. Comparison of the calculated 

g(v) of single-chain polyethylene with the 

infrared spectrum of an e\rtended chain of 

polyethylene. 

sider the experimental infrared spectrum of high-density crystalline polyethylene 

(Figure 3-14). The infrared active phonons at r for an orthorhonibic lattice with 

two chains per unit cell show the predicted crystallinity bands as doublets originating 

from the k = 0 modes of the former regularity bands for a chain with a ID 

symmetry operator $(Q = n, d/2). In Figure 3-14, the infrared spectrum is superposed 

to the singularities of the calculated g(v) while the Ranian active peaks are 

indicated in the figure. It is quite clear that not all the peaks in the observed vibrational 

infrared or Ranian spectrum find a one-to-one correspondence with the 

singularities in the calculated g(v). It follows that, in addition to the ‘perfect’ part of 

the all-trans polyethylene chain, extra structural features must exist whose existence 

is indicated by the additional peaks in the infrared or Raman spectra. It will be 

shown that the extra absorption or scattering are associated with various kinds of 

conformational defects. 

In this section we proceed as follows. 

1. First, we take the perfect and infinite polymer chain and truncate the chain in 

increasingly shorter segments and predict theoretically and watch experimentally 

the effects of such truncation. The structural perfection is retained. 

2. Any kind of chemical and/or structural disorder can be introduced into the 

chain. We chose the various plausible kinds of disorder, their population and 

distribution, and try to predict the infrared, Raman and neutron-scattering 

spectra by suitable theories and computational techniques. 

3. Calculated theoretical and experimental cases will be presented. 

3.9.2. Finite Chains 

The theoretical modeling of the vibrations of molecular chains with finite length has 

been nicely treated by Zbinden [18] and by Snyder and Schachschneider [9, 341. 

While the approach by Zbinden is mathematically complete, but it applies to simplified 

monoatomic chains, quite away from chemical reality, the model by Snyder 

and Schachschneider is more directly applicable to real molecules and will be

3.9 Moving Towards Reality: From Order to Disorder 125 

quickly mentioned here. We restrict our discussion to the most connnon case of 0- 

bonded chains, which, as discussed in Section 3.5, do not permit long-range electronic 

coupling. 

Let y~ be the number of identical chemical units in a chain segment (different 

chemical groups at either ends are neglected) and let each chemical unit consist of n 

atoms; we treat here the case of a model molecule with free ends [18]. Let us take 

one chemical unit and its 3 x n oscillators (which, for sake of clarity, we describe 

here as internal (R) or group coordinates II). Because of elastic coupling with its 

neighbor at either side, each oscillator couples with its identical neighbors and 

generates 11 normal modes which can be described as waves characterized by a 

phase coupling y = jn/ll (where j = 0,1,2,. . . y - 1). Their 11 frequencies y(y) lie at 

finite y values along the dispersion curve of the corresponding infinite polymer with 

identical chemical, stereochemical, and conformational structure. Each vibration 

corresponds to a given phase coupling defined by the number of chemical repeat 

unit which make up the finite chain. The introduction of phase coupling implies the 

fact that within such short chains, nomal modes describe quasi regular waves (quasi 

phonons) similar to those which propagate in an infinite chain. The remaining 

modes which do not lie on the branches of the dispersion curves must be associated 

with the vibrations localized on the end groups and can be identified by the constant 

values of the frequencies when chain length is changed. It follows that: 

(i) if I? is the number of repeat units making up the finite chain (neglecting the two 

end groups) one expects to find a sequence of bands (or band progression) 

which lie on each of the dispersion branches of the corresponding polymer. 

Let us take for example the n-alkane n-nonadecane (CH~(CH~)I~CH~). Let us 

focus on the vibrations of the segment consisting of y~ = 17 CH2 groups all in 

trails conformation. Each CH2 generates 3 x 3 = 9 vibrations (CH2 antisymmetric 

and symmetric stretch, CH? bending, wagging, twisting, rocking, 

antisymmetric and symmetric C-C stretch, CCC bending, and C-C torsion). 

Each of these ‘oscillators’ generates 17 waves, each of which is characterized by 

a phase coupling nj/ 17. The corresponding frequencies lie on the dispersion 

curves of a single chain of infinite all-trans polyethylene. The band progressions 

observed for n-nonadecane will be discussed in Section 3.19. 

(ii) The frequency range spanned by each band progression depends on the extent 

of intramolecular coupling. If the dispersion of the frequency branch is small, 

the progression is squeezed, many bands overlap (e.g., the stretching vibrations 

of C-H groups), and no progression can be observed experimentally. When 

intramolecular coupling is large, band progressions are clearly observed 

(-1 100-720 cm-’ is the frequency range covered by CH2 rockings in all-trans 

n-alkanes). 

(iii) The intensity in infrared and/or Raman of each of the bands within the progressions 

depends on the dipole (polarizability) changes associated with the 

corresponding vibration (quasi phonon with y, z jn/y). We have previously 

seen that for an infinite chain only k = 0 phonons are active and all other are 

inactive, as dipoles or polarizability changes cancel out in the case of perfect 

phonon waves. In going from short to longer chains, optical selection rules

tend to the limit of the infinite case and we expect to find in the spectra of the 

finite, but long, chains clear (generally, but not always strong) quasi k = 0 

phonons while the other members of the band progression will show quickly 

decreasing intensity. This is the typical case observed in the series of n-alkanes 

by Snyder and Schachschneider [9. 341. 

(iv) It is obvious that if normal-mode calculations have permitted calculation of the 

dispersion curves for the infinite polymer, the identification of the band progressions 

in finite chains with identical structure is made easier. In contrast. if a 

systematic study is carried out on a series of inolecules of the same chemical 

class, but with increasing chain length, the experimental identification of the 

band progressions allows experiinental determination of the phonon dispersion 

branches from which a reliable force field can be derived. Vibrational spectroscopy 

then becomes a complementary tool (sometimes unique) to the extremely 

more expensive and elaborate neutron-scattering techniques. The whole work 

has been very clearly and precisely explained and successfully applied by 

Snyder and Schachschneider [9, 341. 

3.9.3 Polymer Chains with Structural Defects 

In order to model the reality of a polymeric material we need first to introduce in 

the calculations the various energetically possible structural defects. The tjFes of’ 

defects to be considered are the following: 

0 Clzenzical defects, e.g., head-to-head linking in an otherwise head-to-tail chain. A 

typical case is the real structure of the chain of polyvinylfluoride ICHz-CFz),, 

which contains a sizeable fraction of undesired head-to-head defects [89]. Once 

the chemistry of a given polymer is approximately known, other kinds of defect 

structures can be envisaged. A typical case is often found for isotopically substituted 

chains when substitution is not ideally complete. 

Stereocheiiiicnl defects, e.g., syndiotactic configurations in an otherwise isotactic 

chain. 

Cotiforwmtiond defects, e.g., gauche conformers in an otherwise all-trnizs chain 

structure. 

It is obvious that, because of intramolecular interactions, the introduction of 

some kind of chemical and/or stereochemical defects forces also the introduction of 


Next, the model requires the definition of the concentrafion mid distribiitiori (e.g., 

Bernoullian, Markovian, etc.) of defects. If a small concentration of defects with a 

random distribution is considered, defects most probably are isolated in the host 

polymer ID lattice. When the concentration increases, even a random distribution 

generates both isolated defects and a distribution of ‘islands’ of various lengths (see, 

for instance, [90, 911). 

When such variables are well defined, calculations require the construction of the 

usual dynamical matrix and the solution of the corresponding eigenvalue equation.

3.9 Moving Towards Reality: Froin Order to Disorder 121 

It becomes apparent that as the translational symmetry of the chain is removed 

by the existence of defects, any periodicity is lost and Eq. (3-431 cannot be used. 

Ratter, Eq. (3-17) must be used which corresponds to the dynamical case of a finite 

molecule. Since in this case our molecular model is huge and has no symmetry, the 

size of the secular equation to be solved becomes extremely large. Moreover, such 

types of study require the freedom to try many models with different kinds, concentrations, 

and distributions of defects. Last, but not least, in order to play a game 

as close as possible to reality in these studies, molecular models must be composed 

of as many monomer units as possible. 

In solid-state physics, the vibrations of very simple lattices containing a small 

concentration of simple defects have been treated with sophisticated analytical 

treatments using Green’s fiinctions 1921. Even if some authors have bravely tackled 

an analytical solution of the dynamical problem of disordered polymers [93], they 

were forced to introduce into the molecule such drastic structural simplifications 

that the flavor of chemistry has been lost and the theoretical molecular models have 

become, again, too unrealistic. 

For our complex molecular systems the problem must be solved by numerical 

methods. The problem of the solution of a huge secular equation producing thousands 

of vibrational frequencies can be solved by the application of the so-called 

‘Negative Eigenvalue Theorem’ (NET) originally proposed by Dean [79] who 

aimed at the calculation of the vibrational spectra of disordered ice. We think that 

the original work by Dean did not receive due acknowledgement; indeed, his 

methods has been widely applied (with little due reference to Dean) in solid-state 

physics and in molecular theories whenever the eigenvalues corresponding to vibrational 

or electronic states of huge and disordered systems had to be calculated. (As 

an example of the calculation on the electronic states of disordered systems, see 

[941). 

Dean’s method calculates the density of states (vibrational or electronic) in a 

given energy range. The whole energy range can be spanned by the calculation and 

the complete g(v) is obtained and can be plotted as an histogram to be compared 

with the experimental spectrum. The accuracy of the histogram can be improved by 

narrowing the steps (Av) in the energy interval in each cycle of the calculation (i.e., 

by improving the resolution of the numerical experiment). 

Dean’s method works well for band-matrices which are always found in the case 

of polymer chains or in the case of tridimensional lattices with short-range interactions. 

The reader is referred to the original papers for a discussion of the method, 

its advantages and limitations [91. 951. 

Once the histogram of g(v) is plotted, one needs to find the eigenvalues for such 

huge matrices corresponding to the approximate eigenvalues comprised in a given 

energy range. The problem has been solved with the application of the Wilkinson’s 

inverse iteration method [96] (hereafter referred as to IIM; see for instance [97]). 

The comparison of the calculated g(v) with the actual experimental infrared, 

Ranian and neutron-scattering spectra requires some care and possibly more refined 

calculations with improved resolution. g(v) gives one information, namely how the 

very many normal modes are clustered in a given frequency range. From IIM we 

may also learn the shape of the normal modes at a chosen frequency. Nothing is

128 3 Vibrational Spectra as a Probe of Stvuctzirnl Order 

known about the dipole transition moments (infrared intensities) or Raman or 

neutron-scattering cross-sections unless specific calculations are carried out using 

other models for the electronic distribution (dipole derivatives, polarizability derivatives, 

and vibrational amplitudes respectively 1971). Care must then be taken in 

carrying out the comparison between calculated g( v) and observed spectra. 

3.10 What Do We Learn from Calculations? 

We have just outlined the way theoreticians in polymer dynamics face the problem 

of the understanding of the vibrational spectra of polymers in their realistic state. 

The reader interested in the theoretical aspects will find in the references quoted all 

indications which will enable him or her to grasp the theory and to carry out the 

calculations. 

However, a larger number of reader do not wish to play with calculations, but 

simply to know what we have learned about polymer structure and spectra using 

these calculations and to apply quickly and directly the knowledge acquired to their 

own problems. We report here the concepts of general validity in polymer spectroscopy 

which were derived from calculations and which can be applied to any polymer 

(for a general discussion, see [ 16, 19, 661). 

The reader is certainly familiar with the traditional group-frequency approach 

generally used in the past 50 years for the chemical application of the infrared 

spectra (Section 3.4). Correlation tables and books have been written which discuss 

group-frequency correlations and provide the way to carry out a chemical diagnosis 

from the vibrational spectrum (so far, the infrared spectra have enjoyed great 

popularity; the reader is advised to extend their interest to the very useful and easily 

available Raman spectra). 

As already discussed in Section 3.4, from the dynaniical viewpoint chemical correlations 

are based on the fact that there are a few vibrations (normal modes) which 

are Imyely localized on the functional group and give rise to medium/strong absorption 

bands in infrared (scattering lines in the Raman) at specific and wellisolated 

frequencies. The occurrence of such bands in the spectrum implies the 

existence of such a functional group in the molecule under study. On the other 

hand, when the vibrations involve a large molecular domain (collectiue motions) the 

observed bands are no longer so characteristic of a given group of atoms. 

The vibrations of disordered polymer chains follow the same kind of conceptual 

path which, on the other hand, was also found earlier by physicists in the study of 

lattice dynamics of very simple disordered lattices [98]. 

Let us make the problem simpler by considering briefly the small molecular unit 

which contains the defect as an isolated entity consisting of n atoms which generate 

3n normal vibrations that occur somewhere in the vibrational spectrum. Next, we 

insert just one defect unit into the otherwise perfect polymer chain and allow the 

whole system to display its own new dynamics through the observed (or calculated) 

vibrational spectra. 

The dynamics (new frequencies and displacements, see Eq. (3-17)) of the defect

3.10 Wlmt Do We Learn from Cnlcailcitioizs? 129 

COz rock 

___c 

I 

GAP W S ~ W ~ 

/ 

Figure 3-15. Example of a dynainical situation of a po1yniel;ylene r chain -(CH2)n- which contains 

defects consisting of CD2 groups and of CH3 units at either ends. The motion of CD:! rocking 

occurs in a gap and does not couple with the host lattice, thus generating localized modes; on the 

contrary the frequencies of both external deformation and umbrella motions of the CH3 group 

occur in the frequency range spanned by the dispersion curves of the host lattice; coupling takes 

place and resonance modes are generated. 

embedded in the polymer chain is determined by the masses and the geometrical 

arrangement of its vibrating atoms (G matrix, Eq. (3-15)) and by the new intramolecular 

forces, i.e., by the new force constants (type, strength, and distance of 

intra-unit and inter-unit coupling (F matrix, Eqs. j3-2), (3-3) and (3-16)). 

Calculations carried out so far on the simplest prototypical classes of polymer 

describe substantially three types of phenomena (Figure 3- 15):

130 3 Vibratiod Spectru as LL Probe of Structural Order 

Phonon mode 

a 

Resonance mode 

b 

Figure 3-16. Dynamics of a one-dimensional lattice containing one defect. Vibrational displacements 

of each unit are rotated of 90” for a better display of the normal modes. (a) uiiperturbcd 

phonon of the host lattice; (b) resonance mode: and icj gap/local mode if!-om [66]). 

1. Some of the normal modes of the defect strongly couple with the normal modes 

(phonons) of the polymer and generate collective motions which lose all the 

characteristic of the motion of the isolated unit. These normal modes are generally 

labeled as ‘resomince ~zodes’ (Figure 3- 16) because their frequencies are close 

enough to those of the host lattice that they generate mechanical resonance with 

some of the normal modes of the polymer. This is generally found when the 

frequency of the separated unit happens to occur within the range of frequencies 

spanned by one of the dispersion curves of the perfect host polymer 1D lattice. 

2. When the frequency of the separated unit differs strongly from the frequencies 

spanned by the dispersion branches of the polymer chain, the motion of the unit 

cannot couple with the motions of the host lattice, it cannot be propagated and 

generates a mode strongly localized in space (at the site of the defect within the 

polymer) and in energy (at a characteristic specific frequency). These modes are 

generally labeled as ‘yap modes’ or ‘out-oflbancl modes’, since the frequency of 

the isolated units occurs either in a gap between dispersion branches 01- above 

the highest frequency branch respectively. Out-of-band modes in organic polymers 

can only be observed in the case of perdeuterated polymers which contain 

H atoms as defects. In this case, C-H stretching modes occur at frequencies 

higher than any other normal modes of the polymer. 

3. Compared with the results from the dynamics of simple impure 3D lattices, in 

the case of 1D lattices (polymers) localized modes (or quasi-localized modes) can 

sometimes also be generated if the frequency of the defect unit occurs within the 

frequencies spanned by a dispersion branch of the perfect host lattice. As the

+ 

3.11 A Very Siriiple Ccrse: Lattice Dyriurriics of HCI-DCI Mi-xed CrjistLiIs 131 

polymer chain is conformationally flexible it may well occur that, for a certain 

geometrical arrangement of the atoms. the motion of the defect unit happens to 

be orthogonal (or quasi-orthogonal) to the displacements of the atoms during a 

certain vibration of the host lattice. It has been shown [99] that in an n-alkane 

chain the CH: wagging of the central unit in a conformational sequence such as 

-(G)-CH?-(G)- generates a pseudo-localized CH2 wagging motion almost fully 

decoupled from the vibrations of the host lattice, even if the frequency occurs in 

the frequency range spanned by the CH? wagging dispersion branch [99]. 

The observation of characteristic gap modes in disordered polymers becomes 

then the basis for the use of the vibrational spectra as useful probes for the structural 

characterization of these systems. 

It should be remembered that the above discussion has been presented by considering 

only one single defect embedded in a host polymer chain. When a distribution 

of defects is considered, defects may occur close enough to couple mechanically. 

Splitting and changes of the gap modes (frequencies and intensities) may occur and 

are indicated by the calculations (see, for instance [ 1001). 

3.11 A Very Simple Case: Lattice Dynamics of 

HCl-DCl Mixed Crystals 

It is considered that the simplest possible realistic case on which our theories can be 

tested, and which could act as a guide to the reader for the discussions which follow, 

are the studies of mixed crystals of HC1 and DCl [97]. Solid HCl and DC1 undergo 

a first-order phase transition at 98.4 OK, and 105.0 OK, respectively. Below the transition 

temperature, X-ray and neutron diffraction studies show the existence of an 

ordered face-centered orthorhombic structure which contains planar zigzag chains 

consisting of HCl (DCl) molecules linked to each other by hydrogen bonds (for a 

list of references on the physical properties of solid HCl, see [97]). Our model of the 

chain consists of a pure infinite isolated chain of the type ... H-Cl.-.H-Cl...H-Cl... 

(Figure 3-17). Using the techniques presented in the previous sections we have 

Figure 3-17. Geometry and internal 

coordinates (in-plane and out-of-plane) for 

crystalline HCI or DC1 in the orthorhombic 

- 

modification. Only one single chain has been 

- + 

considered. =z - + + -

132 3 T/ihrtitional Spectra us LL Probe of Structural Order 

a 

Figure 3-18. Dynaniical data 

for the perfect chain of HCl 

and DCl based on a valence 

force field. (a) Dispersion 

curves; (b) k = 0 normal 

modes; (c) densities of 

vibrational states. 

calculated dispersion curves (Figure 3-l8a), density of vibrational states g( 11) (Figure 

3-l8b) and vibrational displacements for k = 0 phonons both for pure HCI and 

DCI (Figure. 3-18c). Vibrational infrared intensities have also been calculated based 

on a simple fixed point-charge model [97]. 

As a second step, we generated models of chains of HCl containing defects of 

DC1. Calculatioiis are made easier because isotopic substitution does not change the 

force field. Several mixtures were considered and we will focus here ody on three 

cases of mixed chains with 3%, 25% and 45Y” DC1 in HCI. Chains were made up by 

600 chemical units.

3 I1 A Yey) Siiiiple Case: LLittice Dynnniics of HCl-DCl Mixed Cvvstu1,s 133 

Figure 3-18 (b) 

b 

Suitable experimental procedures were envisaged in order to obtain such materials 

and in order to record the infrared spectra to be compared with the calculated 

ones. 

Figure 3-18a immediately indicates that the HC1 stretching of molecules isolated 

in a DCl lattice can generate localized out-of-band modes, while DCl stretching in 

a HCl host lattice generate typical localized gap-modes. This can occur for other 

lower-frequency modes for both molecules. The 03-co4 branches span frequency 

ranges with partial overlapping between HCl and DCI lattices, thus possibly generating 

resonance modes. Also, the vibrations of 07-Wg may couple. No coupling 

can occur between the out-of plane phonons of HC1 (05-06) and the in-plane 

phonons of DC1 (w3-014) since they are orthogonal. 

The histograms calculated for the three isotopically impure chains are given in 

Figure 3-19 and show a complicated and almost inextricable pattern. Some details 

are enlarged in Figure 3-20.

134 

3 I 'ibrutioiwl S'Jectra us LI Probe of' Structud O~tler 

y (w. 

i 56 

I 

HCf 

100 

50 

C 

Figure 3-18 (c) 

The object of the next step is to apply IIM to identify the origin of the various 

peaks calculated in the g(v) of the mixtures. First, we consider the random distribution 

of 25"h DCl defects in a host lattice of HCI (Figure 3-21); the total number 

of units in this calculation is 100. Let Z be the number of chemical units within the 

defect. Moreover, for sake of completeness, let us also associate qualitatively to 

each motion its corresponding dipole moment changes. Let us assume that each bar 

in Figure 3-21 gives not only the vibrational amplitude, but also the dipole moment

3.11 A Vq> Siiiiplc Crrw: Luttice Djwmnics of HC1-DCI Mixed Crystals 135 

100 

glw i 

50 

0 

Figure 3-19. Density of vibrational 

states for a single mixed chain of (a) 3%. 

jb) 25%, and (c) 45% DC1 in HC1 (chains 

of 600 chemical units) 

0 100 

200 300 LOO 500 

Figure 3-20. (a) Experimental 

infrared spectrum in the HCI 

and DCl stretching regions of 

a mixed HCI/DCI crystal 

contaiuiiig 25'X of DC1; (b) 

g(i8) and (c) dipole-weighted 

g[ i t) assuming a Lorentzian 

band shape with Ai8,p = 2 

cm-'. The observed spectra 

have been shifted as indicated 

for sake of qualitative coinparison. 

A. 

L 

-A- 

19.33 1950 1970 

A 

2ioo 2i20 2i~o 27ko

136 3 Vilwtrtionrrl Spectra (1s N Probe of Striictiirul Orrlev 

I 

/I/ 

11 c) 

‘1 ‘1 w=2729.5 cm-l 

di 

I“ w=270u.5 cm-l 

el 

Figure 3-21. Sample eigenvectors 

in the HC1 stretching region for a 

single chain of mixed 25% DCI in 

HC1. The top line gives the distribution 

of defects. Vibrational displacements 

are rotated of 90” for 

motions. 

changes. The vectorial sum of the dipole moment changes is related to the absorption 

intensities. We can then qualitatively predict also the intensity of each of the 

calculated modes. The distribution of defects given at the top of the figure indicates 

the existence in the chain of 14 DCl isolated units (Z = l), 14 doublets (Z = 8) +1 

triplet (Z = 3). Application of NET, IIM, and intensity coinsiderations indicate that 

at 1’ = 1948.76 cn-l only the single DCl niolecules as isolated entities perforin fully 

localized vibrations at the site where they are located with strong intensity. At 

v = 1948.5 and 1948.68 cni-’, the four single units separated only by one HCl 

spacer feel some intramolecular coupling and generate localized modes with some 

phase coupling, the former has strong intensity and the latter zero intensity. Next, 

at 1’ = 1941.5 cm-’ only isolated doublets generate an in-phase motion (strong) of 

the two molecules. The out-of-phase motion occurs at Y = 1956.5 cm-] (inactive). 

Finally, the motions characteristic of the triplet occur at 1’ = 1938.5 (strong) and 

1959.9 cn1-l (weak). It is also apparent that in this frequency range the atoms which 

belong to the host lattice do not move (i.e., the motion of the DC1 defect is not 

propagated throughout the chain of the host HCl lattice). The concept we wish to 

stress here is that the cases just examined are typical examples of modes localized in 

space and energy characteristic of particular ‘defects’ or of ‘islands’ or ‘clusters’ of 

defects. Moving to the HCl stretching range (near 2700 cm-I) the analysis of the 

composition of the chain from localized modes can be completed. For the same 

chain considered in Figure 3-22 the distribution of HC1 sequences is the following: 

4 (Z = I), 4 (Z = 2), 3 (Z = 3), 2 (Z = 4), 2 (Z = 5), 2 (Z = 6), 1 (Z = 7), 1 (Z = 8) 

and 1 (Z = 9). From Figure 3-22, the ‘island analysis’ carried out using the IIM 

method allows the following frequencies Characteristic of each island to be located

Figure 3-22. Sample eigenvectors in 

the in-plane bending region for a 


HCI. The distribution of defects is 

the same as that given in Figure 

3-21. 

dl 

Wz360.6 cm-' 

unquestionably: Z = 1, 2719.5 cm-' (strong); Z = 2, 2708.5 cm-' (weak or inactive) 

and 2729.5 cm-' (inactive); Z = 5-6, 2700.5 cm-' (strong) and Z = 8-9, 

2739.5 cm-' (inactive). 

An example of resonance modes which cannot provide any signal characteristic 

of a given defect is given in Figure 3-23 for the same chain as that considered in 

Figure 3-21 and 3-22. In this case the bending modes are considered where (see 

above) frequency branches overlap. Because of intramolecular coupling, collective 

vibrational waves extending along large domains of the chain are generated and do 

not provide any useful and characteristic signal which may contribute with a weak 

and broad background. 

It is realized that the situation found for impure simple 3D lattices is found 

also for the single chain considered here and very likely may be found for other 

polymers containing an increasing concentration of defect. In the HCl (or DC1) 

stretching region these mixed crystals behave as 'two-mode system' or 'persistencetype' 

system. 'Amalgamation-type' behavior is observed in the bending range where 

sections of density of states overlap both for HCl and DCl [98].

138 3 Vibrcitional Spectru as a Probe oj Stnrcttiral Order 

dl 

w=3a1 a cm-‘ 

el 

w=355.0 cm-’ 

Figure 3-23. Examples of resonance 

modes in the bending region for a 


HCl. 

3.12 CIS-trans Opening of the Double Bond in the 

Polymerysation of Ethylene 

The method just outlined was used, at the earlier stages of these kinds of study 

[loll, for the elucidation of the mechanism of polymerization which was, at that 

time, an unsolved puzzle. The chemical problem can be stated briefly in the following 

way: when ethylene is polymerized using Ziegler-Natta catalysts a linear 

polymer is formed. It was known that the catalysts acts on the double bond of the 

ethylene unit and the polymethylene chain can grow on both sides of the ethylene 

unit once the double bond has been opened. The opening, however, can occur in cis 

or trms configuration. 

H H H I H 

‘c-c’ 

‘c-c’ 

H/I J\H H’ l\H

3.12 CIS-trans Opening @the Double Bond 139 

The question which remained to be answered for inore than a decade was 

whether the opening of the double bond occurs in cis or trans. The understanding of 

this reaction mechanism was of fundamental importance for the development of 

other similar reactions for other classes of materials. 

One way to solve the problem was to polymerize deuterated derivatives of ethylene 

with the same Ziegler-Natta catalysts. If cis di-deutero ethylene (monomer A) 

is polymerized and the opening of the double bond is CIS. the resulting polymer 

should be rich in the so-called ‘cridzro’ units 

HH 

c-c 7 

DD 

c-c-c-c 

HDDH 

DHHD 

a 1 2 

If the opening is cis for the case of monomer B (tram-dideuteropolyethylene) the 

reaction should yield a polymer rich in the so-called ‘threo’ units. 

DH 

c-c 7 

HD 

c-c-c-c 

HHDD 

DDHH 

b 1 2 

The opposite should occur if the catalyst forces the double bond to open in trans 

configuration. 

The polymers were made by chemists, but the identification of the configuration 

of the deuterium and hydrogen atoms remained unsettled for a long time. The 

vibrational spectra were recorded, but their interpretation was impossible on the 

basis of the classical techniques of polymer spectroscopy. 

An example of the procedure followed in the identification of the spectra is the 

following. 

Let the configurations of monomeric units be indicated as follows 

HD DH HH DD 

c-c- 7 

-c-c- 

- -c-c- i -c-c- 4 

DH HD DD HH 

I I1 111 IV 

From the viewpoint of calculations a computer program which generates a sequence 

of 1-andoni numbers was used in order to generate a random sequence g( 1’)

c 

40’ r 

Polymer 8 (100) .? 

Wave number (cm-’ 

a 

Figure 3-24. (a) Calculated 

density of vibrational slates of 

polymers A and B with 100 

monomeric unit. (b) Observed 

infrared spectra of poly(ci.v- 

CHD=CHD) and poly(rrrrris- 

CHD=CHD). 

for both chains with 50150 concentration of a and b. The sequence is the following: 

bbaabbaabbbbaabaaabababa 

bbbbaabaaabbbabbbbaaabba 

aaabababbbaabababbbaabaa 

ababbabbbabbaaababaaaabbb 

The two cases considered here are: (i) polymer A, generated by replacing a with unit 

I and b with unit 11; and (ii) polymer B, obtained by replacing a by unit 111 and b by 

unit IV. 

Each chain contains 600 atoms. The corresponding GR and FR matrices (of dimension 

1800 x 1800) were constructed and NET applied for the calculation of 

got). The calculated g( v) (Figure 3-24a) compared with the experimental infrared 

spectra (Figure 3-24b) clearly shows that the calculated g(v)(A) corresponds to

3.13 Defect Modes cis Structtrr.al Probes in Poljw~eihylriie Chins 141 

0 

.- 

c 

I , / / 

, I / / 

Figure 3-24 (continuedj 

I l l , , 

7350 1300' ' I ' 7250 ' 

Wave number Icm-') 

' 1200 

1 

b 

the infrared spectrum of poly(cis-CHD=CHD) and g(v) (B) corresponds to the 

experimental spectrum of poly( trans-CHD=CHD). 

It can then be concluded that the polymerization by Ziegler-Natta catalysts 

occurs with the cis opening of the double bond. 

3.13 Defect Modes as Structural Probes in 

Polymethylene Chains 

The strongest effort by our group and by other authors was aimed at revealing the 

structure and molecular dynamical evolution of the very many classes of technologically 

and scientifically relevant molecules consisting of n-alkanes or n-alkyl 

long chains. We report below some of the results obtained and will show some 

applications.

3.13.1 Mass Defects 

Let us introduce a CD2 group as a defect in n-alkanes (CH3-(CHl),,CH3) or n-alkyl 

chain (CH3(CH?),X or in any polyniethylene chain. This is a typical case of ‘mass 

defect’ which leaves the force constant matrix (Eq. 3-31) unchanged because of 

Born-Oppenheimer approximation. The site where the CD2 group is located can 

be changed along the chain. Chemically, this can be easily achieved and selectively 

deutei-ated n-alkane or alkyl derivatives are either coininercially available or can be 

(and have been) easily synthesized. 

Calculations have shown that the frequency of the rocking motion of the unit 

CD? embedded in a CH2 host lattice changes by changing the conformation of 

its surroundings [102, 1031. The calculated and observed frequencies for the CD? 

rocking motion when the adjacent torsional angles are changed are reported in 

Table 3-1. These motions are classical well-localized gap-inodes as they occur in the 

frequency gap between the CH2 rocking and the CCC bending dispersion branches 

of the host lattice. 

Table 3-1. Calculated and experimental CD2 ’gap frequencies’ for selectively deuterated nonadecane 

as function of the torsional angle about C-C bonds. 

Calculated 

Observed 

4 3 2 1 

-CH2-CH>-CH>-CD>-CH3 

T T 

G T 

T G 

G G 

615 622 

615 632 

658 658 

654 

13 12 11 10 9 8 7 

CH2-CH2-CH2-CD2 -CH: -CH:-CH? 

T (T T) T 

G (T T) T 

G (T T) G 

G’ (T T) G 

T (T G) T 

G (T G) T 

G (T G) G 

G‘ (T G) T 

T (G G) T 

T (G G) G - 671 

G [G G) G 

-615 622 

- 650 650

3.13 Defect Mo&s cis Structurcil Probe5 iii Poljwietliylene Cliaiiis 143 

3.13.2 Conformational Defects 

The study of the conformational defects in polymethylene chains has been tackled 

some time ago in a systematic way by the school of Snyder by norinal inode analysis 

of short-chain molecules (see, for exainplc (1041) and by our group at Milano 

using NET [95. 1011. The results by both techniques are very similar. In the previous 

discussion in Figure 3-14 we have pointed out that a few peaks in the calculated 

g( 1’) for infinite tims polyethylene do not find corresponding infrared and/or 

Ranian lines to be assigned to k = 0 motions. These additional experimental peaks 

are just an evidence of the existence of extra structures which we can identify as 

conformational defects using the theory we are presently discussing. 

The main characteristic defect modes calculated (and observed) for tram planar 

segments are the following: 

1. GG defect. The wagging vibration of a CH2 group located between two adjacent 

grrirclie conformations -CH-,(G)-CH?(G)-CH-,- wagging. The calculated (and 

observed) frequency lies near 1355 cin-’. 

2. GTG‘ defect. CH2 wagging. -CH2(G)-CH,(T)-CH-,(G’)-. This defect introduces 

two wagging modes because the two CH2 groups are joined by a bond in 

traizs conformation and are partly decoupled from the host lattice by the gauche 

bonds. The intra-defect coupling generates out-of-phase and in-phase modes 

calculated (and observed) near 1375 and 1306 c1ii-I. 

3. GGTGG dc+x-t. This defect is the one which possibly exists on the surface of the 

single crystal lainellae of polyeythylene if the chain re-entry forins a tight fold 

along the [200] crystallographic plane [ 1051. This defect generates practically the 

three lines expected for GG and GTG’ defect in the CH-, wagging region and an 

additional quasi out-of band mode (near 714 cm-I) just outside the k = 0 limiting 

rocking mode of the CH2 group [99]. 

4. Heud-to-hecicl defects of the type. -RCH-CH-,-CH-,-RCH-. The CH2 rocking 

mode of a sequence of two CH-, groups between two heavy boundaries is calculated 

and observed near 850 cm-’ in agreement with empirical correlations 

derived from the study of ethylene-propylene copolymers [ 1061. 

5. Clzaivi end modes it2 pi-alkyl clznins. When the group -CH-,-CH>-CHl-CH3 in 

such chains has TTT conformation, the methyl ‘umbrella deformation’ iU) is 

calculated and observed near 1375 c1n-l and the external deforniation mode 6 

CH3 occurs near 890 cm-’. According to the previous discussion in this chapter 

both motions should be considered end-group modes whose frequency should 

remain constant when the length of the chain increases. This is indeed the case if 

the chains keep the all-trams structure. However, since both motions occur in a 

frequency range spanned by the dispersion cui-ves of the 1D lattice they couple 

with the vibrations of the bulk and their coupling extends for a few CH2 units 

along the chain. If the conformation of the chains near the ends changes, such 

intramolecular coupling changes and the frequencies of the ‘end groups’ may 

change. Table 3-2 reports the values of the calculated and observed frequencies 

when chain ends are conformationally distorted in the TG and GT conformations.

144 3 Vibrutionnl Spcctrtr as N Prohc OJ Structurnl Order 

Table 3-2. Characteristic calculated and expeiimental ‘in band’ frequencies fot the conformers 

-CH~’CHL’~’CH~-CHI and -CHP(~)CH~(~’CH~-CH~ at the end of the n-nonadecdnr molecule 

v (cm-’) 

Calculated 

Observed 

TG 1348 1342 (IR) 

879 866 (R), 866 (IR) 

769 766 (IR) 

GT 1367 

851 844 (R), 845 (IR) 

3.14 Case Studies 

3.14.1 Conformational Mapping of Fatty Acids Through Mass 

Defects 

It becomes apparent that the use of selective deuteration becomes a powerful tool 

for the mapping, site by site, of the coiifoiniation of polymethylene chains. This has 

been applied for the conformational mapping of n-alkane and of fatty acids. The 

idea was first proposed by Snyder and Poor for n-alkanes [lo21 and was later 

applied to the case of fatty acids [lo71 and has settled the issue on the conformation 

in the solid state which has long been matter of great debate [108]. 

Fatty acids are typical of a class of compounds for which the existence of premelting 

phenomena has been claimed with the formation of ‘liquid-like’ structures 

for the molecules while still in the solid. The structure of these systems has been 

studied by many workers. The molecules in the solid are dimeric since pairs of carboxyl 

groups face each other and are held together by hydrogen bonds. Moreover, 

complex polymorphic phenomena have been revealed depending on the way crystallization 

is carried out (from the melt, from solution, froin the kind of solvent, 

etc.) [107, 1091. More recent works based on the study of LAM inodes in the 

Raman suggested that the molecules of fatty acids, independent of the crystalline 

form, are never all-trans planar and contain somewhere along the chain a gauche 

distortion [108]. These information were at odds with our studies on LAM modes of 

the same materials which were shown to be fully trans-planar [110, 11 I]. 

The use of defect gap modes using selectively deuterated materials have definitely 

solved the problem. We have recorded the infrared spectra of solid perhydrogenated 

stearic and palmitic acids and of a few selectively deuterated derivatives. 

For palmitic acid, single CD2 group were placed in position 2, 3, or 4 (carbon atom 

1 is that of the carboxyl group); for stearic acid, the single CD? group was placed in 

position 2, 3, 4, 5, 6, 7, or 13. As an example of the way we proceeded, Figure 3-25 

illustrates the infrared spectra of: (i) normal undeuterated palmitic acid (Figure

0 

3.14 Case Studies 

N 

0 

d 

- Y 

N 

o m 

orm 

v 

I 

0 

U 

c- 

.-- - 0 

I I I I 

ln 

N 

0 

m 

'9 

0 0 

e 

Y 

V 

0 L

146 3 Vibratioiiul Sycctm CIS (I Probe of Structural Ordcr 

3-25a); (ii) palinitic acid ?-Dz (Figure 3-25b); and (iii) pahnitic acid 3-D2 (Figure 

3-25~1. The clear band which appears near 620 cn-' is the gap-niode indicating 

unquestionably the existence of the local conformation CH?(T)-CD?IT)-CH? both 

in positions 3 and 4. Identical spectra have been observed along the whole chain, 

thus confirming the whole lrms structure of palinitic acid in the solid state [107, 

110, 1111. 

3.14.2 Molecular Mobility and Phase Transitions in Therrnotropic 

Liquid Crystals from the Spectroscopy of Defect Modes 

The issue is the description of the molecular changes accompanying the phase 

transitions in systems which show liquid crystalline phases. While much is known 

on the morphological changes, little is known on what happens to the molecular 

structure through the various phase transitions. 

We use defect mode spectroscopy for aiming at the structural changes of prototypical 

thermotropic liquid crystals such as alkylcyanobiphenyl and polyesters. 

3.14.2.1 Case 1: Cyano-alkyl biphenyls 

Let us first consider dodecyl-cyanobyphenyl which shows two phase transitions at 

48 "C (crystalline-semectic A) and at 58.5 "C (smectic A-isotropic). The analysis 

of the whole spectrum is reported in the original work [112] and we focus here only 

on the variation with temperature of the conforiiiational structure of the dodecyl 

side chain which we hope to reveal with the study of the temperature-dependent 

vibrational spectrum in the 1420-1280 cm-' range where defect modes are expected 

to occur. 

Figure 3-26 reports such spectra, together with the difference spectra obtained in 

the following way. Let s(J + 1) be the spectrum recorded at temperature T(J + 1) 

and s(j) the spectrum taken at temperature T(j). Then AT = T(J + 1) - T(j) is the 

temperature gradient in our experiment. From Figure 3-26, while drastic changes 

are observed from the crystalline to smectic A phase, the spectra do not change 

much from sniectic A to isotropic. The difference spectra highlight such changes 

and allow the defect modes GTG', GTG, GG, and end-TG to be identified. The 

following conclusion can be easily reached: (i) the trans-planar chain exists up to the 

crystal-smectic A transition; (ii) at the crystal-smectic A transition the trans-planar 

chain collapses in a conforinationally distorted geometry, as indicated by the appearance 

of all defect modes listed above; (iii) the observed upward shift of the 

umbrella deformation mode of the CH3 group from 1376.5 to 1378 cm-l indicates 

that the environment at the end of the chain is changed; and (iv) no great conformational 

changes occur from the sniectic A to the isotropic phase; the cyanobiphenyl 

core must then supervise this phase transition (as also shown by the 

vibratioiial spectrum). Another case treated is that of 4-cyano, 4'octyloxy biphenyl 

[113].

3.14 Case Studies 147 

OC 

60 

58 

57. 

56 

55. 

54 

52. 

50 

49 

47 

a. 

43 

41 

39 

36 

33. 

21 

Figure 3-26. Dodecylcyanobiphenyl. (a) Temperature-dependent infrared spectra in the defect 

mode frequency range. (b) Experimental spectroscopic evidence from difference spectroscopy that 

at the phase transition crystal - sniectic the conformation of the dodecyl residue changes froin all- 

~YLIII.F to 'liqnid-like' (the most connnon defect modes appear clearly). From sinectic to liquid phases 

no drastic confoiiiiational changes are observed (see text). 

3.14.2.2 Case 2: Liquid Crystalline Polymers: Polyesters 

We consider the case of the polyiner -[(CH2)lo-00C-C~H~-00C-C6HJ-COO- 

CcH4-COO],- (hereafter referred as HTH-10) whose phase transitions are the following: 

K-S 221 "C; S-I 260°C. From the overall aiialysis of the temperature-

148 3 ViLwntional Spectra us N Probe of Striictural Order 

- 

5_1_ 

L 

60-513 5 

- 5135-575 

575-565 

- 

- 

_v_ - 

-- 

56 5-54 

54-526 

52 6-50 5 

505-49 

-49 - 47 1 

47.1 - 45.5 

43-41 

004.- 

- 

41-39 

a, 

C 

0 

__c. 

'? 

39-36 

2 002- 1 

n 

- ---36-33 1 

a 

-3.1 -27 

AT 

0.00 t 

Figure 3-26. (continued) 

dependent Raman and infrared spectra [I 141 it could be concluded that, in the K 

phase, most of the decamethylene segments are in the trans-planar conformation 

and that upon heating gnuclze rotamers are generated. The identification of what 

kind of conformational defects occur in the disordering process when the temperature 

increases can only be attempted by the study of the temperature-dependent 

spectrum in the traditional CH2 wagging range. We restrict here our analysis to 

the temperature-dependent spectra in the usual defect modes range from - 1400 to 

- 1300 cm-'. We aim at revealing the conformational evolution through phase 

transition of the segments consisting of 10 CH-, groups. Figure 3-27 illustrates the 

temperature-dependent infrared spectrum in the CH2 wagging range plotted in


GI A V E N U M BE R 5 

Figure 3-27. Liquid crystalline polyester with decamethylene spacers. Temperature-dependent 

infrared spectrum in the defect modes frequency range (see text). The spectra and shifted one from 

the other as in a tridimensional plot. 

three dimensions for the sake of clarity. It becomes apparent that the tvnizs-planar 

structure (no defect modes in this frequency range) evolves with increasing teniperature 

with the generation of only the defect mode near 1365 cm-' due to GTG' 

defects. No absorption due to GG defects (band at 1353 cm-') is observed, i.e., 

sharp kinks are neither generated in the sinectic nor in the isotropic phases. The

150 3 T’ibratioiinl Spectra L ~S u Probe oj StrzictLir.al Order. 

intensity of the GTG’ defect inodes increases sharply at the K-S transition, remains 

on a smooth platem until the S-I transition occurs, and increases steeply in the 

I phase. It has to be noted that even in the I phase the GG defect mode is still 

practically not observable. 

The mechanism on a molecular level which is derived from the vibrational spectra 

can be suniinarized as follows: at room temperature the sample contains a sizable 

concentration of ordered material in which the decamethylene spacers are 

predominantly in the trms-planar structure. When the I(-S transition is 

approached the main spectroscopic signal observed is that associated to the GTC’ 

modes. It is known that the GTG’ defect introduces a kink in the chain but does not 

change the trajectory of the polyniethylene sequence keeping both arms at either 

side of the defect parallel. The overall effect is that the length of the decaniethylene 

segment simply shrinks. It follows that in the S phase the spacers shrink because of 

the introduction of conformational kinks, inducing a sort of shearing motion between 

the large aromatic mesogenic groups. The lack of any sizable concentration 

of GG structures even in the I phase leads to the interesting conclusion that the 

molecular structure in the I phase is not much different from that in the S phase. No 

indication is found of ‘liquid-like structures’ observed in other polyniethylene systems 

in the liquid phase. It has been proposed [114] that one way to reconcile 

the observation of macroscopically isotropic phase with the observation on the 

molecular level is to envisage that the isotropic phase consists of a collection of 

microdomains whose directors are randomly oriented. However, inside the microdomains 

polymer molecules are organized as they are in the S phase. 

3.14.2.3 Case 3: Chain Folding in Polyethylene Single Crystals 

It is known that chain folding occurs in polyethylene crystals; the fold structure, 

however, has been a subject of interest and controversy [105]. It was obvious for 

polymer spectroscopists to use vibrational infrared and/or Raman spectra to try to 

contribute to the understanding of the shape of the fold (i.e., the molecular conformation) 

within the fold at the surface of polyethylene single crystals. 

The large amount of spectroscopic work carried out by many authors is 

sumarized in [115]. Here, we wish to focus specifically on the contribution given by 

defect mode spectroscopy considering both calculations and experiments. The issue 

is whether the surface of polyethylene single crystals consists mostly of loosely 

looped folds or also of a fraction of chains with a tight adjacent re-entry. In the 

former case, ‘liquid like’ defects such as GG, GTG, GTG’ etc. are expected and can 

be recognized in the vibrational spectrum. If tight fold re-entry exists defects of the 

type GGTGG must occur at the surface of the lainellar single crystal. 

The first question which needed to be answered in a less qualitative way was 

where to locate the vibrations localized on the GGTGG defect. 

Calculations have been carried out on a long polymethylene all-tivm host molecule 

containing evenly spaced GGTGG defects. NET and IIM and the ‘island 

analysis’ were applied [99]. The most meaningful results are presented below. 

Two highly localized modes are calculated to lie near 1361 and 1358 cin-’ and

3.14 Case Studies 15 1 

correspond to the out-of-phase motions of the two CH, groups located between two 

C-C bonds in gcruchc conformation. These two modes are largely localized at the 

defect while the other portion of the chain remains practically still. It is interesting 

to note that the defect modes in the CH2 wagging region near 1350 cm-' show that 

in this spectral range the GGTGG defect behaves as if it were made up of GG and 

GTG defects moving almost independently. 

In the lower-frequency region defect modes are calculated near 1105, 900, and 

714 cm-'. The mode near 714 cni-' is of particular interest since it lies just below 

the limiting k = 0 mode located at 720 cn-l of the CH2 rocking dispersion curve 

and is just at the edge of a large energy gap which extends as far as the cut-off of the 

branch cu5 (see Figure 3-3). From IIM, we learn that the defect mode near 714 cmP1 

originates from a cluster of several CH, groups, including the defect, which is performing 

mainly a rocking motion. From the shape of the nornial mode this nearthe-edge 

mode cannot yet be considered a localized gap mode since it shows a 

nonnegligible degree of cooperativity. Similar cases of near-the-edge resonances 

were found for polyethylene in the lower frequency range and for polyvinylchloride 

containing head-to-head defects [90]. 

The modes calculated near 1105 and 900 cm-l are the results of a complicated 

mixture of several internal coordinates, but can be described mainly as C-C resonance 

modes. They are likely to be observed in the Ranian spectra. 

