LitvinovVictorM. DePrajnaparamita SpectroscopyofRubbersandRubberyMaterials ISmithersRapraPublishing2011!02!28

Spectroscopy
of Rubbers
and Rubbery
Materials
Victor M. Litvinov and Prajna P. De
Spectroscopy of Rubbers
and
Rubbery Materials
Rapra Technology Limited
Shawbury, Shrewsbury, Shropshire, SY4 4NR, United Kingdom

Telephone: +44 (0)1939 250383 Fax: +44 (0)1939 251118
http://www.rapra.net
First Published 2002 by
Rapra Technology Limited

Shawbury, Shrewsbury, Shropshire, SY4 4NR, UK
©2002, Rapra Technology Limited
All rights reserved. Except as permitted under current legislation no part

of this publication may be photocopied, reproduced or distributed in any
form or by any means or stored in a database or retrieval system, without
the prior permission from the copyright holder.
A catalogue record for this book is available from the British Library.
Every effort has been made to contact copyright holders of any material reproduced within
the text and the authors and publishers apologise if any have been overlooked.
ISBN: 1-85957-280-4
Typeset by Rapra Technology Limited

Printed and bound by Polestar Scientifica, Exeter, UK
Contents
Contents
1 Characterisation of Elastomers Using (Multi) Hyphenated
Thermogravimetric Analysis Techniques ......................................................... 1
1.1 Introduction ........................................................................................... 1
1.2 Instrumental Part ................................................................................... 3
1.2.1 Experimental ........................................................................... 12
1.3 Applications of Hyphenated Thermogravimetric Analysis
Techniques ........................................................................................... 14
1.3.1 Thermogravimetry - Fourier Transform Infrared
Spectroscopy ............................................................................ 19
1.3.2 Thermogravimetry-Mass Spectrometry .................................... 25
1.3.3 Thermogravimetry - Differential Scanning Calorimetry ........... 29
1.3.4 Thermogravimetry - Differential Scanning Calorimetry
- Mass Spectrometry ................................................................ 30
1.3.5 Thermogravimetry - Differential Thermal Analysis ................. 30
1.3.6 Thermogravimetry - Differential Thermal Analysis
1.3.7 Thermogravimetry - Gas Chromatography
1.3.8 Thermal Desorption - Gas Chromatography - Fourier
Transform Infrared Spectroscopy/Mass Spectrometry .............. 34
1.4 Future Prospects of Hyphenated Thermogravimetric Techniques
in Elastomer Characterisation .............................................................. 35
1.5 Summary .............................................................................................. 36
2 Photoacoustic Fourier Transform Infrared Spectroscopy of Rubbers

and Related Materials ................................................................................... 49
2.1 Introduction ......................................................................................... 49
2.2 History of Photoacoustic Spectroscopy ................................................ 49
i
Spectroscopy of Rubbers and Rubbery Materials
2.3 Theory of Photoacoustic Spectroscopy ................................................ 50

2.4 Instrumentation for PA-FTIR Analyses ................................................ 52
2.5 Analysis of Carbon-Filled Rubbers ...................................................... 57
2.6 Quantitative Analysis of Polymers ....................................................... 60
2.7 Surface Analysis and Depth Profiling ................................................... 66
2.8 Determination of Orientation Function of Polymeric Materials ........... 71
2.9 Conclusion ........................................................................................... 71
3 Infrared Spectroscopy of Rubbers ................................................................. 77

3.1 Introduction ......................................................................................... 77
3.2 Sample Preparation .............................................................................. 80
3.3 Different Types of IR Spectroscopy ...................................................... 81
3.4 Quantitative Analysis ........................................................................... 81
3.5 Applications of IR Spectroscopy .......................................................... 82
3.5.1 Rubber Blends ......................................................................... 92
3.5.2 Self-crosslinking Blends............................................................ 96
3.5.3 Polyurethanes .......................................................................... 98
3.5.4 Rubber-filler Interaction ........................................................ 102
3.5.5 Milling ................................................................................... 108
3.5.6 Adhesion ................................................................................ 111
3.5.7 Degradation ........................................................................... 113
3.6 Reverse Engineering ........................................................................... 113
3.7 Conclusion ......................................................................................... 114
4 Application of Infrared Spectroscopy to Characterise Chemically

Modified Rubbers and Rubbery Materials .................................................. 125
4.1 Introduction ....................................................................................... 125
4.2 The Infrared Spectra of Commonly Used Diene Rubbers ................... 125
ii
Contents
4.3 Hydrogenation ................................................................................... 126

4.4 Halogenation ..................................................................................... 135
4.5 Isomerisation ..................................................................................... 139
4.6 Cyclisation ......................................................................................... 139
4.7 Hydrosilylation .................................................................................. 142
4.8 Hydroboration ................................................................................... 144
4.9 Hydroformylation .............................................................................. 144
4.10 Oxidation .......................................................................................... 146
4.11 Phosphonylation ................................................................................ 146
4.12 Sulfonation ........................................................................................ 147
4.13 Ionomer Formation ............................................................................ 147
4.14 Ionomeric Blends ............................................................................... 151
4.15 Weathering and Degradation of Polymers .......................................... 153
4.16 Modification by Radiation ................................................................. 156
4.17 Conclusion ......................................................................................... 157
5 Infrared Spectroscopy of Rubbery Materials ............................................... 167

5.1 Introduction ....................................................................................... 167
5.2 Polyethylenes ..................................................................................... 168
5.3 Polyvinyl Chloride ............................................................................. 180
5.4 Thermoplastic Elastomers .................................................................. 188
5.5 Conclusion ......................................................................................... 200
6 Crosslinking of EPDM and Polydiene Rubbers Studied by Optical

Spectroscopy ............................................................................................