The results of the calculations were checked with the experimental infrared and 

Raman spectra of the cyclic hydrocarbon C34H68 whose structure determined by 

Newiiian and Kay [116] indicates that two segments of 12 CH? groups in a transplanar 

zig-zag conforination are joined by two GGTGG defects. Coincidences 

between experiments and calculations are striking. However, a critical analysis 

needed to be made ([99, 1151) on whether these signals are characteristic only of the 

GGTGG defect and may allow such a defect to be disentangled from the signals of 

other conforniational defects possibly existing in a realistic polyethylene single 

crystal. The conclusion reached is that calculations on several model systems and 

experiments on C34H68 suggest [99] that a simultaneous appearance in the infrared 

spectrum of two bands near 1342 cm-' and one near 700 cm-' (certainly below the 

in-phase limiting CH2 rocking mode of the tram-planar chain) can be taken as evidence 

for the existence of GGTGG defects in polyethylene single crystals. The other 

calculated modes occur at frequencies where other modes of the host lattice may 

also occur. Calculations were also extended to predict the intensity pattern for 

the defect modes in the CH2 wagging range [117, 1181. Figure 3-28 compares 

the experimental spectrum with that calculated using the force field proposed by 

Shimanouchi [ 1 191. 

Experiments of difference spectroscopy combined with band deconvolution processes 

were presented and conclusions were not unambiguous [ 1 151. On the other 

hand, the arbitrary decisions taken when applying band deconvolution techniques 

sometimes make issue more complex. We have tried to avoid band deconvolution 

and proceeded directly to careflil experiments of difference spectroscopy on sectored 

polyethylene single crystals. The principle on which the experiments was based is 

the following. From the study of sectored polyethylene single crystals it is known 

that sectors correspond to regions where molecular chains are crystallized in the

1380 1340 1302 

Figure 3-28. Coinparison between the calculated and esperimental 

infrared spectrum in the CH2 wagging region of the 

GGTGG conformational defect. Calculations have been made 

the with force field by Shiinanouchi [I 191. The intensity 

cm-' 

paraiiieters were developed in the laboratory of the author. 

{ 1 lo} and { 100) crystallographic planes. By changing the crystallization conditions, 

sectored crystals with varying ratios between the { 110) to {loo} sectors can 

be obtained. Chain-folding studies has also indicated that tight-fold re-entry with 

GGTGG defects can only occur when chain crystallize in the {loo} planes. 

Sectored crystals were kindly provided to us by Professor A. Keller with the following 

% of { loo} faces: 0, 31, 35,45, and 55. The experiment of difference infrared 

spectroscopy we attempted is sketched Figure 3-29. We aimed at isolating the absorption 

to be associated with the GGTGG possibly existing on the surface of the 

{loo} fraction of the sectored crystals. The experiments were not easy and after 

many attempts the final spectrum obtained (Figure 3-30) is compared with that of 

the cyclic hydrocarbon C34H60. The matching of the two spectra is striking and does 

suggest that, in agreement with calculations, on the surface of polyethylene single 

crystals obtained in particular crystallization conditions tight fold re-entry with 

GGTGG defects may indeed occur. 

3.14.2.4 Case 4: The Structure of the Skin and Core in Polyethylene 

Films (Normal and Ultradrawn) 

In addition to the defect modes presented in the previous sections it has been shown 

recently by both calculation and experiments that generally when yazrclze defects 

exist in polyethylene a broad band arises in the CH2 bending range near 1440 cm-' 

[ 1201. The band is very broad, possibly because it originates from a distribution 

of nonequilibrium gauche structures constrained by the sample morphology (Figure 

3-31). Close to such absorption one observes for crystalline orthorhombic n-alkanes 

and polyethylene the doublet due to crystal field splitting (see Section 3.6). Let a 

and b label the two components observed near 1473 and 1463 cm-' respectively 

(Figure 3-31). It is known that the setting angle B between the planes defined by the 

tr.arzs-planar chain skeleton in the lattice is approximately 42" [121]. Such value of 

0 determines the intensity ratio IJIb = 1.2333 between the two components of the 

crystal doublet of the CH2 bending and CH2 rocking in orthorhombic n-alkane and


(I10 + 200) +BULK y (BULK) 

a 

(I10 .ZOO) 

SURFACE (110) 

b 

(200) 

Figure 3-29. Scheme of the experiments of diference spectroscopy on sectored single crystals of 

polyethylene. (a) single crystal - bulk = (1 10) + (200) surface of the sectored crystals; (b) 

( 110) + (200) surface - ( 1 10) surface = (200) surface. 

Figure 3-30. Comparison between the infrared 

spectrum in the GGTGG defect mode region of 

the crystalline cyclic molecule C34H68 and the 

result of the digerence spectroscopy sketched in 

Figure 3-29. 

I 

ll00 1375 1350 1325 1300 1275 1250 1225 1200 

WAVE N UM BER5

154 3 Vihtional Spectvu as a Probe of Structurd Order 

1500 1460 1440 1420 1 uoo 

Wa v e n urn be r s 

Figure 3-31. Infrared absorption spectrum of a coininercial film of polyethylene before and after 

subtraction ofthe spectrum of the defect mode component near 1440 cm-'. After subtraction of the 

defect mode the ratio between compoiient a and b of the correlation field splitting approached the 

ideal value for a setting angle of42" (see text). 

polyethylene [122]. Certainly the wing at the higher frequency side of the defect 

mode overlaps with at least line b of the doublet, thus adding more intensity to line 

b and removing any physical meaning to be observed intensity ratio between the 

two components. However, by difference spectroscopy it is possible to remove the 

absorption due to the conformationally irregular fraction. 

This observation is the basis for the determination of the amount of material with 

pzdw structure in any sample consisting of polymetliylene chains partially crystalline 

and partially conformationally disordered amorphous. 

This method has been applied to experiments of multiple internal reflection infrared 

spectroscopy on thin films of commercial polyethylene [120]. It is known that 

'optical microtomy' can be achieved by changing the incidence angle of the inconiing 

beam. Structural depth profiling has then be performed thus allowing to probe 

the ordered-disordered structure of such films (Figure 3-32).


1500 1480 1460 1 quo 1420 ! 400 

Wavenurnoer 

Figure 3-32. Multiple internal reflection spectrum in the CH2-bending region at different angle of 

incidence of a 56 pi-thick film of commercial polyethylene. The spectrum shows the increase of the 

fraction of conformationally irregular material by increasing penetration depth.

156 3 Vibratioiicil Sr,ectru CIS a Probe of Stnictiiid Order 

It has been found that the outer skin of the film is much more crystalline (-SO(%) 

and the crystallinity decreases when probing inside the film reaching a concentration 

of -60%) crystallinity approximately 5-8 pm inside the film [130]. 

The above analysis combined with defect mode spectroscopy carried out in infrared 

(absorption in polarized and unpolarized light, and multiple internal reflection) 

and Raman has been applied to the structural characterization at the molecular 

level of ultra-drawn films of high-density polyethylene (UD-HDPE) [ 1331. The 

comparative study has allowed to describe the structural situation of UD-HDPE 

films as follows: 

(i) The skin of the film within - 10 pin is highly crystalline and orthorhombic. 

(ii) UD-HDPE film is a multiphase system with the topological distribution of 

the phases changing in going from the surface into the core. 

(iii) The orthorhoinbic phase exists both on the surface and in the core, while the 

monoclinic phase starts at a penetration depth d, > - 10 pm. 

(iv) Both the orthorhombic and the monoclinic phases are highly anisotropic and 

are oriented along the draw direction. 

(v) An additional phase is identified from d, > 10 pni into the core. This phase 

consists of sequences of CH2 groups in tvrrrzs conformation. The orientation of 

the chains is not parallel to the draw direction. This additional phase exists at 

low draw ratios and decreases its concentration qualitatively at higher draw 

ratios with respect to the orthorhoinbic component. 

(vi) GTG’ kinks are organized in anisotropic domains, seemingly as a result of 

collective phenomena of orientation during stretching. 

(vii) The chain end methyl groups find themselves in a disordered environment 

similar to that found in liquid n-alkanes. The existence of liquid-like droplets 

or of disordered domains around the CH3 groups may be envisaged. 

(viii) GG defects are observed and may originate from intrinsic liquid-like structures, 

from the disorder around the CH3 heads and from some kind of chain 

folding between crystallites or within polymer lainellae. 

(ix) The concentration of conformational defects is comparatively small when 

normal solid polyethylene is considered. The concentration of defects decreases 

when the draw ratio increases, thus showing that conforniationally 

disordered and coiled chains unwind towards a better alignment and perfection 

of the tmns-planar molecules. 

All these information have been compared with other physical data whch help 

in the understanding of the molecular structure of such technologically relevant 

materials. 

3.14.2.5 Case 5: Moving Towards More Complex Polymethylene Systems 

The role of long alkyl chains in natural phenomena and in industrially relevant 

systems is well recognized. Due to the flexibility of the polymethylene segments 

these systems show many phase transitions which are not yet clearly understood at


the molecular level. Many authors have focused on the vibrational infrared and 

Raman spectra as probes for understanding the structure and dynamics of these 

systems. 

We have tried to analyze a few cases with our spectroscopy of defect modes and 

mention here some of the systems studied. 

1. Fattj, acids. Generally, fatty acid crystallize as dimers held together by hydrogen 

bonds between the two carboxyl groups which face each other in a head-to-head 

arrangement. In other sections of this chapter the contributions to the understanding 

of the 'static' structure given by the spectroscopy of mass defects and by 

LAM spectroscopy have been already presented. The case of fatty acids with 

long polymethylene chains has been studied [ 1241 focusing on the conformational 

evolution which originates the phase transitions. FTIR spectra clearly and 

distinctly show that when thermal energy is provided to stearic acid in the crystalline 

state first interlamellar CH? ...CH3 distances increase (as seen in the case 

of n-nonadecane; see Section 3.19) and hydrogen bonds then relax. Simultaneously, 

intermolecular distances increase with resulting lateral expansion of the 

lattice. These phenomena become relevant approximately 10 " below melting. 

Few confonnational kinks are generated below melting and only close to the 

melting the concentration of GTG, GTG', and GG defects increases rapidly 

bringing the system to a conformational collapse in the liquid phase [124]. 

However, most of the conformational disorder occurs on the lamella surface, 

thus adding more evidence to the existence of a general phenomenon of 'surface 

melting' which emerges from our studies on many chain molecules [125]. These 

results give a detailed description of the mechanism of phase transition in fatty 

acids which was vaguely hinted by NMR experiments [ 1261. 

2. Bilajw systenis. The bilayer organic perowskytes were obviously used as simple 

models of biological bilayered phospholipids and biomeinbrane systems. We 

have studied [127] the system [n-CH3( CH2)"NH3]2MnC14 (hereafter referred to 

as C14Mn) which in the solid forms two coexisting, but parallel, layers; a similar 

molecule with Zn (C14Zn) in place of Mn forins two interpenetrating or 'intercalated' 

layers. The conformational flexibility and phase transitions of these two 

classes of molecules turns out to be different. DSC data indicate for C14Mn a 

main phase transition in the range 74-85" which is described by our conformational 

studies as involving the formation exclusively of GTG' kinks which do 

not alter the conformational trajectory of each alkyl stem. No GG defects are 

observed and clusters of ordered and disordered chains coexist over a large 

temperature range (67-80°C). By simulation of the spectra it is found that, on 

the average, at 78°C only one GTG' kink per chain is formed; between 65 and 

78 "C chains with only one kink are formed and arrange themselves in clusters. It 

follows that disorder proceeds cooperatively throughout the system. 

The behavior with temperature of C14Zn is totally different [127]. The main 

phase transition at 100°C is accompanied by an abrupt generation of conformational 

defects at 99 "C. Above melting, chains have reached almost a liquidlike 

structure. Correlations between chains do not exist. However, from the 

temperature dependence of the factor group splitting of the CH2 bending and

158 3 Plhrutional Spectru CIS Q Probe qf'StructLirri1 Order 

CH2 rocking inodes it is observed that the transition is 'prepared' since the 

intermolecular forces between chains in C14Zn [weaker than in C14Mn) quickly 

weaken and the thermal expansion gives more freedom for torsional fluctuations 

of the alkyl chains, thus leading to conformational collapse. 

Two classes of mechanisms which lead molecules to phase transition and melting 

have been found, namely: (i) nucleation of conforniational disorder from clusters 

whose size continuously increases, thus affecting later all chains; and (ii) cooperative 

expansion of the whole lattice caused by an increase with temperature of the mean 

amplitudes of torsional fluctuations which reaches a threshold level after which 

conformational collapse occurs. 

In nature, many other polymethylene systems-also of practical relevance-can be 

found which show phase transitions that need to be accounted for. More theoretical 

and experimental spectroscopic work needs to be done. We have studied the spectra 

of intercalated layered alkylaininonium zirconium phosphates [ 128) and of the solutions 

of decylammonium chloride [ 1291. The spectroscopic data present U~~LISLI~~ 

features with temperature which need further theoretical and experimental studies 

using the techniques presented in this chapter. 

3.15 Simultaneous Configurational and Conformational 

Disorder. The Case of Polyvinylchloride 

The cases treated in the previous sections were restricted to the study of the few 

cases of mass defects and of the very many cases of conformational defects. Most of 

the relevant polymers, however, may contain configurational disorder which carries 

as a consequence also some conformational disorder. This case has been considered 

within the theoretical approach discussed in this chapter. 

The mathematical aspects for a general case of polymers with defects in tacticity 

were first worked out, formulae are available and can be applied to any case. We 

have applied these mathematical techniques to the case of polyvinylcloride (PVC) 

which is known to be strongly configurationally disordered. Our work was done 

when the problem of the structure of PVC was highly controversial. 

First, the dynamics of ordered PVC was treated in terms of the three energetically 

most likely ordered chain structures [go], namely: (i) planar extended syndiotactic 

with (-TTTT-) conformation; (ii) helical isotactic, threefold helical structure 

(-GTGTGT-); and (iii) folded syndiotactic (-TTGGTTGG-) (Figure 3-12). Phonon 

dispersion curves and density of vibrational states were calculated for the three 

models using the most reliable force fields available at that time (Figure 3-12). The 

infrared and Raman activity of the k = 0 phonons for each perfect chain are the 

following: model (i), z = 2, t = 1, 12 atoms per crystallographic unit cell, $(n, d/2), 

point group, 9A1 (ir 11, R) +7A2 (R) t9B1 (ir I, R) +7B? (ir I, R); model (iii 

z = 3, t = 1, 18 atoms per unit cell, $(2x/3,d/3), point group, 16A (ir 11, R) +

17E (ir 1, R): (iii) z = 4, t = 1, 18 atoms per unit cell, D: point group, 17A (R) 

+17B, (ir 11, R) +17B: (ir I, Ri +17Bi (ir 1, Ri. 

The knowledge of the dispersion curves, of the spectroscopic activity of the k = 0 

phonons and their precise frequencies has allowed us to locate the energy gaps, 

where to expect localized defect modes, and to predict where the activation of the 

density of states singularities (strong peaks in g(v) with no activity in IR and/or 

Ranian) due to lack of symmetry because of disorder could generate extra absorption 

or scattering. 

Since we take for granted that the actual samples of PVC contain geometrical as 

well as configurational defects at non-negligible concentrations, calculations were 

extended [90] with the introduction of: (i) conformational defects as isolated units 

or as possibly interacting multiple units; and (ii) isolated or possibly interacting 

configurational defects. Of necessity, when configurational defects are introduced, 

local conformational distortions must be considered so as to justify the existence of 

an energetically plausible defect. NET was applied for chains consisting of 200 

chemical units (i.e., matrices of 3600 x 3600 were constructed), while IIM was 

restricted to shorter chains made up by 50 units (900 x 900). 

The distribution of defects was random and the concentrations ranged from an 

isolated unit to 8'Yo defects. The construction of the large matrices was made from 

smaller submatrices associated to each chemical unit in a given conformation or 

configuration. The detailed algebra and the complete expressions for such submatrices 

are available in [90] and can be applied to any case of polymers made up 

by chiral chemical repeat units. 

The attention was mainly focused on the C-C1 stretching range of the supposedly 

mostly syndiotactic chain. Polymerization in urea clathrates provides the 'purest' 

sample of planar syndiotactic PVC [130]. Such a chain shows a large energy gap in 

the CCl stretching range (-650 to 820 cm-'). It is very likely that defect modes 

from other structures may generate localized gap modes which may be used as 

useful structural probes. 

While we refer to [90] for a thorough analysis of the complex structural situation, 

in Figures 3-33 and 3-34 we give two examples of the frequency and vibrational 

displacements for an isolated GG or an isolated isotactic diad defect in a fillly 

syndiotactic planar zig-zag host chain. It must be noted that gap frequencies are 

different, but that the vibrational displacements associated to these gap modes involve 

many chemical units (-20). The complexity of the problem is shown in Figure 

3-35 where the density of vibrational states of a realistic model of configurationally 

disordered PVC is compared with the experimental spectrum. 

3.16 Structural Inhomogeneity and Raman Spectroscopy 

of LAM Modes 

The problem raised in Section 3.14.1 requires further discussion. From longitudinal 

accordion motion (LAM) spectroscopy of a variety of fatty acids Vergottin et 211.

160 3 Vibmtioriul Spectra us a Probe of Strircturul Order 

Scale 

GG-defect 

gap mode 

GG 

0.05 

0 

Figure 3-33. Gap mode (described in terms of total displacements of the atoms) for a GG defect 

in a syndiotactic PVC chain. Notice that the gap mode is localized over a large section of the 

chain. 

Figure 3-34. Gap mode (described in terms of total displacements of atoms) for an isotactic isolated 

dad. Notice the breadth of the mode which extends over at least 20 carbon atoms.

3.16 Stnictural Znlioi~zogeneity and Ranmn Spectroscopy qf LAM Modes 161 

Figure 3-35. Density of vibratioiial states for 

a realistic model of configurationally disordered 

PVC (Bernoullian paramater 

P, = 0.47). Comparison with the experimental 

spectrum (see text). 

[lo81 suggested that the alkyl chains of fatty acids contain a few conformational 

defects. This conclusion was shown not to agree with the conformational mapping 

side-by-site obtained by defect-mode spectroscopy with the introduction of CD2 

defects in the polymethylene alkyl chain. 

Such discrepancy seemed either to weaken the relevance of LAM spectroscopy 

for the measure of the length of the all-trans segment in a polyethylene chain, or to 

cast doubts on the reliability of defect-mode spectroscopy. Some further studies 

were in order. 

Since dispersion curves have been clearly discussed in Sections 3.5 and 3.6, the 

concepts which form the grounds of LAM spectroscopy can be easily defined. Let 

us consider an all-tmizs polymethylene chain and focus our attention on the branch 

(08 of the phonon dispersion curve. The branch starts as mostly C-C stretching and 

ends as CCC bending. For a finite chain of N chemical units we expect to locate on 

this branch a band progression of N vibrational modes characterized by the phase 

coupling pJ = jn/N. The ‘most in-phase mode’ occurs at very mall frequencies; its 

frequency approaches zero when the chain tends to be of infinite length. Let 

1,2,3 . . . N label the components of the band progression of the modes which lie on 

the Cllg branch. 

Schaufele and Shimanouchi [ 13 I] have shown that for centrosymmetric all-from 

polymethylene chains the lowest frequency mode performs an accordion-like motion 

(LAM) with the atomic displacements describing a wave with one node at the

center of the chain [ 1321. Such a mode has been labeled as LAM 1 mode. The higher 

frequencies components have been indicated analogously as LAM?, LAM3, etc. 

The odd modes give origin to a strong band progression in the Raman spectrum, 

with LAMl extremely strong and the other components with intensity rapidly 

decreasing as the number of nodes increases. 

Schaufele and Shiinanouchi have considered the longitudinal stretching motion 

of an elastic rod of length L, elastic modulus E and density p and have shown that 

the frequency of its ‘according’ motion can be written as 

(3-53) 

The same authors have shown that if E and p are suitably chosen for polymethylene 

chains, VLAM~ of tram-planar finite polymethylene chains also follows the 

same law. In this case L is the length of the all-miis polymethylene chain which can 

be approximated as 

L(A) z 1.25(A)(N - 1) (3-54) 

where N is the number of CH2 units of the chain. 

LAM spectroscopy gained much popularity in polymer characterization and 

morphology since it has been possible to measure from Raman spectra the length of 

the tmns-planar stems in polyethylene laniellae and in many other relevant cases. 

As mentioned above, these concepts were applied in a straightforward manner 

to the case of fatty acids by the French group [108], but the results were highly 

controversial, thus casting some serious doubts on the applicability of LAM 

spectroscopy. 

The dynamical problem has been re-examined with the purpose of evaluating the 

effects of chain ends on LAM modes. Simplified models were employed which 

consider chemical repeat units as point masses joined by elastic springs. Various 

cases of perturbations of the LAM modes were considered namely: 

(i) The effect of heavier masses at one or at both ends of the polymethylene chain 

mimicking systems such as X-(CH2),-X where X = -CH3, F, C1, Br, I, etc. It 

transpires that \’LAMI decreases when the mass of X increases [133]. 

(ii) The effect of joining a polymetliylene chain with another chain with varying 

length and with different and variable spring constants. The realistic case 

studied was the class of eniifluorinated n-alkanes F(CFZ)~ . (CHZ)~H where N 

and M change in a broad range of values. Contrary to expectations, the two 

segments do not generate two independent LAMl modes, nor can a simple 

heavy mass (X as (i)) at the end of a chain account for the observed spectra. 

Calculations show that the molecule vibrates as a whole and generates only one 

LAMl mode whose frequency is critically dependent on the length of both 

segments. Moreover, the node of such LAMl inode shifts along the chain 

depending on the length N and M of the two arms [134]. 

(iii) The above results opened the way to the study of dimeric systems where the 

two monomers are joined head-to-head by springs with varying spring con-

3.1 7 Fermi Resonancrs 163 

stants. This model aimed at reproducing the class of dimeric fatty acids or fatty 

alcohols. This is the most important case both for its dynamical implications 

and for the consequences on structural determination as indicated at the beginning 

of this section. The modeling [133] considers two chains with one 

heavier mass at one end joined head-to-head by a spring with variable spring 

constants. Such a spring mimics the hydrogen bonds which link the two carboxyl 

groups. 

The results of the calculation indicate that LAMl extends over the whole dimer, 

with the node at the center of the dimer between the two carboxyl groups. The frequency 

of LAMl has to be searched at much lower frequencies and the frequencies 

of the band progression of higher-order LAM modes are totally different from those 

identified in the spectra of these materials by the French group. The new band 

progression is easily found in the low frequency Raman spectra (Figure 3-36), thus 

giving further experimental support to what was shown by defect-mode spectroscopy 

(see Section 3.14.1). namely that fatty acids in the crystal do not contain conformational 

kinks and are fully trans-planar. 

3.17 Fermi Resonances 

The next step in OUT analysis is to focus on additional spectroscopic manifestations 

of molecular order and disorder which are not only intellectually interesting as 

physical phenomena, but also become of extreme importances as markers for the 

experimental routine characterization of the very many systems containing segments 

of polyinethylene chains. 

We are faced with the following experimental facts which need to be understood 

and possibly used for analytical and structural determinations. The Raman spectrum 

of polyethylene (or practically of any polymethylene chain) in the solid alltrans 

structure is very similar (or identical) to that given in Figure 3-37a; upon 

melting, the overall appearance of the Raman spectrum is that given in Figure 

3-37b. 

3.17.1 Principles 

The experimental fact is that whenever a molecule contains a few CH? groups the 

Ranian spectrum shows a very characteristic pattern of lines in the CH2-stretching 

frequency range. Since they are strong they can be observed in a variety of materials, 

even under the most awkward experimental conditions. Indeed, they can be 

observed for monolayers of organic or biological inaterials spread on surfaces, in 

aqueous solutions, in microscopic crystals, etc. Moreover, the spectrum shows 

drastic changes in frequencies and intensities when the molecular structure or the 

environment changes. For this reason the Raman spectrum is at present a very

164 3 li'hratioiziil Spectra as a Probe oj'StrzictLira1 Order 

+ 

+ 

+ 

400 200 0 

Frequency (cm-' ) 

Figure 3-36. Rainan spectra (0-400 cm-' ) of (a) hexatriacontane, n-C36H74; (b) stearic acid, 

n-Cl7H35COOH and (c) strearyl alcohol, n-ClgH17OH. Notice that since fatty scids and alcohols 

form head-to-head hydrogen-bonded diniers LAM modes extend throughout the whole dirneric 

system. Thus, the LAM progression shifts toward much lower frequencies than those expected for 

one isolated chain, approaching the values of hexatriacontane. The lowest frequency strong LAM1 

mode has its node at the center of the molecule, inside the hydrogen-bonded diineric structure (see 

text). 

powerful and essential diagnostic tool on the conformatioilally ordered/disordered 

structure of complex organic, biological and industrially relevant polyinethylene 

materials which contain polymethylene chains. 

The reason of such peculiar aid specific spectroscopic manifestations is that 

complex Fermi resonance phenomena occur in molecules containing polymethylene 

chains. We are concerned in this chapter on the Ferini resonances associated with 

the vibrations of polymethylene systems in the trans or gauche states, considered 

either as isolated entities and/or when they are organized in ordered and/or disordered 

supermolecular structures. Let us remember that Ferini resoiiaiice derives

3.1 7 Femi Resorinnces 165 

2 

L 1 

2900 2800 

frcgurlcr I cm-' 1 

I 

[ , , I , ! 

3000 2900 28oC 

Figure 3-37. Examples of the changes with physical states of the Raman spectrum of polyniethylene 

systems: (a) Raman spectrum of crystalline orthorhombic polyethylene in the CH stretching region; 

(b) Raman spectrum of polyethylene in the molten state in the same frequency range. 

from the mixing, because of anharmonic terms in the potential, of two (or more) 

eigenstates (one fundamental and one (or more) overtone( s) or combination( s)) 

which belong to the same symmetry species. As a result, generally the two (many) 

interacting levels repel each other and the generally weak overtone(s) (or combination(s)) 

borrows (borrow) intensity from the much stronger fundamental [2, 31. 

Fermi resonance has for a long time been a well-known and explicable fact in the 

case of small molecules [135]; the case of hydrocarbon chains was simply and 

clearly presented by Lavalley and Sheppard [ I361 for the molecule of hexadeuteropropane 

(CD3-CH?-CD3), which consists of an isolated CH? group between two 

CD3 groups with which little or no mechanical coupling is likely to occur, in a first 

approximation. The molecule belongs to the symmetry. Figure 3-38a gives the 

survey Raman spectrum of this molecule, while Figure 3-38b and 3-38c show an 

expanded section of the Raman spectrum polarized ( 1 1 ) and depolarized (I)[137]. 

Normally, in the CH? stretching range we should reasonably expect two lines to 

be associated with the antisymmetric (d-) and symmetric (d') CH? stretching of 

species BZ (1) and A1 ( 1 1 ) respectively. Instead, we find two strong lines certainly 

with A, symmetry (polarized) and one certainly with B7 symmetry (depolarized). 

The dynamical situation which justifies the observed Raman spectrum is the fol-

CO]CH$O] 

a 

CD3CHzC03 MELT, I I I 

300.K 

L 

I( 

J 

177.K 

L 

101.K 

1400 1500 2800 2900 3000 

AAYAN FREPUENCY (ern-' 

b 

0 1500 2800 2900 xxx) 

RAMAN FREPUENCY (crn-' 

C 

Figure 3-38. (a) Survey Raman spectrum of hexadeuteropropane (CDj-CHz-CDl) in the solid 

state at 15 "K. The species of the vibrational transitions are indicated; (b) expanded temperaturedependent 

Raman spectrum in the 1400-3000 ern.-' range in polarized light (11 polarization) exhibiting 

the transitions of A, species; (c) expanded temperature-dependei~t Raman spectrum in the 

1400-3000 cn1-l range in polarized light (1) showing the transitions of B1 species, free from Fei-mi 

resonances near 2925 cin-'. The dashed scattering is due to 'bleeding' of the 11 polarization due to 

experimental difficulties.

lowing. The CHl-bending mode 6 [certainly of A1 symmetry ( 1 1 ) ) 

3.17 Fermi Resonaiices 167 

(Figure 3-37b) 

observed at 1460 cm-~' should have its first overtone 126 : 2 x 1460 = 2920 cni-', 

A1 x A1 = Al) which can enter Fermi resonance with the d+ mode of A1 symmetry. 

Eigenstates mix, frequencies are split apart with intensity borrowing froin the 

fundamental to the overtone producing two strong lines at 2876 and 2942 cni- ' 

(Figure 3-38a-c'i. 

It should be remembered that the first overtone 26 should generally have vanishing 

intensity, but because of Fernii resonance its frequency is pushed upward and 

it borrows a lot of intensity from the fundamental d+ observed at 2876 cm-' . 

A similar, but seemingly more complex case, occurs for polymethylene chains. 

Let us consider the dispersion curve of polyethylene or the corresponding vibrations 

of finite polymethylene chains (Figure 3-39 j whose frequencies lie on these dispersion 

curves at finite values of the phase coupling (Section 3.5). The vibrations of the 

lower portion of the dispersion branch of CH2 rocking have their first overtone 

levels ( - 720 x 2 zz 1440 cnrl) close to the fundamental levels of the CHZ-bending 

motions. On their turn, the first overtones of the levels along the dispersion curve of 

the CH2-bendings occur just in the frequency region where the fhdaniental CH 

stretching modes occur (2 x -1440 z 2880 cnirl). When symmetry allows, Fernii 

resonances occur and levels repel each other with drastic intensity changes. The 

issue becomes formally more complex since (Section 3.6) when molecules crystallize 

in an orthorhonibic lattice (with two molecule per unit cell) the number of frequency 

branches doubles, thus doubling the number of levels which can generate 

both overtones and combinations which may produce Fermi resonances in the 

proper frequency ranges, as just indicated. 

These Fernii resonance phenomena have found detailed mathematical treatment 

first tackled by Snyder et al. [138, 1391 for single all-tram chains and later reformulated 

by Abbate et al. for ii) single chains, (ii) an orthorhombic lattice [140], (iii) 

all-trans, and (iv) conformationally distorted chains [ 141, 1421. The most interesting 

features to be accounted for are the Raman scattering of polymethylene chains in 

the frequency regions centered near 1450 and 2900 cnirl. 

For the sake of simplicity in the discussion which follows we label with (+) and 

(-) the symmetric and antisymnietric combinations respectively either of internal 

coordinates (e.g., C-H stretches within a CH? group) or of 'group coordinates' between 

two CH2 units within the 'chemical repeat unit' made up by two CH? groups. 

First, we discuss the case of true k = 0 modes for an infinite chain (or the case of 

k 4 0 for finite long chains). Let us, for the sake of simplicity, consider the spectrum 

of traiis polyethylene both as a single chain and crystallized in an orthorhombic 

lattice with two molecules per unit cell. Since the spectrum of polyethylene 

as single chain is not available for this discussion we necessarily use the calculated 

spectra as simulation of reality [140]. Let us first compare the general features of the 

calculated Raman spectra in the CHz-bending centered near 1435 cm-' for a single 

chain and for crystalline polyethylene (Figures 3-40a, b) (point group D?!,). The 

single very strong line near 1435 cm-' (Figure 3-39) is to be assigned to the limiting 

k = 0 totally symmetric in-phase CH?S+ mode of A, species. For the crystalline 

material, such a mode splits into two levels near 1420 (Ag) and 1435 an-' (BI?); an 

additional broad scattering (with a wing towards higher frequencies) is calculated

168 

3 Vibrational Spectra as u Probe oj. Strurtrrrcil Ordu 

30( 

V 

cm -1 

I 

25C 

15C 

1oc 

Figure 3-39. Dispersion curves of 

single-chain polyethylene, indicating the 

scheme of Fermi resonances which can 

occur between phonon of different 

branches. 

near 1464 cm-'. Following the previous procedure we look first at the first overtone 

of the infrared active CH? rocking P- mode for a single chain (line group symmetry) 

(2P-), BzU x Bzu =A,. For the line-group, Fermi resonance can take place 

between 2P- of A, species and the A&+ flindamental, giving rise to the weaker line 

near 1445 an-'. Extension of the treatment to the case of the orthorhombic lattice 

with two molecules per unit cell needs to consider that each level splits into two

3.17 Ferini Re.sonrriire.s 169 

a 

Figure 3-40. (a) Calculated Raman spectra in the CHZ-bending region of single-chain polyethylene. 

The BZ was divided into 72 intervals of 5". Each Lorentzian function has AI~,/~ = 5 cn-': (b) 

Calculated Ranian spectrum in the CHz-bending region of crystalline polyethylene as n convolution 

of Lorentzian lines as in case (a).

15 

Figure 3-41. Calculated Raman spectrum for crystalline polyethylene in the CH-stretching region. 

The spectrum is a convolution of numerous combination and overtone levels in Ferini resonance 

with the nonresonant CH2 antisyinmetric stretchings. This spectrum should be compared with the 

experimental one of Figure 3-37. 

levels; moreover, in addition to the overtone levels, we should also consider their 

combinations. For the space group we can use the observed experimental frequencies: 

(2P-)-, B3u x B3,, = A,, 2 x 731 = 1462 an-'; 2(P-)+, B?,, x Bzu = A,, 

2 x 719 = 1438 cm-'; (P-)- + (P-)', B3u x B?, = Bl,, 731 + 719 = 1450 cn-'. It 

follows that for the space group the mode (af)+ can couple with 2(P-)- and 

2(P-)+ while (6')- can interact with the combination (P-)- + (P-)'. Raman 

spectra of stretch-oriented (extruded rods) in various polarization directions 128, 

1231 show that the symmetry species of the lines at 1464 and 1440 cm-' is certainly 

Big, while the line at 1416 cm-' is certainly A,. 

The full treatment needs to consider in addition to the k 40 modes also the 

k # 0 modes (throughout the whole BZ) due to either the finiteness of the chain, or 

to the whole continuous dispersion curve for an infinite chain. The population of 

overtones and combinations becomes extremely large and many of these levels can 

enter Fermi resonances with a large transfer of intensity from the fundamental 

strong modes. 

Once the bending region has been understood the same approach must be taken 

by considering all line-group or space-group overtones and combinations of the d 

modes which are in Fermi resonance with the CH2-stretching modes. All details of 

the calculations are provided in [140]. Figure 3-41 gives the result for crystalline 

polyethylene. 

The most important and useful feature in the Raman spectra of polymethylene 

chains is that reported in Figure 3-37, which is easily observed every time a sequence 

of CH2 units occurs in a molecule. If the system forms an orthorhombic 

lattice, generally one observes a broad and strong line near 2850 cn-' with a broad 

wing which extends towards higher frequencies foiming a broad background scattering 

that reaches -2950 cm-'. Broad and sometimes ill-defined bands appear 

near 2930 and 2900 cin-l. For the orthorhoinbic system, such a background scat-

3. I7 Fernii Resoiiances 17 1 

tering has been shown experimentally to have A, symmetry [28, 1231 and originates 

from a convolution of a large population of overtone and combination levels which 

are in Fermi resonance with the symmetric CH?-stretching mode which, in principle, 

should occur as an unperturbed strong level near 2880 cm-', but has been 

pushed downward by Fernii resonances with the manyfold of overtones and combinations 

to which it has also lent intensity (Figure 3-41). Floating on top of such 

a broad A, scattering we find the strong and sharp line associated with the CH? 

antisymmetric stretching which remains unperturbed because it does not enter in 

any Fermi resonance coupling. 

3.17.2 Applications 

We need now to translate the above theoi-etical results into practice. First, let us 

start from an orthorhombic lattice with two molecules per unit cell and lower the 

symmetry to a supermolecular organization in which the cell can be considered as 

isolated and decoupled from its neighbors. This situation can be achieved in the 

following case: (i) isotopic solid solutions; (ii) molecule as a guest in a chemically 

different host lattice (e.g., urea clathrates) [ 1431: (iii) liquid crystalline-like systems; 

and (iv) systems with confomiationally distorted chain ends, but with an ordered 

bulk [1031. The following changes of the Raman spectrum are expected (and 

observed): 

(i) The triplet of lines in the bending region becomes almost a singlet (@), since the 

crystal field splitting disappears and the convolution of overtone and conibination 

bands for finite chains (two phonon states for polyethylene) in Fermi 

resonance with the crystal field split fundamentals are removed (Figure 3-40). 

(ii) As the number of overtone and conibination states in Fermi resonance with the 

A, fundamental state (CH2 symmetric stretch) decreases we expect the lowering 

of the total A, scattering and possibly a shift toward higher frequencies 

of the A, fundamental. Generally, the scattering in the valley between the 

antisymmetric and symmetric CH? stretch 2980 and 2850 cm-' respectively) 

is observed to decrease and the two fundamental lines sharpen somewhat. 

Attention should be paid to the fact that for chains with CH3 end groups 

the symmetric CH-stretching mode shows up as a weak line within the same 

frequency range. 

The next step is to interrupt the long trans chain by gauche defects. The following 

changes are expected (and observed) in the Rainan spectrum [ 141, 1421. 

(iii) As both P and 6 inodes of CH? groups in gauche conformation occur at different 

frequencies the population of vibrational levels which enter Fernii resonances 

in the trans interaction scheme decreases. We expect changes of the 

Raman spectrum both in the CH?-stretching and bending regions. Qualitatively, 

the vibrational spectrum should change from that of a full. all-fr~nis 

single chain molecule toward that of a fully decoupled structure, thus tending

172 3 Vibrutionul Spectra as n Pvohe of StrirctLiral Order 

to the situation described in Figure 3-38 for hexadeuteriopropane. It should 

be noticed that necessarily the 2850 cn1-l line should move towards higher 

frequencies approaching the unperturbed value near 2880 cni-' and the 2s line 

near 2930 cm-' should eventually become stronger as it does not share with the 

manyfold of binary combinations and overtones the intensity borrowed from 

the fundamental originally observed near 2850 cm-'. 

(iv) A drastic increase in concentration of gaziche structures occurs when the polymethylene 

system melts. In that case one can assume that the statistics of 

Flory's Rotational Isomeric State Model [48] works and we can describe the 

system as coiisisting of a distribution of gauche and short trans sequences. 

Upon melting, the d+ band near 2850 cm-' becomes the strongest feature in 

the spectrum and d- collapses into a very broad band with its maximum 

shifted upward of 4-10 cni-I. The overtone level near 2940 cm-l increases 

in intensity. This is a very useful marker when the chain is confoimationally 

collapsed (liquid like) (Figure 3-37b). 

(v) There are two possible interpretations of the broadening of d- band observed in 

the liquid phase. The first is that the broadening is the result of a convolution 

of many transitions associated to d- mode of CH? groups in constrained nonequilibrium 

near gauche conformations [143, 1441. Another explanation of the 

broadening refers to the overall dynamics of the whole chain which may occur 

in the liquid (see Section 3.18) [145]. 

(vi) A detailed study has been carried out by Abbate et al. on the molecule 

CD3-CH2-CHa-CD3 11411 as a model in which (a) the possible coupling with 

the methyl end groups are removed by deuteration and (b) only one degree of 

torsional freedom between CH? groups is allowed. It is immediately apparent 

that in gauche conformation the symmetry of the molecule is lowered, thus 

changing the resonance schemes which may involve also the d- mode which 

is unperturbed in the truns case. Calculations show that both d- and d+ 

are allowed to mix, thus reshuffling the intensities in a way that is difficult to 

predict. 

(vii) The experimental intensity ratio R = I2~50/12940 has been proposed as an experimental 

way to determine the relative concentration of trans versus gauche 

conformers in a sample. R has been matter of strong debate (see for instance 

[ 1401). As a result of the calculations by Abbate on hexadeuteriobutane [141], 

because of a mixing of the modes and the unpredictable changes of the extent 

of Feimi resonance, we do not advise the use of R as analytical probe in liquid 

polymethylene systems or in partially disordered solid polyethylene samples. 

3.18 Band Broadening and Conformational Flexibility 

The experimental spectroscopic facts which need to be accounted for in a comprehensive 

way in relation to order-disorder are the following (we report here the 

most striking manifestations, recorded in various ways, for oligo and polyniethylene 

systems):

3.18 Band Broadening and Confonnational Flexibility 173 

Figure 3-42. Hexadeuteropropane: 

10- 

n 

0 " d+(CHZ) 

I I I 

Figure 3-43. Ranian spectrum in the C-Hstretching 

range of C~lH44: (a) solid and ib) melt. 

a 

3050 3000 2950 2900 2850 2800 cm-I 

1. For short chains, the antysimmetric d- mode broadens in the Raman spectrum 

by increasing temperature and sometimes shifts upward in frequency (Figure 

3-42), while d+ neither shifts substantially nor broadens. (See also Figure 3-37.) 

2. For long polymethylene chains and polyethylene at melting, the strong d- 

modes collapses into a featureless very broad scattering and by increasing temperature 

shifts its maximum towards higher frequencies (Figure 3-43). 

3. For crystalline n-alkanes the width of d- does not increase with teniperature up 

to a few degrees below melting (Figure 3-44). 

4. When n-alkanes are inserted as clathrates in urea channels, the width of d- is

174 3 Vilmitionul Slirctru us ci Probe of Stnictiirul Ortlev 

Figure 3-44. Evolution with temperature of the Raman 

spectrum of solid Cl,Hd, in the CH stretching rangc. 

Notice the behavior with temperature of d+ and d- inodes.

3.18 Bmd Brourleiziriy md Cmforrnational Flrxibilitjj 175 

IL-180°C 

-1 

cm 

3050 3000 2950 2900 2850 2800 2750 

Figure 3-45 Raman spectrum of C~jH52 in urea clathrate at 3°C and -180°C. Notice that for 

d-At,,,,2 = 6 cn-' at T= -180°C and Ai~,p = 28 cm-' at T= 25°C; Ai~,p of d' remains practically 

constant at the same two temperatures. 

narrow at low temperatures and broadens substantially at higher temperatures 

(Figure 3-45). 

3.18.1 Principles 

In Section 3.17, experiments have been presented showing that in the case of polyinethylene 

molecules (including polyethylene), in going from the solid all-tlvrizs to 

the liquid phases, the pattern of Raman scattering changes dramatically but in 

a consistent and characteristic way. We focus our attention on the temperaturedependent 

broadening observed for the d- and 8 modes; in contrast, the width and 

shape of d+ modes are practically temperature-independent. In Section 3.17.2 it was 

suggested that one possible explanation of the band broadening of d- could be a 

convolution of bands associated with d- modes in different quasi g~iziclie conformations 

existing in the liquid or 'amorphous' material. 

This interpretation may be only partial since Snyder has shown that band 

broadening of d- and 6 modes is certainly observed also for hexadecane in uiea 

clathrates where the molecule is believed to be (on average) in the tu.m~-planar

176 3 Lri/ibvational Spectrci cis (1 Probe c?f'Stncc.tLiral Ovtlev 

conformation [ 1461. Some additional or alternative processes then must exist. We 

wish to discuss the results of band-broadening studies which may shine more light 

on the conformatioiial dynamics in systems where order-disorder may exist. 

A full discussion of the theory and experiments on band-broadening effects in 

n-alkanes is reported in [137]. We wish to outline here a few basic concepts which, 

in a first approximation, are the tool to be used in the understanding of the complex 

dynamics in conforinationally flexible systems. 

Band broadening [147] both in infrared and Raman spectra of polyatomic molecules 

in the liquid phase may arise from vibrational relaxations (VIBR) (i.e., some 

kind of process dissipates the vibrational energy stored in a vibrational transition) 

and/or from reorientational motions of molecules (REORIENT) (i.e., the energy 

stored in a vibrational transition is used to generate a reorientation in space of 

the molecule). Vibrational relaxation can be due to vibrational energy dissipation 

to the rotational-translational degrees of freedom of the thermal bath, to intramolecular 

energy transfer between vibrational modes, or to several intermolecular 

phenomena such as resonance vibrational energy transfer and vibrational 

dephasing. Theory states that the amplitude-dependent vibrational relaxations are 

temperature-independent, while the angular-dependent reorientational relaxations 

depend strongly on temperature. 

Raman scattering provides a nice way to gather information on these processes. 

For liquid samples, it is possible to separate experimentally the isotropic spectrum 

(ISOT) from the anisotropic one (ANIS), namely 

I(IS0T) = I(ll) - (4/3)I(i) 

(3-55 j 

I(AN1S) = I(i) (3-56) 

From these measurable quantities, under certain approximations [ 1371, it is also 

possible to give precise expressions for the relaxation times: 

where 

z(V1BR) = 1 /(KcAu~~)(ISOT) (3-57) 

z(RE0RIENT) = ~/(KcAv~/~)(REORIENT) (3-58) 

It is coinnionly known that the depolarization ratio for linearly polarized light is 

given by 

p = I(I)/I(II) 1 3P2/(45C(' + 4p2j (3-60) 

where a is the mean polarizability describing the isotropic component of the Ranian 

polarizability tensor and P, the anisotropy, describes the anisotropic component. 

It follows that for totally symmetric modes p 5, 0, I(il)= IfISOT), and the vibra-

3.18 Band Bvoaderiing arid Conjonnationril Flexibility 177 

tional relaxation contribution can be easily evaluated. For all antisymmetric modes 

a = 0 and p = $; only the contribution from coupled vibrational-reorientational 

motions can then be obtained from Raman experiments. The evaluation of the pure 

reorientational contribution can be obtained by suitably subtracting the contribution 

from vibrational relaxations (Eq. (3-59)). 

3.18.2 Short and long n-alkanes 

The Raman spectra of a series of n-alkane suitably chosen have been studied [137] 

with the precise aim of extracting from d- (antisymmetric, - depolarized) and d' 

(symmetric - polarized) modes of the CH2 group information of general validity on 

intra- and intermolecular dynamics using the techniques discussed in Section 3. IS. 1. 

The following molecules have been studied [ 1371 in the solid and liquid states: 

propane-d6 (CD3CH2CD3) butane-d6 (CD3CH2CH2CD3), and pentane-dlo 

(CD~CH~CHZCH~CD~) and n-nonadecane. The following conclusions, which may 

be of general interest, have been derived: 

1. For d+ p = 0, thus allowing us to distinguish the contribution by vibrational 

relaxation processes; for d- CI = 0 and p = %, thus the coupled vibrationalreorientational 

processes are active simultaneously. The reorientational contribution 

was derived by suitable deconvolution methods (Eq. (3-59)). 

2. The temperature independence of the band width of d+ modes has been verified 

and provides an average value of T(VIBR) - 1.3 x 10l2 s for the three small 

deuterated molecules. An amplitude-dependent vibrational dephasiiig process 

(inhomogeneous broadening) best accounts for the observed temperatureindependent 

broadening. 

3. A ~ J of ~ / d- ~ is strongly teniperature-dependent and increases rapidly with 

increasing temperature (Figure 3-41); correspondingly, r(RE0RIENT) decreases 

by at least one order of magnitude from low to higher temperatures (see 

Table 3-3). It is found that, at a given temperature, identical for the three deuterated 

molecules, the band width increases with increasing chain length. Moreover, 

at a constant temperature, the reorientational relaxation time increases 

with increasing chain length. This trend is consistent with the view that the contribution 

of reorientational relaxation decreases with increasing chain length, as 

might be expected based on inertial and frictional effect. Indeed, it is conceivable 

that while the molecule of propane in the liquid phase may easily tumble in 

all directions about its axes, for longer molecules the end-over-end tumbling 

becomes progressively less likely while some sort of rotation about the long 

molecular axis may still occur even for longer n-alkanes. 

4. For chains larger than propane (for which no torsional motion about 

(CHl)-(CH?) bonds exists) it is quite conceivable that the increasing hinderance 

to molecular reorientation which broadens d- mode comes from torsional 

motions of the molecular skeleton, i.e., its overall torsional flexibility. In other 

words, when a flexible molecule tries to tumble it necessarily drives the motions 

of the torsional angles. We expect these contributions to increase with iiicrcasing

178 3 Vihmtional Spectra ITS CI Probe of Structural Order 

Table 3-3. Frequencies, bandwidths (full widths at half inaxiriiumi and reorieiitntioiiizl-rel~ixalion 

times of d- of propane-dh, butane-d6 and pentane-dl,] as a function of temperature. 

Temperature A1'1,2(d-) s(RE0RIENT) d- 

iK) (cm-') (cm-l) x 1012 s 

CD3CH:CD? 101 

177 

235 

288 

300 

CD1 CHzCH2 CD3 135 

185 

231 

300 

CD3CD2CH2CD2CD3 143 

191 

240 

300 

2919.0 15.7 

3920.5 24.8 

2925.0 70.0 

2923.5 79.0 

2896.5 15.0 

2896.5 24.2 

2899.5 40.0 

2898.0 65.0 

2896.5 18.1 

2898.0 20.1 

2899.5 24,s 

2901.0 30.9 

1.3 

0.58 

0. I7 

0.15 

1.7 

0.66 

0 33 

0.1s 

1.2 

0.93 

0.67 

0.48 

temperature and chain length. The threshold temperature at which for a given 

chain length coupled rotations and torsion are able to generate conformational 

kinks in a more or less nonequilibrium structure (either pinned at a given site or 

mobile along the chain) remains an unsolved problem faced by many authors 

who tackled the problem by numerical simulations [ 1481. 

5. For the short-chain models it has been found that, by increasing the temperature 

(i.e., when the coupling between rotations and torsions increases), the frequency 

of d- increases consistently by a few wavenumbers (see Table 3-3); this suggests 

that either the diagonal or off-diagonal force constants within each CH2 unit 

change when the niolecule is (even slightly) conformationally distorted or the 

onset of torsional distortions activates iiitraniolecular coupliiig which extends at 

an unknown distance s at either side of the CH2 group (see Eq. (3-45)). 

3.18.3 Collective Chain Flexibility 

In this section our attention is focused on attempts to describe in molecular terms 

the broadening and the upward shift of d- mode as discussed in Sections 3.18.1 and 

3.18.2. 

Let us first consider the case of a long 11-alkane molecule in the solid crystalline 

phase. Froin lattice dynainical calculations it is known that collective librational 

phonons (about the chain axis) exist in the lattice, some giving rise to scattering 

and/or absorption in infrared [70]. Let us take a n-alkane molecule with an even 

number of carbon atom which belongs to the C?h point group. The rigid rotation 

about its axis is of B, species, the same as the Raman active d- mode. When the

3.18 Band Broadening and Confornmtioiud Flesibility 179 

molecule is embedded in a medium which transforms the free rotation into a librational 

motion d- and the libration can, in principle, couple because they belong 

to the same symmetry species. A similar observation was made by Snyder for the 

CH2 rocking mode 11461. 

The contribution toward the understanding of chain flexibility comes from the 

following experiments and calculations which are described in detail in 11371 and 

[145]. 

1. The first experiment has been that of recording the Rainan spectra of long 

n-alkanes in urea clathrates [ 1451 and in another host lattice (perhydrotriphenylene) 

where each guest n-alkane molecule sits in a channel, isolated from 

the other chains and retaining, on average, the all tram-planar structure. Similar 

experiments are also reported in [146] and [149]. By changing the temperature, 

the host lattice changes the average diameter of the columns where the guest 

molecule sits, thus allowing a varying degree of librational freedom for the guest 

molecule. While at low temperature the librational motion is strongly hindered, 

at relatively higher temperatures the lattice expands anisotropically allowing the 

guest molecule to perform librational motions of large amplitude. 

We expect the band of the d- mode to be narrow at low temperatures and to 

broaden at higher temperatures. Figure 3-45 shows that this is indeed the case. 

The band not only broadens but its frequency also shifts upward by a few 

wavenumbers. It is concluded that the almost rigid (and possibly small amplitude) 

collective librations about the chain axis at low temperature evolve with 

increasing temperature towards a collective libro-torsional (or libro-twisting) 

motion with nonnegligible angular distortions more or less hindered by the 

environment. 

2. It must be pointed out that the broadening and upward frequency shift of the d- 

mode cannot be ascribed to the generation of gauche defects inside the host lattice. 

It has been verified experimentally that the frequencies of LAMl, LAM3, 

and LAM5 of the guest molecule embedded in the host lattice are identical to 

those observed for the same n-alkane in the crystalline state [145]. Hence, the 

molecule does not change its length from that of the tvaizs structure [145]. 

3. In going from the solid to ‘softer’ systems, certainly the frequency of the d- 

mode shifts upward by a few wavenumbers. No shifting (nor broadening) is 

observed for crystalline n-alkanes until the so-called soft ‘rotatory phase’ is 

reached before melting. In the so-called a phase (see Section 3.19) some shifting 

and broadening is observed. 

Levin [150] and Gaber and Peticolas [151] have shown that the upper shift 

and broadening occurs also in bioinembrane materials. Such a shift has been 

used by Gaber and Peticolas to monitor the ‘lateral interactions’ in phospholipids. 

We notice that for these phospholipid-type material the long-chain alkyl 

residues are fixed at one end to polar head groups and thus cannot perform rigid 

librational motions about their axes. Forcefully they can only perform librotwisting 

or libro-torsional motions. 

4. The last step is to account for the dynamical origin of the fact that when the long 

polyniethylene chain increases its libro-torsional flexibility the frequency of d-

180 3 Vibrational Spectra as ci Probe of Strtrctirrcil Ordu 

modes shifts very selectively upward, while the frequency of d+ mode remains 

unchanged. There must be some specific process which should be discovered. 

With this purpose various dynamical possibilities have been considered [ 1451. 

It has been concluded that a rational explanation is that there exists a selective 

intramolecular coupling between d- modes and the torsional modes of the carbon 

backbone. Symmetry consideration tell us that for a trans-planar molecule 

the existence of such coupling will affect d- and not d+ modes. In other words, 

within the quadratic approximations interactions of the type fCH stretch/C-C torhions 

seem to exist and are modulated by the amplitude of the collective torsional 

flexibility. 

Another contribution to the upper shift may come from a complex scheme 

of Fermi resonances which are activated because of the lowering of the instantaneous 

symmetry during the libro-torsional motion. This difficult problem 

remains unsolved. 

5. The final observation, relevant for the discussion which follows in Section 3.19, 

is that in crystalline a-alkanes the band width of d- is relatively narrow and the 

frequency does not change throughout the temperature range of existence of the 

crystalline lattice. This suggests that libro-torsional motions are not allowed in 

the crystal and only small amplitude collective librational phonons seem to be 

allowed. Some additional observations will be added in Section 3.19. 

3.19 A Worked Example: From n-Alkanes to 

Polyethylene Structure and Dynamics 

3.19.1 Description of the Physical Phenomena 

We wish to guide the reader through a realistic case worked out step-by-step using 

most of the concepts and methods discussed in this chapter. We consider the case of 

n-nonadecane (CH~-(CH~)I~-CH~) as a prototypical case and try to derive from 

the vibrational spectra all possible information about its structure and dynamics. 

Both order and disorder will be considered. The methods applied and the conclusions 

reached form the background knowledge to be used for the understanding ot 

(i) the structural properties of many molecules containing long-chain alkyl derivatives; 

and (ii) the structural features of polyethylene or other polymers containing 

polymethylene segments. 

The basic physics we wish to understand is the mechanism at the molecular level 

which generates the phase transitions of relatively short n-alkanes. The experimental 

nonspectroscopic facts to be accounted for are the following (for a collection of 

references and a full discussion of many of the data presented in the first part of this 

section see [ 1031): orthorhombic n-alkanes first undergo a solid-solid phase transition 

(Toicc) to a phase variously described as 'rotatory phase' or 'c1' phase whose 

structure has been matter of controversy in the literature. This phase characterizes a

3.19 A Worked E.utinple 181 

pre-melting state for n-alkanes; just a few degrees above ToLx the crystal melts. The 

temperature range of existence for the t( phase decreases with increasing the length 

of the chain and is also pressure-dependent. Even polyethylene seems to show a 

similar pre-melting state just a few degrees before melting. 

The generality of the phenomenon to be understood was the reason which has 

convinced us of the need to perfom a detailed spectroscopic study on the simplest 

system, n-nonadecane which shows the largest temperature range for the existence 

of the a phase. 

We have studied the following molecules: n-nonadecane, n-nonadecane-2D2 and 

n-nonadecane- 1 OD? (hereafter referred to as C19, 2D2 and 1 OD2 respectively). 

The known structural features of C19 are the following: crystalline C19 below 

22.8 "C crystallizes in the orthorhonibic lattice with two molecules per unit cell 

[152]. Chains are trans-planar and perpendicular to the plane formed by the methyl 

groups, thus being perpendicular to the lamella surface. At 22.8 "C a phase transition 

is observed to the a phase and melting occurs at 32 "C. The measured heat of 

transition is 3300 cal/mol, while the heat of melting is 10950 cal/mol [153]. The 

volume changes determined with a dilatometer at the solid-solid phase transition 

and melting are 2.5% and 10.5% respectively. The structure of the r phase below 

melting is reported to be 'hexagonal' with some indication of rigid rotations of the 

chains [154]. 

We add here inforination for the two selectively deuterated materials which were 

prepared in the authors' laboratory. There is no reason to think that 2D10 may 

not have the same structure as the parent molecule. The perturbations by the two 

deuterium atoms should be negligible regarding the static structure while they will 

affect the vibrational spectra. As a result of the 'regular' packing, the lamellae 

formed by these all-trans 10D2 molecules contain precisely in the middle a 'layer' of 

CD2 groups. From the viewpoint of vibrations we expect normal crystal field splittings 

for the vibrations of the CD2 since intermolecular phase coupling within the 

lattice may occur regularly. In contrast, 2D2 molecules in the crystal, while retaining 

the orthorhombic structure, are arranged at random with the heads labeled 

with CD2 randomly distributed up or down within the crystal. Within Borii- 

Oppenheimer approximation the intermolecular potential cannot distinguish between 

deuterated or undeuterated ends of the hydrocarbon chain. However, because 

of the random mechanical perturbations of the vibrations at the site of chain heads 

in this molecule, crystal field splitting of the CD2 vibrations is not expected (in 

spite of the orthorhombic packing). These predictions are indeed verified in the infrared 

and Raman spectra of both lOD2 and 2D2. 

3.19.2 Vibrational Spectra of pure n-Nonadecane 

N = 59 is the number of atoms in the C19 molecule. The symmetry point group of the 

single trcms-planar chain of C19 is C?" and its 59 x 3 - 6 = 171 vibrations can be 

classified in the following irreducible representations with their spectroscopic activity:

182 3 T7ibmtional Spectra CIS a Probe of Structurd Order 

I- P 

u+w 

I I I I I 1 

IW 1400 1200 loo0 000 600 

v cm-' 

Figure 3-46. Survey infrared spectrum of n-nonadecane in the orthorhonibic phase. Band progressions 

are indicated. For the labeling of the modes see text. R = skeletal stretching, p = CHj 

deformation. 

The infrared spectrum of C19 in the solid orthorhonibic phase is shown in Figure 

3-46; we use this spectrum as reference in the present analysis. The Raman spectrum 

of the same material is given in Figure 3-47. 

The identification of the large number of normal inodes in the infrared and 

Raman spectra of these molecules becomes indeed unfeasible. Life becomes easier 

if we consider C19 as a segment of polyethylene made up by 17 CH2 units and try 

to located the 17 normal modes lying in each frequency branch of the dispersion 

curves of infinite polyethylene (Figure 3-1). Each mode is then characterized by a 

phase shift q, between adjacent CH2 units, qj = 2nj/17, (here j = 0,. . . , 17 - 1). 

The possibility of observing all noniial modes depends on their symmetry and on 

their intrinsic intensity in infrared or Raman, as well as on the dispersion of each 

phonon branch. As discussed in Section 3.9.2, we expect to observe in the infrared 

spectrum band progressions with decreasing intensity for branches with reasonable 

dispersion, while for flat branches band progression will be strongly compressed 

with strong overlapping. 

The experimental observation predicted by theory has been discussed previously 

in detail by Snyder and Schachtaschneider [9, 341. From the dispersion curves 

of Figure 3-1 it follows that 110 clear band progressions are expected for CH2 

stretching (d, overlapping) and CH2 bending (6, overlapping), while progressions 

are expected for CH2 wagging (w, very weak in IR); in the CH? wagging progression 

we clearly locate Ws, W7, WS, and W11; the other CH2 wagging motions are

3.19 A Wovh-ecl E.uanzple 183

184 3 T/ihrutionul Spectra as u Probe of Strzicturul Order 

overlapped by the stronger CH3 U mode near 1375 cm-' . The expected progression 

of CH2 twisting (t) must be very weak in IR while CH? rocking (PI must be strong/ 

medium and clearly noticeable from 720 to near - 1100 cni-'. C-C stretchings are 

weak in IR. For the band progression of the P modes we expect to observe nine 

components of B2 species corresponding to quasi-phonons with odd j values. 

J = I, 3, and 5 turn out to be almost overlapping and the others must he observed in 

the frequency range -750-1 100 cm-' . The observed band progressions are clearly 

displayed in Figure 3-46. 

The infrared bands at 1375 and 890 an-' are unquestionably assigned to the 

umbrella (U) and rocking ([I) modes of the methyl groups respectively (Figure 3-46). 

The Raman spectrum provides further information. It is generally known that in 

the Raman spectra only the limiting k i 0 modes for the infinite chain appear as 

strong or easily observable transitions. This is also the case for C19 where only the 

k 2 0 modes are observed. 

The most important information conies from the observation of the band progression 

of the LAM modes (Section 3.16) with k = 1,3, and 5 components 

observed at 124.5, 342, and 490 cni-'. Using equations 3.53 and 3.54, the frequency 

of LAM1 at 124.5 c1n-l proves that the chain comprises exactly 19 C atoms. 

The conclusions are as follows: 

Conclusion no. 1 

The infrared and Raman spectra provide unquestionable evidence that C 19 below 

22.8"C consists of trnns-planar molecule with 17 CH2 units capped by two CH3 

groups. Further observation of the infrared spectrum for the orthorhombic phase 

shows the splitting (factor group splitting) of the 6 modes and of many components 

of P modes into doublets with a measurable intensity ratio (Figure 3-48). Analogously, 

the Raman line associated to k - 0 of the 6 mode near 1460 cni-' is split 

into a doublet (Figure 3-47). The number of components of the splitting is related to 

the number of molecules (2) in the unit cell. If the orthorhombic lattice is accepted, 

the intensity ratio of the components of the splitting in the infrared provides the 

value of the setting angle of the hydrocarbon chains in the unit cell (8 e 42"). The 

shape of the Ranian scattering in the 3000 and 1450 cni-' range can be interpreted 

in terms of Fermi resonances (Section 3.17) uniquely between trans-planar chains. 


IR and Raman spectra provide evidence that two trms-planar molecules are 

organized in the orthorliombic lattice, with setting angle 8 = 42". 

3.19.3 Temperature-Dependent Spectra 

Next, we study the temperature-dependent infrared and Raman spectra of C19 

through phase transition (Ta) (T, = 22.8 "C) and melting (T,) (TI,, = 32 "C). We 

first focus on the spectroscopy of C19. Each experimental observation is labeled and 

will be included in the overall analysis. In the IR spectrum:

3.19 A Worked Example 185 

19 5°C 

215°C 

28°C 

Figure 3-48. Temperature dependence of 

the infrared spectrum of a-nonadecane 

in the CW2 rocking region showing that 

the crystal field doublets coalesce into 

singlets at the orthorhombic 4 c( phase 

transitioiis indicating the existence of 780 760 740 720 700 

‘single chains’. 

v cm-1 

(a) At T, the band progression of the P modes which shows factor group splitting 

at lower temperature loses all the splitting, but remains as prominent feature in 

the IR spectrum until T, (Figure 3-48). 

(b) The progression of the P modes disappears at T,, and is replaced by a broad 

band near 722 crn-l.

186 3 Vibrational Spectra cis a Probe of Structural Order 

I I I I I I 

1000 900 800 700 

v cm-' 

Figure 3-49. Band progression 

due to CH2 rocking 

modes for n-nonadecane in 

the orthorhornbic phase (at 

0°C) and in the a phase. 

Notice at 25 "C the appearance 

of a few weak bands 

and the disappearing of the 

crystal field splitting. 

(c) At T, the mode at 1375 cni-' (U) shifts to 1378.5 cm-'. 

(d) At T, a weak, clear and broad absorption appears at 766 cni-* and remains a 

clear, weak and broad band in the melt (Figure 3-49). 

(e) At T, additional peaks are observed in the infrared at 766, 804, 845, 866, 877, 

940, and 956 cm-' (Figure 3-49). Some of these modes may be assigned to the 

even j values of the P band progression. 

(f) At T, in the CH2 wagging progression we clearly locate WS, W7, WS, and W11 

as in the case of the solid, but the frequencies of these components are rigidly 

shifted toward higher frequencies for a few wavenumbers (Figure 3-50). The 

other CH2 wagging motions are overlapped by the CH7 U inode which shifts to 

-1378 cm-l. 

(8) At T, the CH; ,B mode shifts from 891 to 889 cin-l. 

(h) At T, a weak, but unquestionable peak arises at 1342 c1n-l (Figure 3-50). 

Upon melting: 

(i) At T, the band progression of the CH2 wagging disappears (Figure 3-50). 

(j) At T,ll the peak at 1342 c1ii-l gains considerable intensity (Figure 3-50). 

(1) At T,, the two well-known defect modes absorptions near 1353 cm-' and 1370 

cni-' appear together with the broad peak centered near 1306 cm-' (Figure 

3-50). 

In the case of the Rainan spectrum the most meaningful observations are the following: 

(m) Above T, the features of the k = 0 modes of the infinite trans-planar molecule 

are still the dominant features (Figure 3-51b). 

(n) At T, the lattice band at 98 cm-' disappears. 

(0) At T, the LAMI frequency shifts from 124.5 to 123.2 cm-' (Figure 3-52).

3.19 A Workrd Example 187 

Figure 3-50. Ternperaturedependent 

infraied spectra of 

n-nonadecane showing changes 

through the solid-cc (T - 22OC) 

and r-liquid (T - 29 "C) phase 

transitions 

i I , I 

1400 1360 1320 I290 1240 

v cm-' 

(p) At T, the A, component of the factor group splitting of the CH2-bending 

mode near 1417 cm-' practically disappears, even if not completely (Figure 

3-53). 

(q) At T, the valley between the A, (2845 cm-') and Bl, (2882 cm-') modes of the 

C-H-stretching region becomes emptier and the band at 2845 cm- ' sharpens. 

(r) At T, a slight increase in the scattering near 1090 cm-l is observed. 

(s) At T, on the lower frequency side of the well-isolated peak near 890 cm-' (/3 

mode) two weak bands are observed at 866 and 844 cm-' (Figure 3-54).

188 3 Vibrational Spectra as a Probe ojStrticttirci1 Order 

ORTH. 

MELT 

I , 1 I 

2wQ 2800 ” 1200 800 coo 0 

v cm-1 

I 

Figure 3-51. Changes of the Raman spectrum of n-nonadecane in the various phases: (a) ortliorhombic, 

(b) a, (c) liquid. 

The Raman spectrum in the melt shows the following features: 

(t) At Tm the LAM1 mode disappears completely and is replaced by the so-called 

‘pseudo LAM’, a broad and weak scattering near 320 cn1-l common to all 

liquid n-alkanes, melt polyethylene or any molecule containing long CH2 

chains in the ‘melt’ state (Figure 3-51c). 

(u) At T, the Raman lines at 866 and 844 cm-’ gain in intensity and, together 

with the line near 890 cm-l, form a strong characteristic triplet with components 

of similar intensity (Figure 3-51c). 

(v) At Tm the Fermi resonance line at 1461 cmP1 decreases while the Fermi resonance 

line at 2930 cm-‘ increases.

3.19 A Worked E.rcnniplr 189 

LAM I 

I 

Figure 3-52. Details of the changes of the 

Ranian spectra of ti-nonadecatie at 0 "C 

(orthorhombic) and 29.7 "C ( M phase). The 

I 

lowering of the frequency of LAM 1 shows 200 170 140 110 

that interlamellar forces have weakened. 

v crn-1 

I 1 I I I I I 1 

1470 1440 1410 1380 1350 1320 1290 1260 

cm-1 

Figure 3-53. Details of the Raman spectrum of n-nonadecane just above and just below the solid 

- il phase transition. The crystal field splitting near 1430 cnr' disappears and only the single chain 

bending deformation remains at 1440 cin-'. The Fernii resonance line near 1460 cm-' is also 

slightly perturbed.

j&890 860 830 

29,JT 

~.~ 

Figure 3-54. Details of the Raman spectrum of n-nonadecane 

in the orthorhombic (0°C) and c( phases 179.7‘C). The line 

near 890 cn-’ indicates that all chain ends are in the TT confomiations. 

In the u phase a small poptilation of chaiii~ has 

tilted their heads in TG arid GT conlormations. 

(w) At TI, in the CH stretching range the line at 3882 cin-’ is submerged by the 

broad scattering near 3000 cn-’ from which two peaks emerge at 2850 (d’) 

and 2875 cm-‘ (d-). The typical Raman spectrum of the “melt” is observed. It 

should be compared with Figure 3-43. 

(x) At T, near 720 cmP1 a clear stepwise increase of the diffuse broad background 

scattering is observed. 

(y) At T, the line near 1090 cin-’ becomes strong at the expenses of the k = 0 

singularities of the infinite trans-planar chain at 1135 and 1074 cm-’ in the 

orthorhombic and a phases. This line is characteristic of the stretching of the 

C-C bonds in gauche conformation (Figure 3-51c). 

Conclusioii no. 3 

At the orthorhombic to a phase transition the disappearance of the factor group 

splitting (observations (a) and (n)) indicates that the orthorhombic lattice with two 

molecules per units cells disappears and is replaced either by a unit cell with one 

molecule per unit cell or by a lattice where the tridimensional order is practically 

lost. 

Chains, however, remain mostly straight tmns-planar (observations if), (in),(o), 

(p) and (9)). Perturbations appear at chain ends (observations (c), (g) and (o)), 

indicating that the medium surrounding the CH3 group has changed, in particular 

the inter-lamellar forces have weakened (observation (0)). The following structures 

appear at the end of the molecule -CH?(T)-CH2(G)-CH?-CH3 and 

-CH?(G)-CH~IT)-CHI-CH~(observation (s)). This accounts for the lowering of 

the overall symmetry which causes the appearance of a few of the coinponents of 

the P modes inactive for the orthorhombic stiucture (observation (e)). It seems 

likely that some sort of disordering at the surface of the lamella takes place with 

some chains having their ends tilted in gauche conformations (observation (r)). The 

occurrence of a few GTG, GTG’ and GG conformational defects (or unspecified 

gauche defects) is indicated (observations (11) and (1)); the topology of these defects 

along the inoIecuIes cannot, so far, be defined.

3.19 A Worked E.xuniple 191 

Figure 3-55. Sketch of assumed 

locations of selectively deuterated 

chains of n-nonadecane (2D2 arid 

10D2, see text) in the orthorhombic 

crystal. 

- 

-5Hz- $ D z - ~ H ~ 


The process of melting can be described in the following way: the tvuizs-planar 

structure of the fraiu chains collapses (observations (i), (j) and (t)) and conformational 

disorder takes place accompanied by the appearance of the following 

conformational defects: GG, GTG, GTG’ (observations (j) and (l)), end-gauche 

(observation (u)) and gauche in general (observations (b), (t), (y), (v), (w), and (x)). 

In practice we find characteristic signals of liquid-like structures. 

With the aid of the selectively deuterated species 2D2 and 10D2 further details on 

the processes of solid-solid phase transition and melting can be derived. 

First, we consider the 2D2 molecule. Of necessity, the process of packing of traizsplanar 

chains to forni an orthorhonibic crystalline lamella cannot distinguish 

deuterated and undeuterated chain ends; the resulting lamella can be described by 

a random distribution of ‘up’ or ‘down’ chains (Figure 3-55). With such chains we 

can distinguish the conformational behavior of the heads and/or the tails of the 

chains. It becomes thus possible to be sure of the origins of the signals coming from 


The most direct information comes from the frequencies of the CD? rocking 

which have been calculated with normal coordinate calculations and are reported in 

Table 3.1. 

The infrared spectrum of 2D2 in the 670-600 cm-’ range shows only a band 

near 620 cm-’ below To (Figure 3-56). The band is not split, as mentioned above, 

because the precise phase coupling is lost by the random distribution of ‘up’ and 

‘down’ chains. At the solid-solid transition a very weak additional band appears 

near 655 cm-’ (Figure 3-56) indicating the formation of gauche conformations at 

the deuterated chain ends, thus reiterating what has been already learned from C19. 

From a quantitative measure of the absorption intensities of the these modes it 

transpires that approximately 25% of chain ends are conformationally distorted. 

The infrared spectrum of 10D2 provides further information (Figure 3-57]. For 

the orthorhonibic phase the line near 620 c1n-l is split because in this case the phase 

coupling is achieved since translational symmetry is obtained within the crystal. At

30.5 'C 

Figure 3-56. Temperature dependence of the infrared 

\..:-A 

3, 6.C spectrum of n-nonadecane-2D2 through orthorhombic 

+ c( + phase transitions. Notice that (i) in the a phase a 

small number of deuterated chain heads has taken up the 

L.:-32Pc 

TG conformation which increases dramatically in the 

I " " ' J 

melt, (ii) in the melt a certain number of deuterated chains 

700 660 620 580 

v cm-1 do not tilt their heads but still keep the TT conformation. 

T, the doublets disappear since chains move and become uncorrelated (isolated). 

Since no absorption is observed near 655 and 677 cn-' it means that in the middle 

no conformational distorsions are generated in the a phase. The liquid-like structure 

is, however, observed at the melting. 


The final model emerging from this part of the study is the following: starting from 

a regular and ordered orthorhombic structure in which all tmns-planar chains are

3.19 A Worked E.vnn~yle 193 

Figure 3-57. Temperature dependence of the infrared 

spectrum of n-nonadecane-1 OD: through orthorhonibic 

+ c( phase transitions. Notice that (i) the orthorhombic 

crystal disappears at 22.4 "C, (ii) in the c( phase at the center 

of the lamella no confomiational kinks are formed 

while they certainly appear in the melt (33.2"C) Combination 

of the data of this figure and of Figure 3-56 are direct 

evidence of 'surface melting'. 

arranged in a crystal with a setting angle of 42" at T, the lattice expands anisotropically, 

interlamellar forces weaken and a conformational disordering takes 

place at the surface of the lamella. Such disordering is generated by the conformational 

deformation of approximately 25% of the chains which tilt their head with 

gauche conformation either in position 2-3 or 3-4. The inside of the lamella is not 

perturbed and all chains are still trans-planar (Figure 3-58). 

The next question which remains to be answered is whether surface disordering is 

generated by the longitudinal motion of the chains. The following experiment has 

been made [155].

I94 

3 k'ibrcitionul Specfra CIS a Probe of Structural Order 

a 

ORTH 

a 

\N\ 3 ~ 

y.\\\ \ \ ' 

b 

Figure 3-58. Sketch of the mechanism of phase transition of n-nonadecaiie based on detailed 

spectroscopic evidence. (a) Schematic overall view; (b) details on a molecular level. 

Let us take two samples of crystalline orthorhombic n-alkanes with an odd 

number of carbon atoms. They show properties identical to those observed for C19. 

Let us make by suitable methods a mechanical mixture say of C21 and C23. The 

infrared spectra of the mechanical mixtures at low temperature are just the superposition 

of the spectra of the two independent components. Let us focus on the 

crystal field splitting, say of the doublets near 940 cn-' (C23) and 925 cin- ' (C21 j 

(Figure 3-59). If the mechanical mixture is brought near the solid-solid phase transition 

the infrared spectrum evolves with time. Time- and temperature-depeiideiit 

spectra show that the doublets become singlets thus showing that chains have 

moved (Figure 3-59). The final product is a mixed crystal of C21 and C23 as also 

proved by parallel DSC studies. These studies were first reported by Ungar and 

Keller [ 1561 and were re-examined and interpreted as reported in [ 1551.

3.19 A Worked E.vnrqde 195 

009.1 

0084 

0 

n 

b 0074 

$00641 

00541 I I 1 I I I 

960 950 94C 930 92C 910 900 

&a ve 1 eng t n (c m ’’ 1 

Figure 3-59. Time-dependent infrared spectruni in the 960-900 cm-’ range showing the Pi5 component 

of the CHz rocking progression band of n-CrlH44 and the P17 component of n-Cr3HdZ. 

Evolution with annealing time. Annealing temperature 30T, A(t = 0); B (t = I hr); C (t = 2 h); D 

(t = 4 h); E (t = 6 h). At t = 0 the sample consists of a mechanical mixture of the two materials. At 

t = 0 both solids are crystalline orthorhombic as shown by the correlation field splitting doublet. By 

increasing the time of annealing both doublets coalesce into a singlet indicating that chains have 

migrated and formed a mixed crystal. 

Figure 3-60 shows the kinds of kinetic studies of the mixing process which have 

been made. 


In the a phase chains move longitudinally and are ‘squeezed out of the lamella’; the 

onset of such motion occurs near the solid-solid transition temperature. 

This is in agreement with recent measurements of quasi-elastic neutron scattering 

by Gullaume et al. on C19 at 300°K who determined the existence of translational 

longitudinal diffusion with discrete steps of one or two CH2 units with a measured 

diffusion coefficient D = 6 x lop6 cm’s-l [157]. Statistical thermodynamics calculations 

give theoretical support to this kind of phenomenon [ 1581. These conclusions 

find a striking support by the calculations of molecular dynamics by Klein [159] 

which describe precisely all the details that we have found from spectroscopy. 

Moreover, recent experiments of temperature-dependent scanning tunnelling microscopy 

by Rabe et al. [160] give just ‘photographic’ support at the molecular level of 

the process as described by spectroscopy, namely: (i) first the chains organized in 

an ordered lattice; (ii) then the onset of longitudinal diffusion which generates

196 3 IG!wtioiinl Specfrri us N Probe of Sfiw-tiird Order 

tan,, t Hours) 

Figure 3-60. 1 : 1 molar mixture 

of n-CzJHqg and n-C.5HF2. Kinetics 

of the process of diffusion 

at four different temperatures. 

Data from calorimetric 

measurements 

conformational deformations at chain ends; (iii) diffusion which generates the 

interpenetration of chains; and (iv) the full collapse towards a liquid-like structure. 


As a consequence of such longitudinal motion, if crystallites of n-alkanes are suitably 

brought together thermally activated transport of matter can take place and 

diffusion of molecules from one crystallite to another can occur. The kinetic of 

diffusion is function of temperature, length of the chain, contacts between crystallites 

(ie., pressure put on the sample), etc. 

We can proceed further and try to understand the mechanism at the molecular 

level which drives the chains outside their own lamella by a thermally activated 

process. 

We refer the reader to [ 1031, [ 1551 and [ 1451 for a first analysis of the various data. 

We simply mention the line of thought followed by various authors on this rather 

complex problem. 

Several theoretical models have been proposed in an attempt to account for the

3.19 A Worked Example 197 

existence of the c( phase and for the longitudinal motion. First, the general idea 

was that in the a phase chains rotate rigidly [152-1541. Next, models of rotational 

diffusion coupled with longitudinal translations of one CH2 were proposed based 

on NMR data [ 161, 1621. Elastic neutron-scattering experiments introduced the 

idea of diffusional cooperative rotations or of independent translational motions 

with jump length of one CH;! unit followed by 180" rotational jump [163]. A 

translation of two CH? units was proposed [164] in agreement with calculations 

by McCallough [165]. Most of the models proposed consider that in the solid state 

chains perform large-amplitude collective rototranslational motions. McCullough 

has presented an interesting model which shows by calculations that chain can find 

a favorable energy path if they librate rigidly with large amplitude librations and 

translate along the chain axis. 

The knowledge at the molecular level of the mobility of chain molecules in the 

solid state is of basic importance for the understanding of the general processes of 

annealing of chain molecules. Attempts have been made to derive information using 

band shapes in the Ranian spectra, especially focusing on the CH-stretching vibrations 

(Section 3.18). 

As already stated at the end of Section 3.18, we have measured the temperaturedependent 

band widths and the frequency of the CH2 d- mode for several crystalline 

n-alkanes and polyethylene and found no sizable broadening or upward frequency 

shifts until they approach the melting point. All the theoretical models that 

imply large amplitude librotorsional motions are not supported by the spectroscopic 

criteria discussed in Section 3.18. We found evidence of large-amplitude 

libro-torsional oscillations only when the n-alkane chains are included as clathrates 

in various systems. 


Based on the analysis of teniperature-dependent Ranian band shape and frequency 

shift of the Raman active antisymmetric CH1-stretching modes (Section 3.18), 

contrary to the case of urea clathrates, in the solid crystalline state (including the CI 

phase), n-alkane chains do not perform large-amplitude libro-torsional motions. 

We must search for alternative molecular models which may account for longitudinal 

diffusion with transport of matter without implying large amplitude librotorsional 

oscillations. A conceptually and, at least, intellectually exciting model 

which may help in this search has been presented by Mansfield and Boyd [166] who 

attempted to account for molecular relaxation phenomena. We label this model as 

Utah-twist. The main message by these authors is that calculations suggest that an 

orthorhombic lattice of n-alkanes can host a chain distorted by a gentle and long 

twist. The long twist is achieved with a uniform succession of small torsions about 

the C-C bonds which is extended over approximately 12-15 CH2 units. At the end 

of the gentle twist the chain is flipped by 1800 with respect to the head of the twist. 

Such a long twist (which we label as 'twiston') can propagate along the polymethylene 

chain without finding a barrier to its propagation. Based on the aton-atom 

potential chosen by the two above authors the activation energy for generating a

198 3 JJiber-ulioiiiil Spectra as a Probe oj Structural Ordeer- 

Figure 3-61. Projection onto the ab face of the structuic 

I 1 

of the orthorhombic cell of polyethylene (two molecules 

per unit cell, setting angle o = 42"). 

twiston is estimated to be -10 kcal/mol. The twiston model implies for a polymer: 

(i) the initial formation of a disordered region near the surface of the crystal; (ii) the 

propagation of the twiston through the crystal along the chain axis; and (iii) the 

disappearance of the twiston at the opposite surface leaving the chain rotated by 

180" and translated by one CH2 unit. 

The idea is formally very appealing since, from the spectroscopic viewpoint, 

it implies a collective mobility of the polymethylene chain which accounts for the 

observed longitudinal mobility, surface melting, or disordering and transport of 

matter without requiring large-amplitude libro-torsional motions which are not 

detected by vibrational spectroscopy, as previously discussed. 

The twiston model has been further expaiided by Mansfield [I671 and later by 

Skinner and Wolynes [168]. We shortly mention the physics behind since it may be 

applied to other cases in polymer physics. The whole problem of polynielhylene 

chains is reviewed in [169]. 

Let us take the structure of the orthorhombic unit cell of polyethylene (Figure 

3-61) and consider the CH2 units as point masses of mass 117 at a distance ci from the 

axis of the molecule. The units can rotate by an angle 0 about the molecular axis. 

When they rotate they are subject to the elastic response C by the torsions between 

two adjacent (CH?) units. From Figure 3-61 it is apparent that when the chain 

rotates as a rigid body about its axis in an orthorhombic lattice it probes a twofold 

periodic potential V(0) from the crystal field. 

The Hamiltonian proposed by Mansfield has the following form 

In Eq. (3-61) the first term represents the contribution from the crystal potential, 

the second term the kinetic energy, and the third is the quadratic elastic contribution. 

In a gentle twist it can be assumed that Oi do not change much from one unit 

to the other. According to Mansfield one can make the continuous approximation 

and reach a solution of the type: 

@(x - xo - vt) = 4 arctan exp [ k y(x - xo - vt)] (3-62)

3.1 9 A Worked Example 199 

The soliton solution gives the values of Oi which describes a twist in the chain 

which, in the absence of friction, propagates keeping the same shape and velocity v. 

The Utah-twist (twiston) is then approximated to a solitonic wave whose velocity 

of propagation is estimated by Mansfield (using some molecular parameters) to be 

- 10j cm s-I. 

The final concept which emerges is that in the orthorhonibic crystal one can 

conceive the existence of conforniational collective excitations thermally generated 

which extend over approximately 20 CH2 units and move along the chain with a 

certain velocity. 

The main problem is to prove the existence of such mobile twistons. The idea of 

using spectroscopy implies the fact that we should be able to detect somewhere in 

the vibrational spectrum absorption and/or scattering associated to vibrations 

localized on the twiston. Calculations have been made (i) of the density of vibrational 

states of a long polymethylene chain containing a random distribution of 

twistons and (ii) of the infrared spectrum (frequencies and intensities) of the molecule 

C19 containing a twiston. A comparison has been made with the infrared 

spectrum of C19 in the all-frans structure [169] (Figure 3-62). 

in these calculations the twiston is assumed not to be mobile, but pinned at a 

given site of the chain. The fact that the twiston is mobile along the chain in principle 

does not deny the possibility, under certain conditions, to observe its vibration 

in the infrared/Raman spectra. The conditions are the following: 

(a) Since the band width is inversely proportional to the vibrational lifetimes of 

each normal vibration, if the velocity of the twiston is small the vibrational 

lifetime is long enough to allow the formation of its own vibrational mode and 

can give rise to an observable IR or Raman band. if the twiston moves too fast, 

the vibrational lifetimes are too short and the band broadens and flattens in 

a broad continuum. Simple calculations show that if a band of the twiston has 

to be observed with Ai~,p M 20 cin-' it sets an upper limit to the velocity of the 

twiston of = lo4 cm-' which is not too different from the velocity estimated by 

Mansfield. 

(b) The concentration of twistons has to be large enough to generate a detectable 

absorption in IR or scattering in the Raman. The concentration of twistons 

depends on their energy of formation which has to be low if twistons are to 

be seen spectroscopically. The energies calculated for the Utah-twist [ 1661 are 

much too high and would make the twiston undetectable by spectroscopic 

techniques. 

The results of the theoretical calculations of the spectra of twistons and a few 

attempts to catch spectroscopic signals possibly associated to twistons are reported 

in [169]. The evidence collected so far is neither compelling nor completely negative. 

The model needs to be supported or dismissed by other physical techniques. 

The model of twistons which move as 'solitary waves' was also proposed by 

Dwey-Aharon et al. [ 1701 in the case of polyvinyljdene fluoride (CHz-CF?),, which 

upon poling transforms from the helical structure (TGTG') (labeled either as p 

form or form 11) into a planar all-tram one (a, or form I). Form I is technologically

200 3 Vibrutional Spectra (IS CI Probe of Str'zictural Order 

Figure 3-62. Searching for twistons: section of the calculated spectruin of ti-nonadecane (-) in the 

all-trrrns planar conformation; (- - -) with the conformational distortion like a soliton according to 

Mansfield and Boyd. Both frequency and intensity were calculated.


very important, since having a strong dipole perpendicular to the chain axis shows 

remarkable pyro and piezoelectric properties widely used in modern technology. 

According to Dwey-Aharon when the conformationally disordered material existing 

in form I1 is placed in a strong electric field (poling) conformational twistons are 

generated which propagate and yield form I. The spectroscopy of polyvynilidene 

fluoride and the search for twistons has been treated in [171]. 

Finally, the concept of twiston has been again introduced in a spectroscopic work 

to account for the set of peculiar experimental data collected in the study of the 

phase transition in ethylene-tetrafluoroethylene alternating copolymers [ 1721. 

Acknowledgments 

The contributions to knowledge presented in this chapter are the result of an 

enthusiastic collaboration of many researchers and students who worked in our 

group at the Politecnico of Milano. Our warmest thanks are extended to Dr. M. 

Gussoni, who laid the background for the development of this science. 


[l] H. H. Nielsen, Ha?idbzrclz der Plzvsik, Springer, Berlin, Vol. 37/1, 1959. 

[2] G. Herzberg, Infrared and Ranian Spectra of’ Polyatonzic Molecules, Van Nostrand, Princeton, 

USA, 1963. 

[3] E. B. Wilson, J. C. Decius and P. C. Cross, Moleculur I/ibrutioiis, McGraw Hill, New York, 

1955; V. Volkenstein, M. A. Eliashevich and B. Stepanov, Kolebuni.vu Molecul 11, Moscow, 

1949; S. Califano, Vibrational Stutes, Wiley, New York, USA, 1976. 

[4] For a general discussion see: W. Person and G. Zerbi (Eds.), Vibrotional61teizsities in Infrared 

and Ranm Spectroscopy, Elsevier, Amsterdam, The Netherlands, 1984; L. A. Gribov, Intensity 

Theory for IIfrared Spectra of Polyatoniic Molecarles, Consulting Bureau, New York, 

USA, 1964. 

[5] M. Gussoni in Aduances in hfiared and Ru~nan Spectroscopy, (Eds. R. J. H. Clark and R. H. 

Hester) Heyden, London, England, 1979, vol. 6 p. 96. 

[6] M. Gussoni, C. Castiglioni and G. Zerbi, J. Mol. Strzrct, 1989 198, 475 M. Gussoni, C. 

Castiglioni, M. N. Ramos, M. Rui and G. Zerbi, J. Mol. Strirct. 1990 224, 225 C. Decius, J. 

Mol. Spectry, 57, 384 1975; A. J. Van Straten and W. Smit, .I Mol. Spectry., 1976 62, 297 M. 

Gussoni, C. Castiglioni and G. Zerbi, J. Phys. Clienz. 1984 88, 600. 

[7] S. Trevino and H. Boutin, J. Macronzol. Sci., 1967 Al, 723. 

[8] E. Curtis, Thesis, University of Minnesota, 1958. 

[9] J. H. Schachtschneider and R. G. Snyder, Spectrocki/n. Acro, 1963 19, 17. 

[lo] The most known set of original computing programs which has later originated many other 

programs for dynamical and force constant calculations is that written by J. H. Schachtschneider, 

Shell Rep. 57/65. 

[ 111 See, for instance, J. L. Duncan, Force Curisfunt Calcirlcrtions ik hfokecirles, Specialist Report, 

Chemical Society, London, 1975, vol. 23, p. 104: G, Fogardsi and P. Pulay, in I’ihr.trtiorm/ 

Spectru cind Strircfure, (Ed. J. R. Durig) Elsevier, Amsterdam, 1985, Vol. 14, p. 125.

202 3 Vibrutioncil Spectra CIS a Probe of Striicturnl Odcr 

[I21 M. Gussoni, C. Castiglioni and G. Zerbi, J. Phys. Clzenr., 1984 88, 600. 

[I31 M. Gussoni, M. Ramos, C. Castiglioni and G. Zerbi, J. Mol. Struct., 1988 174, 47. 

[ 141 P. Pulay, X. Zhou and G. Fogarasi, in Rewit E~~~~ei~inientrd and Co/iif~/ittrtion[/I Atluirrrc,cc. i~7 

Molecular Spectroscopy, (Ed. R: Fausto), NATO AS1 Series C, Vol. 406, p. 99; W. B. Person. 

K. Szczesniak and J. E. Del Bene, ibidem, p. 141. 

[I51 K. Palmo, L.-0. Pietila and S. Krimm, Conzpurers Clienz., 1991 f>, 249; K. Palmd, L.-0. 

Pietila and S. Krirniii. .I. Comp. Cl~eni., 1991 12, 385; K. Palmo, N. G. Mirkin, L.-0. Pieti13 

and S. Krimni, M~momulecziles, 1993 26, 6831, K. Palmo, L.-0. Pietila and S. Krimm, 

computer.^ Clieni., 1993 17, 67; K. Palmo, L.-0. Pietila and S. Krimni, J. Comp. Clwnr.. 1992 

13, 1142. 

[16] G. Zerbi, Vihrcitional Spectra qf High Pol~wers, (Ed. E. G. Brame), Applied Spectroscopy 

Reviews. Dekker, New York, USA, 1963, vol. 2, p. 193; H. Tadokoro; Srrirc.trrw uf'Crj~stir1- 

lii7e Pol~~ers, Wiley, New York. 1979. 

[I71 P. C. Painter, M. M. Coleman and J. L. Koenig. Tlie Tlreury of t'ihutiond S)c,c.fro.cc'opj7' ctr7d 

its Applications to Polymeric Materials, Wiley, New York, USA, 1982. 

[I81 For a general discussion of the dynamics of simple and simplified chain molecules see: R. 

Zbinden, Znjiaved Spectroscopy of High Poljvners, Academic, New York, USA. 1964. 

[ 191 G. Zerbi, Vibrational Sliectroscopy of Very Largt, molerirles, in Advances in Infrared and 

Rainan Spectroscopy (Eds. R. J. H. Clark and R. H. Hesterj Wiley-Heyden, New York, 

USA, vol. 11, p. 301. 

[20] J. C. Decius and R. M. Hexter, Moleciilar I4bratians in Crystcils, McGraw-Hill, New Yorb, 

1977; S. Califano, V. Schettino and N. Neto, Lattice Dynnmics of i2.loleciilrrr Cr~.sfa/s. Lectut-e 

Notes in Chemistry, Springer, Berlin, 1981. 

1211 M. Gussoni, G. Dellepiane and S. Abbate, J. Mol. Spectry., 1975 57, 321. 

[22] For a discussion on the problem of redundancies see: M. Gussoni and G. Zerbi, Accademia 

Nazionale dei Lincei, Rendiconti serie VZZI 1966 40, 843; M . Gussoni and G. Zerbi, ibid. 

p. 1032; M. Gussoni and G. Zerbi, Clieni. PIiys. Lett., 1968 2, 145: I. M. Milk Clwni. 

Pliys. Lett. 1969 3, 267; G. Zerbi, in Modern Trends in I'ibrationul Spectroscop?', (Eds. 

A. J. Barnes and W. J. Orville-Thomas) Elsevier, Amsterdam. 1977. p. 261: M. V. Volkenstein, 

L. A. Gribov, M. A. Eliashevich and B. Stepanov, Kolehzniyci Molt~kul, Moskow, 1972, 

p. 203. 

[23] N. B. Colthup, L. H. Daly and S. E. Wiberley, Introdwtioii to Znfifi.arei/ and Xtnncirr S,ut*c./roscopy 

Academic, New York, 1975. 

[24] R. N. Jones and C. Sandorfy, Clienrical Applicntions uf Spec~troscopj~, in Teclmiques of 

Organic Chemistry, (Ed. A. Weissenberger) Interscience, New York, USA, 1956, vol. IX. 

[25] L. Bellaniy, The hfr~irt~l Slwctriz of COn711kY Moleciiles, Wiley, New Yo]-k, USA, 195s. 

1261 K. Nakamoto, /nj?ared Spectru of Polyatornic Molecules, Wiley, New York, 1963. 

[27] See the series of papers such as, for example, T. Shinianouchi, Tables qf Moleczrltrr Frcquencies, 

Part 7, J. Physicul and Cheniical Reference Dota, 1973 2, 225. 

[28] For the case of a polymer see, for instance, G. Masetti, S. Abbate, M. Gussoni and G. Zerbi. 

J. Clzern. Pliys. 1980 73, 4671. 

[29] M. Gussoni and G. Zerbi. J. &lo/. Spectry., 1968 26, 485. 

[30] G. Dellepiane, M. Gussoni and G. Zerbi, J, Chern. Plqis. 1970 53, 3450; M. Gussoni, G. 

Dellepiane and G. Zerbi, in Molecular Structures and I'lbrations (Ed. S. J. Cyvin) Elsevier, 

Amsterdam, 1972, p. 101. 

[31] P. Torkington, J. Cliern. Phylys., 1949 17, 347. 

[32] Y. Morino and K. Kuchitsu, J. cheni. Plz-vs., 1952 20, 1809. 

[33] J. H. Schachtschneider and R. G. Snyder, Spectrochini. Actu 1963, 19, 17. 

[34] R. G. Snyder and J. H. Schachtschneider, Sj7ectrocliim. Acta. 1963 19, 117. 

[35] J. H. Schachtschneider and R. G. Snyder, Spectrocliim. ACIU, 1965 21, 1527. 

[36] R. G. Snyder, J. Clwrn. Phys., 1967 47, 1316. 

1371 R. G. Snyder and G. Zerbi, Spectroclrini. Acta, 1967 23 A, 391. 

[38] W. T. King and B. L. Crawford, J. Mol. Spectry., 1960 5, 429. 

[39] J. Overend and J. R. Scherer, Spc.c.trocAini. Actrt, 1960 16, 773. 

[40] Handbook of Conducting Polyniers, (Ed. T. A. Skotheini) Dekker? New York. 1986. Vols. I 

and 2.


[41] M. Gussoni, C. Castiglioni and G. Zerbi. in SjJNti'o.SCOpy qf rlduarrced Mciterinls (Eds. R. J. 

H. Clark and R. E. Hester) Wiley, New York, 1991, p. 251. 

[42j G. Zerbi. M. Gussoni and G. Castiglioni. in Coi+tyaterl Po/ym,r.s, (Eds. J. L. Bredas and R. 

Silbey) Kluwer, Netherland, 1991, p. 435. 

[43] C. Castiglioni, M. Del Zoppo and G. Zerbi, .I R~mm Spectry.. 1993 24. 485. 

1441 We consider as almost unique example of the existence of long range interactions in covalent 

systems the superlinear increase with conjugation length of the second order hyperpolarizability 

y in long chain polyenes as reported by: I. D. W. Samuel. 1. Ledoux, C. Dhenaut, J. 

Zyss, H. M. Fox; R. R. Scrock and R. J. Silbey, Science, 1994 256, 1070. 

[45] G. Natta and P. Corradini, Nuotjo Cinzeiito, Siqpl., 1960 15, 40. 

[46] B. Wunderlich. hlcicrortzoleciilur Physics, Vols I and 2, Academic, New York, 1976. 

(471 See, for instance, P. De Sanctis, E. Giglio, A. Liquori and A. Ripanionti, J. Po/yrn. Sci., 1963 

Al; 1383. 

[48J P. Flory. Statisticcil Mechanics of Chciirz Mokecdes, Interscience, New York, 1969. 

[49] M. Tasumi and T. Shimanouchi, J. Cheiii. Phys, 1965 43, 1245; M. Tasumi and S. Krimm, J. 

Cherii. Phys., 1967 46, 755. 

[50] L. Piseii and G. Zerbi, J. Mol. Spctry., 1968 26, 254. 

[51] A. A. Maradudin, E. W. Montroll and G. H. Weiss, Solid State Pliys., Si/pp/., 1963 3, 1. 

[52J L. Brillouin, HJaoe Propagation in Periodic Strrrctures, Dover, New York, 1953. 

[53] G. Zerbi, in Advances in Infrared and Raman Sprcfroscop-v, (Eds. R. J. H. Clark and R. E. 

Hester) Wiley, 1984, p. 301. 

[S4] G. Zerbi and M. Sacchi, Mucromolecrilrs, 1973 6, 692. 

[55] P. Bosi, R. Tubino and G. Zerbi, J. Clzeriz. Phys.? 1973 59, 4578. 

[56] L. Piseri and G. Zerbi, J. Chern. Phys., 1968 4S, 3561; G. Zerbi and L. Piseri, J. Chew?. Phys., 

1968 49. 3840. 

1571 For Calculations of dispersion curves of Hydrogen bonded systems in one dimension: R. 

Tubino and G. Zerbi, J. Chem. Phj..s., 1969 51, 4509; A. B. Dempster and G. Zerbi, J. Cliem 

Pliys., 1971 54, 3600; R. Tubino and G. Zerbi, .I. Cheni. Phys. 1970 53, 1428. 

[58] G. Zerbi, C. Castiglioni and M. Del Zoppo. in Conducting Polymer: The Oligorner Approuch 

(K. Mullen and G. Wegner Eds. ) VCH, Keinheim (1998). 

[59] C. Castiglioni, M, Gussoni and G. Zerbi, J. Cheni. Phis., 1991 95, 7144. 

[60J R. Tubino, L. Piseri and G. Zerbi. J. Chenr. Phys., 1972 56, 1022. 

1611 P. W. Higgs, Proc. Roy. Soc., London, 1953 11220, 472. 

[62] T. Miyazawa, J. Clieni. Phys., 1961 35, 693. 

[63] G. Masetti. S, Abbate, M. Gussoni and G. Zerbi. J. C/zenz. P/z,vs., 1980 73, 4671. 

[64j B. Orel, R. Tubino and G. Zerbi, Mo1ec~rlrrr Plz~x, 1975 30. 37. 

[65] G. Zerbi, F. Ciampelli and V. Zaniboni, J. PO/.VJJ. Sci., Ptrrt C, 1964 141. 

[66] G. Zerbi, Probiiig the Real Struetiire of Chciin Molecules by Vibrational Spectroscopy? American 

Chemical Society, Specialist Reports. 1983 203, 487. 

I671 T. Miyazawa and T. Kitagawa, J. Polwn. Sci., 1965 B2, 395; S. Enomoto and M. Asahina, 

Pol.vm. Sci., 1964 A2, 3523. 

[68] R. G. Snyder, J. Mol. Spectry., 1961 7, 116; R. G. Snyder, in Methods of E.xperinieiitd 

Physics (Ed. R. A. Fava). Academic, New York, Vol. 16, part A, p. 73. 

[69] S. Krimm, Achmce.r 117 Poljvrier Scieme, 1960 2, 51. 

[70] D. Dows, in Plivsics mid Chrrnisfry of Orgarlic Solid State, (Eds. 1. Fox, M. M. Labes and 

A. Weissberger) Wiley-Interscience, New York, 1963, Vol. 1; W. Vedder and D. F. Hornig, 

Aehmzces in Spec~troscoly, 1962 2, 189; M. I. Bank and S. Krimm, J. Appl. Pliy,~., 1969 10, 

248; S. Kriinin and T. C. Cheam. Fmwluy Disc. Chem. Soc. 1979 68, 243. 

[71J V. Zamboni and G. Zerbi, J. Po!,?n. Sci., 1964 C7, 153: G. Zerbi and G. Masetti, .J. Mol. 

Spectry, 1967 22, 284; G. Zerbi and J. P. Hendra. J. Mol. Spectry., 1968 27, 17. 

[72] For Polytetrafluoroethylene: F. J. Boerio and J. L. Koenig, J. them. Phj.s., 1971 54, 3667 

and C. J. Peacock, P. J. Hendra, H. A. Willis and M. E. A. Cudby, J. C/ic/i/. Soc. A. 1970 

2943. 

[73] For Isotactic Polypropylene, ref. 66, p. 501, and G. Masetti, unpublished work. 

1741 M. Peraldo, Giizz. Chi~tr. Ztal, 1959 89, 798; M. Peraldo and M. Farina, Chi///, Did iMilaii), 

1960 4-3, 1349.

204 3 Vibrational Spectra [is (i Prohe of Structiirul Ordeer, 

[75] G. Zerbi, M. Gussoni and F. Ciampelli, Spectr~ochini. Actci, 1967 23, 301. 

[76] G. Masetti, F. Cabassi and G. Zerbi, Polymer, 1980 21, 143. 

[77] G. Natta and P. Corradini, Atti Accad. Nazle. Lincei, Mein. Ckrisse Sci. Fis. Mu/. Nor., 1955 

4, 73: G. Natta, P. Corradini and M. Cesari, ibici.. 1960 21, 365. 

[78] G. Zerbi, Sperialist Reports 1983, 203, 501. 

1791 H. JeKreys and B. S. Jeffreys, Method3 of Mertliemuti~d Physics, Cambridge, New York, 

1950, p. 40; P. Dean and J. L. Martin, PIX Roy. Soc. (London), 1960 A259, 309; P. Dean, 

Rev. Mod. Pliys., 1972 44, 127. 

[80] A. Rubcic and G. Zerbi, il/lnrrorrio/ecirles. 1974 7, 754. 

[81] W. Myers, J. L. Donovan and J. S. King, J. Cheni. Phys., 1965 42, 4299; G. C. Summerfield, 

J. Chem. Pliys., 1965 43, 1079. 

[82] A. Sjolander, Arkio. Fysi/c, 1958 14, 315. 

[83] L. V. Tarasov, Soviet P1iys.-Solid State, English Transl., 1961 3, 1039. 

[84] S. Trevino and H. Boutin, J. Macrornol. Sci., 1967 Al, 723. 

[Sj] W. Myers, G. C. Summerfield and J. S. King, J. Chem. Phys., 1986 44, 184. 

[86] S. F. Parker J. Chem. Soc., Faradq Tram> 1996 92, 1941. 

[87] Faraday Discussions of the Chemical Society Orpnixifion of Miicronio/ec,/iles in the Condensed 

Phase, No. 68, Roy. Soc. Chem. 1979. 

[88] G. Zerbi in Phonons, (M. A. Nusimovici Ed.), Proceedings of the international conference, 

Rennes, France 1971, p. 248. 

[89] G. Zerbi and G. Cortili, Chern. Comnt, 1965 13, 295; G. Cortili and G. Zerbi, Spectrochini. 

Actrr, 1967 33A, 285. 

1901 A. Rubcic and G. Zerbi, hfncrornolecules, 1974 7, 759. 

[91] M. Tasumi and G. Zerbi, J. Cheni. Phys. 1968 48, 3813. 

[92] B. Fanconi, Ann. Rev. Pliys. Clien?. 1980 31, 265. 

[93] C. Schmid and K. Holzl, J. Polyni. Sci., Pliys. Ed.., 1972 10, 1881; C. Schmid and K. Holzl, J. 

Phys. C, 1973 6, 2401; V. N. Kozyrenko, L. Kumpanenko and V. Mikhailov, J. Polyrn. Sci.. 

Polpi. Pliys. Ed., 1977 IS, 1721. 

[94] G. Zannoni and G. Zerbi, J. Mol. Struct., 1983 1000, 385. See also J. J. Ladik, Qucrntuni 

Theory qfSolids, Plenum, New York, 1987, p. 140. 

[95] G. Zerbi, L. Piseri and F. Cabassi, Mo/. Plzys., 1971 2-3, 241. 

[96] J. H. Wilkinson, Computer J. 1958 I , 90. 

[97] M. Gussoni and G. Zerbi, J. Cheni. Phys., 1974 60, 4862. 

[98] I. F. Chang and S. S. Mitra, Atlo. Phys., 1971 20, 359. 

[99] G. Zerbi and M. Gussoni, Polymer, 1980 21, 1129. 

[ 1001 A. Rubcic and G. Zerbi, Macromolecules, 1974 7, 759. 

[I011 M. Tasumi and G. Zerbi, J. Cheni. Phys., 1968 48, 3813. 

[I021 R. G. Snyder and M. Poore, Macromolecules, 1973 6, 709. 

[103] G. Zerbi, R. Magni, M. Gussoni, K. Holland-Moritz, A, Bigotto and S. Dirlikov, J. Chem 

Phys., 1981 75, 3175. 

[I041 For the works by the school of R. G. Snyder in this field see, for instance, M. Maroncelli, S. 

P. Qi. H. L. Strauss and R. G. Snyder, J. Am. Chem. Soc., 1982 104, 6237. 

[I051 A. Keller, Furaday Disc. Cllem. Soc. 1979 68, 244. 

[lo61 G. Bucci and T. Simonazzi, J. Polynz. Sci., port C. Polymer Symposia, 1964 7, 203. 

[107] G. Zerbi, G. Minoni and A. P. Tulloch, J. Chn. Phys.. 1983 78, 5853. 

[I081 G. Vergottin, G. Fleury and I. Moschetto, Advvrrnces i7 hqiared and Ramari Spectroscopy, 

(Ed. R. J. H. Clark and R. E. Hester) Heyden, London, 1978, vol. 4, p. 195. 

[I091 G. Zerbi, G. Conti, G. Minoni, S. Pison and A. Bigotto, J. Phys. Chem., 1987 91, 2386. 

[I101 G. Minoni and G. Zerbi, J. Phys. Cliem., 1982 96, 4791. 

[ 11 I] M. Del Zoppo and G. Zerbi, Poljmer, 1990 31, 658. 

[I 121 E. Galbiati and G. Zerbi, J. Chem. Phys. 1986 84, 3500. 

[I131 E. Galbiati and G. Zerbi, J. C/iem. Plzys., 1987 87, 3653. 

[114] E. Galbiati, G. Zerbi, E. Benedetti and E. Chiellini, Polymer, 1991 3-3, 1555. 

[ 1151 S. J. Spells, S. J. Organ, A. Keller and G. Zerbi, Polymer. 1987 28, 697. 

[116] H. Kay and B. A. Newman, .4cta Crystollogr. 1968 B24, 615.

3.20 Refeeveiicrs 205 

[I 171 P. Jona, M. Gussoni and G. Zerbi. J. Appl. Phys., 1985 57, 834. 

[118] M. Gussoni, P Jona and G. Zerbi. J. Mol. Struct., Theochem, 1985 119, 329; ibid. p. 347. 

[119] T. Shimanouchi, J. Phys. Chem. Ref: Data, 1973 2, 225. 

[120] G. Zerbi, G. Gallino, N. Del Fanti and L. Baini, Polymer.. 1989 30, 2324. 

[I211 G. Avitnbile, R. Napolitano, R. Pinozzi, K. D. Rouse, W. Thomas and B. T. M. Willis, 

Polvn1. Lett.. 1975 13, 371. 

[I221 S. Abbate, M. Gussoni and G. Zerbi, J. Cheni. Phj~x, 1979 70, 3577. 

[I731 E. Agosti, G. Zerbi and I. M. Ward, Polymer 1992 33, 4219. 

[124] G. Zerbi, G. Conti, G. Minoni, S. Pison and A. Bigotto. J. Phys. Chenz. 1987 91, 2386. 

[125] M. Del Zoppo and G. Zerbi, Polymer, 1990 31, 658. 

[ 1261 D. Chapmann, Thc Structi/re of Lipids, Methuen, London, 1965. 

11271 C. Almirante, G. Minoni and G. Zerbi, J. Phys., Cheni., 1986 YO, 852. 

11281 L. Basini, A. Rafhelli and G. Zerbi, C11ein. Motel-. 1990 2, 679. 

[129] G. Ferrdri and G. Zerbi, J. Rrmiczn Spectroscopy, 1994 25, 713. 

[I301 For a discussion on the structure of PVC see, for instance, F. A. Bovey, F. P. Hood, E. W. 

Anderson and R. L. Kornegay, J. Phys. Clzem. 1967 71, 312; L. Cavalli, G. C. Borsini, G. 

Carraro and G. Confalonieri, J. Po/jm Sci., Part A-1. 1970 8, 801. 

11311 R. F. Schaufele and T. Shimanouchi, J. Chem. Phys. 1967 47, 3605. 

11321 T. Shimanouchi, Piwe and Appl. Chem., 1973 36, 93. 

11331 G. Minoni and G. Zerbi, J. Phys. Chem., 1982 86, 4791. 

11341 G. Minoni and G. Zerbi, J. Pdym. Sci. Polyni. Lett., 1984 22, 533. 

[I351 E. Fermi, Z. Physik, 1931 71, 250. 

[136] J. Lavalley and N. Sheppard, Spectrochim. Acta, Part A. 1972 28, 845. 

[137] S. Wunder, M. Bell and G. Zerbi, J. Chem. Phys., 1986 85, 3827. 

11381 R. G. Snyder, S. L. Hsu and S. Krimm. Spectroclzini. Am, Part A, 1978 31, 395. 

11391 R. G. Snyder and J. R. Scherer, J. Chem. Phys., 1979 71, 3271. 

[140] S. Abbate, G. Zerbi and S. L. Wunder, J. Pkjx Chenz., 1982 86, 3140. 

[I411 S. Abbate, S. L. Wunder and G. Zerbi, J. Phys. C/zeni.. 1984 88, 593. 

11421 L. Ricard, S. Abbate and G. Zerbi. J. Phys. Cheni., 1985 89, 4793. 

[143] G. Zerbi and S. Abbate, Chem. Phys. Lett., 1981 80, 455. 

[144] M. Del Zoppo and G. Zerbi, Polymer, 1991 32, 1581. 

11451 G. Zerbi, P. Roncone, G. Longhi and S. L. Wunder, J. Chem. Phys., 1988 89, 166. 

11461 R. G. Snyder, J. R. Scherer and P. Gaber. Biochinz. Biophjx Acta, 1980 601, 47. 

[147] W. G. Rotschild, Dyirtiniics of'Molecukrr Liqiiitls. Wiley, New York, 19S4. 

[148] B. G. Sumpter, D. W. Noid and B. Wunderlich, J. Chem. Phys. 1990 93, 6875; G. L. Liang, 

D. W. Noid, B. G. Sumpter and B. Wunderlich, Actu Polymer., 1993 44, 319. 

[149] M. Kobayashi, T. Kobayashi, Y. Cho and F. Kaneko, J. Chem. Phys. 1986 84,4636. 

[I501 1. W. Levin, in Adoances in IR arid Rantan Spectroscopy, Wiley, New York. 1984, Vol 11. 

[151] B. P. Gaber and W. Peticolas, Biochim. Biophys. Acta 1977 465, 260. 

11521 A. Muller, Proc. Royrl Soc., London Ser. 1930 A127, 417. 

11531 M. Broadhurst, J. Res. Nut/. Btrr. Stand Sect 1962 A66, 241. 

[ 1541 K. Larsson, Nature (London) 1967 213, 383. 

[155] G. Zerbi, R. Piazza and K. Holland-Moriytz, Polymer, 1982 23, 1921. 

[156] G. Ungar and A. Keller, Colloid Pol~vi. Sci., 1979 257, 90. 

[157] F. Guillaume, C. Sourisseau and A. J. Dianoux, J. Cheni. Phys., 1990 93. 3536. 

[158] P. Jona, B. Bassetti, V. Benza and G. Zerbi, J. Chenz. Phys., 1987 86, 1561. 

11591 M. L. Klein, J. Client Soc. Farada~ Trans., 1992 13, 1701. 

11601 L. Askadskaya and J. P. Rabe, Phys. Reo. Lett., 1992 69, 1395. 

[lhl] For review on model calculations see D. W. McClure, J. C/iern. Phys., 1968 49, 1530. 

[162] H. G. Olf and A. Peterlin, J. Poly. Sci., 1970 A2, 791. 

[ 1631 B. Ewen and D. Richter, J. Cheni. Pliys. 1978 69, 2954. 

[164] J. D. Barnes, J. Chern. Phys., 1973 58. 5193. 

[165] R. L. McCullough, J. Macromol. Sci.. Phys., 1974 9, 97. 

11661 M. Mansfield and R. H. Boyd, J. PO/~YJI. Sci.. Polyni. Pliys. Ed.., 1978 16, 1227. 

11671 M. L. Mansfield, CIIenz. Plijx Lett., 1980 69, 383.

206 3 Vibratiord Spectra cis a Prohr of StrirctLiral Order 

[168] J. L. Skiiiiiei- and P. J. Wolynes, Chm. Phys, 1980 73, 4015. 

[169] G. Zerbi aiid M. Del Zoppo, J. Clzern. SOC. Fur.crday T~nr7s., 1992 88, 1835. 

[ 1701 H. Dwey-Aharon, T. S. Slucking, P. L. Taylor and A. J. Hopfinger, Phys. Rev. Sw. B, 1980 

21. 

11711 G. Borionetti, G. Zaiiiioni and G. Zerbi, J. Mol. Srvircf. 1990 12.1, 425 

[I721 S. Radice, N. Del Fanti and G. Zerbi, Po/ynirr, 1997 35, 2753.

4 Vibrational Spectroscopy of Intact and Doped 

Conjugated Polymers and Their Models 

Y. Firrukawa and M. Tusumi 


Most organic polymers are electrical insulators, a property which distinguishes 

them from metals. However, the development of a new class of organic polymers 

with quasimetallic electrical conductivities has been actively pursued during the 

past 18 years, following the discovery in 1977 of high electrical conductivities 

for doped polyacetylenes [ 1, 21. The novel concept of a conducting organic polymer 

has aroused the interest of a large number of researchers in various areas such as 

chemistry, physics, electrical engineering, inaterial science, etc. Detailed discussions 

in each area can be found in many review articles [3-171. In particular, a new field of 

physics has been opened for the purpose of understanding their electrical properties. 

A goal of basic research on conducting polymers is to understand the metallic 

properties of such polymers. Since conducting organic polymers have conjugated 

7c-electrons in common, chemists intuitively believe that studies on the physical and 

chemical properties of conjugated systems would lead to an elucidation of the 

mechanism of charge transport. On the other hand, new concepts, solitoris [18, 191, 

yolui.oizs [20-221, and bipolaroils [21-231 have been proposed by solid-state physicists 

as elementary excitations in conjugated polymers, in order to explain the 

physical properties of these polymers. They are collectively called self-loculized 

e-xcitutions. These concepts and terminologies are unfamiliar to molecular spectroscopists. 

Thus, it seems desirable to build a bridge between solid-state physicists and 

molecular spectroscopists. 

The structures and properties of conducting polymers have been studied by 

various spectroscopic techniques. Ainong them, vibrational ( Rainaii and infrared) 

spectroscopy is a powerful tool for elucidating the molecular and electronic structures 

of conducting polymers as described in previous reviews [13-151. In spite of 

numerous spectroscopic studies, discussions on polarons, bipolarons, and solitons 

were not based upon reliable evidence from vibrational spectroscopy until very 

recently. Thus, the aims of this review are: (i) to provide a general introduction to 

the concepts of polarons, bipolarons, and solitons from the standpoint of molecular 

spectroscopists; (ii) to describe the methodology of Raman studies on these selflocalized 

excitations; (iii) to review the results of studies on poly( p-phenylene) and 

other polymers; and (iv) to discuss the mechanism of charge transport in conducting 

polymers.

208 4 T/ihmtional Spectroscopy of Ivitact arid Doped Coiijirgnted Polymers 

4.2 Materials 

The conductivity of iodine-doped polyacetylene first reported by Shirakawa et al. 

[I] in 19'7'7 was 30 S cm-'. Since then, the conductivity reported for doped polyacetylene 

has kept increasing, the highest conductivity obtained so far for an iodinedoped 

stretched polyacetylene film [1?] being > 10' S cm-', a value comparable 

with that of copper (6 x 10' S cni-I). 

A film of intact polyacetylene usually shows a conductivity lower than lo-' 

S cm-' . However, the conductivity increases dramatically when the film is exposed 

to oxidizing agents (electron acceptors) such as iodine, AsF5, H'S04, etc. or reducing 

agents (electron donors) such as alkali metals. This process is referred to as duping, 

by analogy with the doping of inorganic semiconductors. The polymers that are not 

doped are referred to as ititact polymers in this review. The main process of doping 

is a redox reaction between the polymer chains and acceptors (or donors). Upon 

doping, an ionic complex consisting of positively (or negatively) charged polymer 

chains and counter ions such as 13-, AsF6-, etc. (or Naf, K+, etc.) is formed. 

Counter anions or cations are generated by reduction of acceptors or oxidation of 

donors, respectively. The use of an acceptor causes the p-type doping, and that of 

a donor the n-type doping. The electrical conductivity can be controlled by the 

content of a dopant. A sharp increase in conductivity is observed when the dopant 

content is < 1 mole% per C2H2 unit. After this sharp increase, the conductivity becomes 

gradually higher with further increase in the dopant content. At low doping 

levels, polyacetylene does not exhibit metallic properties, whereas its conductivity is 

high. When the dopant content is more than about 13 mole% per C2Hz unit, polyacetylene 

shows metallic properties such as a Pauli susceptibility [24, 251, a linear 

temperature dependence of the thermoelectric power [26], and a high reflectivity in 

the infrared region [27], though the temperature dependence of conductivity is not 

like that of a metal. The origin of such properties of heavily doped polyacetylene is 

not yet fully understood. 

Following polyacetylene, a large number of conducting polymers have been 

reported. The chemical structures of typical conducting polymers are depicted in 

Figure 4-1. These conducting polymers have conjugated n-electrons in common, 

(a) (b) (C) (d) 

pq..J-@--q)n 

s \ i n 

(e) 

(f) 

p, // 

aniline 

Figure 4-1. Chemical structures of conducting 

polymers. ia) Trurrs-polyacetylene; 

(b) cis-polyacetylene; (c) polyj p- 

phenylene); (d) polypyrrole; (e) polythiophene: 

if 1 poly( p-phenylenevinylene); 

(g) poly(2,5-thienylenevinylenei; (h) poly- 

(leucoemeraldine base form); (i j 

(h) (0 polyisothianaphthene.

4.3 Georrietvy oj Intuct Po1vniev.r 209 

and all are either insulators or semiconductors in their intact states, and they 

become conductors upon doping. Metallic properties are observed for poly( p- 

phenylene) [28], polythiophene 129-331, polypyrrole [3 11, and polyaniline [34, 351 at 

heavy doping levels, although reported data depend on samples and preparation 

methods. The origin of the metallic states of heavily doped polymers is one of the 

major unresolved problem in the field of conducting polymers. 

4.3 Geometry of Intact Polymers 

The question of whether the C-C bonds in an infinite polyene chain are equal or 

alternate in length has been discussed by quantum chemists since the 1950s [36, 371. 

Various experimental results on oligoenes have shown the existence of alternating 

single and double bonds. The bond alternation in rrnns-polyacetylene has been 

confirmed by X-ray [38] and nutation NMR [39] studies. However, it is difficult to 

determine exact structure parameters for conducting polymers from X-ray diffraction 

studies, because single crystals of the polymers are unavailable. It is then useful 

to examine the geometries of the polymers and model oligomers by the molecular 

orbital (MO) method, especially at nb biitio Hartree-Fock levels or in higher 

approximations. 

Conducting polymers can be divided into two types, degenevnte and nondegenerate, 

according to the structure of the ground-states of intact polymers. The 

total energy curve of a degenerate system in the ground state is shown schematically 

as a function of a structural defoiination coordinate, R, in Figure 4-23, and that 

for a nondegenerate system in Figure 4-21 Let us consider an infinite transpolyacetylene 

chain as a prototype of the degenerate ground-state polymers. There 

are two degrees of freedom in bond lengths: rc-c and YC-C, which correspond, respectively, 

to the lengths of the alternating single and double CC bonds. The coordinate 

for structural deformation is simply expressed by the following equation. 

This coordinate reflects the degree of bond alternation. This coordinate is used in the 

effective-conjugation-coordinate model proposed by Zerbi et al. [15] for the purpose 

Figtire 4-2. Total energy as a function 

of a structural deformation coordinate. 

(a) Trans-polyacetylene 

> 

u 

LT 

(degenerate ground-state system); ib) + 

poly( p-plienylenei (nondegenerate 

ground-state system). 

6 

\ A. /

210 4 Vihvatiotial Specti~oscopy of hztuct ~ nd Doped Corijuyuted Po!vnrers 

of explaining the Raman spectra of intact polymers and doping-induced infrared 

bands. (This model will be mentioned again in Section 4.5.) There are two stable 

structures (A and B in Figure 4-2a) with alternating C-C and C=C bonds, while the 

structure with equal C-C bond lengths (C in Figure 4-2a) is unstable. These two 

structures are identical to each other and have the same total energy; in other 

words, they are degenerate. According to a nutation NMR study [39], the lengths of 

the C-C and C=C bonds in tv~ztzs-polyacetylene are 1.44 and 1.36 A, respectively. 

An ab initio MO calculation at the Hartree-Fock level with the 6-31G basis set for 

a model compound, C22H24, has shown that the C-C and C=C bond lengths of the 

central unit are 1.450 and 1.338 A? respectively [40]. Similar results have been 

obtained from calculations which take into account the effects of electron correlation 

[41]. From these experimental and theoretical results, Ro (equilibrium value of 

R, see Figure 4-2a) is calculated to be 0.06 and 0.079 A, respectively. 

A nondegenerate polymer has no two identical structures in the ground state. Let 

us consider poly( p-phenylene) as an example of nondegenerate polymers. In order 

to express the structural defoimation in this polymer, the following coordinate may 

be chosen, as proposed by Castiglioni et al. [42]. 

where T,S are shown in Figure 4-2b. The total energy of this polymer has only one 

minimum at Ro as shown in Figure 4-2b. If we use the bond lengths, vl = v3 = 

v4 = 1’6 = 1.388 A, 12 = 1’5 = 1.382 A, and v7 = 1.492 A, of the central unit of neutral 

p-terphenyl obtained by an ab initio MO calculation at the Hartree-Fock level 

with the double-zeta quality 3-21G basis set [43], Ro is calculated to be 0.490 A. As 

shown in Figure 4-2b, an increase in R means the structural defoimation toward a 

quinoid structure from a benzenoid structure. 

4.4 Geometric Changes Induced by Doping 

The main process of doping is a charge-transfer reaction between an organic polymer 

and a dopant. When charges are removed from (or added to) a polymer chain 

upon chemical doping, geometric changes occur over several repeating units and 

the charge is localized over this region. Structure parameters such as bond lengths, 

bond angles, etc. are changed in this region. For example, the bond lengths [44] of 

neutral 1.7-terphenyl, its radical anion, and its dianion calculated by the PM3 method 

are shown in Table 4-1. From this table, we can see readily that the bond lengths 

of the neutral and charged species are different from each other. The phenyl and 

phenylene rings in neutral p-terphenyl have benzenoid structures. In the radical 

anion and dianion, an increase in R occurs as the result of shortening of the 1’23, v45, 

and v67 bonds and lengthening of the ~ 1 21-34, , and 1’56 bonds; in other words, geometric 

changes from benzenoid to quinoid occur upon ionization. The geometric

1'45 

4.4 Geometric Chunges I?zditced bv Doping 21 1 

Table 4-1. Bond lengths (in A) calculated by the PM3 method for p-terphenyl, its radical anion, 

and its dianion. 

Species Bond length" Coordinateb 

1'11 123 1'34 745 r56 y67 R 

Neutral 1.391 1.390 1.397 1.469 1.397 1.389 0.507 

Radical anion 1.397 1.385 1.417 1.426 1.423 1.366 0.580 

Dianion 1.400 1.371 1.438 1.388 1.446 1.351 0.640 

Numbering of carbon atoms is shown below. Dihedral angles between neighboring benzene rings 

are 48", O", and 0" for y-terphenyl, its radical anion, and its dianion, respectively. The data are 

taken from [44]. 

The structural deformation coordinate defined in Eq. 4-2 is calculated by using the equation, 

R = 11 fi(4~56 ~ 2167 ~ ). 

changes from the neutral species are larger for the dianion than for the radical 

anion. Since the localized charges, which are accompanied by geometric changes, 

can move along the polymer chain, they are regarded as charge carriers in coiiducting 

polymers. These quasiparticles are classified into polarons, bipolarons, and 

solitons according to their charge and spin. 

4.4.1 Polarons, Bipolarons, and Solitons 

Self-localized excitations and corresponding chemical terminologies are listed in 

Table 4-2. Schematic structures of the self-localized excitations in poly( p- 

phenylene) and trans-polyacetylene are depicted in Figure 4-3. In these illustrations, 

the charge and spin are localized on one carbon atom. In real polymers, however, 

they are considered to be localized over several repeating units with geometric 

changes. 

When an electron is removed from an infinite polymer chain, charge +e and spin 

1/2 are localized over several repeating units with geometric changes. This is a selflocalized 

excitation called a positive polaron (Figure 4-3a, e). Since a positive 

polaron has charge +e and spin 1/2, it is a radical cation in chemical terminology. 

When an electron is added to a polymer chain, a negative polaron, which is a 

radical anion, is formed (Figure 4-3b, f). 

When another electron is removed from a positive polaron, charge +2e is localized 

over several repeating units. The charge +2e is considered to be localized in a 

region narrower than that of a positive polaron. This species is called a positive 

bipolaron (Figure 4-3c). Since a positive bipolaron has charge 1% and no spin, it 

is a dication in chemical terminology. A negative bipolaron corresponding to a 

dianion can also be considered (Figure 4-3d).

2 12 4 T4hvritionnl Spectvoscopy of Intact nnd Doped Conjugrited Poljmiers 

Table 4-2. Self-localized excitations and chemical terminologies. 

Self-localized excitation Chemical term Charge Spin 

positive polaron 

negative polaron 

positive bipolaron 

negative bipolaron 

neutral soliton 

positive soliton 

negative soliton 

radical cation 

radical anion 

dication 

dianion 

neutral radical 

cation 

anion 

+e 

--E 

+ 2e 

- 2e 

0 

+e 

-e 

Figure 4-3. Schematic structures of self-localized excitations in poly( p-phenylenej and tucms-polyacetylene. 

(a) Positive polaron; (b) negative polaron; (c) positive bipolaron; id) negative bipolaron; 

(e) positive polaron; (fj negative polaron: (g) neutral soliton: (11) positive soliton; (i) negative soliton;. 

D, donor; A, acceptor; +, positive charge: -, negative charge; 0, unpaired electron. 

In trans-polyacetylene having a degenerate ground-state structure, soliton excitations 

can be formed between the A and B phases (see Figure 4-2a). Solitons are 

classified into neutral, positive, and negative types according to their charge. A 

neutral soliton has no charge and spin 1/2 (Figure 4-3g). A positive (or a negative) 

soliton has charge +e (or -e) and no spin (Figure 4-3h, i). A neutral soliton, a 

positive soliton, and a negative soliton correspond, respectively, to a neutral radical, 

a cation, and an anion of a linear tvans-oligoene with an odd number of carbon 

atoms, CnHn+2 (where n is an odd number). 

Bond-alternation defects or misfits in polyenes were studied using the Huckel MO 

method in the early 1960s [36, 45, 461, though these are nothing but solitons proposed 

by Su, SchrieHer, and Heeger [lS, 191 about two decades after. These workers 

have studied solitons in tvnns-polyacetylene by using a simple tight-binding Hamil-

4.4 Geometric Chmges Iiid~lrcecl bjj Dopiiiy 21 3 

Figure 4-4. Schematic molecular orbital energy levels. (a) Neutral conjugated system; (b) its radical 

cation; (c) its dication; (d) neutral radical of an odd oligoene. HOMO, highest occupied molecular 

orbital; LUMO, lowest unoccupied molecular orbital; SOMO, singly occupied molecular orbital; 

0. electron. 

tonian with an electron-lattice interaction teiin (SSH model), their model being 

essentially the same as the Huckel approximation. The continuuin versions of the 

SSH niodel are very useful for describing polarons and bipolarons as well as solitons, 

because analytic solutions can be obtained for physical parameters (creation 

energy, electronic absorption, etc.) relating to the excitations [21, 47-49]. 

4.4.2 Electronic States of Poiarons, Bipoiarons, and Solitons 

We will first discuss the electronic states of linear conjugated molecules at the 

Huckel level with electron-lattice interaction [7], as models of self-localized excitations. 

The MO energy levels of these systems are schematically shown in Figure 4-4. 

In a neutral conjugated molecule (e.g., p-terphenyl), the bonding and antibonding 

levels are formed as the result of interaction between n-electron levels (Figure 4-4a). 

In its radical cation, geometric changes lead to an upward shift of the highest 

occupied molecular orbital (HOMO) and a downward shift of the lowest unoccupied 

molecular orbital (LUMO), and one electron is removed from the HOMO 

level (Figure 4-4b); the singly occupied molecular orbital (SOMO) is then formed. 

In its radical anion, the HOMO level is occupied by two electrons and the LUMO 

has one electron. Since geometric changes in the dication, where another electron 

is removed from the HOMO level (Figure 4-4c), are larger than those in the radical 

cation, the HOMO and LUMO levels shift further. In the dianion, each of the 

HOMO and LUMO levels is occupied by two electrons. In the MO energy levels of 

an odd trans-oligoene radical, CllHJ1+2' (where n is an odd number), a nonbonding 

level, which is occupied by one electron, is formed in the center of the band gap as 

shown in Figure 4-4d [36]. In its cation and anion, the nonbonding level IS occupied 

by null and two electrons, respectively.

CB 

I 

B; 

VB 

r-7 

Figure 4-5. Schematic electronic band 

structures. (a) Neutral polymer; (b) posi- 

IWS tive polaron; icj negative polaron; id) 

positive bipolaron; (e; negative bipolaron: 

!f! neutral soliton; (g) positive soliton; (11 

negative soliton. CB, conduction band: 

VB. valence band; 0. electron; arrow, 

(e) (f) (9) (h) electronic transition. 

Next, we discuss the electronic transitions due to polarons, bipolarons, and solitons 

in a polymer chain on the basis of a theoretical study reported by Fesser et al. 

[48] on a continuum electron-phonon-coupled model. The electronic energy levels 

of a neutral infinite polymer and those of polarons, bipolarons, and solitons are 

shown schematically in Figure 4-5. 

In an infinite neutral polymer, interaction between repeating units leads to the 

formation of electronic energy bands. The bonding and antibonding levels constitute, 

respectively, the valence band (VB) and the conduction band (CB) as shown in 

Figure 5-5a. The band gaps (WI in Figure 5-Sa) of most intact polymers are in the 

range between 35000 and 8000 cm-' (4.3 and 1.0 eV, 1 eV = 8066 cm-I). 

For a positive polaron (Figure 5-5b). two localized electronic levels, bonding and 

antibonding, are formed within the band gap. One electron is removed from the 

polaron bonding level. Thus, a positive polaron is expected to have the following 

three intragap transitions: 

01, polaron bonding level +- valence band 

02, polaron antibonding level + polaron bonding level 

(03, polaron antibonding level t valence band, and conduction band t polaron 

bonding level. 

When a negative polaron is formed, one electron is added to the polaron antibonding 

level. In this case also, three transitions are expected (Figure 4-5c). 

The electronic structure of a positive bipolaron is shown in Figure 4-5d. Since the 

geometric changes for a bipolaron are larger than those for a polaron, localized 

electronic levels appearing in the band gap for a bipolaron are farther away from 

the band edges than for a polaron. Two electrons are removed from the bipolaron

4 5 Mer1iodolog.v of Raiii~11 Studies of Polarons, B@olmons, and Solitons 215 

bonding level. Thus, a positive bipolaron is expected to have the following two 

iiitragap transitions: 

01 ’, bipolaron bonding level t valence band 

(oj’, bipolaron antibonding level - valence band 

When a negative bipolaron is formed, two electrons are added to the bipolaron 

antibonding level (Figure 4-5e) and thus two transitions are again expected. The 

intensities of the electronic transitions for polarons and bipolarons will be discussed 

in Section 4.7.1.2. 

When a soliton is formed, a nonbonding electronic level is formed at the center of 

the band gap (Figure 4-5f-11). For a neutral soliton, the nonbonding electronic level 

is occupied by one electron, while for a positive soliton and a negative soliton, the 

level is occupied by null and two electrons, respectively. Thus, all the neutral, positive, 

and negative solitons are expected to have only one intragap transition, us, as 

shown in Figure 4-5f-h. 

Polarons, bipolarons, and solitons are different from each other as described 

above. It is expected that we can detect these excitations separately by electronic 

absorption, ESR, and vibrational spectroscopies. In particular, geometric changes 

induced by doping can be studied by vibrational (Raman and infrared) spectroscopy. 

4.5 Methodology of Raman Studies of Polarons, 

Bipolarons, and Solitons 

Doping induces strong infrared absorptions that are attributed to the vibrational 

modes arising from charged domains generated by doping 110, 14, 15, 501. In the 

case of trans-polyacetylene, Horovitz et al. [5 1-53] have proposed the aniplitudemode 

model for explaining doping-induced infrared absorptions. According to 

their theory, the normal vibrations that induce oscillations in bond alternation are 

strongly observed in the infrared spectra of doped polyacetylene. Castiglioni et al. 

[ 541 have reformulated the amplitude-mode theory in tenns of the GF matrix 

method [55] used in molecular spectroscopy. In their treatment, they have proposed 

an effective conjugation coordinate (expressed by Eq. 4- 1) which reflects bond 

alternation. They have also explained the doping-induced infrared absorptions 

of other conducting polymers such as polypyrrole, polythiophene, poly( p-phenylenevinylene), 

etc., by using effective conjugation coordinates [50]. However, the 

types of self-localized excitations (polarons, bipolarons, and solitons) existing in 

these doped polymers have not been identified. 

Raman specti-oscopy of conducting polymers has mainly treated the structures of‘ 

intact polymers (geometry of polymer chains, conjugation length, force field, etc.) 

[ 13-15]. Although the Ramaii spectra of doped polymers have been reported 

113-151, the observed Raman bands have not been correlated to the types of self-

2 16 4 Vilmitioizal Spectroscopy cf Intact nnd Doped CorzjLiyuted Poluviner=r 

localized excitations created by doping. Zerbi et al. [15] have drawn a conclusion 

from the effective-conjugation-coordinate model that the Raman bands arising 

from the charged domains are extremely weak and would never be observed. 

However, a new approach to Raiiian identification of the self-localized excitations 

will be described in this review. 

Most of intact polymers show electronic absorptions in the region from ultraviolet 

to visible. Upon doping, new absorptions with several peaks appear in the 

region from visible to infrared. These new absorptions are associated with selflocalized 

excitations created by doping. T~LIS, we can observe vibrational spectra 

arising from the charged domains, by using resonance Raman spectroscopy with 

a wide range of excitation wavelengths from visible to infrared [56, 571. So far, 

although most of the Raman spectra of doped polymers have been measured by 

using visible laser lines for excitation, it is also desirable to use laser lines with 

longer wavelengths. 

A useful method for analyzing the vibrational and electronic spectra of polymers 

is to compare them with the spectra of oligomers. In chemical terminology, 

polarons, bipolarons, and charged solitons are nothing but radical ions, divalent 

ions, and ions of oligomers, respectively (see Section 4.4.1). If these charged 

oligomers have geometries similar to those of self-localized excitations, these compounds 

would give rise to the electronic and Raman spectra similar to those of the 

self-localized excitations. We can identify self-localized excitations on the basis of 

the spectroscopic data for charged oligomers (charged oligomer approach) [56, 571. 

4.6 Near-Infrared (NIR) Raman Spectrometry 

Since the 1970s, Raman spectrometry has been using visible laser lines as the most 

common excitation sources. It was generally believed that NIR laser lines were not 

suitable for obtaining high-quality Raman spectra, though since Hirshfeld and 

Chase [58] and Fujiwara et al. [59] reported successful NIR Raman spectrometry 

measurements in 1986, this technique has advanced dramatically and NIR Raman 

spectrophotonieters are now commercially available. 

We will describe two NIR Raman spectrophotonieters, one with a diffraction 

grating, and the other with an interferometer. A Ti : sapphire laser (Coherent Radiation 

890) pumped with an argon ion laser (Coherent Radiation Innova 90-6) is used 

as an NIR source for Raman excitation. This laser covers the wavelength range 

between 680 and 1000 nm. Since the output intensity of the Ti: sapphire laser is very 

stable for a long time in comparison with a NIR dye laser, it is a good excitation 

source for NIR Raman spectrometry. Continuous-wave lasers are used for measuring 

the Raman spectra of doped polymers, because the doped samples are liable 

to damage due to high peak powers of pulsed lasers. Raman spectra are usually 

measured on a single polychromator (Spex 1870) equipped with a charge-coupleddevice 

(CCD) detector (Princeton Instruments LN/CCD-l024TKBS, back-illuminated 

type, 1024 x 1024 pixels) operated at -120 "C. The CCD detector functions

4.7 Poly(p-pheiq.lene) 217 

with an ST-135 controller (Princeton Instruments) linked with an NEC PC-9801FX 

computer through an IEEE-488 interface. The spectroscopic response of the CCD 

detector extends to about 1000 nin. Optical filters (holographic notch filters) are 

used to eliminate the Rayleigh scattering. The advantage of this system lies in its 

high optical throughput due to the use of the single polychromator and optical 

filter. 

We have modified an NIR Fourier-transform spectrophotometer (JEOL JIR- 

5500) for the nieasurements of Rarnan spectra excited with the 1064-nm line of a 

continuous-wave Nd : YAG laser (CVI YAG-MAX C-92) [60]. Rainan-scattered 

radiation is collected with a 90" off-axis parabolic mirror in a back-scattering 

geometry. Collected radiation is passed through two or three long-wavelength-pass 

filters to reduce the Rayleigh scattering. 

4.7 Poly( p-phenylene) 

As a typical example, we will first discuss the results of the electronic absorption 

and Raman spectra of Na-doped poly( p-phenylene). Poly( p-phenylene) consists of 

benzene rings linked at the para positions as shown in Figure 4-lc. Since poly(pphenylene) 

has a nondegenerate structure, polarons and/or bipolarom are expected 

to be formed by chemical doping. As model compounds of poly( p-phenylene), 

y-oligophenyls (abbreviated as PPn, 11 being the number of phenylene and phenyl 

rings) such as biphenyl (PP2), p-terphenyl (PP3). p-quaterphenyl (PP4), p-quinquephenyl 

(PP5), and p-sexiphenyl (PP6) are used. It has been shown by X-ray 

diffraction studies that the two phenyl rings of PP2 [61, 621 in crystal at room temperature 

are coplanar, and PP3 [63] and PP4 [64] in similar conditions deviate 

slightly from coplanarity. However, these deviations are not significant in analyzing 

their vibrational spectra. Accordingly, the phenylene and phenyl rings of PP5 and 

PP6 in solids, and poly( p-phenylene) are assumed to be coplanar. 

4.7.1 Electronic Absorption Spectra 

4.7.1.1 Intact and Doped Poly(ppheny1ene) 

The electronic absorption spectra of undoped and electrochemically Bu4N+(ntype)-doped 

poly( y-phenylene) films are shown in Figure 4-6a, b, respectively [65, 

661. The undoped poly( p-phenylene) film shows the electronic absorption band at 

27400 cm-' (365 nm), which is associated with the interband transition from the 

valence band to the conduction band. The value of the band gap (0101 in Figure 4-521) 

is estimated to be 24200 cm-' from the onset of the electronic absorption [66]. The 

electronic transition energies of PPn [67, 681 are plotted in Figure 4-7. The transition 

energy decreases as the chain length becomes longer, an observation explained 

by the fact that increased conjugation occurs for longer chains. The observed tran-

30 20 10 0 

ENERGY / lo3 cm-l 

Figure 4-6. Absorption spectra 

of !a) undoped and (b) electrochemically 

Bu4N +(n-type)- 

doped poly~,y-phenylene) films. 

Arrows indicate the positions of 

excitation wavelengths (488.0. 

514.5, 632.8, 720. and 1064 nm) 

used for Ranian measurements 

[66J. 

40 

. 

0 

30 

> 

u 

(r 

w 

z 

; 

0 

20 

u 

0 

10 

2 3 4 5 

NUMBER OF RINGS 

Figure 4-7. Observed electronic transition energies 

of p-oligophenyls. (a) Neutral species: (b) band I of 

the radical anions; (c) band 11 of the radical anions; 

(d) band I' of the dianions. The data of neutral 

p-oligophenyls are taken from 1671. The data of the 

radical anions and dianions are taken from [69, 701. 

sition energy of poly(p-phenylene), 27400 cn-', is lower than that of PP6, 31 500 

cni-' , indicating that the conjugation length of poly( p-phenylene) is much longer 

than 6. Upon Bu4N+ doping) two broad absorptions appear at about 5600 cm-' 

(1 800 nm) and 19 400 cm-' (5 15 nm). The assignments of these bands will be discussed 

in the next section on the basis of the spectra of the radical anions (PPn-)

4.7 Pol)~(p-pheiz~~lene) 219 

ENERGY i 1 0 ~ ~ ~ - ' 

ENERGY I oZcm 

Figure 4-8. Absorption spectra ofa THF solution ofp-terphenyl at (a) early and ib) later stages of 

reduction. In (a), the 35600 cni-' band decreases and the 20800 and 10900 cm-' bands increase in 

intensity. In (b), the 20800 and 10900 cm-' bands decrease and the 15400 cm-' band increases in 

intensity. 

and dianions (PPn-) of p-oligophenyls, which are models of negative polarons and 

bipolarons, respectively. 

4.7.1.2 Radical Anions and Dianions of p-Oligophenyls 

The electronic absorption spectra of PPn- and PPH- have been systeinatically 

studied [69, 701. When a tetrahydrofuran (THF) solution of PP3 is exposed to a 

sodium mirror, its absorption spectrum gives new bands (Figure 4-8). At an early 

stage, the band observed at 35700 cm-' (280 nm) arising from neutral PP3 

decreases and the bands observed at 20800 and 10900 cm-' (481 and 916 nm) 

increase in intensity. These new bands are attributed to the radical anion of PP3 

(PP3'-). In the course of this spectral change, an isosbestic point is observed at 

about 31 500 cm-', indicating that PP3 is quantitatively converted into PP3'-. 

When the solution is further exposed to the sodium mirror, the 20 800 and 10 900 

cm-' bands decrease and the band observed at 15400 cm-' (650 nm) increases in 

intensity. The 1.5400 cm-' band is ascribed to the dianion of PP3 IPP3'-). During 

this reaction, three isosbestic points are observed at about 25 800, 20400, and 

13400 cm-'. The THF solutions of PPiz (n = 4,.5, and 6) show similar reduction 

reactions which may be described as 

PPrz + Na 4 PPi7'- + Na' 

PPi7'- + Na i PPn2- + Na' 

Since PPn- is stable for a long time, we can conclude that the following disproportionation 

reaction does not occur. 

2PPn'- 4 PPiZ + PPd

220 4 Vibvofionnl Slwctvoscopy oj Intact and Doped Conjugated Po1ynier.s 

band I1 

- 

‘6 b?, @6 d hg 

4 -fs- 4 *- 

$4 ‘4 

tJ3 tJ3 

(a) 

(b) 

I 

I 

I 

I 

I 

I 

I 

I 

Figure 4-9. Schematic energy level diagrams. (a) The 

radical anion of biphenyl IPP2.-); (b) the dianion of 

biphenyl (PP2’-). 0, electron; arrow. electronic transition. 

The inolecular orbital levels are taken from 1771. 

This reaction is similar to the process of the intramolecular formation of a bipolaron 

from two polarons. 

The electronic transition energies [69, 701 of the radical anions and dianions of 

p-oligophenyls are plotted in Figure 4-7. Two strong bands are observed for the 

radical anions and are called bands I and I1 (curves b and c in Figure 4-7, respectively). 

On the other hand, one intense band is observed for the dianions, which 

is called band I’ (curve d in Figure 4-7). In the electronic absorption spectra of 

the radical cations and dications of a-oligothiophenes [71-73], two strong bands 

and one strong band are commonly observed for the radical cations and dications, 

respectively. In the case of the radical cations of oligophenylenevinylenes, two 

bands are observed [74, 751, although different results are also reported [76]. 

The absorption spectra of conducting polymers including poly( p-phenylene) have 

been discussed within the frame work of the Hiickel approxiination [7, 101. In order 

to correlate the spectra of the radical anions and dianions of y-oligophenyls with 

those of doped poly( p-phenylene), we will discuss the observed absorption spectra 

of the charged p-oligophenyls on a similar theoretical basis. The energy diagram of 

PP2’- is shown schematically in Figure 4-9a. Molecular orbital levels in this figure 

are taken from the results calculated by the Pariser-Parr-Pople-SCF-MO method 

for PP2 (021, symmetry) [77]. A Pariser-Parr-Pople-SCF-MO-CI calculation for 

PP2‘- [78] shows that band I is assigned to the transition mostly attributable to 

d6. These transitions are 

the electronic excitation of dl0 t d7 and band I1 to d7 +- 

polarized along the long molecular axis. The a, and d9 levels are nonbonding with 

respect to the inter-ring CC bond. It is reasonable to consider that these assignments 

hold also for PPn- (12 = 3-6). Moreover, band I’ of PP2’- is assigned to the transition 

mostly attributable to the electronic excitation bl,,

is worthwhile pointing out that the cu3 and (03' transitions are symmetry forbidden. 

According to a continuum electroii-phonon-coupled model reported by Fesser et 

al. [48], the (01 and 032 transitions are dominant among the three expected for a 

polaron, and the 031' transition is more intense between the two expected for a 

bipolaron. Finally, it is expected that a polaron hus two intense intrugup tramitions 

and a bipolavoir one intense tmnsition [79]. 

The absorption spectrum of Bu4Nf-doped poly( p-phenylene) can now be 

explained as follows. Two broad bands centered at about 5600 and 19400 or1 

(Figure 4-6) are attributable to the 031 and cc)? transitions of negative yolarons, 

respectively. These assignments will be discussed again in Section 4.7.3 on the basis 

of the Raman results. 

4.7.2 Raman Spectra 

4.7.2.1 Intact Poly(p-phenylene) and p-Oligophenyls 

The observed infrared and Raman spectra of intact poly( p-phenylene) [80-841 have 

been analyzed by normal coordinate calculations [43, 84-87]. The factor group of a 

coplanar polymer is isomorphous with the point group &. When the JZ plane is 

taken in the plienylene-ring plane (s axis along the polymer chain) and the x axis 

perpendicular to the phenylene-ring plane, the irreducible representation at the zone 

center (k = 0) is as follows: 

rvlb 

rvlb 

in-pldne = 5ug(R) + 4b1u(IR) + 4hu(IR) + 5hg(R) 

out-of-plane = 2 ~u + lbl,(R) + 3hg(R) + 2hu(IR) 

where R and IR denote Raman- and infrared-active vibrations, respectively. The 

1064-nm excited Raman spectrum and infrared absorption spectrum of intact 

poly( p-phenylene) prepared by dehalogenation polycoiidensation of y-dihalogenobenzene 

are shown in Figure 4-10a, b, respectively. The observed and calculated 

vibrational frequencies of poly( p-phenylene), "C-substituted analog, and perdeuterated 

analog are listed in Table 4-3, except for CH (CD) stretches [q(ug), i'g(hlu), 

v16(bzu), and v20(03~)]. Major Ranian and infrared bands have been assigned. 

In Figure 4-10b, the 809 cm-' infrared band is assigned to the CH out-of-plane 

bend of the phenylene rings, and the 765 and 694 cm-' bands are assigned to the 

CH out-of-plane bends of the terniinal phenyl rings. 

The Raman spectrum of poly( p-phenylene) has been compared with those of 

p-oligophenyls. The 1064-nm excited Rainan spectra of p-oligophenyls in the solid 

state [88] are shown in Figure 4-11. The bands at 1605-1593, 1278-1275, 1221- 

1220, and 795-774 cmrl correspond to those at 1595, 1282, 1222, and 798 ,nip' of 

intact poly( y-phenylene), respectively. These bands are assigned to the ug vibrations 

(v2-vg). The atomic displacements of these modes obtained by normal coordinate 

calculations based on the PM3 method [84] are depicted in Figure 4-12a-d. The 

1595 cm-' mode (Figure 4-12a) is contributed mainly by the CC stretch of the 

phenylene ring. The 1282 cm-' mode (Figure 4-12b) is mainly contributed by the

m 

0 

m 

I 

1500 1000 5 

WAVENUMBER icrn~l 

Figure 4-10. Vibrational spectra of intact poly(yphenylene). 

(a) Raman spectrum taken with the 

1064-nni line in powder sample; (b) infrared spectrum 

in a KBr disk [83]. 

inter-ring CC stretch. The 1222 cm-' mode (Figure 4-12c) is a mixture of CC 

stretch and CH in-plane bend. The 798 cm-' mode (Figure 4-12d) is a mixture of 

CC stretch and CCC deformation. The intensity of the 1278-1275 cn-' band relative 

to that of the 1221-1220 c1n-l band decreases as the chain length becomes 

longer. Thus, this ratio is a marker of the length of conjugated segments [81, 82, 881. 

The wavenumbers of the Raman bands of poIy( p-phenylene) and y-oligophenyls 

are almost independent on the excitation wavelength [Sl]. It is well known that the 

wavenumbers of the two strong Rainan bands of tmm-polyacetyleiie depend greatly 

on the wavelength of the excitation laser. These large dispersions are explained by 

the existence of segments of various conjugation lengths that give rise to the Raman 

bands at different wavenumbers [13, 141. However, the Raniail spectra of other 

conducting polymers do not show such dependence on the excitation wavelength. 

The (resonance) Raman intensities of conjugated molecules have been analyzed 

by the classical vibronic theory due to Tang and Albrecht [89]. Inagaki et al. 1901 

have shown that resonance Ranian intensities of totally symmetric modes of p- 

carotene (an oligoene with eleven conjugated C=C bonds) arise from the A term 

(Franck-Condon term) of the Albrecht theory. Probably, the Franck-Condon 

factor plays an important role in determining the resonance Rainan intensities of 

almost all conjugated molecules. The resonance Ranian intensity of each totally 

symmetric mode is proportional to the square of the shift of the potential minimum 

in the resonant excited electronic state from that in the ground electronic state along 

the normal mode giving rise to the resonance Raman band, when the shift is small 

[91-941. The effective conjugation coordinate represents approximately the change 

of equilibrium geometry between the ground electronic state and the first dipoleallowed 

excited electronic state [15]. Thus, the resonance Raman intensity of a 

normal mode is determined by the contribution of the effective conjugation coordinate 

to the mode. Details are described in [15]. It has been demonstrated that the

~ 

~ 

~ 

~ 

~ 

~ 

~ 

~ 

~ 

~ 

~ 

801 

Table 4-3. Vibrational frequencies (in cm-') of neutral polyi p-phenylenei and its "C-substituted 

and perdeuterated analogs. 

Sym- No. Normal species '3C-substituted Perdeuterated analog 

metric 

analog 

species 

Observed Cnlculated Observed Calculated Observed Calculated 

1831 [84] [43] [86] [87] [83] [841 ~ 4 1 1841 1871 

1595 

1282 

1222 

798 

- 

1482 

1048 

1000 

- 

1401 

I254 

1081 

- 

624 

809 

~ 

~ 

~ 

~ 

1599 1626 1661 1601 1541 

1295 1282 1289 1290 1239 

1226 1233 1127 1184 1209 

776 792 846 802 768 

969 951 961 ~ 

427 367 401 - - 

831 823 834 

1477 1487 1510 1490 1452 

1060 1045 1051 1044 - 

991 984 968 1004 965 

945 968 945 ~ 

768 754 760 - 

441 410 402 

1407 1399 1440 1412 1361 

1232 1275 1268 1343 ~ 

1119 1162 1075 1118 ~ 

1603 1601 1654 1652 - 

1290 1325 1328 1343 

633 625 602 616 602 

487 488 460 418 ~ 

198 797 790 ~ 

497 457 458 - - 

1547 1563 

1246 1255 

1216 892 

741 764 

960 ~ 

415 ~ 

825 ~ 

1445 1354 

1035 817 

958 978 

936 ~ 

742 ~ 

426 ~ 

1375 1338 

1 I93 - 

1092 ~ 

1546 ~ 

1277 - 

610 606 

470 - 

789 650 

48 1 ~ 

1563 1572 

1281 1253 

881 863 

751 776 

775 ~ 

378 - 

647 ~ 

1353 1364 

812 811 

984 977 

824 ~ 

663 - 

414 ~ 

1334 1347 

1185 1268 

854 855 

1579 1620 

1016 1044 

600 593 

450 383 

661 ~ 

435 - 

Rainan spectrum of neutral poly( p-phenylene) is well explained by the effectiveconjugation-coordinate 

model [95]. 

4.7.2.2 Doped Poly(p-phenylene) and the Radical Anions and Dianions 

of p-Oligophenyls 

The Raman spectra of heavily Na-doped poly( y-phenylene) measured with the 

488.0, 514.5, 632.8, 720, and 1064 nm laser lines are shown in Figure 4-13 [70]. The 

Raman spectra of Na-doped poly( p-phenylene) are apparently different from that 

of intact poly( p-phenylene), indicating that these bands arise from charged domains 

(i.e., negative polarons and/or negative bipolarons) created by Na-doping. The 

assignments of these bands will be discussed on the basis of the data for the radical 

anions and dianions of p-oligophenyls.

224 4 Vibrcitioiznl Spectroscopy of Intact and Doped Conjugated Polymers 

RAMAN SHIFT/cm-' Excitation wavelength is 1064 nin [88] 

Figure 4-12. Atomic displacements of inajor vibrational modes 

(k = 0) of poly(p-phenylene). (a) 1599 an-' [v2(ng)]; (b) 1295 

cm-' [q(ug)];(c) 1226 cm-I [114(a,)]: (d) 776 cm-I [v5(az)]: 

(d) (el (f) (e) 1477 c1n-l [1~,0(b,~)]; (f) 991 c1n-l [iy(hlu)][84] 

The Raman spectra of the radical anions of PP2 [92, 96, 971, PP3 [70, 981, PP4 

[70, 981, PP5 [70] and PP6 [70], and the diaiiions of PP3 [70, 981, PP4 [70, 981, PP5 

[70], and PP6 [70] have been reported by several authors. The Raman spectra of the 

radical anions (PPnP) and dianions (PPTz~~) in THF solutions are shown in Figures 

4-14 and 4-15, respectively. The excitation wavelengths used are rigorously or 

nearly resonant with band I1 of the radical anions and band I' of the dianions. The 

Raman spectra of the radical anions are similar to each other, and so are the 

Rainan spectra of the dianions. However, the two groups of the Raman spectra are 

different from each other and from those of neutral 12-oligopheiiyls. These results 

suggest that the negative polarons and bipolarons, corresponding respectively to the 

radical anions and dianions, can be identified by Raman spectroscopy. 

The Ranian spectra of the radical anions in Figure 4-14 show siiiall changes with

4.7 Poly(y-phenylene) 225 

Figure 4-13. Raman spectra of Na-doped poly( - 

p-phenylene) taken with various wavelengths of excitation 

lasers. (a) 488.0; (b) 514.5; (c) 632.8: (d) 720, (e) 

1064 nm [70]. 

1500 1000 

RAMAN SHIFT / Cm-' 

Figure 4-14. Resonance Ranian spectra of the radical anions 

of p-oligophenyls in THF solutions. (a) p-Terphenyl; (b) 

p-quaterphenyl; (c) p-quinquephenyl; (d) p-sexiphenyl. Excitation 

wavelengths are 458.0 and 514.5 nm for (a) and (b-d), 

respectively. The bands of the solvent are subtracted [70]. 

t 

t 

(I) 

2 

W 

b 

z 

2 

< 

2 

4 

LT 

1500 1000 

RAMAN SHlFT/cm-' 

the chain length. The intensities of the bands observed in the 1032-991 and 780- 

753 cm-' regions decrease as the chain length becomes longer, indicating that these 

bands are associated with the terminal rings. The bands observed in the 780-753 

cni-' region show an upshift in wavenumber, as the chain length becomes longer. 

This result indicates that the wavenumbers of these bands can be used as a measure 

of the localization length of polarons. The bands observed in the 1325-1305 c1n-l 

region are assigned to the inter-ring CC stretch [70, 92, 981. The corresponding 

bands in neutral PPIZ are observed in the range between 1278 and 1275 cm-'. About 

40-cm-' upshifts are explained by the increased TC-bond orders of the inter-ring CC 

bonds, i.e., shortening of the inter-ring CC bond lengths. In PP?, the calculated

226 4 k'ilmtioiiul Spectroscopy of Iiitact aid Doped Conjirgtrted Polyinem 

1500 7 000 

RAMAN SHIFTicm' 

Figure 4-15. Resonance Raman spectra of the dianions of 

p-oligophenyls in THF solutions. (a) p-Terphenyl; h) y- 

quaterphenyl; (c) p-quinquephenyl; (d) psexiphenyl. Excitation 

wavelengths are (a) 632.8, (b) 720, (cj 1064, and (d) 1064 nni, 

respectively, The bands with x are due to the solvent. Fluorescence 

background is subtracted in all the spectra 1701. 

length of the inter-ring CC bond (rq5') shrinks from 1.469 to 1.426 A in going from 

neutral PP3 to PP3'- (see Table 4-1). 

In the Raman spectra of the dianions in Figure 4-15, the bands in the 1479-1468 

cm-' and 1188-1 162 cm-' regions are strongly observed. Corresponding bands 

are weakly observed or are not observed in the spectra of the radical anions. This 

reflects structural differences between the electronic ground state and the excited 

state associated with band I1 of the radical anions and band I' of the dianions. The 

bands observed in the 1357-1320 cm-l region are assigned to the inter-ring CC 

stretch [70, 981. About 70 cm-l upshifts in going from neutral PPn to PPiz'- are also 

explained by the shortening of the inter-ring CC bond lengths. These upshifts of 

PPn- are larger than those of PPn'-. This observation suggests that the degrees of 

shrinkage of the inter-ring CC bonds in PPn'- are larger than those in PPn-. This 

is consistent with the results of calculation for PP3 given in Table 4-1 : 1.469 A (PP3) 

and 1.388 A (PP3'-). It is worth pointing out that the frequency of inter-ring CC 

stretch increases with the value of structural deformation coordinate R (column 8 

in Table 4-1) for PP3, PP3'-, and PP32-. In the 1000-940 cm-' region, there is a 

series of bands at 943 (PP3'-), 951 (PP4'-), 974 (PP5'-), and 992 cm-' (PP6'-). 

This band series may be useful in estimating the localization length of bipolarons. 

Next, we will analyze the Raman spectra of Na-doped poly( p-phenylene) (Figure 

4-13) on the basis of those of the radical anions and dianions ofp-oligophenyls. The 

Raman spectra of Na-doped poly( p-phenylene) taken with the 488.0- and 5 14.5-nm 

lines (Figure 4-13a, b) are different froin those with the other three laser lines 

(Figure 4-13c-e), and are similar to those of the radical anions of y-oligophenyls 

(Figure 4-14). This observation can be understood by considering that the Raman 

bands arising from charged domains formed by Na-doping in the polymer chains 

are quite similar to those of the radical anions. Since the radical anions are viewed 

as negative polarons in the polymer, negative polarons exist in Na-doped poly(pphenylene) 

[70]. These Raman spectra of Na-doped poly( p-phenylene) are observed 

under resonant conditions. Thus, the results described above indicate that the

Table 4-4. Observed Raman frequencies (in cn- ' ) of negative polarons and negative bipolarons in 

Na-doped poly( p-phenylene). 

No. Negative polaron" Negative bipolaron" Assignmentb 

A 1591-1590 1594-1586 syni. in-phase mode cousistirig of vz(~i0) 

B 1477-1371 sym.. VlO(h,) 

c 1315-1 314 1352- 1344 sym., in-phase, 1'; ((ig) 

D 1297-1190 Sylll., OUt-of-phdse, IQ[ag) 

E 1221-1219 1222-1219 sym., in-phase, 1'4(Ug) 

F 1179 1190-1 I79 sym., out-of-phase, 1'4 (ag) 

G 980-974 992-951 sym., 1'12(blu) 

H 783-777 syni., in-phase, 1'5 ((1,) 

Data are taken from [83]. 

Numbering and symmetry correspond to those of neutral poly( pphenylene). Sym. denotes a 

symmetric mode. 

absorption around 488.0 and 514.5 nm (20500 and 19400 cm-') is attributed to 

polarons. 

The Raman spectra excited with laser lines in the 632.8-1064 iim region (Figure 

4-15-e) are quite similar to those of the dianions of p-oligophenyls (Figure 4-1 5), 

which correspond to negative bipolarons in the polymer. Accordingly, negative 

bipolarons also exist in Na-doped poly( y-phenylene), and the absorption in the 

632.8-1064 nm (15 800-9400 cni-I) region arises from bipolarons [70]. In the 1000- 

950 cm-' region of the Raman spectra, a few bands are observed at 979 and 957 

c11i-l (632.8-11111 excitation), at 977 and 952 Gin-I (720 nm), and at 992, 974, and 

956 cm-' (1064 nm). By comparing these bands with those of the Raman spectra of 

the dianions, it may be concluded that negative bipolarons localized over four, five, 

and six phenylene rings coexist in Na-doped poly( p-phenylene) . 

The Ranian bands arising from negative polarons and bipolarons, together with 

the tentative mode assignments based on the wavenumber shifts upon I3C-substitution 

[83], are listed in Table 4-4. The observed bands are called bands A-H as 

shown in Table 4-4. Bands A, C, E, and H correspond to the ag modes (vz, i = 2-5) 

of neutral poly( 13-phenylene), respectively. The 1282 cm-l Raman band of neutral 

poly(y-phenylene) upshifts to 13 15-13 14 cn-l in the spectra of polarons and 1352- 

1344 cm-' in the spectra of bipolarons. According to normal coordinate calculations 

[43, 84, 86, 871, the 1282-cmP' mode is mainly attributed to the inter-ring 

CC stretch. The observed upshifts upon Na-doping reflect the increasing z-bond 

order of the inter-ring CC bond, i.e., structural changes froin benzenoid to quinoid 

[70]. The upshifts for bipolarons (70-62 cm-') are larger than those for polarons 

(33-32 cm-l), indicating that the geometric changes in bipolarons are larger than 

those in polarons. On the other hand, bands A and E show little shifts in wavenumber 

and band H shows a moderate shift. 

In the Raman spectra of negative bipolaroils (Figure 4-13c-e), bands B, D, F,

228 4 Ii’Dratioizal Spectroscopy of Intact and Doped Coiijuyatecl Polytners 

and G are strongly observed. Band B is assignable to a symmetric mode consisting 

of the v,~(bl,,) mode (Figure 4-12e) of each benzene ring in the charged domain. 

Band G is also attributed to a symmetric mode consisting of the i112(61~,) mode 

(Figure 4-12f). Bands D and F are assignable to out-of-phase symmetric inodes 

consisting of the I I ~ ( L ~ and ~ ) i ~ ~ ( modes ~ i ~ )(Figure 4-12b, c) of each benzene ring, 

respectively. The appearance of bands B, D, F, and G indicates the loss of translational 

symmetry of neutral polyi p-phenylene) chain, wliich is consistent with the 

forniatioii of localized charged doniains upon doping. Bands B, D, F, and G are 

not observed or weakly observed for negative polarons, but strongly observed for 

negative bipolarons. These observations seem to reflect different geometric changes 

occurring between polarons and bipolarons 011 going from their respective ground 

electronic states to their respective excited electronic states. 

4.7.3 Assignments of Electronic Absorption Spectra of Doped 

Poly( p-phen ylene) 

In Section 4.7.1, two broad bands centered at about 5600 and 19400 cm-l in the 

electronic absorption spectrum of Bu4N+-doped poly( p-phenylene) have been 

attributed to the QI and 02 transitions of negative polarons, respectively, on the 

basis of the data of the radical anions and dianions of p-oligophenyls. Raman 

spectroscopy provides information on the electronic levels associated with polarons 

and bipolarons as well as on their geometries. The Ranian results indicate the coexistence 

of polarons and bipolarons in Na-doped poly( p-phenylene). The Ranian 

bands due to polarons are resoiiantly enhanced by the use of the 488.0 and 514.5 

11111 laser lines, which are located within the electronic absorption around 19 400 

cin-l (51 5 nin). These observations confirm the proposed assignment that the 

19400 c1n-l band is due to the cuz transition of polarons. The Ranian bands due to 

bipolarons are resonantly enhanced by the use of the 632.8, 720, and 1064 nm laser 

lines. However, in the absorption spectruin of Bu4N+-doped poly( p-phenylene), no 

peaks are observed in the 15800-9400 cm-’ (632.8-1064 nm) region. This fact can 

be explained by considering that the content of the bipolaron is very small and the 

QI’ absorption due to the bipolaron is overlapped by the strong to1 absorption band 

due to polarons. 

4.8 Other Polymers 

The Raman spectra of doped states of polyaniline [99], poly( p-phenylenevinylene) 

[75, 100-1021, and polythiophene [72, 731 have been also analyzed on the basis of 

the data of charged oligoiners (charged oligomer approach). The detected selflocalized 

excitations are suniniarized in Table 4-5. Some important conclusions can 

be drawn from these Rainaii studies.

4.8 Other Polwieys 229 

Table 4-5. Types of self-localized excitations detected by Raman spectroscopy. 

Polymer Dopant Species Reference 

Type Reagent Content 

poly( p-phenylene) 11 

poly ( p-phenylenevmylene) n 

polythiophene 

polyaniline (emeraldine base 

form, 2A)” 

polyaniline (leucoemeraldine, 1A) 

11 

n 

P 

P 

P 

P 

P 

p 

heavy 

heavy 

light 

heavy 

heavy 

heavy 

light 

heavy 

heavy 

heavy 

polaron, bipolaron 




polaron 

polaron 

polaron 

polaron 

polaron 

polaron 

a HC1-doped emeraldine base form is called emeraldine acid form or 2s form 

1. Only polarons are detected for p-type doping. The Ranian spectra of the polymers 

doped with acceptors do not show large changes with various excitation 

wavelengths [72, 73, 751. The polymer-chain structures of p-type-doped polymers 

are more homogeneous, and more regular arrays of polarons are formed. 

2. Polarons and bipolarons coexist in n-type doped polymers in contrast to the 

results of p-type doping. The Raman spectra of the polymers doped with donors 

depend on the excitation wavelength [70, 1001. These observations are explained 

by the existence of various localization lengths of polarons and/or bipolarons 

[70, 1001. The polymer-chain structures of the n-type-doped polymers are inhomogeneous. 

Possibly, Na-doped polymers are easily hydrogenated, even in the 

presence of a small amount of water during the sample preparation, and negative 

polarons and/or bipolarons are formed on the short conjugated segments. 

The Raman spectra of doped polyacetylene have been reported with the excitation 

of visible laser lines [103-1101. There have been some discussions about the 

question of whether these bands arise from charged domains or undoped parts 

remaining in the doped films. Recently, the Ranian spectra of maximally Na-doped 

polyacetylene excited with NIR lasers have been reported [57, 60, 1111 and conipared 

with those of the radical anions of u, w-diphenyloligoenes and a, co-dithienyloligoenes 

[ 11 11. It has been concluded that the observed bands arise from charged 

domains (charged solitons and/or polarons), because the observed Raman spectra 

of the doped polymer are similar to those of the radical anions of x, oAiphenyloligoenes. 

However, further studies are required for characterizing doped polyacetylene 

in more detail.

230 4 Vilwntional Slwctvoscopjl of Intact mitl Doped Conjzrgutecl Poljwievs 

4.9 Electronic Absorption and ESR Spectroscopies and 

Theory 

It is widely accepted [7, 101 that the major species generated by chemical doping it? 

nondegenerate polymers is bipolarons, except for polyaniline. However, the result^ 

obtained from Raman spectroscopy indicate the existence of polarons at heavily 

doped polymers. These results are inconsistent with the established view. The 

experimental basis for claiming the existence of bipolarons has been electronic 

absorption and ESR spectra [7, 9, lo]. We will comment on these experimental 

results . 

The electronic absorption spectra of p-type-doped polypyrrole have been interpreted 

in terms of polarons and bipolarons for the first time 11121. Electronic 

absorption spectra of doped polypyrrole in the range from visible to NIR show 

three bands due to intragap transitions at low dopant contents and two bands at 

high dopant contents [ 1131. The three bands are attributed to polarons and the two 

bands to bipolarons [112], because a polaron is expected to have three intragap 

transitions and a bipolaron is expected to have two intragap transitions (see Section 

4.4.2). A quantitative ESR study [114] has confirmed the interpretation of the electronic 

absorption experiments. This rule of thumb for band assignment has so fiiialso 

been applied to other conducting polymers. 

However, the two electronic absorption bands of doped poly( p-phenylene) are 

not explained by bipolarons, as demonstrated in Section 4.7.3. Furthermore, 

p-type-doped polythiophene shows two absorptions at about 12000 and 5200 cnir ’ 

[72, 73, 115, 1161 and p-type-doped poly( p-phenylenevinylenej shows two broad 

absorptions at about 19000 and 7300 cmP1 [75, 117, 1181. In these two systems, 

only positive polarons are detected by Raman spectroscopy (see Table 4-5). These 

results of Rainan spectroscopy suggest that the two-band pattern in the electronic 

absorption spectrum is attributed to polarons. As discussed in Section 4.7.1.2, it is 

expected that a polaron has two intense intragap transitions and a bipolaron one 

intense transition. From these findings, we propose a new rule: a two-hmd pattern 

correspondy to yolavons mid one-barid puttern to hipolarons [79], when these species 

do not coexist. When polaroils and bipolarons coexist, and their electronic absorptions 

are overlapped, a more careful examination of the observed spectrum is of 

course necessary. 

Recently, Hill et al. [119] have proposed the existence of a singlet radical-cation 

dimer (i.e., polaron dimer) as an alternative to a spinless bipolaron to explain weak 

ESR signals from doped polymers. Their proposal is based on the observed dimerization 

of the radical cation of 2,5”-dimethylterthiophene. They have found by 

electronic absorption and ESR measurements that the above radical cation is 

dinierized at low temperature even at a low concentration (4 x lor4 mol/l). Singlet 

intrachain polaron pail-s and interchain polaron pairs give no ESR signals. Thus, 

the absence of ESR signals or observation of weak ESR signals does not directly 

lead to the conclusion that the doped polymer contains only bipolarons. 

The stability of polarons, bipolarons, and solitons has been studied theoretically

under various models. According to the SSH model and siinilar approximations 

with electron-lattice coupling for a single polymer chain [7, lo], the creation energy 

of one bipolaron in a nondegenerate system (or two charged solitons in a degenerate 

system) is lower than that of two polarons. Thus, it is considered that two separate 

polarons are always unstable, and changed into a bipolaron (or two charged solitons). 

For doped polyacetylene, the effect of the electron--electron interaction and 

the dopant potential, which are not taken into account in the SSH model, have been 

studied 11201. On the other hand, the existence of polarons has been theoretically 

proposed by several authors. Kivelson and Heeger [121] have proposed a polaron 

lattice (regular infinite array of polarons) for explaining the metallic properties of 

heavily doped polyacetylene. Mizes and Conwell [ 1221 have reported that polarons 

are stabilized in short conjugated segments for trmzs-polyacetylene and poly( p- 

phenylenevinylene) from the results obtained by a three-dimensional tight-binding 

calculation. Shimoi and Abe [123] have studied the stability of polarons and bipolarons 

in nondegenerate systems by using the Pariser-Parr-Pople method combined 

with the SSH model. A polaron lattice is stabilized by the electroii-electron 

interaction at dopant contents lower than a critical concentration, whereas a bipolaron 

lattice is stabilized at dopant contents higher than the critical concentration. 

Thus, importance of the electron-electron interaction and three-dimensional interaction 

as well as the electron-lattice coupling has been demonstrated. However, 

the nature of the metallic states of heavily doped polymers remains an unresolved 

theoretical problem. 

4.10 Mechanism of Charge Transport 

Overall electrical conduction in doped polymers is considered to have contributions 

from various processes such as intrachain, interchain, interdomain, interfibril 

transport, etc. The contributions of individual processes would depend on the solidstate 

structure and morphology [8]. Charge transport has been discussed in terms of 

polarons, bipolarons, charged solitons, and their lattices (6, 8, 101. Even when the 

dopant content is very small, electrical conductivity increases dramatically. At this 

doping level, discrete localized electronic levels associated with self-localized excitations 

are formed within the band gap, and the temperature dependencies of d.c. 

conductivity have been explained by the hopping between localized electronic states 

181. It is conceivable that the energy gap between the lower (or upper) localized state 

and the valence (or conduction) band for a positive (or a negative) polaron is 

smaller than that of a positive (or a negative) bipolaron. Then, it is expected that 

polarons play a more important role in hopping conduction than bipolarons. 

Heavily doped polymers are of special interest, because these polymers show 

metallic properties such as a Pauli susceptibility, a linear temperature dependence of 

the thermoelectric power, a high reflectivity in the infrared region, etc., as described 

in Section 4.2. The temperature dependence of d.c. conductivity in heavily doped 

polymers is not like a metal, whereas metallic properties are observed. These results

Figure 4-16. Schematic structures of polythiophene chains doped with electron acceptors (dopant 

content, 25 mole'% per thiophene ring) and bonding electronic levels of positive polarons and 

bipolarons. (a1 Polaron lattice: (bi bipolaron lattice. A. acceptor: +. positive charge; -, negative 

charge; 0, electron; -, electronic energy level. 

Ferrni level .-.. 

-.... Ferrni level 

Figure 4-17. Schematic band structures of 

dn infinite polymer chain doped with 

acceptors (dopant content, 25 mole%) per 

thiophene ring). (a) Polaron lattice; 

(b) bipolaron lattice Shadowed areas are 

(a) (b) filled with electrons [ 1261. 

can be explained as follows. Although a single chain itself behaves like a metal, 

macroscopic electrical conduction is limited by interchain, interdomain, or interfibril 

charge transport. Here, we will discuss the intrachain metallic charge transport. 

Let us suppose an infinite nondegenerate polymer chain (e.g., polythiophene) 

doped heavily with electron acceptors. At a high dopant content, the polymer-chain 

structure and electronic structure of the doped polymer are radically different from 

those of the intact polymer. As typical cases, we will describe two kinds of lattice 

structures of doped polythiophene (dopant content, 25 mole'%, per thiophene ring): 

a polaron lattice and a bipolaron lattice. They are the regular infinite arrays of 

polarons and bipolarons. The schematic polymer-chain structures are shown in 

Figure 4-1 6. Band-structure calculations have been performed for polaron and/or 

bipolaron lattices of poly( p-phenylene) [ 1241, polypyrrole [ 1241, polyaniline [ 1251, 

polythiophene [ 124, 1261, and poly( p-phenylenevinylene) 11271, with the valenceeffective 

Hamiltonian pseudopotential method on the basis of geometries obtained 

by MO methods. The schematic electronic band structures shown in Figure 4-17

4. I0 Mechnriism oj Charge. Transport 233 

are based on the results calculated with this method for a polaron and a bipolaron 

lattice of p-type-doped polythiopheiie (dopant content, 25 inole%, per thiophene 

ring) [ 1261. 

A polaron has a localized electronic level occupied by one electron (i.e., SOMO). 

In a polaron lattice formed at a heavy doping level, electronic wave functions of 

neighboring polarons are overlapped and interact with each other. As a result, a 

half-filled metallic band is formed from the polaron bonding levels occupied by one 

electron (Figure 4-17a). The Fernii level is located at the center of the polaron 

bonding band. The electronic states of the polaron lattice are not localized, whereas 

the electronic states of a polaron are localized. The calculated bandwidths of the 

polaron bands are 8900, 7900, and 5600 cm-l (1.1. 0.98, and 0.69 eV) for heavily 

doped polyaniline [ 1351, polythiophene [ 1261, and poly( p-phenylenevinylene) [ 1271, 

respectively. If the conjugatioii length of the polymer chain is not sufficiently long, 

the metallic band is not formed. In this case, discrete levels are formed and hopping 

conduction is expected to take place. In contrast to the polaron lattice, no metallic 

band is formed for the bipolaron lattice, even if the conjugation length is long 

enough (Figure 4-17b). This is because a bipolaron has no singly occupied electronic 

level, whereas a polaron has a singly occupied level. In addition, the differences 

in the electronic band structure between the polaron and bipolaron lattices 

suggest that the bipolaron lattice is the result of a dinierization of the polaron lattice 

like the Peierls transition [126]. 

In the case of rrnns-polyacetylene (a degenerate polymer), band-structure calculations 

[126] indicate that a polaron lattice has a metallic band whereas a charged 

soliton lattice has no metallic band. An alternating polaron-charged soliton lattice 

[ 1281 has been proposed for explaining the metallic properties of heavily doped 

states. 

We will discuss polymer-chain structures obtained by Raman spectroscopy and 

metallic properties. The Raman measurements indicate the existence of positive 

polaroiis for p-type-doped polythiophene, poly( p-phenylenevinylene), and polyaniline 

at heavy doping levels (Table 4-5). These results indicate that positive polarons 

are formed on the polymer chains upon acceptor doping. If conjugation length 

is sufficiently long and the dopant content sufficiently high, a large number of 

polarons are formed and interact with each other. Then, a polaron lattice is formed. 

Therefore, the results of p-type-doped polythiophene, poly( p-phenylenevinylene), 

and polyaniline are consistent with the formation of polaron lattices at heavy 

doping levels. Tlie rnetullic properties obs'eriled fbr cloped polpers probubly origincite 

from the poluroii lattice. From the results obtained so far, we can point out some 

important conditions for obtaining metallic polymers: (i) a long conjugation length; 

(ii) a high dopant content (large degree of charge transfer); and (iii) formation of 

polarons. In the case of n-type doping, both negative polarons and bipolarons are 

detected for doped poly( y-phenylene) and poly( p-phenylenevinylene) (Table 4-5). 

Polarons and bipolarons with various localization widths are detected, and electronic 

states are probably localized, even at heavy doping levels. It is considered 

that such an inhomogeneous structure does not lead to formation of a metallic 

band. 

It is well known that physical properties of conducting polymers depend on

234 4 I i’brationul Spectroscopy of Intact and Doped Conjugared Polynzers 

synthetic conditions. doping conditions, and other treatments of samples. Thus, in 

order to confirm tlie proposal that the metallic properties originate from the 

polaron lattice, it is requisite to use the same polymer film or the films from the 

same batch for measuring various physical properties and Raman spectra. 

4.11 

It has been demonstrated that Raman spectroscopy is a powerful tool for studying 

self-localized excitations such as polarons and bipolarons in doped conducting 

polymers. Resonance Rainan spectroscopy by using exciting laser lines in a wide 

range between visible and near-infrared gives valuable information on self-localized 

excitations; especially, near-infrared Raman spectroscopy developed recently is iniportant. 

Spectroscopic data on the radical ions and divalent ions of oligomers are 

useful for analyzing the electronic and vibrational spectra of doped polymers. New 

assignments of the electronic absorption spectra of polarons and bipolarons are 

proposed: a two-band pattern is attributed to polarons, and a one-band pattern 

to bipolarons. Polarons and/or bipolarons have been detected for doped poly( y- 

phenylene), polythiophene, poly( p-phenylenevinylene), and polyaniline by Raman 

spectroscopy. On the basis of tlie Rainan results, metallic properties at heavily 

doped polymers is attributed to formation of a polaron lattice with a half-filled 

metallic band. 

Postscript: After the manuscript of this Chapter was completed, a review book 

treating the polaron/bipolaron in conjugated polymers was published [ 1291. 


[l] Shirakawa, H., Louis, E. J., MacDiarmid. A. G., Chiang, C. K., Hccger, A. J., J. Cheni. Soc., 

Clieni. Coi72iiitiii. 1977, 578-580. 

[2] Chiang, C. K., Finchcr, Jr., C. R., Park, Y. W., Heeger, A. J.. Shirakawa, H.. Louis, E. J., 

Gau, S. C., MacDiarmid, A.G., Phys. Reti. Lett. 1977, 39, 1098-1101. 

[3] Wegner, G., Angeiv. Clzetii. ht. Ed Engl. 1981, 20, 361-381. 

[4] Etcmad, S., Heegcr, A. J., MacDiarmid. A. G., Ann. Rev. Phys. Clzeni. 1982, 33, 443-469. 

[5] Baughman, R. H., Bredas, J. L.. Chance, R. R., Elsenbaumer, R. L., Shacklette, L. W., 

Clzein. Rev. 1982, 82, 209-272. 

[6] Chien, J. C. W., Pol!~acetyletie-Clzerriistry, Physics, mzd hlatevial Science. New York: 

Academic Press, 1984. 

171 Brkdas, J. L., Street. G. B. Acc. Clzcrii. Rex 1985, 18, 309-315. 

[8] Roth, S., Blcier, H., Ah. Phys. 1987, 36, 385-462. 

[9] Patil, A. O., Hecger, A. J., Wudl, F., Cllen7. Rev. 1988, 88, 183-200. 

[ 101 Hecger, A. J., Kivclson, S., Schricffer, J. R., Su, W.-P., Reu. Mod Php. 1988, 60, 781L850. 

[ll] Kuzmaay, H., Mchring, M., Roth, S. (Eds.) Springer Ser. Solill-State Sci. 1987, 76; 1990, 91 ; 

1992, 107. 

[I 21 Lu, Y., Soktons arid Pokzr~or?~. Singapore: World Scientific, 1988.

4.12 Reftrences 235 

[ 131 Kuzmany, H.. Mokromol. Chern., iMacronzo/. svriip. 1990, 37, 8 1-97. 

[14] Harada, I., Furukawa; Y., in: l’ibrcitiorzul Spectiti atid Structure: Durig, J. R. (Ed.). Amsterdam: 

Elsevier, 1991; Vol. 19, pp. 369-469. 

[I 5) Gussoni, M., Castiglioni, C., Zerbi, G., in: Athwices in Spectro.mp,v: Clark, R. J. H., Hester, 

R. E. (Eds.). Chichester: John Wiley & Sons, 1991; Vol. 19, pp. 251-353. 

[ 161 Bredas, J. L., Silbey. R. (Eds.), Conjirgured Polymers. Netherlands: Kluwer Academic Publishers, 

1991. 

[I71 Tsukamoto, J., Ad;. Ph~,.s. 1992, 41, 509-546. 

1181 Su. W. P., Schrieffer, J. R., Heeger. A. J., Phys. Rev. Lett. 1979. 43, 1698-1701. 

(191 Su, W. P.. Schrieffer, J. R., Heeger, A. J., Phys. Rev. B 1980, 22, 2099-2111. 

[20] Su, W. P., Schrieffer, J. R., Proc Ntrtl. Actid Sci. 1ISA 1980, 77, 5626-5629. 

[211 Brazovskii, S. A., Kirova, N. N., Sozi. P/i]’s. JETP Lett. 1981, 33, 4-8. 

[22J Bishop, A. R., Campbell, D. K., Fesser, K., Mo/. Cr~ist. Liq. Crysl. 1981, 77, 253-264. 

[23] Bredas, J. L., Chance, R. R., Silbey, R., iMol. Crjvt. Liq. Cryst. 1981, 77, 319-332. 

[24] Ikehata, S., Kaufer, J., Woerner, T., Pron, A,, Druy, M. A., Sivak, A., Heeger, A. J., Mac- 

Diarmid, A. G., P/ij.s. Rev. Lett. 1980, 45, 1123-1126. 

[25] Chung, T.-C., Moraes, F., Flood, J. D., Heeger, A. J., Pliys. Rev. B 1984, 29, 2341-2343. 

1261 Park, Y. W., Denenstein, A,, Chiang, C. K., Heeger, A. J., MacDiarinid, A. G., Solid Sttrte 

CO/~IIIIUTI. 1979, 29, 747-75 1. 

[27] Tanaka, M., Watanabe, A,, Tanaka, J., Bid. Ckm. Soc. 4117. 1980, 53, 3430-3435. 

[28] Kume, K., Mizuno, K., Mizoguchi, K., Nomura, K., Maniwa, Y., Moo(. Cryst. Liq. Cryst. 

1982, 83, 285-290. 

[2Y] Moraes, F., Davidov, D., Kobayashi, M., Chung, T. C., Chen, J., Heeger, A. J., Wudl, F., 

Synth. Met. 1985, 10. 169-179. 

[30] Kaneto, K., Hayashi, S., Ura, S., Yoshino, K., J. Pl7ys. So(,. Jpn. 1985, 54, 1146-1153. 

[31] Mizoguclii, K., Misoo, K., Kume, K., Kaneto, K., Shiraishi, T., Yoshino, K., Sjwfh. Met. 

1987, 18, 195-198. 

[32J Logdlund. M., Lazzaroni, R., Stafstrom, S., Salaneck, W. R., Bredas, J. L., Phys. Rev. Lett. 

1989, 17, 1841-1844. 

1331 Masubuchi, S., Kazania, S., Swzth. Met., 1995, 74, 151-158. 

[34] Ginder, J. M., Richter, A. F., MacDiarmid, A. G., Epstein, A. J., Solid Store Conimziri. 1987. 

63, 97-101. 

[35] Mizoguchi, K., Obana, T., Ueno, S., and Kume, K., Syiith. Met. 1993, 55, 601-606. 

[36] Salem, L., The Molecultu Orbitol Tlzcorp of” Co@pted Si1sterir.s. New York: Benjamin, 1966. 

[37] Kohler, B. E., in: Coi?jtigated Po1yiiier.s: Bredas, J. L. and Silbey, R. (Eds.). Dordrecht: 

Kluwer Academic Publishers, 1991; pp. 405-434. 

[38] Fmcher, Jr., C. R., Chen, C.-E., Heeger, A. J., MacDiarniid, A. G., Hasting, J. B., Plzys. Rev. 

Lett. 1982, 48, 100-104. 

[39] Yannoni, C. S., Clarke, T. C., Phys. Rev. Lett. 1983, 51, 1191-1193. 

[40] Villar, H. O., Dupuis, M., Watts, J. D., Hurst, G. J. B., Clementi. E., J. Clzenz. Phys. 1988, 

88, 1003-1009. 

[41] Hirata, S., Torii, H., Tasumi, M., J. Cl7enz. Pliys. 1995, 103, 8964-8979. 

1421 Castiglioni, C., Gussoni. M., Zerbi, G., Syntli. Met. 1989, 29, ElLE6. 

[43] Cuff, L., Kertesz, M., Mocroii7olecuke.s 1994, 27, 762-770. 

[44] Ohtsuka, H., Master’s Thesis, The University of Tokyo, 1993. 

[45] Longuet-Higgins, H. C., Salem, L., Proc. Roj~. Soc. A 1959, 251. 172-185. 

[46] Pople, J. A,, Walmsley. S. H., Mo/. Pliys. 1962, 5, 15-20. 

[47] Takayama, H., Lin-Liu, Y. R., Maki. K., Plzys. Reo. B 1980, 31, 2388-2393. 

1481 Fesser, K,, Bishop, A. R., Campbell, D. K.: Pkys. Reu. B 1983, 27, 4804-4825. 

1491 Onodera, Y., Pliys. Rex:. B 1984, 30, 775-785. 

[50] Zerbi, G., Gussoni, M., Castiglioni, C., in: Copjrgated Pdyrners: Bredas, J. L. and Silbey, R. 

(Eds.). Dordrecht: Kluwer Academic Publishers, 1991; pp. 435-507. 

[S 11 Horovitz, B., Solid Stntc. Contmi//7. 1982, 41. 729-734. 

1521 Vardeny, Z., Ehrenfreund, E., Brafman, 0.. Horovitz, B., PliI,s. Rea. Lert. 1983.51,23176-2329. 

[53] Ehrenfreund, E., Vardeny, Z., Brafman, O., Horovitz, B., Pliys. Rev. B 1987, 36, 1535- 

1553.

236 4 I’ihratioiznl Spectroscopy of Intact nizd Doped Conjirpted Polj~iiwrs 

[54] Castiglioni, C., Navarrete, J. T. L., Zerbi. G., Gussoni, M., Solid St~ite Coninirrn. 1988, 65. 

625-630. 

[55] Wilson, E. B., Decius, J. C., Cross, P. C., Molecular Vibra/ions. New York: MacGraw Hill. 

1955. 

[56] Furukawa, Y., Springer Ser. Solid-Stute Sci. 1992, 107, 137-143. 

[57] Furukawa, Y., Sakanioto, A., Ohta, H.. Tasumi, M., Svnrh. Ale/. 1992, 49, 335-340. 

[58] Hirschfeld, T., Chase, B., Appl, Spectrosc. 1986, 40, 133-137. 

(591 Fujiwara. M., Hamaguchi, H., Tasumi, M., Appl. Spectrosc. 1986, 40, 137-139. 

1601 Furukawa, Y., Ohta, H.. Sakamoto, A,, Tasumi, M., Spectrochirn. Actrr 1991,474 1367-1373. 

[hl] Trotter, J.. ilctri Cryst. 1961, 14. 1135-1340. 

[62] Hargreaves, A,, Rizvi, S. H., Act0 Cryst. 1962, 15, 365-373. 

[63] Baudour, J. L., Cailleau, H., Yelon, W. B., Acta Cryst. 1977, B33, 1773-1780. 

[64] Delugeard, Y., Desuche, J., Baudour, J. L., Actci Cribst. 1976, B32, 702-705. 

[65] Tabata, M., Satoh, M., Kaneto, K., Yoshino, K., J. Phj?,~. Soc. Jpn. 1986, 55, 1305-1310. 

[66] Satoh, M., Tabata, M., Uesugi, F., Kaneto, K., Yoshino, K., Sj’nth. Met. 1987. 17, 595-600. 

[67] Gillam, A. E., Hey, D. H., J. Chern. Soc. 1939, 1170-1177. 

1681 Dale, J., rlctu Clien?. Scnnd. 1957, 11, 650-659. 

[69] Balk, P., Hoijtink, G. J., Schreurs, J. W. H., Rec. Trrrv. Chiin. 1957. 76, 813-823. 

[70] Furukawa. Y., Ohtsuka, H., Tasumi, M., SJnth. Met. 1993, 55, 516-523. 

[71] Fichou, D., Horowitz, G., Xu, B., Garnier, F., Synth. Met. 1990, 39, 243-259. 

[72] Furukawa, Y., Yokonuma, N., Tasumi, M., Kuroda. M., Nakayama, J., Mol. C~J’.S/. Lfq. 

Crysf. 1994, 256, 113-120. 

[73] Yokonuma, N., Furukawa, Y., Tasumi, M., Kuroda, M., Nakayania, J., C/ie/n. Phys. Lett. 

Chein. Pliys. Lett. 1996, 255, 431-436. 

[74] Deussen, M., Bassler, H., Chenz. Phjx 1992, 164, 247-257. 

[75] Sakamoto, A,, Furukawa, Y., Tasunii, M., J. Phys. Chem. 1994, 98, 4635-4640; J. P/ij,x 

Chenz. 1997, 101, 1726-1732; Furukawa, Y., Sakamoto, A,, Tasunii, M. Mucrornol. Symp 

1996, 101, 95-102. 

[76] Spangler, C. W., Hall, T. J., Synth. Met. 1991, 44, 85-93. 

[77] Dewar, M. J. S.,TrinajstiC, N., Czech. Cheni. Cowviitm. 1970, 35. 3136-3189. 

[78] Zahradnik, R., Carsky, P., J. Phys. Chen7. 1970, 74, 1240-1248. 

[79] Furukawa, Y., SJint/7. Mct., 1995, 69, 629-632; J. Plrys. Chenz. 1997, 100, 15644-15653. 

[SO] Shacklette, L. W., Chance, R. R., Ivory, D. M., Miller, G. G., Bauglman, R. H., Spth. Met. 

1979. 1, 307-320. 

[81] Krichene, S., Lefrant, S., Froyer, G., Maurice, F., Pelous, Y., J. Plzys Colloq. 1983, 44, 733- 

736. 

[82] Krichene, S., Buisson, J. P., Lefrant, S., Slvth. Met. 1987, 17, 589-594. 

[83] Furukawa, Y., Ohtsuka, H., Tasumi. M., Wataru, I., Kanbara, T, Yaniamoto, T., J. Runicin 

Spectrosc. 1993, 24, 551-554. 

[84] Ohtsuka, H., Furukawa, Y., Tasumi, M., Wataru, I., Yamanioto, T., unpublished work. 

[85] Rakovic, D., Bozovic, I., Stepanyan, S. A,, Gribov, L. A,, Solid Stare Commun. 1982, 43, 

127-129. 

[86) Zannoni, G., Zerbi, G., J. Chern. Phys. 1985, 82, 31-38. 

1871 Buisson, J. P., Krichene, S., Lefrant, S., Synrh. Met. 1987, 21, 229-234. 

[88] Ohtsuka, H., Furukawa, Y., Tasumi, M., Spectrochinq. Act0 1993, 49A, 731-737. 

[89] Tang, J., Albrecht, A. C., in: Rnniaiz Spectroscopy: Szynianski, H. [Ed.). New York: Plenum, 

1970; Vol. 2, pp. 33-68. 

[90] Inagaki, F., Tasumi, M., Miyazawa, T., J. Mol. Spectrosc. 1974, 50, 286-303. 

[91] Warshel, A., Dauber, P., J. C/ierii. Ph?~s. 1977, 66, 547775488, 

1921 Yaniaguchi, S.? Yoshimizu. N., Maeda, S., J. Phys. Clzeni. 1978, 82, 1078-1080. 

1931 Peticolas, W. L., Blazej, D. C., Chein. Plzys. Le/t. 1979, 43, 604-608. 

[94] Kakitani, H., Clienz. Phys. Lett. 1979, 64, 344-347. 

[95] Cuff, L., Kertestz, M., J. Phys. Chern. 1994, 98, 12223-12231. 

[96] Takahashi, C., Maeda, S., Clieni. Phys. Lett. 1974, 24, 584-588. 

[97] Aleksandrov, I. V., Bobovich, Ya. S., Maslov, V. G., Sidorov, A. N., Opt. Spectrosc. 1975. 

38, 387-389.


[98j Matsunuma, S., Ydmaguchi, S., Hirose, C., Maeda, S., J. Pliys. Chem. 1988, 92, 1777-1780. 

[99] Furukawa. Y., Ueda, F., Hyodo, Y., Harada, I., Nakajima. T., Kawagoe, T., Macro- 

7nolecule.s 1988, 21, 1297- 1305. 

[loo] Sakamoto, A,, Furukawa, Y., Tasumi, M., J. P/zJ~s. C/7eni. 1992, 96, 3870-3874. 

[lo1 J Sakamoto, A., Furukawa, Y., Tasunii, M., SJW//Z. Mct. 1993, 55, 593-598. 

[I021 Sakamoto. A,, Furukawa, Y., Tasumi, M., Noguchi, T., Ohnishi, T., Syrrtl7. Met. 1995. 69, 

439-440. 

[I031 Harada, I.. Furukawa, Y.. Tasunii, M., Shirakawa, H., Ikeda, S., J. Clzeni. Phys. 1980. 73, 

1746-4757. 

[I041 Kuzmany, H., Plzys. Stat. Sol. B 1980, 97, 521-531. 

[lo51 Furukawa. Y., Harada, I., Tasumi. M., Shirakawa, H., Ikeda, S., Clzetri. Lett. 1981, 1489- 

1492. 

[lo61 Schiigerl, B., Kuzniany, H., Phys. Stat. Sol. B 1982, 1//, 607-617. 

[I071 Faulques, E., Lefrant, S., J. P/i>i.r. Colloq. 1983, 44, 337-340. 

[lo81 Eckhardt. H., Shacklette, L. W., Szobota, J. S., Bauglunan, R. H., Mol. Cryst. Liq. Cryst. 

1985,1/7,401-409. 

[lo91 Tanaka, J., Saito, Y., Shimizu, M., Tanaka, C., Tanaka, M., Bzrll. Chem. Soc. Jpn. 1987, 60, 

1595-1605. 

[110] Mulazzi, E., Lefrant, S., Swtlr. Met. 1989, 28, D323-D329. 

11111 Furukawa, Y., Uchida. Y., Tasumi, M., Spangler, C. W., Mol. Cryst. Liq. Cryst. 1994, 556, 

721-726; Kim, J.-Y., Ando, S., Sakamoto, A., Furukawa, Y., Tasumi, M., Synth. Met. 1997, 

89, 149-152. 

[lly Brtdas, J. L.. Scott, J. C., Yakushi, K., Street, G. B.? Phy. Reu. B 1984, 30, 1023-1025. 

[I131 Yakushi, K., Lauchlan, L. J., Clarke, T. C., Street. G. B., J. Chm?. Plzys. 1983, 79, 4774- 

4778. 

11141 Nechtschein, M., Devreux, F., Genoud, F., Vieil, E., Pernaut, J. M., Genies, E., Synth. Met. 

1986, 15, S9-78. 

[ 1151 Chung, T.-C., Kaufman, J. H., Heeger, A. J., and Wudl, F., Phys. Rev. B 1984, 30, 702-7 10. 

[I161 Kaneto, K., Kohno. Y., and Yoshino, K., Mol. Cryst. Liq. Crj).~l. 1985, /IS, 217-220. 

[117] Bradley, D. D. C., Evans, G. P., and Friend, R. H., Syntlr. Met. 1987, 17, 651-656. 

[I181 Voss, K. F., Foster, C. M., Smilowitz, L., MihailoviC, D., Askari, S., Srdanov, G., Ni, Z., 

Shi, S., Heeger, A. J., and Wudl, F., Phys. Rev. B 1991, 43, 5109-5118. 

[I191 Hill, M. G., Mann, K. R., Miller, L. L., Penneau, J.-F., J. Am. Cliem. Soc. 1992, 1/4, 2728- 

2730. 

[I201 Stafstrom. S., in: Corzjligated Polyrners: Brtdas, J. L. and Silbey, R. (Eds.). Dordrecht: 

Kluwer Academic Publishers, 1991; pp. 113-140. 

[I211 Kivelson, S., Heeger, A. J., Plzys. Rev. Lett. 1985, 55, 308-31 1. 

[122] Mizes, H. A., Conwell, E. M., Pl7y.s. Reu. Lett. 1993, 70, 1505-1508. 

[I231 Shimoi, Y., Abe, S., Phys. Re[). B 1994, 50, 14781-14784. 

[124] Bredas, J. L., Thtmans, B., Fripiat, J. G., Andrt. J. M., Chance, R. R., Plzys. Rev. B 1984, 

29, 6761-6773. 

[125] Stafstrom, S., Brtdas, J. L., Epstein, A. J., Woo. H. S., Tanner, D. B., Huang, W. S., Mac- 

Diarmid, A. G.. Phys. R ~P. Lett. 1987, 59, 1464-1467. 

11261 Stafstrom, S., Bredas, J. L., Phys. Reu. B 1988, 35, 4180-4191. 

11271 Brtdas, J. L., Beljonne, D., Shuai, Z., Toussaint, J. M., Syntlz. Met. 1991, 43, 3743-3746. 

[I281 Tanaka, C., Tanaka, J.. Syntlz. Met. 1993, 57, 4377-4384. 

[ 1291 Sariciftci, N. S. (Ed.) Primary Plzotoescitatioiis I/7 Conjugated Polymers: Molecular Evciton 

I)P~SIISemicoiithrc~tor Bud Model. Singapore: World Scientific, 1997, Chapter 17.


5 Vibrational Spectroscopy of Polypeptides 

S. Kriinm 


The polypeptide chain, (NHCH(R)CO),l, is the basic backbone chemical unit of 

proteins, and knowledge of its possible structures and dynamics is crucial to an 

understanding of the functions of these important biological macromolecules. In 

synthetic polypeptides, all the side chains, R, are usually the same; in proteins, they 

can be any of 20 ainino acid residues whose specific sequence is determined by the 

base sequence in the DNA of the corresponding genome [I]. The unusual property 

of these chains is their ability to adopt a wide range of three-dimensional structures 

121, which accounts for the richness of biological functions that they can perforin [f]. 

This presents a special challenge to mastering the spectroscopy of these polymeric 

molecules. 

While vibratioiial spectroscopy is not capable of the structural resolution of X- 

ray diffraction, it nevertheless has some important advantageous features. First, it 

is not generally limited by physical state: samples can be in the form of powders, 

crystals, films, solutions, membranous aggregates, etc. Second, a number of different 

experimental methods probe the structure-dependent vibrational modes of the 

system: infrared (IR), Raman (both visible and UV-excited resonance), vibrational 

circular dichroism, and Raman optical activity, many of these with time-resolution 

capabilities. Finally, in addition to providing structural information, vibrational 

spectra are sensitive to intra- and intermolecular interaction forces, and thus they 

also give information about these properties of the system. 

To gain such insights, however, requires that we achieve the deepest possible 

understanding of the vibrational spectrum. Correlations using characteristic band 

frequencies, particularly with resolution-enhanced spectra [3], can be usefd, but it is 

increasingly being recognized that this is a fundamentally limited approach [4]. For 

the kind of understanding we seek it is necessary to determine the detailed normal 

modes of the molecule. Only by achieving an accurate description of these modes 

can we expect to explain fully all features of the vibrational spectrum and thereby to 

validate the structure and force field inputs to the normal-mode calculation [S]. This 

is now beginning to be possible for the polypeptide chain. 

In this chapter we review the present state of the calculation of accurate normal 

modes of polypeptides. Since the burden of such calculations falls primarily on the 

quality of the force fields, we first review the state of present methods for obtain-

240 5 l i’hrrrtionnl Spectroscopy of Polypeptides 

ing such force fields, which include empirical, ~ i h initio, and molecular mechanics 

approaches. We then discuss applications to characterizing the aniide modes of 

the peptide group. Finally, we describe some of the results that have been obtained 

on characteristic polypeptide conformations, namely extended-chain and helical 

structures. 

5.2 Force Fields 

The normal modes of a molecule are determined by its three-dimensional distribution 

of atomic masses and its intra- and intermolecular force fields. The spectroscopic 

validation of the structure and these interactions derives from a satisfactory 

prediction of the frequencies and intensities of the observed IR and Raman bands, 

with the proviso that the bands are appropriately assigned to the calculated eigenvectors. 

Because of the complexity of the spectra this part of the process cannot be 

treated lightly, since more than one band match-up may be possible. A number of 

methods are available to verify band assignments [5] (isotopic substitution, dichroism, 

etc.), but since these are usually limited in practice, the success of the normal 

mode approach will ultimately depend on the reliability of the force fields. We 

therefore devote part of this section to a discussion of the development of such 

accurate force fields. This development has proceeded primarily through empirical 

refinement of force constants but is relying increasingly on inputs from ab initio 

calculations and will probably ultiinately be based on inolecular mechanics (MM) 

energy functions. 

The development of an empirical force field requires selecting a physical niodel 

for the potential, choosing which terms are important, and optimizing the force 

constant parameters by a least-squares fitting of calculated normal modes to 

observed bands. (Since there have been summary [5] as well as detailed [6-8] 

treatments of the methods for obtaining the frequencies (eigenvalues) and forms 

(eigenvectors) of the normal inodes for a given force field, we will not repeat this 

formalism here but will refer only to those aspects that are relevant to our discussion.) 

It is useful to recall that while the computational challenge becomes quite 

demanding for polypeptide structures lacking symmetry, such as globular proteins, 

for those structures that have helical or translational symmetry, such as the helical 

and extended-chain forms of synthetic polypeptides, the calculational problem is 

much simplified. This is because the only modes that can exhibit IR or Raman 

activity are those in which equivalent vibrations in each helix unit differ in phase by 

0, &[ , or ?2

5.2 FOYCC Fields 241 

such calculations can only be done for relatively small molecules, and the force 

constants generally need to be scaled so that calculated frequencies agree with 

experimental data. Nevertheless, valuable information can be obtained by this 

approach. 

The MM method will be seen to be the most useful one for obtaining force constants 

that can be used for different conformations of the molecule. In this approach 

the force constants are obtained from the second derivatives of an assumed potential 

energy function consisting of quadratic bonded terms and non-quadratic nonbonded 

terms. Although present MM functions are too crude to be spectroscopically 

reliable, a new method for deriving such functions holds the promise of 

providing a reliable vibrational force field for the polypeptide chain. 

5.2.1 Empirical Force Fields 

The most general force field of a molecule would include anharmonic as well as 

harmonic terms. However, with the limited experimental information generally 

available for refining an empirical force field for complex molecules, the harmonic 

approximation is the only feasible one at present. This means that, for the isolated 

molecule, we need to know the force constants, FQ, in the quadratic term of the 

Taylor series expansion of the potential energy, V: 

where r is, for example, an internal displacement coordinate and IZ = 3N - 6, N 

being the number of atoms in the molecule. If interactions with other molecules are 

involved, analogous intennolecular interaction energy terms must be included. 

Since the maximum number of observable frequencies is n, whereas the general 

number of F,s is n(n + 1)/2, it is necessary to find a physically meaningful model 

for T7 that brings the number of empirically determined F,Js into reasonable relation 

to the number of observable frequencies, even including those of isotopic derivatives. 

Two main models have been used to accomplish this: the Urey-Bradley 

force field (UBFF) and the general valence force field (GVFF). Both models incorporate 

the same diagonal (i.e., Fll) valence-type ternis, involving bond stretch, angle 

bend, and torsion coordinates, but differ in how they treat the off-diagonal (i.e., F,,) 

terms. 

In the unmodified UBFF, elaborated by Shimanouchi [lo], off-diagonal terms in 

T’ are represented by atom pair 1,3 non-bonded interactions. Because this introduces 

inherent redundancies in the coordinates, the equilibrium configuration of the 

molecule represents a state with internal tensions, and therefore linear terms must 

be included in the force field. By introducing the redundancy condition, the linear 

terms can be eliminated from V, although the coefficients of the quadratic terms 

then contain ‘intramolecular tension’ as well as the valence and nnn-bonded force 

constants. The important characteristic of such a UBFF is that only a limited

242 5 Vibrcitional Spcc.ii.o.scoyj~ of Po1jpt.ytitlt.s 

number of valence-type off-diagonal terms are in effect included in the potential. 

Such simple UBFFs can therefore be physically inaccurate, and as B result they 

have had to be modified by the explicit inclusion of some additional F;,s [lo]. 

In the GVFF there is no inherent limit to the number of Fi,s that are included. 

There is only the restraint that the number of Fs should not exceed the number or 

observable frequencies; in fact, it should generally be much smaller so as to overdetermine 

the force field. Since an independent set of coordinates can be chosen, 

e.g.. local symmetry coordinates, and the redundancy conditions explicitly given. 

there is no need to include linear terms in V, and Eq. (51) is the most general 

representation of the force field. Our discussion will focus on applications of the 

GVFF. 

The refinement of an empirical force field for a niacroinolecule usually starts with 

the vibrational analysis of a smaller model system of known structure followed by a 

transfer of these force constants to the larger system, with possible subsequent force 

constant adjustments. In the case of the polypeptides, the preferred model system 

for the peptide group has been trans-N-methylacetamide (NMA), CH3CONHCHi. 

Synthetic polypeptides of known structure such as P-sheet polyglycine (PG), R = H, 

and a-helical and P-sheet poly(L-alanine) (PLA), R = CH3, have served as basic 

examples of the polypeptide systems. 

The force field refinement of the model system introduces important requirements 

and constraints [5, 7, 81. 

1. The types of F, to be included in V must be selected. In practice, this has been 

based on prior experiences with the GVFF, but nevertheless the choices are, in 

general, arbitrary. 

2. A starting set of force constant values, from which are calculated a set of normal 

coordinates, Pa, and normal frequencies, vx, needs to be chosen. These are 

refined to observed data by a least-squares optimization procedure. We note that 

a normal mode Q.* can be visualized in terms of the local, usually symmetry ($1, 

coordinates since 

S;=CL 1% p % 

1 

(5-7) 

and the normal mode calculation provides the elements L,, of the eigenvector 

matrix. Thus, for a given QLy the relative S, are given by the relative Lla. Visualization 

in Cartesian coordinates is also possible. An analogous description is given 

by another characteristic of the normal mode, namely the potential energy distribution 

[PED). This consists of the relative contributions of each S, to the total 

change in potential energy during QE, the fractional contribution of a particular 

S, being given by 

where I,, = 47c'c'v~. vu in cm-'. (Contributions from F,j elements are typically 

small and ai-e usually neglected. However, they can be large and negative, which

5.2 Force Fields 243 

accounts for occasional unusually large contributions in a PED. In such a case, 

the diagonal elements alone do not give a sufficiently complete description of the 

norinal mode.) 

3. Correct band assignments must be made in matching calculated to observed v,. 

This is probably the most important part of the procedure, since acceptable frequency 

agreement (usually 5-10 cm-' ) does not necessarily insure that correct 

eigenvectors have been obtained. There are a number of properties that help 

to judge the validity of band assignments: molecular symmetry, which specifies 

the activity of a mode, and in appropriate situations predicts the IR dichroisni 

(parallel, 11, or perpendicular, I, to the chain axis) and Raman polarization 

characteristics of the bands; isotopic substitution, which identifies the general 

nature of a mode, and by the reproduction of specific observed frequency shifts 

substantiates the correctness of the eigenvectors; overtone and combination 

bands, which should be consistent with the assignments of fundamentals; and 

predicted band intensities, which are very sensitive to the eigenvectors. The 

intensity agreement is particularly important since it provides a test of the 

eigenvectors that is independent of the frequencies. For example, the IR 

absorbance, A,, associated with normal mode Qa is given by [l 11 

A, = 42.2547 (g) (5-4) 

where dG/aQ, is the normal coordinate derivative of the dipole moment, viz., the 

transition dipole moment (Dk' amup'l2), and A, is in km mol-'. Since from 

Eq. (5-2) 

= Li, (g) 

(5-5) 

Qu 

we see that, through the eigenvector matrix, the IR intensities sample another 

physical property of the molecule, namely the dipole derivatives, which arise 

from the charge re-distributions during the vibrations. Such dipole derivatives 

can also provide a measure of the coordinate contributions to the IR intensity of 

a band through a dipole derivative distribution (DDD)[ 121. In addition, relative 

intensities of different symmetry species of a mode can provide information on 

structure. Thus, in a sample with a molecular chain axis aligned (say mechanically) 

in a known direction, since the orientation of a specified dc/aQ, with 

respect to the chain axis can be related mathematically to an experimentally 

observed dichroic ratio, DR-A,( ll)/Av(L), a DR measurement can help to 

establish the dG/L7Q, orientation and therefore provide a test of a structural 

hypothesis. 

The refinement of the macromolecular force field from that of the model system 

follows concepts similar to those described above, with three important additions. 

The first, and simplest, is that since the structures are different, new kinds of force

244 5 Vibrational Spectroscopy oj Polypeptides 

constants must be introduced. This is not a major problem since initial values, for 

example for polypeptide side chains, are generally available. Second, the macromolecular 

system generally experiences inter- as well as intramolecular interactions, 

and these have to be incorporated into the force field. For polypeptides, these include 

hydrogen bonds and non-bonded interactions. Third, since for polypeptides in 

particular we are especially concerned with describing different molecular conformations, 

attention has to be given to incorporating the conformation dependence 

of the force constants. This is not easy to do in the fi-amework of an empirical force 

field. 

With respect to the intermolecular interactions, the hydrogen bonds can be 

treated like the other internal coordinates, i.e., in term of bond stretching, angle 

bending, and torsion force constants. However, since hydrogen bonds vary in their 

geometry, we should ideally know how these force constants depend on the structure 

of the bond. While preliminary studies have been done for the NH hydrogen 

bond of the peptide group [13], there is at present no comprehensive description of 

this relationship. 

The itnportant non-bonded interactions are of two kinds, dispersion and transition 

dipole moment interactions. The former correspond to the so-called van der 

Waals forces and are generally given by a Lennard-Jones type of potential, viz. 

where R,, is the distance between atoms i and j, the exponent of the repulsive term 

is traditionally 12 (although 9 has been found to be more effective in some cases), 

the exponent of the attractive term is for basic reasons 6, and A, and B, are respectively 

the repulsive and attractive van der Waals parameters of atom i. Such 

potentials have generally not been added to a GVFF for polypeptides since, in 

distinction to polyethylene [ 141, they do not affect midrange and high frequencies 

significantly. However, they probably are important in describing low-frequency 

modes. 

Transition dipole moment interactions between peptide groups can influence 

higher frequency modes, as was shown by their effects on the splittings of amide I 

and amide I1 modes of a-helix and P-sheet polypeptides [15, 161. Such transition 

dipole coupling (TDC) arises from the potential energy of interaction, V&, between 

transition moments, Aji, in different peptide groups, and is given by [17] 

where LY4, is a geometrical factor given by 

with E being the direction of the transition moment and YAB the distance between the


centers of the moments. In terms of normal coordinate dipole derivatives, 

A@ = (4.1058/~:’~) ~ 

(5-9) 

sg 

(3 P a 

where v, is the unperturbed frequency. In view of Eq. (5-51, a calculation of Vlfcl will 

require a summation over terms in 1 dcA/dSi I 1 ?yB/?5’, 1 . The force constant associated 

with F’(/(/ is 

in mdyn k’, and if we take the effect of TDC as a perturbation on va, the frequency 

shift is given by 

(5-1 1) 

Early applications of this theory [18, 191 parametrized d@/dQ from observed band 

splittings. Ah initio calculations of these dipole derivatives [20] showed that the 

empirical TDC parameters were in agreement with nb initio as well as IR intensity 

results [ 171, thus emphasizing that the TDC mechanism is an important intermolecular 

interaction in polypeptides, particularly for modes with large d@/aQ such as 

amide I and ainide 11. 

The matter of the conformation dependence of the force field of a polypeptide 

is one that has received little attention. We know that the differences in vibrational 

frequencies of different conformers derive mainly from the differences in their threedimensional 

structures, with force constant differences playing a secondary role. 

Nevertheless, certain spectral features cannot be understood without taking into 

account variations in force constants with conformation (a good example is the 

tvans versus gnzrche angle-angle interaction force constant [213). Independent 

empirical refinements of a-PLA and /I-PLA also demonstrate such force constant 

differences [5]. The problem is how to express such variations in a general explicit 

form in V. This is difficult to achieve simply from knowledge of independently refined 

force fields of different conformers. We return to this question below in our 

discussion of MM force fields. 

Empirical force fields have been developed for NMA, for PG, and for CY-PLA and 

/I-PLA. In the case of PLA, it has also been possible to refine an ‘approximate’ force 

field in which the CH3 side chain is replaced by a point mass (of 15 amu) [22]. This 

is useful in studying modes of the backbone chain that do not couple significantly 

with side-chain modes. 

5.2.2 Ab Iizitio Force Fields 

Ab initio molecular orbital calculations, by giving the potential energy surface of 

a molecule at its minimum energy conformations, provide a complete vibrational

246 5 Vibrationcil Spectroscopj.1 of Pol.vpeptides 

force field, i.e., all diagonal and off-diagonal force constants [23]. Since the force 

constants are given by the second derivatives of the potential in Cartesian, and 

therefore non-redundant, coordinates, the force field is unique. Transformation is 

then always possible into an internal or local symmetry coordinate basis. Dipole 

and polarizability derivatives are also obtained from the ah iiiitio calculation, and 

therefore IR and Ranian intensities can be predicted. 

Aside from the previously mentioned problem of doing such calculations on very 

large molecules, there are still some limitations at present in directly implementing 

ab iriitio force field calculations on model systems. At the Hartree-Fock (HF) 

level, force constants can be 10-30% too large because of basis set limitations and 

the neglect of electron correlation. This necessitates the empirical scaling of force 

constants so that calculated frequencies agree with observed. While systematic 

scaling procedures have been developed [24], there is still variability in the number 

and kind of scale factors that are used. In any case, a careful comparison of basis 

sets in terms of their geometry and frequency predictions is still very important. For 

example, the structure of NMA has non-planar symmetry at the HF/6-31G* level, 

but it exhibits planar symmetry at the HF/6-31 +G* level, i.e., with the addition 

of diffuse functions [25]; some well-assigned normal modes of malkanes cannot 

be reproduced in the proper order at the scaled HF/6-31G* level but are correctly 

calculated with a scaled HF/6-31G basis set [26]. With the inclusion of electron 

correlation, for example by the Mdler-Plesset (MP) perturbation method, the force 

constants are much closer to experiinentally compatible values, and such calculations, 

though very computer-intensive as the molecule gets larger, can provide 

usable force fields with minimal scaling. 

Despite these problems, ab irzitio force fields provide an improved route to 

avoiding the arbitrariness and incompleteness of empirical force fields. Although 

such calculations on model systems do not translate into force fields for macromolecules, 

they can provide important components of such force fields. For example: 

an ab initio force field for glycine hydrogen bonded to two HzO molecules [27] 

provided the necessary starting point for refining the carboxyl group part of the 

empirical force field for glutathione [28]; an ah iiiitio force field for diethyl disulfide 

[29] permitted detailed correlations to be made between SS and CS stretching frequencies 

and the conformations of disulfide bridges in proteins [30, 311; an ab iiiitio 

force field for ethanethiol, together with an empirical force field for the peptide 

group, provided detailed structure-spectra correlations between CS and SH stretch 

frequencies and the conformation of the cysteine side chains in proteins [32]. 

Ab iizitio force fields also demonstrate that force constants vary with conforination, 

a factor that needs to be taken into account if we are to obtain an accurate 

description of the normal modes associated with the many possible conformations 

of the polypeptide chain. This is clearly illustrated by a study of the scaled ah irzitio 

force fields of four iion-hydrogen-bonded conformers of the alanine dipeptide, 

CH3CONHCaH(CPH3)CONHCH3 [33, 341. It was found that diagonal force constants 

such as the torsion (t) change by very large amounts, and even the stretch (s) 

and deformation (d) force constants are significantly dependent on the conforniation. 

Off-diagonal force constants can also change significantly with conformation. 

In principle this is not surprising since we must expect changes in electronic struc-


ture with conformation, In terms of conveniently designated components of the 

energy of the molecule, which of course the ah initio calculation does not provide, 

we can conceive of these changes as arising from changes in the charge distribution 

with conformation, from changes in the dispersion energies as interatomic distances 

change, and from any iiitrinsic valence force constant variations with conformation. 

While we can show from ab initio calculations that the force field is not independent 

of molecular conformation, it is much more difficult to specify the explicit 

nature of this dependence. This is precisely because the method provides only a 

total energy for a given structure. If we wish to infer the varying contributions of 

different physical components, we must assume a model for the potential energy 

and strive to make its behavior mimic the ah initio results as closely as possible. It is 

toward this goal that some present efforts are directed. 

5.2.3 Molecular Mechanics Force Fields 

An MM energy function, whose second derivatives at a minimum are the spectroscopic 

force constants, is generally represented as a sum of quadratic and nonquadratic 

terms. The foimer encompass the valence-type deformations, such as 

bond stretching and angle bending, while the latter involve torsions, dispersion 

interactions, electrostatic interactions, and possible hydrogen-bond terms. 

In order to serve as a spectroscopic force field, such an MM function would at 

least have to reproduce vibrational frequencies to spectroscopic standards, viz., to 

root-mean-square (rms) errors of the order of 5-10 cnir'. Some early MM potentials 

came close to this mark [35], but since frequency agreement was not their primary 

goal (but rather reproduction of structures and energies), subsequent efforts 

resulted in functions with unacceptable spectroscopic predictability (rms errors of 

>50 cm-'). One of the reasons for this was the lack of proper attention to crossterms 

in the potential. Initial efforts to improve such potentials have led to only 

marginally better frequency agreement. For example, MM3 gives an rms error for 

frequencies below 1700 cnir' of 38 cm-' for butane [36a] (reduced to 22 cm-' in 

MM4 [36b]) and 47 cm-l for NMA [37]; a Hessian-biased force field [38] gives a 

similar rms error of 29 cn1-l for polyethylene [39]; for CFF93, the ims error for n- 

butane is 25 cmrl 1401 and is 34 cnrl for a group of four acyclic and three cyclic 

hydrocarbons; and a second generation version of AMBER gives 43 cm-' for 

NMA [41]. Some force fields have been specifically designed to give improved 

spectroscopic predictability and these [42, 431 indeed show better agreement, although 

improvement is still desirable: the rms error for the above NMA frequencies 

is 24 cn-l for SPASIBA [44]. 

What is clearly needed is an approach that systematically incorporates spectroscopic 

agreement in the initial stages of the optimization of the MM parameters. 

This has been achieved by a procedure designed to produce a so-called spectroscopically 

determined force field (SDFF) [45, 461. 

The elements in the SDFF methodology are the following. First, a form is selected 

for the MM potential, one that is hopefully inclusive enough to incorporate all the

248 5 Vibrational Spectroscopy qf Polypeptides 

physically important contributions. The first iinplementation of this approach, viz., 

for linear saturated hydrocarbon chains [47], used a potential of the form 

where the y,s are internal coordinates whose intrinsic (i.e., equilibrium) values are 

ql0 and the flJ are the intrinsic MM quadratic (i.e., valence-type) force constants. At 

this stage all possible cross-terms are included. The intrinsic torsion potential is 

represented by a Fourier cosine series, 

(5-13) 

where V, is the barrier to a rotation of periodicity m associated with torsion angle 

x. (In general, Vr,,. is a sum over terms in a Fourier series, but in the hydrocarbon 

case only the V3 term was found to be necessary.) The non-bonded potential consists 

of the dispersion term (an exponent of 9 was found to be more suitable for the 

repulsive term) plus a coulomb teim that is usually represented by the interactions 

between partial charges Qi and Q, on atoms i and j, i.e., in the case of the hydrocarbons 

(5-14) 

where K is the dielectric constant and EO the permittivity of free space. The summation 

of Vf& is over all atom pairs 1,4 and higher. The Q1 can be fixed charges or they 

can include charge fluxes, i.e., dQ1/dSJ [42]. (Inclusion of charge fluxes also permits 

the calculation of TR intensities.) 

The second element in the procedure is the choice of a spectroscopic force field, 

i.e., a set of F,J. This could be an empirical force field, but since we wish to provide 

as complete a description as possible at this stage, a scaled ab initio force field is the 

one of choice. The scaling process also connects the force constants to experimental 

frequencies and band assignments, and avoids the problems of having incorrect 

eigeiivectors because of using inappropriate basis sets [26]. 

The most important part is the third element, viz., the ability to make a transformation 

from the spectroscopic to the MM force field [46]. Since this transformation 

is analytic, it preserves in the SDFF the frequency as well as structure 

agreement of the scaled ah initio calculation. Although the transformation is initiated 

by assuming a starting set of p:lt, parameters, these are subsequently optimized 

in the refinement procedure [48]. In addition to providing the flJ, the transformation 

procedure also gives the ql0 [46]. 

The fourth, and also important, element is the application of this transformation

5.3 Amide Modes 249 

to the nh initio force fields of a set of conformers of the model molecules for the 

macromolecular system. In the case of the hydrocarbons, this consisted of the four 

stable conformers of M-pentane and the ten stable conformers of n-hexane [26, 471. 

At this stage, the non-bonded parameters are optimized under the standard MM 

assumption that the intrinsic force constants (fl,) and intrinsic geometry parameters 

(ql0) are the same for all conformers 1481. The relative (ah initio) conformer barriers 

and energies are incorporated in optiniizing the V, and Q1. At this point, specific 

forms of the conformation dependence of some fiJ can be determined and incorporated 

in V. 

Finally, since it is neither desirable nor necessary to include in I/ the huge number 

off,,, most of which are very small and do not significantly affect the vibrational 

frequencies, the set of force constants is systematically reduced based on a preassigned 

limit on the frequency error [46]. In this process, the remaining fl, are 

optimized, usually requiring minimal change, to compensate for the effects of the 

force constants that were dropped. 

There are a number of advantages to the SDFF approach to refining an MM 

force field. By utilizing a complete (i.e., ab iizitio) spectroscopic force field scaled 

to experimentally assigned bands, frequency agreement is assured at the start. This 

not only incorporates highly accurate information about the minima in the potential 

surface, but it avoids the possibility of biasing the non-bonded parameters to 

compensate for an incomplete formulation of the fl, terms in V. In addition, since 

the optimization is done in a nonredundant coordinate basis [49], the uniqueness of 

the force field is assured; subsequent transformation to a redundant basis of choice 

is always possible. Finally, since force fields for many different conformers are 

involved in the optimization, there is a better chance of revealing the nature of 

specific conformation dependences of the cross-terms. 

The application of this procedure to an SDFF for the hydrocarbon chain has 

given excellent results [47]. With a set of 50 force constant and geometry parameters 

the observed frequencies of tt-pentane and ttr-hexane below 1500 c1n-l could be 

reproduced with an rms error of N 11 cm-'. This force field has now been extended 

to include branched saturated hydrocarbons [50]. An SDFF for the polypeptide 

chain is now under development [51]. 

5.3 Amide Modes 

Since the peptide group, -CONH-, is the most distinctive component of the backbone 

of the polypeptide chain, it is not surprising that the normal modes of this 

group, the so-called aniide modes, were first studied in the simplest representative 

molecule, viz., NMA. It is useful to examine the nature of these modes in NMA and 

in blocked dipeptides since they have served as important guideposts in the analysis 

of polypeptide spectra in terms of structure.

250 5 Vibuntiorid Si7ectvoscop.y of Polypeptides 

5.3.1 N-Methylacetamide 

Normal mode calculations on NMA have been done with both empirical and scaled 

ab iizitio force fields. In the empirical force field calculations, the CHj group has 

been treated as a point mass, using a UBFF [52] as well as a GVFF [51, and as an 

all-atom structure, again with a UBFF [53] and a GVFF [54, 551. In the ah iizitio 

calculations, a variety of basis sets and scaling methods have been used: 4-31G, 

with adjustment of mainly diagoiial force constants to liquid-state fi-equencies of 

NMA and its isotopomers [56]; 4-31G and 4-31G*, with no scaling [57]; 4-21, with 

four general scale factors and some approximated force constants [58]; and 4-31G*, 

with 10 scale factors [59], optimized to matrix-isolated frequencies [60]. 

All of the above calculations have been on the isolated molecule, even though in 

many cases comparisons have been made with experimental results on hydrogenbonded 

systems (such as the neat liquid). The first noimal mode calculation on 

a hydrogen-bonded system was at 4-31G* on NMA-(H20)2, in which one H2O 

molecule was hydrogen bonded to the CO group and the other to the NH group, 

with scale factors optimized to IR and Raman data on the aqueous system [61]. In 

view of the absence of infomiation on the detailed water structure associated with 

the NMA molecule in aqueous solution, this was considered to be a satisfactory 

minimal model for representing the normal modes of hydrogen-bonded NMA. 

(A 4-31G calculation 011 NMA-(H20)3 has also been done [62].) Subsequently, a 

3-21G calculation was done on NMA-(fomiamide)z, NMA-(FA)?, using six 

general scale factors [63]. In view of the determination that diffhe functions are 

necessary to give a planar symmetric peptide group in the isolated molecule 1251, 

the NMA-(H20)2 calculations have been repeated at 6-31 + G* [64], and these are 

discussed below. 

Early studies [52, 65, 661 sought to describe the characteristic vibrational modes 

of the peptide group. In addition to the localized NH s mode, which is subject to 

Fermi resonance interactions [5], there are other relatively localized modes, labeled 

amide I-VII (although it was recognized [66] that some of these, because of their 

delocalization, would be more variable than others). These amide modes have been 

used as a basis for discussions of peptide spectra, and in Table 5-1 we describe them 

and compare the PEDs for isolated and aqueous hydrogen-bonded NMA. It should 

be noted that these results are on fully optimized structures (all frequencies real), 

in both cases the conformation being N,Ct. (The symbol indicates that the in-plane 

N-methyl hydrogen is cis to the (N)H and the in-plane C-methyl hydrogen is trans 

to the (C)O.) 

The amide I mode is primarily CO s, both in NMA and NMA-(H20)2. However, 

the additional contributions to the eigenvector depend very much on the 

specifics of the calculations (only contributions to the PED of 2 10 are given in 

Table 5-l), as well as, of course, on the absence or presence of hydrogen bonding. 

For NMA, the next largest contributor in a CH3 point-mass calculation is CN s 

[52, 51, whereas in an all-atom calculation it is CCN d [55, 591. The situation for 

NMA-( H20)2, where the lower frequency is an obvious consequence of hydrogen 

bonding, is complicated by the experimentally demonstrated coupling of CO s 

and HOH b 167, 681, which makes the latter the next largest contributor. In NMA-

5.3 Ainide Modes 251 

Table 5-1. Amide modes of isolated and aqueous hydrogen-bonded N-Methylacetamide" 

Mode v(obsib v(calc)c Potential energy distributiond 

I 1706 

1646 

1626 

I1 1511 

1584 

111 1265 

1313 

IV 658 

632 

V 391 

(7501" 

v1 626 

(600)' 

VII ~ 

195g 

1708 

1632 

1626 

1512 

1575 

1264 

1305 

648 

640 

391 

744 

627 

605 

- 

20s 

CO ~(83) CCN d(l1) 

CO s(30j [O)HOH b(30) fH)HOH b(18) 

CO s(38) (0)HOH b(28) (H)HOH b(1S) 

NH ib(5l) CN s(28) 

CN s(48j NH ib(36) 

NH lb(21) CO lb(21) CN ~(18) CC S(10) CCH3 sb(l0) 

NH lb(26) CN ~(20) CCHi sb(20) CO lb(12) 

CC ~(36) CO ib(34) 

CO ib(35) CC s(32) 

CN t(118) NH ob(51) CO ob(23) 

CN t(43) H 0 b2128) NH ob(21) 

CO ob(67) CCH3 or(26) CN t( 11) 

CO ob(50) 0 . H b2(30) CCH3 OI(19) NH ob(l0) 

CN t(38) NH ob(21) 

First line: NMA; second line(s): NMA-(H20)2. 6-31+G* basis set. 

NMA: from [59]; NMA-(H20)2: from 1611 and 1681, except where noted. 

NMA: from [73]: NMA-(H20)2: from [64]. 

s: stretch, d: deformation. b: bend. ib: in-plane bend, sb: synmetric bend, ob: out-of-plane bend, 

b2: essentially ob, or: out-of-plane rock, t: torsion. Group coordinates are defined in [59]. Contributions 

> 10. 

Possible value, from 1691. 

Observed in neat NMA [65]. 

From 1621 (at 206 cm-' from [70]1. 

(FA)2 this contribution is NH in-plane bend (ib) [63], but such a result is uncertain 

because of the absence of experimental data on this complex and the resulting lack 

of optimization of the FA force constants. Nevertheless, we see that even for as 

localized a mode as CO s the specific nature of the eigenvector can depend on the 

details of the calculation. 

The amide I1 mode is a mixture of NH ib and CN s, with the former dominating 

in NMA and the latter, at a higher hydrogen-bonded frequency, dominating in 

NMA-(HzO)?, (Table 5-1). It is interesting that the CN s predominance does 

not appear at 3-21G for NMA-(FA)> [63], nor at 4-31G [62] or 4-31G" [61] for 

NMA-(H20)2, showing the evident influence of basis set on the predicted eigenvectors. 

Nor is CN s predominant in polypeptides [5]. 

The amide I11 mode is also a mixture of NH ib and CN s, although other coordinates 

make significant contributions. Since CH3 symmetric bend (sb) is an important 

one, CH3 point-mass calculations [52, 51 distort the relative importance of 

CC s. As seen from Table 5-1, CO ib is a relatively large contributor in NMA and a

252 5 Vibvritional Spectroscopji of Polvpeptides 

small one in NMA-(HzO)?. (It is not clear why NH ib makes no contribution in 

NMA-(FA)z [63].) The position and nature of this mode is variable in polypeptides 

[5], since NH ib contributes significantly to a number of normal modes in the 1400- 

1200 cn-' region, with tlie mixing being dependent on tlie side-chain structure. 

The amide IV mode is a mixture of CO ib and CC s in NMA, although the latter 

contribution is often replaced by other coordinates in polypeptide structures [j]. This is 

an example of a peptide group mode that is not in general localized and well defined. 

The ainide V mode, on the other hand, seems to be a characteristic band of 

hydrogen-bonded peptide groups, although an unusual feature seems not to have 

been noted previously. This mode, which is generally a combination of CN t and 

NH out-of-plane bend (ob), has been assigned to a band near 725 cm-' in neat 

NMA [65, 661 and possibly -750 cni-' in aqueous NMA [69]. It is found to have a 

counterpart, the amide VII mode, calculated at -200 cm-' [54, 55, 51 and observed 

near this frequency [62, 701. However, ah iriitio calculations on an isolated molecule 

[57, 591 produce only a single CN t, NH ob mode at -400 cn-', consistent with an 

observed Ar-matrix band at 391 cnir' [71]. The increase in the NH ob force constant 

of an isolated molecule needed to give a hydrogen-bonded frequency in the 

700 cm-' region apparently divides the contributions of these coordinates between 

the two frequency regions. Isolated molecules also exclude the contributions from 

hydrogen-bond deformation modes (see Table 5-1). 

The amide VI mode is predominantly CO ob, with a significant CCH3 outof-plane 

rock (or) contribution. This, of course, cannot be represented in a CH3 

point-mass model [5]. The nature of this mode is also significantly dependent on 

conformation in polypeptide structures [5]. 

5.3.2 Blocked Dipeptides 

Although the analysis of NMA serves to introduce the representative types of 

characteristic modes of the peptide group, it does not provide insights into how 

these modes might be influenced by varying conforniational interactions between 

adjacent groups, such as can occur in polypeptide chains. For this purpose it is 

useful to examine the modes of a blocked dipeptide, such as N-acetyl-L-alanine-Ninethylamide, 

the Ala dipeptide. This was first done for just the amide I and amide 

I11 modes using a fixed empirical force field and no TDC [72], so the conclusions are 

necessarily limited. Better insights can be obtained from an ab iriirio study of four 

non-hydrogen-bonded conformers of this molecule [33]. 

In Table 5-2 we show the calculated aniide I, 11, 111, and V frequencies and PEDs 

of the four non-hydrogen-bonded conformers of the Ala dipeptide, which differ 

mainly in the values of the p(NCa), $(C"C) torsion angles (see Figure 5-1). (The 

amide IV and VI modes are no longer distinguishable as relatively pure modes, CO 

ib and CO ob contributing significantly and mixing with other coordinates in the 

region of -830-515 cm-' [33].) It should be noted that in this calculation the 4-21 

ah iizitio force constants were scaled with six general scale factors, so the frequencies 

are not necessarily representative of true experimental values. However, the changes 

in frequency should be indicative of the effects of changes in conformation.

Table 5-2. Some amide modes of four non-hydrogen-bonded conformers of the alanine dipeptidea. 

Mode 

Conformerb 

I 1719 

CO sI(78)" 

1694 

CO s2(81) 

1726 

CO sl(84) 

1679 

CO s2(84) 

1724 

CO sl(79) 

1703 

CO s2(83) 

1720 

CO s2(83) 

1710 

CO sl(82i 

I1 1550 

NH ib2(53) 

CN s2(27) 

1532 

NH ibl(61) 

CN sl(18) 

1552 

NH ib2(57) 

CN s2(27) 

1526 

NH ibl(61) 

CN sl(18j 

1545 

NH ib2(61) 

CN ~2128) 

1536 

NH ibl(68) 

CN sl(19) 

1540 

NH ib2(65) 

CN ~'(25) 

1529 

NH ibl(68) 

CN sl(22) 

I11 1265 

CN sl(28) 

CN s2(12) 

CO ib(12) 

1252 

H" b2(27) 

CN s2( 16) 

CN sl(14) 

1292 

CN s2(29) 

C'C s(22) 

NH ib2(22) 

CO ib2112) 

1253 

H" bl(33) 

CN sl(24) 

NH ibl( 11) 

1265 

CN sl(39) 

CO ibl(l7) 

CC s( 14) 

NH ibl(13) 

1247 

CN ~2126) 

H" b2172) 

H" bl(13) 

NC s(l1) 

1295 

C"C s(28) 

CN ~327) 

NH ib2( 12) 

CO ib2(11) 

1260 

Ha b2(37) 

CN sl(19) 

NH ibl(l2) 

v 672 

NH ob2(42) 

NH ob1135) 

CN t2(20) 

CN tl(17) 

NC" t(13j 

558 

NH ob2(48) 

NH obl(38) 

CN tl(12) 

CN t2(21) 

NC" t( 14) 

CO ibl(13) 

657 

NH ob2(28) 

NH ob1124) 

CN t2111) 

547 

NH obll61) 

CN tl(47) 

NH ob2(36) 

NC" t(28) 

CN tl(18) 

585 

CN t2(44) 

NH ob2(40) 

573 546 

NH ob2(40) NH obl(ll3) 

CN t2(13) CN tl(67) 

C"C s(l2) NC" ti65) 

NH obl(12) 

CO ob2( 10) 

605 

NH ob2(47) 

NH obl(451 

CN t2(29) 

CN tl(2l) 

NC' t(16) 

CO oblill) 

504 

NH obl(711 

NH ob2155j 

CN tl(S3) 

NC" t(43r 

CN t2(38) 

a From [33]. 

q. li/ - p2: -134", 38"; C(R: -92", -6"; C(L: 61", 41"; a': -162", -55". 

s. stretch. ib: in-plane bend, b: bend, ob: out-of-plane bend, t: torsion. Group coordinates are defined in [34]. Contributions 2 10.

254 5 Vibrational Spectvoscopj? of Polypeptides 

Keeping in mind that the peptide groups are not identical, because of the different 

chemical surroundings, we find that aniide I frequencies, because of their highly 

localized nature, generally follow the changes in CO s force constants with conformation 

[33]. These in turn are determined by the changes in CO bond length (the 

relationship is essentially linear [33]), which reflect the changing bonded and nonbonded 

interactions in the molecule with conformation. The larger variation in CO 

s2 (41 cm-') than in CO sl (16 cni-'1 is probably a result of its more variable conformational 

environinent adjacent to C", and therefore the varying (even if small) 

contributions of other coordinates. It is interesting that in a' the frequency order is 

reversed compared with the other conformers. 

The aiiiide I1 modes are less sensitive to conformation: the variation in either 

mode is of the order of 10 cm-', although the splittings do vary. As with amide I, 

the different frequencies of the two groups are mainly a reflection of their nonidentical 

chemical environments. 

The latter fact is demonstrated dramatically by the aniide I11 mode: the eigenvectors 

for the two peptide groups are significantly different, that of the higher fre- 

quency mode containing no PED contribution (210) of H" bend (b) although this 

coordinate is a major contributor to the lower frequency mode. Nor is NH ib a 

major contributor, as it was in NMA (in fact, it does not appear in the p2 modes). 

The large variations in frequency separation between the two amide I11 modes 

reflect the effect of conformation through H" b, since if H" is replaced by D" this 

variation disappears (the bands are now at 1294 & 1 and 1266 k 6 cin-'j [33]. This 

agrees with previous observations [5] that the amide I11 mode in polypeptides contains 

a significant conformation-dependent contribution from H" b. The interaction 

between H" b and NH ib is illustrated by the fact that in an (ND)? Ala dipeptide 

H" b occurs at 1300 and 1296 (p2), 1320 and 1272 ( a~), 

1305 and 1283 ( a~), and 

1324 and 1284 (a') cm-' [33]. The H" b coordinate also mixes with CN s to predict 

significant IR bands at 1277 cm-' in p2 and 1291 cm-' in a' [33]. Thus, with its 

added sensitivity to side-chain composition [5], amide I11 in polypeptides is no 

longer a simple characteristic NH ib, CN s mode. 

The amide V mode, which appears as a single band near 400 cn-' in NMA (see 

Table 5-l ), is split in the nonhydrogen-bonded conformers into at least two bands 

with predominant NH ob, CN t contributions. (These paired coordinates can make 

significant contributions to other modes, and each contributes individually to still 

others.) It might be thought that this splitting (and higher average frequency) is 

a result of the arbitrary scale factors that were used [33]. However, 4-31G" calculations 

with scale factors transferred from NMA [59] give similar results [73]. Thus, 

we must conclude that kinematic as well as potential coupling between the two 

peptide groups is responsible for the nonlocalized contribution of these coupled 

coordinates. 

The presence of hydrogen bonding obviously introduces a new element into the 

modes of the Ala dipeptide. This can be examined from the results on the three 

stable hydrogen-bonded conformers, C5, C7E, and C7A (Figure 5-1) [34]. In Table 

5-3 we show their calculated amide I, 11, 111, and V frequencies and PEDs. 

The relative amide I mode frequencies again generally follow the relative values 

of the CO s force constants [34], although mixing of the two CO s displacements can

5.3 Amidc Modes 255 

Figure 5-1. Optimized confonxatioiis 

of Ala dipeptide 

(p, ti/ in parentheses). Peptide 

groups-1: C6N7, 7: C12N17. 

Hydrogen bonds are indicated by 

broken lines. (From [33]) 

alter this: for C5, even though the CO sl force constant (1 1.040) is larger than the 

CO s2 force constant (10.922), the predominantly CO sl mode is at a lower frequency 

(1 668 cm-') than CO s2 (1 698 cm-I). This is also counterintuitive in terms 

of the (admittedly weak) hydrogen bond to C02, whereas the relative frequencies 

are in the expected order for C7E and C7A. In distinction to the nonhydrogenbonded 

structures, NH ib is now a significant component of the amide I modes of 

the hydrogen-bonded structures (C7E and C7A). (To a lesser extent, this is also true 

for the Ala dipeptide in a crystal conformation similar to that of C5 but in which 

both peptide groups are fully hydrogen bonded [74].) 

The amide I1 mode frequencies generally reflect the influence of hydrogen bonding, 

which manifests itself primarily through the magnitudes of such force constants 

as NH ib and CN s. (It should be noted that explicit hydrogen bond coordinates are 

not included in the coordinate basis [33, 341, so the force constants implicitly reflect 

the influence of this interaction.) As expected, the higher frequency modes are 

associated with the peptide group having a hydrogen-bonded NH. In addition, the 

modes in C5 involve coupled NH ib displacements [34], which makes it difficult in 

general to infer the explicit dependence of amide I1 on conformation. 

The amide I11 modes demonstrate an increasing complexity compared with the

~ 

Table 5-3. Some aniide modes of three hydrogen-bonded conformers of the alanine dipeptide". 

N 

ul 

Mode 

Conformer" 

I 1698 

CO s2(47)" 

CO sl(34j 

I1 1557 

NH ib2(49) 

CN s2(28j 

111 1276 

CN sl(23) 

NH ibl(20) 

CN s2(14) 

CO ibl(l1) 

c5 

1668 

CO sl(52) 

CO s2(32) 

1525 

NH ibl(49) 

CN sl(251 

NH ib2(12) 

1229 

H" b2(40) 

CN s2( 14) 

C7E 

1704 1676 

CO ~2164) CO sl(70) 

NH ib2(26) NH ibl(15) 

C7 A 

3 

1699 1682 2- 

CO s2(51) CO sl(55) E. 

NH ib2(231 CO ~2119) CO sl(14) NH ibl(l51 2 

$ 

1587 1534 

1594 1551 

P" 

c1 

NH ib2(59) NH ibl(60) NH ib2(59) NH ibl(601 Y 

CN s2(27) CN sl(28) CN s2(28) CN sl(27) 

2 

c, 

co s2(21) co sl(14) CO s2( 18) CO sl(14) 

0 

?$ 

1279 1258 1335 1325 1292 

H" bl(28) CN s2(31) H" b2(30) H' bl(26) CN ~1137) 

CN ~1123) H" b2120) CN s2( 13) CN s2(19j CO ib(16) 

NH ibl(16) H" b1116) H" bl(12) H' b2(18) MC s(12) 

NM s(13) C"C s(13j 1 

bl 

% 

2 

c 

V 693 

NH ob1132) 

CN tl(23) 

CO ib21 15 i 

NC" t( 14) 

683 

NH obl(39) 

CN tl(35) 

NC" t(24) 

CO ob2(14) 

585 754 577 809 58 1 

NH ob2(111) NH ob2(51) NH obl(102) NH ob2(51) NH oblill6) 

CN t2( 57) CN t2(40) CN tl(76) CN t2( 19) NC' t(951 

NC" t(67) 

CN tl(8ll 

r; 

a From [34]. 

q, $ 4 5: -166", 167": C7E: -85", 73'; C7A: 75", -62". 

s: stretch, ib: in-plane bend, b: bend, ob: out-of-plane bend, t: torsion, d: deformation, M: methyl C. Group coordinates defined in [34]. Contributions 

2 10.

5.4 Polypeptides 257 

nonhydrogen-bonded conformers. While NH ib contributes (at the level of 2 10) 

to the higher frequency C5 and C7E modes, it is absent from the C7A modes. 

Furthermore, if we ascribe amide I11 character to H" b, CN s modes with minimal 

NH ib, then C7A has such modes at frequencies well over 1300 cnrl, again 

reflecting the conformation-dependent contribution of H" b. (The situation is 

somewhat reversed in the crystalline Ala dipeptide [74], with a calculated mode at 

1325 cni-' showing traditional CN s, NH ib character and modes at 1272 and 1258 

cm-' having H' b, CN s character.) Since NH ib can contribute throughout the 

-1400-1200 cn1-l region, and can mix variably with CN s and Ha bend, it is not 

surprising that the amide I11 modes are a complex function of conformation. 

The amide V modes are dominated by the effects of hydrogen bonding. Thus, the 

free amide V mode is at -580 cni-' in all three conformers, while the bonded 

modes reflect the increasing hydrogen bond strengths: -690 cm-l in C5 (the splitting 

probably due to an accidental resonance with another mode), 754 cm-' in C7E, 

and 809 cni-' in C7A. Hydrogen bonding thus seeins to mask conforinational 

differences in amide V. 

These studies on NMA and the Ala dipeptide illustrate that, although it is useful 

to characterize peptide spectra in terms of supposedly localized amide modes, it is 

the subtle differences in the detailed nature of these vibrations with varying conformation 

and hydrogen bonding that provide the spectral clues to differences in 

structure. This is demonstrated by analyses of tripeptides [28, 75-77], which also 

involve interactions between two peptide groups. Although done with an empirical 

force field [5], these analyses show the sensitivity of vibrational spectra to specific 

structural features. This emphasizes the importance of highly accurate force fields 

in enabling vibrational analysis to realize its potential for being a powerful tool to 

study the structures and forces in macromolecules. 

5.4 Polypeptides 

As will be obvious from our previous discussion, a comprehensive understanding of 

the vibrational spectra of polypeptides depends on having reliable normal mode 

analyses to combine with appropriate experimental studies. Such analyses in turn 

rest on the quality of the force field. At the present time, an SDFF for the polypeptide 

chain is not available, the most detailed normal mode analyses being based 

on extensive empirical force fields [5]. These force fields have been tested on tripeptides 

of known structure with excellent predictive results [28, 75-77], and have 

proven to be very satisfactory for polypeptide systems. For reasons given in Section 

5.2.1, however, we must be careful about a final acceptance of all of their detailed 

predictions. Acknowledging these reservations, the general features of polypeptide 

spectra provided by the present detailed empirical force fields can nevertheless be 

accepted with confidence. 

In this section, we describe results of some of the vibrational analyses that 

have been done on the main secondary conformations found in polypeptides, viz.,

258 5 Vibrationul Spectroscopy oj' Polypeptides 

extended-chain and helical structures. Not all known structures have been the subjects 

of such analyses, but this approach will certainly be extended to them in the 

future. 

5.4.1 Extended-Chain Polypeptide Structures 

5.4.1.1 General Features 

Stereochemically acceptable extended-chain (so-called p-sheet) structures were first 

described by Pauling and Corey from molecular model building (78, 791. In some 

cases, specific structures of synthetic polypeptides have been determined from X-ray 

diffraction studies. These have provided the bases for the vibrational analyses of this 

class of polypeptide chain conformations. 

In these analyses, some structural constraints have had to be imposed in order to 

achieve a reasonable degree of transferability in the force fields. These constraints 

have consisted of adopting a standard peptide group geometry, which is probably 

not a serious problem, and accepting a completely planar peptide group (i.e., the 

CN torsion angle, Q, equal to 180"), which may be more serious since departures 

from planarity of 5-10' require relatively little energy (and are not uncoininon in 

proteins) and may have an important effect on some of the modes. The standard 

geometry of the peptide group is given in Table 5-4, based [XO] on the refined X-ray 

structure of P-PLA [Sl]. (The dimensions for PGI are slightly different [19].) We 

therefore assume that chain conforinations differ only in their q, $ torsion angles. 

As Pauling and Corey first pointed out [78, 791, adjacent hydrogen-bonded chains 

in a P-sheet can occur in two arrangements, parallel and antiparallel (with respect to 

the direction of the chemical sequence along the chain). In addition, depending on 

the axial stagger between adjacent chains, the sheet can be 'pleated' or 'rippled', the 

latter being possible only if all-L and all-D chains alternate in the sheet. (In practice, 

only PG can satisfy this requirement, since it has not as yet been possible to 

Table 5-4. Standard geometry of the peptide group 

Bond length 

('4") 

Bond angle 

(degrees) 

C"C 1.53 C"CN 115.4 

co 1.24 C"C0 121.0 

CN 1.32 CNC" 120.9 

NH 1 .oo CNH 123.0 

NC ' 1.47 NCT 

1 

C"H 1.07 NC"H 

109.5 

CC"H 

OCNH 180.0 

a From 180, 811.

5.4 Polvpeptida 259 

Figure 5-2. Antiparallelchain 

pleated sheet of 

polyi L-alanine). The CH; 

group is represented by 

the largest circle. (From 

[51) 

Figure 5-3. Antiparallelchain 

rippled sheet of 

polyglycine. (From [5]) 

produce such a regular alternation with other side-chain-residue polypeptides.) 

Thus, in principle we can have antiparallel-chain pleated sheet (APPS) (Figure 5-2), 

antiparallel-chain rippled sheet (APRS) (Figure 5-3), parallel-chain pleated sheet 

(PPS), and parallel-chain rippled sheet (PRS) extended-chain structures. Their 

structural parameters are given in Table 5-5 [82]. 

In describing these structures, and in the vibrational analyses, we assume the 

sheets to be infinite in two dimensions and therefore calculate only the in-phase unit 

cell modes. Although modes of only a single sheet are treated, the TDC interactions 

have been evaluated over a number of sheets (7-11). The effect of finiteness of a 

single sheet on just the TDC interactions for the amide I mode has been examined 

[83, 841. 

Extended-chain regions are a major component of protein structures. They can 

occui- in mixed antiparallel and parallel arrangements even within a sheet, the sheets 

can be twisted, and complex ‘barrel’ arrangements are also found [2]. It is therefore

Table 5-5. Structural parameters of ,&sheet StrUCtUreS". 

Paraiiieterh APPS' APRSd PPSC PRS' 

4.73 

3.445 

138.4 

135.7 

2.73 

1.75 

9.8 

164.6 

4.77 

3.522 

149.9 

146.5 

2.91 

2.12 

31.4 

134.4 

4.85 

3.25 

119.0 

114.0 

2.74 

1.82 

5.5 

172.0 

4.80 

3.25 

- 119.0 

114.0 

2.82 

1.75 

4.9 

172.0 

" From [82]. 

Lengths in A. angles in degrees. a (or a/2): lateral separation between adjacent chains. 17: axial 

separation between adjacent units in one chain. 

Antiparallel-chain pleated sheet of poly(L-alanine). From [80, 81 1. 

Antiparallel-chain rippled sheet of polyglycine. From [82]. 

Parallel-chain pleated sheet of polyil-alanine). From [83]. 

Parallel-chain rippled sheet of polyglycine. From [83]. 

of major importance to understand the vibrational dynamics of this type of polypeptide 

chain conformation. 

5.4.1.2 Antiparallel-Chain Structures 

Antiparallel-Chain Pleated Sheet 

The APPS structure is the predominant one in synthetic polypeptides, and is a 

prevalent motif in proteins [2]. (Polyglycine is an exception, which xe treat below.) 

An X-ray diffraction study of the simplest representative polypeptide, /I-PLA [811, 

has provided the geometric parameters of the APPS. In addition to those given 

in Table 5-5, we note that the axial shift between adjacent chains, measured from 

the position where the hydrogen bond is linear, is 0.27A (as obtained from a TDC 

analysis of the amide I mode [80]). The vibrational analysis of /I-PLA thus provides 

the basic spectral characteristics of the APPS. 

A PLA sample can be easily oriented, and dichroic IR spectra are therefore 

readily obtainable. Such spectra [85], together with far-IR spectra [86], are shown in 

Figure 5-4. Together with spectra of the ND inolccule [87, 881, as well as Rainan 

spectra [89] (Figure 5-5), a large amount of spectral information is available, both 

to optimize an empirical force field and to provide a reasonable description of the 

normal modes. 

The symmetry of the p-PLA structure determines that the normal modes will 

be distributed among four symmetry species with optical activity and IR dichroism 

as follows [SO]: A-30 modes, Raman; B1-29 modes, Raman, IR (11); B2-29 modes, 

Raman, IR (I); B3-29 modes, Raman, IR (I). The results of the most recent nor-

I 


)I 

.- + 

$ 800 1000 1200 1400 1500 1600 1700 

0 

1 1 1 

2800 2900 3000 3100 3200 3300 3400 

Figure 5-4. Infrared spectra of /l-poly(Lalanine). 

Upper two panels: mid-IR 

region. (-) Electric vector perpendicular 

to the stretching direction. (- - -) 

Electric vector parallel to the Stretching 

direction. (From [SS]) Lower panel: 

far-IR region. (From [86]) 

Frequency (ern-') 

2 

0) 1 

Figure 5-5. Ranian spectrum of /l-poly(L 

alanine). (From [SS]) 

900 800 700 600 500 400 300 200 100 

FREQUENCY (cm-' 

ma1 mode analysis [22] are given in Table 5-6 (the NH s modes, which are influenced 

significantly by Fernii resonances 1901, and the CH s modes are discussed in 

an earlier paper [SO]). 

There are some general points to note about this normal mode analysis. First, the

Table 5-6. Observed and calculated frequencies (in cm-' ) of APPS poly(L-alanine). 

Observed" Calculated Potential energy distributionh 

Raman IR A Bi B2 Bi 

1669s 

1553VW 

1538W 

1451s 

1451s 

1399W 

1368W 

1335W 

1311W 

12433 

1694W 1 1 

1632VS I 

1555MW I 

15248 1) 

1454s / / 

1446s i 

1454s 11 

1446s _L 

1402MW 11 

1386W i 

1372MW / / 

1330W br I 

1309 sh 

1670 

1539 

1455 

1452 

1402 

1372 

c 

1236 

1695 

1528 

1455 

1452 

1399 

1372 

1630 

1562 

1452 

1451 

1383 

1317 

1299 

1698 

1592 

1453 

1451 

1385 

1337 

1305 

CO s(78) CN sf141 

CO sf76) CN si19) 

CO sf73) CN s(21 I 

CO sf70) CN ~(21) 

NH ib(571 CN sf211 '2°C s[l01 

NH ibi53) CN sf17) C"C ~(14) 

NH ibi48j CN sf22) CO ibi121 C*C $11) 

NH ibi41j CN si26) CO ib~141 C"C s(13) 

CH3 abl(34) CH7 ab2(39) 

CH3 ab1146) CH7 ab2137) 

CH-, ablf88) CH? rlfl0) 

CH; ab1186) CHI rlfl0) 

CH3 ab2145) CHi dblf43) 

CH3 ab2(48) CH? ablf41) 

CHI ab2i861 CH7 12(10) 

CH? ab2(89) CH7 12j10) 

H" b2(32) CH7 sb(17) C*C sfl4) NH lbi13) 

Ha b2f33) CHI sb120) NH ib(l5) C'C sill) 

CH7 sb(70i H' blil6I CYCD S! 111 

CH3 ~b174) H" bl (16, C^Ci'~i10) 

CH7 sbi64) H'b? (161 

CH1 sbi6lI H'b2 119) 

H" b2132) NH lb(25) C"C ~(12) CH3 sbfll) 

H" b2164) NH Ibf 15) 

H" b2130) CN ~(181 CO ibfl9 

NH M23) CO ibfl8) C'C si14) CN sf131 

H' b2134) NH ibil9) NC' ~(19) CN sil3i

1226M 1222s (1 

1165W 

1167s (1 

1120VW 

1092s 

10S4W 1 

1052M I 

1065M 

967M 

909VS 

837VW 

775M 

698VW 

966s 11 

925M I 

902 W 11 

778MW 

706SI / / 

1198 

1162 

1093 

c 

1055 

969 

913 

{ 

704 

668 

1231 

1195 

1161 

1092 

1054 

970 

706 

662 

1195 

1125 

1085 

1064 

918 

844 

775 

708 

1196 

112s 

1086 

1065 

913 

846 

755 

705 

H' b2(?8j NC' ~(241 NH lb(l8) CN s( I I I 

NC" s(361 CH; rl(13) C'C $11) NH ib(10) 

C'CB s(331 H" bl(27) NC' s(22) CH3 sb(l1) CH; rl(10) 

NC" s(30) NH ib(13j C"C 412) H' b2(ll) CH3 rlill) 

C"CP s(33) H' bl(261 NC' ~(221 CH3 rl(10 CH; sbt 10, 

Hi bl(56) CH3 sb(19) C'CP ~(13) 

H" bl(541 CH3 sb(18j C"C1 ~(131 CN ~(10) 

CH3 r2(27] Ha bl(26) CH3 Sb(11) 

H" b1176) CHI 12/76) CHI sbflli 

CH3 12(55\ C"C1' ~(21) 

CH3 r2(541 C"CP $2) 

CH3 rl(271 CH3 r2(251 C'CD ~(101 CN s(l0) 

CH? rl(26) CHI r2126j C"Cp dl31 CN dl01 

C'C/'s(23! HI blL2l) CH3 rl(18) CH3 12(17) 

C"Cp s(22) H" bl(21) CH3 r2(19) CHI rl(17) 

C"Cp ~(501 CH3 rl(141 H" bl(l1) 

C"Cp s(50) CHI i-1113) HY blillj CHI i-2110, 

CH3 rl(50) NC' ~(251 

CH3 rl(501 NC" s(2Y 

NC" s(27) CH3 rl(23) CN s(15) 

C"C s(15) CN s(14) CH; r2(14) CNCd d(13j CO s(l1) 

NC" ~(30) CH? 11f23) CN ~(l4! 

'2°C s(15) CN s(14) CHI r2(14) CNC' d(13) CO s(12) 

C"C s(34) NC" s(13) CN s(12) CTp ~(12) 

C"C s(31) NC' s(19j C"CB s(14j 

CO lb(19) C'C s(l5) NC's(131 CI'b2i13j NC'Cdi12) 

CNC" 411) 

CO ib(181 C"C s(16) CB b2(14) NC' $12) NC"C d(l1) 

-+ 

CNC" d( 11) 1 

CN t(74) NH ob(2Sj NH 

3 

0 lb(23) 

P 

CN t(41) NH ob(41) NH 0 lb(19~ CO ob1181 H" bl(l0) 

CN t(74) NH ob(28) NH 0 ib(72) 

CN t(48) NH ob(41) NH . 0 Ib(l9) CO ob(131 H" bl(10) N 

m 

w 

CO ob(50i CN ti351 Cb bliloj 

CO ob(47r CN t(41) CD bl(10) 

b 

a 

2 

e

Table 5-6. (coiztimied) 


Raman IR A Bi B? B3 

628VW 624 

629W 11 

437w 

332W 

300M 

266VWsh 

615WI 

448M I 

432M )I 

326W // 

440 

327 

i 

27 1 

235Msh 240 

240M / / 

18SVW 177 

62 1 

440 

328 

279 

240 

211 

151 

626 

591 

447 

286 

286 

252 

238 

CO ib(46) (2°C si16) CO obil3) 

CO ib(49) C"C s(16) CO ob(12) 

CTN d(38) CO ob(l8) NCYC di16) NH ob(l6) CB b2(15) 

622 CTN d(31) CO ob(30) NH ob(22) Cb b2(161 NCT d(l4) 

592 CO ob(52) C"CN d(19) 

CO ob(62) C"CNd(l1) NH ob(1l) Ha b2(101 

CB bl(55) NH ob(15) 

447 Cp b1154) NH obil3) 

CP b2(75) NC"C d(13) 

CB b2174) NC"C d( 12) 

NC'C d(25) CB b2(11) CO ob(l1) CB bl(10) 

NC"Cd(24) Cp b2(11) COob(l1) CB blilO) 

CO ibf33) NC"C d(19) CNC" d(15j CB b2(15) CB bl(13) 

CO ob(l1j 

CO lb(33j NC"C d(19) CB b2(15) CNC" d(14) CB bl(13) 

CO ob( 10) 

CTN d(39) CaCP t(26r NCT d(1li 

C"CN d(38) C"CB t(25) NC"C d(17) 

CB b2( SO) 

252 Cp b2( 50) 

C'CP t(671 NC"C d(11) 

C'C'' t(49) NC"C d(16) 

C"C'7 t(91) 

238 C'CB t(90) 

CNC" di35) C'CD t123) H. 0 ~ (23) 

CNC' d(621 C"CN d(15) 

156 H 0 s(58) NC'C dl 101 

NH ob(79) CO obi131 NC" dl3) H" b2110j

135s 

9 1 Ms1i 

122Wbr 

147 NH ob(37) CO ob(16) NC' s(13) H" b2(101 

118 

CNC" d(21) NC" C d(18) CTN d(13) NH ob(l?i 

104 

NH...O ib(27) CN t(21) Cp b1115~ 

103 NH.,.O ibi331 CN t(21) NH obi2O) C"C t(15) CO,..H t(l1) 

91 

Cp bl(26) NH ob( 19) H-.O s( 14) C"C t( 11) 

87 H...O ~(33) CNC" d(17) CN t(l1) CN'C d(10) 

74 

C'C t(24) NC" ti22)H.,.O ~(22) CN t(15) 

41 CO...H ib(21) NH ob(70) Cpblilhj NH...O ib(l5) 

NH...O ti 12) 

32 

NH.,.O ib(371 CO-.H ib(22) CN t(l5) 

31 

H...O s(15) CO...H t(l5) CO...H ib(14) NH...O ib(l4) 

NH...O t( 12) NH ob( 11 j CN t( 10) 

19 

CO,.,H ti35) NH...O t(27) CO.,.H ib(l4) H". . .H' sil3j 

a Data taken from [88]. S: strong, M: medium, W: weak, V: very, sh: shoulder, br: broad, 11: parallel dichroism, I: perpendicular dichroism. 

b s: stretch, as: antisymmetric stretch, ss: symmetric stretch, b: bend, ib: in-plane bend, ob: out-of-plane bend, ab: antisymmetric bend, sb: symmetric 

bend, r: rock, d: deforniation, t: torsion. Contributions 2 10.

966 5 Vibr-atioiial Slm-troscopj, of' Pol.vpeptides 

empirical force field is capable of giving very acceptable frequency agreement: 

the rms error between observed and calculated frequencies is 5.9 cm-'. Second, the 

character of the symmetry species is well predicted: A species modes exhibit only 

Rainan activity, B1 and B? species modes exhibit 1) and I IR dichroism. respectively. 

Third, while the agreement is somewhat poorer, band shifts on N-deuteration 

are reasonably well accounted for [88] (for amide I the observed downshift of 

4 cn-' is reproduced exactly by the calculation). Finally, TDC is crucial to 

explaining the large splittings in the amide I and amide I1 modes: in its absence the 

ainide I modes are calculated at 1670 (A), 1673 (B,), 1665 (B2), and 1670 (B3j cin-.', 

and the amide I1 modes are predicted at 1545 (A), 1549 (BI), 1550 (B?), and 1554 

(B3) c111-l [5]. 

With respect to the amide modes, a comparison with the results of the ah initio 

calculation on the hydrogen-bonded Ala dipeptide is of interest, even though the 

force fields are different. Although localized primarily in CO s, the amide I modes 

of /I-PLA are constrained by symmetry to be coupled modes of the four peptide 

groups in the unit cell. Interestingly, CN s is now the second major component 

compared with NH ib foi- the Ala dipeptide. The coupled amide I1 modes contain, 

in addition to NH ib and CN s, C"C s and CO ib contributions not present (at the 

level of 2 10) in the Ala dipeptide. With respect to the strong amide 111 modes, at 

1243 and 1224 cni-', H" b dominates and NC" s also contributes in addition to NH 

ib and CN s. As noted above, NH ib is a contributor to many other modes in 

the 1400-1200 cm-l region. The amide V modes are well accounted for, and in 

particular the observed absence of any significant dichroism in this band 1911 is 

explained by the near coincidence in frequency of the predicted B1 and B3 species 

modes. Besides the aniide modes, local as well as other skeletal modes are very well 

accounted for. 

The excellent agreement obtained for /I-PLA would indicate that this force field 

should be transferable to other APPS polypeptides, of course taking due account of 

side-chain differences. This has been done for Ca-poly( L-glutamate), Ca-PLG (921, 

which X-ray and electron diffraction data had indicated to have an APPS structure 

but which were not extensive enough to provide a definitive conclusion. A proposed 

model [93] was used as the basis for the vibrational analysis, and the good prediction 

of the observed bands [92] supports both the model as well as the viability 

of the force field. The subtle differences between the P-PLA and /I-Ca-PLG spectra 

are accounted for by the normal mode calculation, and emphasize the influence 

of the side chain on the main-chain modes, particularly on amide 111. A similar 

successful vibrational analysis has been done on an alternating copolymer, APPS 

poly( L-alanylglycine) [ 191. 

Antiparallel-Chain Rippled Sheet 

Although early studies assumed that PG also adopts an APPS structure, electron 

diffraction studies of 'single crystals' and oriented thin films [94] suggested that 

extended-chain PG, PGI, has an APRS structure. This was supported by conforniational 

energy calculations [95], and was the basis for a detailed vibrational 

analysis of the PGI spectra [19, 961, strongly confirming this proposal. (Since the


glycine residue is achiral, it can adopt the D or L configuration required of alternate 

chains in the APRS.) This was aided by extensive IR spectra of isotopic derivatives 

(NH, CD2; and ND, CD2) [97] and by Raman spectra [9S], which were important 

because macroscopic samples of oriented PGI could not be prepared. 

The conclusion from vibrational analysis that PGI has an APRS structure was 

based on a detailed TDC study of the aniide I mode splittings for the APPS as 

compared with the APRS as a function of the axial shift between adjacent chain 

[19]. No reasonable shift values for the APPS were in agreement with experiment, 

whereas the APRS gave excellent agreement for a shift of 0 A, exactly that indicated 

by the electron diffraction analysis [94]. This structure (Figure 5-3) has a 

center of inversion and therefore the mutual exclusion rule applies, giving the 

following symmetry species and activity [ 191: As-21 modes, Raman; A,-20 modes, 

IR( 11); Bg-21 modes, Rainan; Bu-19 modes, IR(I). The results of the normal mode 

analysis [96] give an overall rnis error of 6.1 cm-', comparable with that for P-PLA. 

(The force field for p-PLA was in fact refined from that for PGI [80].) A description 

of the amide modes of PGI is given in Table 5-7. 

The splitting of the amide I modes, somewhat smaller than for the APPS structure, 

is again due to TDC: in its absence the calculated frequencies are 1684 (Ag), 

1677 (Au), 1676 (Bg), and 1674 (B,) cm-I. Compared with APPS PLA, CTN d 

now makes a noticeable contribution to the eigenvector. 

The amide I1 modes, which are probably at a lower frequency because of the 

weaker hydrogen bonds, are also split by the TDC interactions: in their absence 

the calculated modes are at 1534 (Ag), 1535 (A"), 1559 (Bg), and 1559 (B,]) cni-'. 

The absence of clearly observed high-frequency modes, particularly an IR band 

near the expected frequency of 1572 cm-I, raises a very important point: it is necessary 

to be able to predict intensities as well as frequencies. Intensities require 

knowledge of dipole derivatives (Eq. (5-4)), but since they also depend on correct 

eigenvectors (Eq. (5-5)) they provide an independent test of the force field. Such an 

1R intensity calculation has been done on PGI [20], and the results show that the B, 

amide I1 mode is indeed predicted to be very weak, in agreement with observation. 

The amide I11 mode region is quite complex, with mainly NH ib, CN s contributing 

near 1300 and 1200 ciii-', where only weak bands are observed. In the 

'expected' region (1300-1200 cm-'), the strong bands are due to CH? twist (tw)- 

dominated vibrations. The NH ib contributions now extend to bands about 1400 

and below 1200 cm-' . 

The amide V mode is at a similar frequency to that in ,!I-PLA, even though the 

PGI hydrogen bond is weaker. This may be because of the large CN t contribution 

and/or the somewhat different nature of the modes. 

5.4.1.3 Parallel-Chain Structures 

Parallel-Chain Pleated Sheet 

Although there are no known PPS polypeptides, this general structure is found 

in proteins [2] and in some tripeptides. in two of which it has been the subject of 

detailed vibrational analyses [75, 771. The excellent prediction of observed I R and

Table 5-7. Amide mode frequencies (in cm-') of APRS polyglycine I. 

Observeda Calculated Potential energy distributionb 

Raman IR A, B" B, B u 

1695 CO s(77) CN si15) CTN d(l1) 

1685M 1689 CO 475) CN si2Oi C'CN d( 11) 

1674s 1677 CO s(74) CNs(211 C"CNd(ll1 

1636s 1643 CO s(69) CN si22) C"CN d(ll'1 

1602 

NH ib(56) CN s(19i C T ~(12) 

1572 NH ib(51) C"C s(16) CN s!14! 

1517s 1515 NH ibi35) CN s(28i C T s(171 CO ib(14) 

1515W 1514 NH ib(35) CN s(271 C'C s(17) CO ibil4) 

1410M 1415 CH? w(44) CH? b(3l) NH ib(l4) 

1408W 1415 CH? w(40) CH? b(33) NH ib(13) 

1304 

NH ib(30) CO ib(19) CN si18j C"C sil6) 

1295W 1286 NH ibi39) (2°C s(17) CO ib(1hi CN ~(12) 

1220w 1213 NC" s(29) NH ib(23) CH2 w(18) CH2 tw(l6) CN s(13j 

1214W 1212. NC' s(29) NH ib(23) CH: w(18) CH2 tw(l5j CN ~(13) 

1162M 1153 NC" s(50) C"C ~(13) NH ib(l2.) 

1152 

736 

NC' s(50) C'C s(14) NH ib(l2.i 

CN t(63) NH...O ib(15) NH ob(l1) H...O s(l1) 

708s 

718 CN t(75) NH...O ib(19) NH ob(16) H...O s(l0) 

{ 

718 

CN t(791 NH ob(36) NH...O ib(23) H...O s(10) 

701 

CN t(79) NH ob(29) NH...O ib(25r H...O ~(15) 

Data taken from [95]. S: strong, M: medium, W: weak. 

s: stretch, b: bend, ib: in-plane bend, oh: out-of-plane bend, d: deformation, w: wag, tw: twist. t: torsion. Contributions 2 10.


Raman bands in these cases, using the same empirical force field as for APPS PLA 

[5], led to the belief that it would be useful to obtain the frequencies of PPS PLA in 

order to characterize the vibrational spectrum of this structure. The unperturbed 

normal modes have been obtained, and the influence of TDC interactions on the 

aniide I and ainide I1 modes, as determined by the intersbeet separation, has been 

assessed [82]. While unperturbed aniide I modes occur at 1678 and 1675 cm-I, TDC 

interactions within one sheet shift these to 1663 and 1642 cm-' and a multisheet 

array ( 11 x 11 x 11 residues) with a separation of 5.3 A (comparable with that in 

APPS PLA) leads to modes at 1669 and 1677 cm-'. A similar effect is predicted for 

ainide 11: single sheet modes are at 1553 and 1586 c~ii-', with the above niultislieet 

separation giving 1549 and 1583 cin-' . 

Parallel-Chain Rippled Sheet 

A PRS structure is only possible for PG, and is ruled out by the evidence for an 

APRS structure for this polypeptide. However, the norinal, modes of PRS PG have 

also been calculated (821. It is interesting that, when multisheet TDC coupling is 

included, the strong Raman amide I mode is predicted to be at a much lower frequency 

than the strong IR amide I mode, just the opposite of what is observed and 

predicted for the APRS structure. This demonstrates the power of detailed vibrational 

spectroscopic analysis in revealing fine details of polypeptide structure. 

5.4.2 Helical Polypeptide Structures 

5.4.2.1 General Features 

The concept of a helical structure as the most general one for a long polymeric 

chain was developed quite early [99, 1001, and possible polypeptide helices having 

an integral number of residues per turn were described [loll. However, it was only 

when this constraint was relaxed and well-defined stereochemical criteria invoked 

[lo21 that a specific conforination, the or-helix, was discovered [102, 1031 that is in 

fact found in polypeptides and proteins [2]. A number of other helical conforniations 

have been found since then, and their structural parameters, are given in Table 

5-8. Vibrational analyses of some of these are described in this section. 

Polypeptides are found in both single-stranded intramolecularly hydrogenbonded 

helices as well as in conformations in which hydrogen bonds are fornied 

between helices. In the former case, the structure is conveniently described by 

several parameters: the number of residues per turn, n (positive if right-handed, 

negative if left-handed); the axial translation per residue, h; and the backbone torsion 

angles, p, $, and o. Sometimes the number of atoms in the 'ring' formed by the 

hydrogen bond, in, is specified [loll (and the residue connectivity via the hydrogen 

bond can be given; thus, 5 - 1 iiieans the NH group on residue 5 is bonded to the 

CO group on residue 1). In such a case, the helix will be designated PI,?, [loll. 

Among such helices. the a-helix (3.6213) structure has been obtained froin a 

careful analysis of X-ray fiber diffraction patterns of PLA [104]. As is general for 

helical structures [ 1051, this helix, designated, a ~, has a counterpart, designated all,

~~ ~ 

270 5 Vibrational Sl~ectroscop~~ oj Polypeptides 

Table 5-8. Parameters of some observed helical polypeptide structures 

Structure 

hlrvrlnloleL 1rltrr 

~1-Hel1x 

q1-Helix 

3lo-Hel1~ 

to-tieli~ 

R-HKIIX 

n” nib hL Yd id cod H-bonde Ref. 

3.62 13 1.496 -57.4 -47.5 180 5-1 104 

3.62 13 1.496 -70.5 -35.8 180 5-1 106 

2.99 10 2.01 -45 -30 176 4- 1 111 

-4.00 13 1.325 62.8 54.9 175 541 1 I ? 

-4.25 16 1.17 51 74 177 6-1 115 

Intevniolec iiliw 

Polyglycine I1 

Polypioline 11 

Polypioline 1’ 

-3.00 3.10 -76.9 145.3 180 

-3.00 3.10 -78.3 148.9 180 

3.33 1.90 -83.1 158.0 0 

122 

128 

135 

Number of residues/turn, + for right-handed and - for left-handed helices 

Number of atoms in hydrogen-bonded ‘ring’. 

Unit axial translation, in A. 

‘’ Dihedral angle, in degrees. 

Hydrogen bonding pattern, N-H-O=C. 

Cis peptide groups. 

(j 

b 

Figure 5-6. Structure of a-helix of poly(L-alanine). (a) al-helix; ib) q1-helix. Methyl groups are 

represented by large circles. (From 151) 

which, though having different p, $, has the same IZ and 11, and is comparably 

allowed energetically [ 1061 (and may be present in bacteriorhodopsin [ 1071). These 

structures are shown in Figure 5-6, where it can be seen that the plane of the peptide


Figure 5-7. Structure of 3lo-helix of poly(a-amiiioisobutyric acid). 

Methyl groups are represented by large circles. (From [S]) 

group in a11 is inclined to the helix axis whereas it is almost parallel to the axis in MI. 

The a-helix conformation is not only coniinoii in polypeptides, and is a basic component 

of proteins [2], but its topology forins the basis for two-stranded coiled-coil 

structures such as tropomyosin [108]. The 3lo-helix is similar to the a-helix, but has 

a 4 i 1 hydrogen-bonding pattern. Although segments are found in proteins [2], 

the only polypeptide in which it has been found, as suggested by electron diffraction 

[ 1091 and supported by vibrational analysis [ 1101, is poly(a-aminoisobutyric acid), 

PAIB (Figure 5-7). Its structure, as indicated by conformational energy calculations 

[111], has a nonplanar peptide group. The o-helix is a left-handed 5 - 1 

helix, -4.013, proposed from X-ray analysis for the structure of poly(j?-benzyl-Laspartate) 

[ 1121. Coiiformational energy calculations have indicated somewhat different 

parameters [ 1131, with other calculatioiis suggesting a further variant [114]. 

The n-helix, said to be present in poly(j?-phenethyl-L-aspartate [115], is also a lefthanded 

helix, -4.2516, although a right-handed form with a 4" distorted NC"C 

angle has been suggested [116]. The 0- and n-helices also have nonplanar peptide 

groups. Other helical conformations have been proposed, a y-helix [lo?, 1031 and a 

2.27-helix [117], but neither of these has as yet been observed in polypeptides. When 

L and D residues alternate along a chain, which occurs in some transinembrane 

peptides, a new class of helices results, the so-called /?-helices[I 18, 1191. Such chains 

can also form double-stranded structures with interchain hydrogen bonds. 

The simplest polypeptide structure that has intermolecular hydrogen bonds 

throughout its lattice is that of PGII, which basically has threefold helical symmetry 

[120]. (Because of its achiral C", PGII can be a right- or left-handed helix.) Although 

the originally proposed structure was modeled with parallel chains, the ease 

of mechanical conversion of PGII to PGI and the evidence for antiparallel chains 

in single crystals of PGI led to the suggestion [ 1211 that an antiparallel-chain structure 

should exist. Such a model was subsequently proposed 11221, and a detailed

272 5 ViDrutiomil Spectroscopy oj Polvpepticles 

Figure 5-8. Antiparallel-chain structure of polyglycine 11. (From [ 1231) 

vibrational analysis of the two structures 11231 showed that the antiparallel-chain 

structure (Figure 5-8) was indeed preferred. In these studies, the suggestion that 

C"-H ... O=C hydrogen bonds were present [ 1241, which was supported by spectroscopic 

studies [ 125, 1261, was convincingly confirmed by normal-mode analysis 

[123]. The same chain symmetry, in a required left-handed helix, is found in polyproline 

II, which has trans peptide groups 1127, 1281, a short contact again suggesting 

the possibility of a C"-H...O=C hydrogen bond [129]. It is interesting that, 

from circular dichroisrn studies [130], a polyproline I1 type of structure, with 2.5- 

3.0 residues/turn [131], was indicated to be a preferred local conformation in many 

'unordered' polypeptides in solution. This proposal has since received strong support 

from Rainan and normal-mode studies [ 1321 as well as from a more detailed 

analysis of the circular dichroism results [133]. A triplet of left-handed PGII or 

polyproline 11 helices can be thought of as the topological parent of the triplestranded 

right-handed coiled-coil structure of collagen [ 134, 1351. When the peptide 

groups in polyproline take a cis configuration, i.e., polyproline I, the chain adopts a 

right-handed 10 residue/3 turn helical structure [ 1361. 

The variety of helical as well as extended-chain and turn [5] structures emphasizes 

the flexibility of the polypeptide chain, and thereby the importance of achieving a 

full understanding of its vibrational dynamics as a function of conformation.

Frequency (cm-'1 

100% 

b 

700 500 300 100 (ern-') 

Figure 5-9. Infrared spectra of a-poly(L-alanine). (a) Mid-IR region. (From [85]).(b) Far-IR 

region. (From [I 371). (-) Electric vector perpendicular to stretching direction. (- - -) Electric vector 

parallel to stretchin2 direction. 

5.4.2.2 a-Helix 

The a1-PLA structure is well determined from X-ray diffraction [ 1041, and therefore 

permits a secure refinement of an empirical force field. The helical symmetry results 

in three optically active modes, defined by the phase difference between similar 

vibrations in adjacent uiiits of the helix: A(0") - 28 modes, Raman, IR( 11); El (99.6') 

- 29 doubly degenerate modes, Raman, IR (1); E2 (199.1")-30 doubly degenerate 

modes, Raman. Extensive IR and Raman studies on a-PLA and a-PLA-ND [5] 

have provided good experimental data. Polarized IR spectra [85, 1371 are shown in 

Figure 5-9 and Raman spectra [138] in Figure 5-10. The force field for the normalinode 

calculation [lo61 was refined using that of /?-PLA [22] as a starting point. 

The frequencies and PEDs [lo61 are given in Table 5-9; the iins frequency error is 

1.6 cm-'.

I I I I I I I I I I I I I 

1600 1400 I200 loo0 800 Mxl 400 (crn-'l 

Figure 5-10. Ratnan spectra of a-poly (L-alanine) (-) 

381) 

Undeuterated. ( ..' ) N-deuterated. (From 

The observed and calculated amide I mode splittings agree quite well, and are 

much smaller than for p-PLA. (The specific A-Ez splitting, due to TDC, remains to 

be verified; this will depend on having good polarized Raman spectra.) The mode 

now contains less CN s and more C"CN d than P-PLA, and a change in the small 

(< 10) NH ib contribution is largely responsible for the observed 8 cni-' downshift 

(calculated 10 cm-' ) on N-deuteration [ 1061. The calculated unperturbed amide I1 

modes at 1529 (A), 1532 (El), and 1536 (Ez) cn1-l again show how important TDC 

is in explaining the observed 29 cn1-l A-El splitting. The NH ib contribution to 

amide I11 now dominates over Ha b as compared with p-PLA, and CN s does not 

appear. The increase in these frequencies, from the 1240-1220 cm-I region in p- 

PLA to the 1280-1260 cm-' region in a-PLA, must be due to the difference in 

conformation (and the resulting change in interaction between H" b and NH ib), 

since the relevant force constants are in fact smaller in the a-helix than in the p-sheet 

[106]. The lower amide V frequencies in a-PLA, 658 and 618 cm-l, than in /I-PLA, 

706 cin-', are undoubtedly a result of the weaker hydrogen bond in the helical 

structure (N...O = 2.86 A in a-PLA versus 2.73 A in P-PLA). 

Other synthetic polypeptides adopt the al-helix structure, and it is of interest to 

see how their normal modes are influenced by a change in the side chain. A polypeptide 

that has been analyzed with the same main chain force field as a-PLA is 

a-poly(L-glutamic acid), a-PLGH [139]. Some of its modes are compared with those 

of q-PLA in Table 5-10. The amide I mode is hardly affected. The observed amide 

I1 mode A-El splitting increases somewhat, from 29 cm-l in a-PLA to 40 cn-I in 

a-PLGH; since the eigenvectors are identical [139], this may reflect a very subtle 

change in structure that alters the TDC interactions. The amide I11 modes are 

somewhat more affected, and this is a result of the side chain: CitH? tw now contributes 

to the E2 and El modes. Analogous changes influence skeletal modes in the 

1200-900 cm-' region, in the case of the CN s, CNC" d modes even altering the 

frequency order of the species. The aniide V modes are also shifted by such changes 

in eigenvectors, and these have an even greater impact on the low-frequency skeletal 

modes. The constellation of such changes may well be characteristic of the side 

chain.

5.4 Polypptidt~s 215 

Table 5-9. Observed and calculated frequencies !in cin-') of ccr-poly(L-alanine). 

Observed" Calculated Potential energy distributionb 

Ramaii IR A EI E2 

1655s 

1658VS 1 1 

1543VW 1545VS I 

15 16Msh 

1458s 1458VS 1 ) 1 

1377W 

1 3 38Msh 

13268 1328sh / / 

1308M 1307 /I 1 

1278W 

1271 W l270M I 

1261 W 1265Msh 

1167M 

1105s 

1050W 

1017W 1016W 1 1 

910W 968M 11 

940VW 

908VS 909M 1) 

8933 I 

882W 

173VW 

156W 

1657 CO $32) CN s( 10) CTN d(10) 

1655 CO ~(82) CN s( 11) C"CN di 10) 

1519 

1452 

1538 

1645 CO s(83) CN s!12) CTN d(10) 

1540 NH ib(46) CN s(31) CO ib(12) C"C ~(11) 

NH ib(46) CN s(33) CO Ib(l1) C"C silo) 

NH ib(45) CN si34) CO ib( 11) 

CHI abl(47) CH; ab2(41) CH3 11(10) 

1452 CHI ab2(54) CH3 abI(31) 

1452 CH3 db2(59) CHI abl(26) 

1451 CH3 ab2j45) CH3 abl(41) CH3 r2(10) 

1451 CH3 abl(57) CH3 ab2(31) 

1451 CH3 abl(62) CHI ab2(26) 

1381SA 1379 1379 1379 CH; sb(100) 

1349 Ha bl(34) NH ib(17) (2°C s(17) C"Cp s(10) 

1345 H" bl(25) H" b2(20) C"C s(17) NH ib(12) 

1334 H" b2158) C"C s(16) 

1314 H" b2(81) 

H" b2(56)H" bl(23) 

{ 1301 I3O8 H" bl(62) 

1287 NH ib(23) NC" s(18) H" bl(l8) 

1278 NH ib(28) H" b2(15) NC' s(14) H" bl(l1) 

1262 NH ib(40) Ha b2(28) 

NC" ~(32) CH: 11(20) C"C" ~(6) 

1170s 11 I 

{1178 1167 C"CD s(32) NC" 421) CH3 rl(15) H" bl(10) 

1162 C"CD s(41) NC" ~(16) H' bl(12) CH3 rl(12) 

1115 CTL' s(64) CH; r2(15) 

1108S1 

1103 C"Cfl s(39) CH? r2(18) 

1094 C'CD s(26) CHI r2(21) 

1043 CH3 rl(47) Ha bl(20) 

1051VS i 1037 CH3 rl(39) H" bl(22) CH3 i2(15) 

1026 

CH; r2(28) H" bl(25) CH: rl(24) 

982 CHI il(34) NC" ~(29) CH3 12(17) C"C ~(15) 

955 CH; r2(32) NC" 420) CH; rl(15) 

951 CH; r2(41) NC" s(16) 

910 

CNs (23) CNC" d( 16) CO 1b( 12) CH3 r2( 1 I ) CO s( 10) 

901 

CN s(18) C"C s(12) CO ibi12) C'CD $10) 

896 C'"'s(l9)C"C a(l4) CN s(14) CO ib(12) NC" s(l1) 

714M 

780 CO ob(42) Cp bl( 10) CN t( 10) 

767 CO ob(57) Cp bl(10) 

754 CO obi38) CN t(30)

276 5 Vibrational SiJectroscopy of Polypeptides 

Table 5-9. (contiriued) 


Raman IR A El El 

- 

693M 691Wsh 11 700 

675 

662W 658s L 660 

637 

618s I 608 

589 

I537 

530VS 

375s 

328W 

310s 

294M 

260M 

209VW 

189M 

165M 

159s 

87W 

375s i 374 

-366Msh 367 

3248 I 326 

290M 11 307 

259Wsh 264 

240W 245 

223VW 230 

188M i 197 

l63M I 155 

120s 1 1 136 

113MshI 96 

84W 94 

40 

492 

366 

310 

244 

205 

151 

87 

49 

38 

"2°C d(36) C"CN d(16) 

CN t(32) CO ib( 18) C"C s( 14) 

CN t(37) NH ob(21) NC"C d(12) 

CN t(59) NH ob(43) NH...O ib(13) 

CN t(47) NH ob(23) CO ob( 15) CO ibi 12) 

CN t(68) NH ob(36) CO ob(26) NH.,.O ibill) 

CO ib(29) C"CN d(21) c"C s(17) CB b2(17) 

NH ob( 11) 

NC'C d(31) C'T s(14) CTN d(1l) CO ib(1l) 

NCT d(37) CO ib(17) (2°C s(16) 

CO ob(16) NH ob(16) Cp b2(15) C"CN d(15) 

CO ib( 15) CNC" d( 10) 

CO ob(21) Cp bl(17) NC"C d(14) NH ob(13) 

d b2(11) 

Cp b2(49) C"CN d(22) Cp bl(16) 

Cp b2(42) Cp bl(19) CO ib(16) 

CNC" d(30) CO ib(20) Cp bl(20) CO ob(17) 

Cp b2(34) CO ib(29) CNC" d(15) 

C"C0 t(35) Cfi b2(14) 

caCD t(91 j 

co t(95) 

C'CB t(63) 

CTN d(27) Cp b2(25) Cp bl(12) CO ob(l0) 

CaCN d( 15) Cp b2( 12) CO ob( 12) 

CNC"d(33)C'CNd(19) Cp b1(14)NHob(14) 

NC"C d( 12) 

NH ob(43) CNC" d(20) NC'C d( 11) 

CNC" d(34) CTN d(21) NC"C d(16) Co bl(12) 

NH ob(24) C'C t(20) NC" t(l8) CN t(18) H...O ~(10) 

CN t(27) NC" t(23) C"C t(22) H...O ~(15) NH ob(l0) 

NH ob(33) NH-.O ib(20j Cp bl(12) NCT d(10) 

CNC" d(l0) 

NH ob(38j NC" t(24) H-.O s(23) CN t(15) Cp bl(13) 

NH ob(58) Cp bl(20) H...O ~ (ll) 

(2°C t(53) H...O ~(18) 

a S: strong, M: medium, W: weak, V: very, sh: shoulder, 1): parallel dichroism, 1: perpendicular 

dichroism. 

s: stretch, as: antisyrnnietric stretch, ss: symmetric stretch, b: bend, ib: in-plane bend, r: rock, 

d: deformation, t: torsion. Constributions 2 10.

~~ ~~ ~~~ ~ ~ ~~ 


Table 5-10. Comparison of some modes of ccl-poly( L-glutaniic acid) and q1-poly(L-alanine) with 

those of ccl-poly(L-alaniiie). 

Moded u1-PLAb ~I-PLGH' MII-PLA 

Observed Calculated Observed Calculated Observedd Calculatedb 

I 

1658 1657iA) 

1655 1655(El) 

1653 

1652 

1657(A) 

1655(El) 

1667 1667(A) 

1661 166O( El) 

I1 

1545 1538(El) 

1516 1519(A) 

1 550 

1510 

1537(Ei) 

1517(A) 

1547 1540( Ei ) 

1515(A) 

I11 

1338 1345(El) 

1278 1287iE2) 

1270 1278(E1) 

1265 1262(A) 

1340 

1296 

1283 

1 326(El) 

12991 E2) 

1287(Ei) 

1263(A) 

1346(E1 

) 

1281 (E2) 

1272(El) 

1260( A) 

NCn s, C"CP s 

1178(A) 

1 167(El) 

1118 

1 l29(A) 

1129(El) 

1175(A) 

1 169(EL) 

CN s. CNC" d 

V 

909 910(A) 

893 901 (€5 

658 660(E1) 

618 608(E1) 

928 

924 

670 

618 

929( El) 

922(A) 

678(Ei) 

626iE1) 

909(A) 

900(E1) 

1 

615(Ei) 

CO ib, C"CN d 

528 537(A) 

565 

549(A) 

536(A) 

NC'C d 

528 522(El ) 

515(E1) 

518(E1) 

CO ob, Cb b 

Cb b 

CTP t 

375 374(Ei) 

366 367(A) 

292 307(A) 

260 264(A) 

409 

318 

280 

402(E1) 

368(A) 

334(A) 

265(A) 

376(A) 

369(E1) 

302jA) 

274(A) 

a I, 11, 111. V: amide modes, in cm-'; s: stretch, b: bend, ib: in-plane bend, ob: out-of-plane bend, 

d: deformation, t: torsion. 

From [106]. 

From [139]. 

From [140]. 

Small changes in backbone structure, such as are represented by the aII-helix, also 

lead to predictable spectral changes, as seen from Table 5-10. The amide I mode, 

found about 10 cm-' higher in (putative [107]) qr-helices [140], is expected to be 

particularly affected because of the weaker hydrogen bond resulting from the tilted 

peptide group. The other higher-frequency modes do not change very much, but 

there are some possible characteristic changes in the low frequency region: the frequency 

order of the symmetry species of the CO ob, Cb b modes inverts, and the 

frequency separation between Cb b and C*Cb t is expected to decrease significantly 

in an. Thus, vibrational spectra can be sensitive to even very small changes in 

structure.

218 5 Vibratioid Spectroscopy oj Pollyeprides 

In addition to frequencies, dichroic properties can be altered by small structural 

changes. Since the peptide group in the uII-helix is more inclined to the helix axis 

than it is in the a,-helix, the DRs of amide modes are expected to differ significantly 

between the two structures. To calculate the DR of, for example, ainide I of an 

a-helix we need to know (?g/dQ~),, and (l@/?Q,),, which involves knowing the L,, 

and dc/dS, (see Eq. (5-5)). Since the L,r are known from the normal mode calculation 

and the d$/ciSl have been obtained from ab azirio calculations [63], it becomes 

possible to synthesize a band intensity profile for a given (even finite) ol-helix structure. 

Such a calculation has been done for the oll-helix and the a11-helix for amide I 

and aniide I1 by computing the IR intensity profiles for the two polarization components 

of these two bands [141]. The results show that the DRs are significantly 

different enough to pennit identification of these structures from this property. 

The ability to predict a band intensity profile opens up an important additional 

dimension in vibrational analysis. It means that we will be able to relate subtle 

spectral differences to small structural changes with a greater degree of confidence. 

Thus, it has been possible to confirm that an observed three-component contour of 

the amide I band of tropomyosin is indeed expected for a coiled-coil a-helix [142]. 

Extensions to understanding the normal modes of proteins become possible [ 1431. 

The systematic incorporation in an SDFF of dipoles and dipole fluxes to calculate 

IR intensities [ 1441 will finally bring to the vibrational analysis of polymeric niolecules 

the completeness and flexibility needed to make it a much more powerful 

structural tool. 

5.4.2.3 3lo-Helix 

The 3lo-helix represents a different topology than the a-helix, being 4 - I rather 

than 5 - 1, and is therefore of importance in studying the vibrational dynamics 

of polypeptide helices. At present, its only clear identification has been in PAIB, so 

its characterization from this polypeptide may lack soine generality. On the other 

hand, the excellence of the experimental data [I 101 makes its vibrational analysis 

secure. 

Since the structure of PAIB has not been determined in detail by diffraction 

methods, the normal-mode studies [110] were based on a 310 structure obtained 

from conforniational analysis [l 1 11 (Figure 5-7). The vibrational studies compared 

experimental data with predictions for two helices, and the results clearly favored 

the 310- over the a-helix. The threefold screw symmetry of this structure results in 

El and E2 species modes reducing to doubly degenerate E species modes. The main 

chain force field was the same as that for oll-PLA, with additional force constants 

refined for the (CH3)2 group. Some inodes from the full analysis [110] are compared 

with those of q-PLA in Table 5-1 1. 

The observed amide I frequencies are slightly lower for the 3lo-helix than for 

the cq-helix because of the slightly stronger hydrogen bond (N...O = 2.83 A versus 

2.86 A, respectively). The increased splitting is undoubtedly due to the different 

TDC interactions as a function of conformation. This probably also accounts for 

the large change in the A-E splitting of the aniide 11 modes. The nominal aniide 111 

modes (they contain no CN s and NH ib dominates only in the 1312 cn-' mode!

5.4 Poliyeptides 219 

Table 5-11. Comparison of some modes of 3 ~~,-poly(cc-arninoisobutyri~ acid) and ccc-poly( L- 

alanine j . 

Mode" 3 !"-PA 1 B" XI-PLA 

Observed Calculated Calculated 

I 

II 

III 

Raman 

1647 

1531 

1339 

1313 

1280 

CN s, CNC' d 908 

V 

NC"C d 594 

C"CN d, CO ib 568 

IR 

1656 11 1665(A) 

1661iE) 

1545 i 1547(E) 

1533(Aj 

1346(E) 

1312(A) 

1280 )I 1287iA) 

905 I/ 905(A) 

694 1 701(E) 

680 I1 676(A) 

595 II 594(A) 

557(A) 

1657(A) 

1655iEl) 

1538( El) 

15 19(A) 

1345(El) 

1287(E1) 

1278( E2) 

1262(A) 

910(A) 

660(E1) 

608(E) 

589(A) 

522(E1) 

537iA) 

a I, 11. 111, V: ainide modes in cm-l; s: stretch, d: deformation, ib: in-plane bend. 

From [110]. 11: 

parallel dichroism, I: perpendicular dichroism. 

From [106]. 

are well accounted for, and exhibit a quite different pattern from that of the MIhelix. 

This is undoubtedly due in part to the different involvement of side-chain 

structures in the two polypeptides. Since no such major difference is seen in the 

aniide V eigenvectors, the large frequency differences between the two helices must 

be due to differences in hydrogen-bonding geometry and/or main-chain conformation. 

(The disappearance of the 694 and 680 cm-l bands on N-deuteration [110] 

unambiguously confirms their assignments to ainide V.) The skeletal N CT d and 

CTN d, CO ib modes have reversed frequency order (and in one case changed 

species) in 3l"-PAIB, and again, this is probably partly due to the involvement of 

side-chain coordinates in the eigenvectors. Future studies of, for example, a 3 10- 

helix of PLA should help to establish the possible generality of these results. 

5.4.2.4 Polyglycine I1 

The PGII structure is the simplest one that is representative of polypeptide helices 

without intramolecular hydrogen bonding. In this case, the antiparallel chains of

280 5 Vibrational Spectroscopy of Polvpeytides 

essentially threefold helices [ 122, 1231 are completely intermolecularly N-H . .. O=C 

hydrogen bonded (Figure 5-8), with strong spectroscopic evidence [ 1231 for additional 

C"-H...O=C hydrogen bonds. The basis for this conclusion is of interest 

since it demonstrates the sensitivity of the spectroscopic technique to subtle features 

of the structure. 

The vibrational analysis [ 1231 was based on comparing the normal modes of 

a parallel-chain and an antiparallel-chain structure of PGII. In both structures all 

N-H... O=C hydrogen bonds can be made. However, whereas in the parallel-chain 

arrangement all CH2 groups participate equivalently in C"-H ... O=C hydrogen 

bonds (implying that all C=O groups have bifurcated hydrogen bonds), in the 

antiparallel-chain arrangement only every third CH2 group can be involved in such 

interactions, and only between like-directed chains. The result is that the chain no 

longer has strict threefold symmetry, and as a result the degeneracy of the E species 

modes is broken. This leads to the expected, and observed, presence of a larger 

number of bands than predicted by a strictly threefold-symmetric chain structure, 

and also of bands that occur in regions where no modes are predicted for the 

parallel-chain structure. Such symmetry-based conclusions are powerful constraints 

on structural possibilities. 

The antiparallel-chain structure (Figure 5-8) has only a screw axis of symmetry, 

and its modes are distributed as follows: A-62 modes, Raman, IR; B-61 modes, 

Raman, IR. As for PGI, extensive IR spectra on isotopic derivatives (NH, CD2; 

ND, CH?; and ND, CD2) were available [97] as well as Raman spectra on the N- 

deuterated molecule [98]. The PGI force field [96] was used as a starting point, and 

refinement required small adjustments in 10 of -70 intramolecular force constants 

[123]. Amide mode frequencies are given in Table 5-12; the overall rnis frequency 

error is 5.4 cm-'. 

The hydrogen bond in PGII is stronger than that in PGI (N...O=2.69 b, vs 

2.91 A), so it is not surprising that the average amide I mode frequency is lower in 

PGII (the unperturbed modes are at 1651.0 f 1.7 cm-I). The very different splittings 

are due to the different TDC interactions for the two structures. The somewhat 

higher amide I1 modes are also consistent with stronger hydrogen bonds. The 

amide I11 modes with significant NH ib, CN s contributions are found near 1280 

cm-', strongly mixed with CH2 tw, quite different from the situation in PGI. The 

isolated NH ib contribution is now found in the 1400-1300 cm-' region. The aniide 

V modes in PGII are divided between a higher frequency group near 750 cmP1 and 

a lower frequency group near 670 cm-'. This is a much larger separation than 

found for other helical structures, and may be a consequence of the interchain 

hydrogen bonding. 

5.5 Summary 

The main message conveyed by the body of present vibrational spectroscopic 

studies of peptides and polypeptides is that detailed information on their structures

Table 5-12. Aniide mode frequencies (in cm-' ) of antiparallel-chain polyglycine 11. 

Observed' Calculated Potential energy distribution 

Rainan IR A B 

(1656 CO s(71) CN ~(20) C"CN d(l0) 

1654 CO s(74) CN s(20) C"CN d( 10) 

1654VS 1655W 

1653 CO s(72) CN ~(20) CTN dil0) 

( 1651 

CO s(73) CN ~(19) C"CN d(10) 

1649 

CO 473) CN s(20) CTN d(10) 

1640VS 1645 CO 474) CN s(19) CTN d(l0j 

1560W 156OW 1565 1565 NH ib(59) CN s(18) 

1550s 

1555 1552 NH ib(52) CN s(20) C"C s(10) 

1548 1548 NH ib(54) CN s(21) '2°C s(12) 

1383MS 1377M 1353 1380 CH? ~(58) C"C ~(15) NH ib(13) 

1350 1350 CH2 w(46) CH2 b(16) NH ib(16) C"C s(14) 

1334VW 1332VW 1344 1345 CH2 w(50) NH ib(16) CH2 b(14) C" C ~(13) 

1303 1304 CH1 tw(31) NH ib(14) CH2 w(13) CN s(12) 

1283M 1283M 1290 1290 CH? tw(29) CH2 w(23) NH ib(13) CN s(12) 

752VW 751W 760 759 CN t(20) NH ob(19) NH-.O ib(14) NC"C d(13) 

CO ib( 12) 

742VW 740 CN t(53) NH.-.O ib(22) NH ob(l8) 

673M 

740M 738 

678 

661 

CN t(49) NH...O ib(20) NH ob( 16) CO ib( 11) 

CN t(77) NH ob(27j NH t(l1) CO ob( 10) 

CN t(83) NH ob(28) NH t( 11) NH.-O ib( 10) 

CN t(88) NH ob(30) NH t(14) 

{ 

a Data taken from [123]. S: strong, M: medium, W: weak, V: very. 

b s: stretch, b: bend, ib: in-plane bend, ob: out-of-plane bend, d: deformation, w: wag, tw: twist, 

t: torsion. Contributions 2 10. 

and interactions can be obtained if experimental data are interpreted by careful 

normal-mode analysis. Although the systematic application of extensive empirical 

force fields has revealed the validity of this approach [5], the full power of this 

technique remains to be exploited. Our discussion has concentrated, because of 

space limitations, on the amide modes, but of course the sensitivity to conformation 

and forces resides in the entire spectrum. The key to extracting this information will 

be the physical reliability of the force fields used in the normal-mode calculations. 

We have come a substantial distance toward this goal with existing force fields [5], 

but improvements can be made and some are already in sight. These presently involve 

empirical force fields, such as the development of an ab initio-based MI-PLA 

force field [145] utilizing better IR [145] and polarized Raman [146] data as well as 

an understanding of helix vibrations [ 1471 that avoids the weak-coupling and perturbation 

approximations 191. However, the ultimate goal will be achieved when 

we have a conformation-dependent force field as represented by an SDFF for the. 

polypeptide chain [51].

282 5 Vibrational Spectroscopy of Polypt>ptiu'c..c 

Acknowledgments 

The author's contributions to the development of this field have been supported by 

the National Science Foundation through grants from its Biophysics and Polymers 

Programs. Helpful discussions with Noemi G. Mirkin and Kim Palmo are gratefdly 

acknowledged. 


Stryer, L., Biocheniistry. Fourth Edition. New York: W. H. Freeman & Co., 1995. 

Richardson, J., Ad[). Protein Cliem. 1981, 34, 167-339. 

Surewicz, W. K., Mantsch, H. H.. Bioihini. Biopliys. Acto 1988, 9-72, 115-130. 

Surewicz, W. K., Mantsch, H. H., Chapman, D., Biochemistry 1993, 32, 389-394. 

Krinim, S. and Bandekar, J., Ado. Protein Clieni. 1986, 38, 181-364. 

Herzberg, G., Molecular Spectra and Struelure. I1 Injkved cind Rninan Spectin oj'Pulyatornic 

Molectiles. New Jersey: Van Nostrand-Reinhold, 1945. 

Wilson, E. B., Decius, J. C., Cross, P. C., Molecular. Vibrations. New York: McGraw-Hill, 

1955. 

Califano, S., Vibrational States. New York: Wiley, 1976. 

Higgs, P. W., Proc. Roy. Soc. London Ser. A 1953,220, 472-485. 

Shimanouchi, T., in: Physical Chemistry: An Aducmceti Treatise: Eyring, H., Henderson, D., 

Jast, W. (Eds.). New York: Academic Press, 1970; Vol. 4, pp. 233-306. 

Person, W., Zerbi, G., Eds., Vibrationcil Intensities in IrIfitiretl aIid Rtirnnn Siiectroscopy. 

Amsterdam: Elsevier, 1982. 

Qian, W., Krimni, S., J. Plzys. Chem 1993, 97, 11578-1 1579. 

Cheam, T. C., Krinim, S., J. Mol. Struct. 1986, 146, 175-189. 

Tasumi, M., Krimm, S., J. Chem. Plzys. 1967, 46, 755-766. 

Krimm, S., Abe, Y., Proc. Natl. Act7d. Sci. U.S.A. 1972, 69, 2788-2792. 

Moore. W. H., Krimm, S., Proc. Nrrtl. Acntl. Sci. U.S.A. 197.5, 7-7, 4933-4935. 

Cheam, T. C.! Krimm, S.. Chem. Plijx Lett. 1984, 107, 613-616. 

Abe, Y., Krimm, S., Biopo/ytners 1972, 11, 1817-1839. 

Moore, W. H., Krimm, S., Biopolymers 1976, 15, 2439-2464. 

Cheam, T. C., Krimm, S., J. Cliern. Phys. 198.5, S2, 1631-1641. 

Schachtschneider, J. H., Snyder, R. G., Spectrochiin. Actu 1963, 19, 117-168. 

Dwivedi, A. M., Krimm, S., J. Phy.~. Clzein. 1984, 88, 620-627. 

Hehre, W. J., Radoni, L., v. R. Schleyer, P., Pople, J. A,, Ab Iizitio Moleciilrir Orbited Theory. 

New York: John Wiley & Sons, 1986. 

Fogarasi, G., Pulay, P., in Vibrational Siiectm mid Structure: Durig, J. R. (Ed.). Amsterdam: 

Elsevier, 1985; Vol. 14, pp. 125-219. 

Mirkin, N. G., Krimm, S., J. Mol. Strtrct. (Theockern) 1995, 334, 1-6. 

Mirkin, N. G., Krimm, S., J. Phys. Cheni. 1993, 97, 13887-13895. 

Qian, W., Krinim, S., J. Phj>s. Chem. 1994, 98, 9992-10000. 

Qian, W., Krimm, S., Biopolyiiievs 1994, 34, 1377-1394. 

Zhao, W., Bandekar, J., Krimm, S., J. Mol. Struct. 1990, 238, 43-54. 

Qian, W., Zhao, W., Krimm, S., J. Mol. Struct. 1991, 250, 89-102. 

Qian, W., Krirnm, S., Bio&niers 1992, 32, 1025-1033. 

Qian, W., Krimm, S., Biopolyniers 1992, 317, 1503-1518. 

Cheam, T. C., Krimm, S.. J. Mol. Struct. ( Theoclzeni) 1990.206, 173-203. 

Cheam. T. C., Kriinm, S., J. Md Struct. (Tlzeoclieni) 1989, IS8, 15-43.

1351 Lifson, S., Stern, P. S., J. Cliew. Phys. 1982. 77. 4542-4550. 

[36] (a) Lii, J.-H., Allinger, N. L., J. Am. Cliem. Soc. 1989, 111, 8566-8575. (b) Allinger, N. L., 

Chen, K., Lii, J.-H.. Corny. Chi. 1996, 17, 642-668. 

[37] Lii, J.-H., Allinger, N. L., J. Cony. Cl~ct~i. 1991, 13, 186-199. 

[38J Dasgupta, S., Goddard 111, W. A., J. Cliem. Plij~s. 1989, 90, 7207-7215. 

1391 Karasawa, N., Dasgupta, S., Goddard 111, W. A., J. Phys. Chem 1991, 95, 2260-2272. 

1401 Hwang, M. J., Stockfisch, T. P., Hagler, A. T., J. iln.1. Clieni. Soc. 1994, 116, 2515-2525. 

1411 Cornell, W. D., Cieplak, P., Bayly, C. I., Gould, 1. R., Merz, Jr., K. M., Ferguson, D. M., 

Spellineyer, D. C., Fox, T., Caldwell, J. W., Kollman, P. A., J. Atit. Clieui. SOC. 1995, 117, 

5179-51 97. 

1421 Miwa, Y., Machida, K.? J. Am. Cl7e117. Soc. 1988, 110, 5183-5189. 

[43] Derreumaux, P., Dauchez, M., Vergoten, G., J. Mol. Str~rct. 1993,295, 203-221. 

1441 Derreumaux, P., Vergoten. G., J. Clwrn. Plz.vs. 1995, 10-1, 8586-8605. 

1451 Palmo, K., Pietila, L.-O., Kriinm, S., Coniputer.~ Clieiv 1991, 15, 249-250. 

1461 Palmo, K., Pietila. L.-O., Krimm, S., J. Cornp. Chem. 1991, 12, 385-390. 

1471 Palmo, K., Mirkin, N. G., Pietila, L.-O., Krimm, S. Macromolecules 1993, 26, 6831-6840. 

[48] Palmo, K., Pietila, L.-O., Krimm, S., Cowipicters Clzeriz. 1993, 17, 67-72. 

[49] Palnio, K., Pietila, L.-O., Krimm, S. J. Con7p. Clieni. 1992, 13, 1142-1150. 

[SO] Pahno, K., Mirkin, N. G., Krimm. S., J. Phys. Cliem. A 1998, 102, 6448-6456. 

1511 Palmo. K., Mirkin, N. G., Krimm, S., to be published. 

1521 Miyazawa, T., Shimanouchi, T., Mizushima, S., J. Clzeni. Phys. 1958, 29, 611-616. 

[53] JakeS, J., Schneider, B., Coil. Czech. Clieni. Cornrn~iri. 1968, 33, 643-655. 

1541 JakeS, J., Krimm, S., S/~ecrrochitn. Actn 1971, 27A, 19-34. 

[55] Rey-Lafon, M.. Forel, M. T., Garrigou-Lagrange, C., Spectrochitii. Act0 1973, 29A, 471- 

486. 

[56] Sugawara, Y., Hirakawa, A. Y., Tsuboi. M., J. Mol. Spcctrosc. 1984, 108, 206-214. 

1571 Sugawara, Y., Hirakawa, A. Y., Tsuboi, M., Kato, K., Morokuma, K., J. Mol. Spxmxc. 

1986,115, 21-33. 

[58] Balazs, A., J. Mol. Stvuct. (Tlleoch7i) 1987, 153, 103-120. 

[59] Mirkin, N. G., Krimm, S., J. Mol. Struct. 1991, 242, 143-160. 

[60] Ataka, S., Takeuchi, H., Tasumi, M., J. Mol. Stixct. 1984, 113, 147-160. 

[61] Mirkin, N. G., Krinim, S., J. Am. Clieni. Soc. 1991, 113, 9742-9747. 

1621 Williams, R. W., Biol~olyiiiers 1992, 32, 829-847. 

[63] Cheam. T. C., J. Mol. Struct. (Tlieoclieni) 1992, 257, 57-73. 

[64] Mirkin, N. G., Krimm, S., J. Mol. Struct., 1996, 377, 219-234. 

[65] Miyazawa, T., Shimanouchi, T., Mizushima, S., J. Clien?. Plzys. 1956, 24, 408-418. 

[66] Miyazawa, T., in: Polyaniiiio Acids, Po/.ypeptides, and Proteins: M. A. Stahmann (Ed.). 

Madison: University of Wisconsin Press, 1962; pp. 201-217. 

[67] Chen, X. G., Schweitzer-Stenner, R., Krimni, S., Mirkin, N. G., Asher, S. A., J. Ani. Che’hrnz. 

SOC. 1994,116, 11111-11142. 

1681 Chen, X. G., Schweitzer-Stenner, R.? Asher, S. A., Mirkin, N. G., Krimm, S., J. Plz~vs. Clzeni. 

1995, 99, 3074-3083. 

[69j Song, S., Asher, S. A., Krimm, S., Bandekar, J., J. Am. Chem Soc. 1988, 110, 8547-8548. 

[70] Miyazawa, T., Bull. Clieni. Soc. Japan 1961, 34, 691-696. 

[71] Fillaux, F., de Loze, C., Clzern. Phys. Lett. 1976, 39, 547-551. 

1721 Hsu, S. L., Moore, W. H., Krimm, S., Biopolyniecr 1976, 15, 1513-1528. 

[73] Mirkin, N. G., Krimm, S., private communication. 

[74] Cheam, T. C.! J. Mol. Sttwt. 1993, 295, 259-271. 

1751 Bandekar, J., Krimm, S., Biopolyiners 1988, 27, 885-908. 

[76] Sundius, T., Bandekar. J., Krimm, S., J. Add. Sfnrct. 1989, 214, 119-142. 

(771 Qian, W., Bandekar, J., Krimm, S., Biopo/.viiienr 1991, 31, 193-210. 

1781 Pauling, L., Corey, R.B.. Proc. Natl. Acnd. Sci U.S.A. 1951. 37, 729-740. 

[79] Pauling, L., Corey, R. B., Proc. Natl. Acad. Sci. US. A. 1953, 39, 253-256. 

[80] Moore, W. H., Krimm, S., Biopolyrners 1976, 15, 2465-2483. 

1811 Arnott, S., Dover, S. D., Elliott, A,, J. Add Biol. 1967, 30, 201-208. 

[S2] Bandekar, J., Krimm. S.. Biopo1ynier.r 1988, 27, 909-921,

284 5 Vibi~itionnl Spectroscopy qf Polypeptides 

[83] Chirgadze, Yu. N., Nevskaya, N. A.. Biopo/,vnwrs 1976, 15, 607-625. 

[84] Chirgadze, Yu. N., Nevskaya, N. A,, Biopolynevs 1976, 15, 627-636. 

[85] Elliott, A., Puoc. ROY. Soc. Loridoil Ser. A 1954, 226, 408-421. 

[86] Itoh, K., Katabuchi, H., Biopolwwrs 1972, 11, 1593-1605. 

[87] Masuda, Y., Fukushima, K., Fujii, T., Miyazawa. T., Biopo/wners 1969, 8, 9 1-99, 

[88] Dwivedi, A. M., Krimm, S., Macrornoleciiles 1982. 15, 186-193. 

[89] Fanconi, B., Biupo&r?ws 1973, 12, 2759-2776. 

[90] Krimm, S., Dwivedi, A. M., J. Rmm7n Spcctrosc. 1982, 12, 133-137. 

1911 Itoh, K., Nakahara? T., Shimanouchi, T., Oya, M., LJno, J., Iwaknra, Y., Biopo/jwwr.y 1968, 

6, 1759-1766. 

[92J Sengupta, P. K., Krimm, S., Hsu, S. L., Bioyolpiners 1984, 23, 1565-1594. 

[93] Keith, H. D., Padden, F. J.. Giannoni, G., J. h4o/. Biol. 1969, 43, 423-438. 

[94] Lotz, B., J. Mol. Bid. 1974, S7, 169-180. 

[95] Colonna-Cesari, F., Premilat, S., Lotz, B., J. Mol. Bid. 1974, 87, 181-191. 

[96] Dwivedi, A. M., Krimm, S., Mucroniolecules 1982. 15, 177-155. 

[97] Suzuki, S., Iwashita, Y., Shimanouchi, T., Tsuboi, M., Bioyolyrners 1966, 4, 337-350. 

[98] Small, E., Fanconi, B.. Peticolas. W. L., J. Clzeiri. Pliys. 1970, 52, 4369-4379. 

[99] Huggins, M. L., Chem Reos. 1943. 3-7, 195-218. 

[loo] Crane, H. R., Sci. Mon. 1950, 70, 376-389. 

[loll Bragg, L., Kendrew. J. C., Perutz, M. F., Proc. Roy. Soc. London Ser.. '4. 1950, 203, 321-357. 

[102] Pauling, L.. Corey, R. B., Banson, H. R., Proc. Nutl. Acud, Sci. U.S.A. 1951, 37, 205-211. 

[I031 Pauling, L., Corey, R. B., Proc. Nutl. Acurl. Sci. U.S.A. 1951, 37, 235-240. 

[I041 Amott, S., Dover, S. D., J. Mol. Bid. 1967, 30, 209-212. 

I1051 Miyazawa, T., J. Polynz. Sci. 1961, 55, 215-231. 

[lo61 Dwivedi. A. M., Krimm, S., Biopo/,ymers 1984, 23> 923-943. 

[lo71 Krimm, S., Dwivedi, A. M., Science 1982,216, 407-408. 

[lo81 Whitby, F. G., Kent, H., Stewart, F., Stewart, M., Xie, X., Hatch, V., Cohen, C., Phillips, 

Jr., G. N., J. Mol. Bid 1992, 227, 441-452. 

[lo91 Malcolm, B. R., Biopo/ymers 1983, 22, 319-321. 

[110] Dwivedi. A. M., Krimm, S., Malcolm, B. R., Biopol-vn7er.c 1984, 23, 2025-2065. 

[ill] Prasad, B. V. V., Sasisekharan, V., Macroniolecu/es 1979, 12, 1107-11 10. 

[112] Bradbury, E. M., Brown, L., Downie, A. R., Elliott, A,, Fraser, R. D. B., Hanby, W. E., J. 

Mol. Bid. 1962, 5, 230-247. 

[I 131 McGuire, R. E., Vanderkooi, G., Momany, F. A,, Ingwall, R. T.. Crippen, G. M., Lotan, N., 

Tuttle, R. W.. Kashuba, K. L., Scheraga, H. A,, M~/croniolecules 1971, 4, 112-124. 

[114] Nambudripad, R., Bansal, M., Sasisekharan, V., Int. J. Peptide Protein Re.7. 1981, 18, 374- 

382. 

[115] Saski, S.: Yasumoto, Y.. Uematsu, I., Mucrornolecult~s 1981, 14, 1797-1801. 

[I 161 Low, B., Grenville-Wells, H. J., Proc. Narl. Acud Sci. US. A. 1953, 39, 785-801. 

11171 Donohue, J., Froc. Nutl. Acud Sci. U.S.A. 1953, 39, 470-478. 

[I181 Urry, D. W., Pvoc. Nutl. Acad. Sci. U.S.A. 1971, 68, 672-676. 

[I191 Lotz, B., Colonna-Cesari, F.? Heitz, F.. Spach, G., J. Mol. Biol. 1976, 106, 915-942. 

[I201 Crick, F. H. C., Rich, A,, Nuture 1955. 176, 780-781. 

[121] Krimm, S., Nature 1966,212, 1482-1483. 

[ 1221 Ramachandran, G. N., Ramakrishnan, C., Venkatachalam, C. M., in: Corzforinution of Biopolymers: 

Ramachandran, G. N. (Ed.) New York: Academic Press, 1967; Vol. 2, pp. 429- 

438. 

[123] Dwivedi, A. M., Krinini, S., Biopolymeus 1982, 21, 2377-2397. 

11241 Ramachandran, G. N., Sasisekharan, V., Ramakrishnan, C., Biochirn. Biophys. Actu 1966, 

112, 168-170. 

[125] Krimm. S., Kuroiwa, K.. Rebane. T., in: Conformi~tion of' Biopoljwiers: Ramachandran, G. 

N. (Ed.) New York: Academic Press, 1967; Vol. 2, pp. 439-447. 

[126] Krimm. S., Kuroiwa, K., Biopol-viners 1968, 6, 401-407. 

[I271 Cowan, P. M., McGavin, S., Ni7ture 1955, 176, 501-503. 

[IZSJ Arnott, S., Dover, S. D., Actu Cr~~mllogr. 1968, B24, 599-601. 

11291 Sasisekharan. V., Acfu Cry~t. 1959, 12, 897-903.

5.6 Refeverices 285 

[130] Tiffany, M. L., Krimm, S., Biopo/ynzer.s 1968, 6, 1379-1382. 

[131] Krimin, S., Mark. J. E., Proc. Nof/. Acnd Sci. U.S.A. 1968, 60, 1122-1129. 

[I321 Sengupta. P. K., Krimm, S., Biopo/i~rnt~rs 1987, 26, S99-Sl07. 

[133] Woody, R. W., Ado. Biop/ivs. Chetn. 1992. -7, 37-79. 

[134] Ramachandran, G. N., Kartha, G., Norm 1955, 176, 593-595. 

[135] Rich, A., Crick, F. H. C., J. Mol. Biol. 1961, 3, 483-506. 

[136] Traub, W., Shmueli, U., in: Aspecfs qf Protciii Structzirr: Ramachandran. G. N. (Ed.) New 

York: Academic Press. 1963; pp. 8 1-92. 

[ 1371 Itoh, K.. Shirnanouchi, T.. Bioyoljwwrs 1970, 9, 383-399. 

[138] Frushour, B. G., Painter, P. C., Koenig, J. L., J. Macrorml. Sci. Reo. hIacr.onzol. Cheni. 1976, 

C15, 29-1 15. 

[I391 Sengupta, P. K., Krimm, S., Biopo/.yn?er.r 1985, 24, 1479-1491. 

[140] Rothschild, K. J., Clark, N. A,, Biophys. J. 1979, 35, 473-488. 

[ 1411 Reisdorf, Jr., W. C., Krimin, S., Bioyh-vs. J. 1995, 69, 271-273. 

[142J Reisdorf, Jr., W. C., Krimm, S. Biocken~istr,~~, 1996, 35, 1383-1386. 

[143] Krimm, S., Reisdorf, Jr., W. C., Faraday Discuss. 1994, 99, 181-197. 

[144] Palmo, K., Krimm, S., J. Coinp. Chenz. 1998, 19, 754-768. 

[145] Lee, S.-H., Krimm, S., Biopolyniers 1998, 46. 283-317. 

[146] Lee, S.-H., Krimm, S., .I. Rarnnrz Spwrosc. 1998, 29, 73-80. 

[147] Lee, S.-H., Krimm, S., Clzern. Phys. 1998, 230, 277-295.


Index 

Amide modes 249 

Amorphous polymers 16 

Anisotropic surfaces 37 

Assignments 243 

Asynchronous spectrum 9ff 

Autopeaks 11 

Band-widths 177, 180 

Bilayers systems 157 

Biopolymers 27 

Biphenyl cyano-alkyl vibrational 

spectra 146 

Bipolaron 21 1, 213, 229 

- lattice 232 

Block copolymers 26 

Broadening of vibrational bands 172 

Chain folding 150 

Charge transport 231 

Chemical regularity of polymer 

chains 99 

Conducting polymer 207 

Conformational periodicity 99, 111 

Conjugated polymer 207 

Coordinates 

- local symmetry 93 

- localized 97 

- vibrational internal 91 

Cross peaks 11 

Cross-cosl-elation function 9 

p-Cyanophenyl p-n-hexylbenzoate 

(6CPB) 40 

- bandassignment 45 

- 2D-correlation analysis 48 

2DIR 1 

Defects 

- chemical, stereochemical 126 

- conformational 143 

- mass 142 

- vibrations of 129f, 142 

Density of states 117,127 

Dichroic difference 3, 6 

Dichroic ratio 6 

Dichroism 243 

- infrared 3 

Dipeptides 246,252 

Dipole transition moment 3 

Disordered polymers 122 

Doping 208,210 

Dynamic IR linear dichroism (DIRLD) 3 

Dynamic spectrum 2 

Electric field-induced orientation 38, 42 

- variable-temperature cell 43 

Electronic measurement set-up 44 

Elongation-recovery 77 

Fatty acids conformation 144, 157 

Ferini resonances 163.168 

Ferroelectric side-chain liquidcrystalline 

polymer (FLCP) 55 

- 2D-correlation analysis 60 

- polarization geometry and mesogen 

alignment 57 

- poling experiment 58 

- schematic representation of the 

segmental motions 61 

Force constants 89, 91 

- group 92 

Force fields 240 

- ab iiiitio 24.5 

- empirical 241 

- molecular mechanics 247 

- spectroscopically determined 237 

Fourier-transform infrared (FTIR) 33

288 Index 

Glass transition 18, 26 

HCI dynamics 13 1 

In-phase spectrum 5 

Intensities, infrared and Raman 91 

Inteiphase 26 

Laser-irradiation 63 

- apparatus for on-line measurements 63f 

Lattice dynamics 88 

Longitudinal Accordion Motion 159 

Methylene group coordinates 95 

Microdomains 26 

Na-doped poly@-phenylene) 217, 223 

Near-infrared spectroscopy 75 

Nematic liquid-crystalline guest-host 

system 52 

Nematic liquid-crystalline side-chain 

polymer (NLCP) 40f, 5 1 

- 2D-cori-elation analysis 5 I 

- switching process 52 

Neutron Scattering 119f 

NIR Raman spectrometry 2 16 

Nonndecane 180 

Non-bonded interactions 244, 248 

Normal modes 90 

- alanine dipeptide 253, 256 

- N-methylacetamide 251 

- 3,,-poly(a-aminoisobutyric acid) 278 

- polyglycine I 266 

- polyglycine I1 279 

- a-poly(L-alanine) 273 

- P-poly(L-alanine) 260 

- a-poly(L-glutamic acid) 277 

Normalized intensity 47 

p-Oligophenyl 218,221 

- dianion 219, 223 

- radical anion 219, 223 

Order parameter 36 

Orientation 3, 5ff 

Orientation factor 6 

Phonon dispersion curves lOlf 

Polaron 211,213,229 

- lattice 232 

Polyacetylene 208 

Poly(acetylene), ~ IZIIZS 105,197 

Polyaniline 229 

Polyesters 147, 139 

Polyethylene 18, 20 

- films 152 

- reflection spectrum 155 

- vibrations 102, 110, 112, 122, 124, 150, 

172 

Polymer blends 20, 22 

Polymerisation of ethylene, mechanism 138 

Poly(methy1 vinylether) 23 

Polypeptide structures 

- helical 269 

- sheet 258 

Poly@-phenylene 217, 221 

Poly(pphenyleneviny1ene) 229 

Poly(propy1ene) , isotactic 113 

Polystyrene 16. 20, 22 

Poly(tetrafluoroethy1ene) 102 

Polythiophene 229 

Poly(vinylch1oride) 118,158 

Potential energy distribution 95, 242 

Proteins 27 

Quadrature spectrum 5 

Rapid-scan FTIR spectroscopy 34 

Regularity bands 113 

Relative absorbance 46 

Reversible optical data-storage 67 

Rheo-optical elongation-recovery cycle 76 

Rheo-optical FTIR spectrocopy 75 

- variable temperature stretching machine 75 

Secular equation 

- for molecules 88 

- forpolymers 101 

Selection rules 109, 125 

Semicrystalline polymers 18 

Shear-induced orientation 69 

Side-chain liquid-crystalline 

copolysiloxanes 69, 72 

Side-chain liquid-crystalline 

polyacrylates 37 

Side-chain liquid-crystalline polyester 65 

- band assignment 65 

- irradiation-induced segmental alignment 66 

- specifically deuterated 65f 

Soliton 211, 213 

- conformational 197, 199

Step-scan FTIR spectroscopy 14,34 

- data collection sequence 44 

Stereoregularity 99 

Strain, dynamic 4 

Stress-strain diagram 77 

Structural absorbance 7. 36 

Synchronous spectrum 9f 

Tacticity bands 113 

Thermal erasure of laser-induced 

alignment 68 

fmris-cis-from Isomerization 76 

Transition dipole 3 

- coupling 244 

Twistons 197 

Two-dimensional correlation analysis 9 

Two-dimensional infrared spectroscopy 1 

Vibrational coupling 95, 97 

Vibrations 

- of chain polymers 100 

- of finite chains 124 

- of one-dimensional lattices 197 

- of tridimensional lattices 11 1

Modern Polymer Spect..

Create successful ePaper yourself

Delete template?

Save as template?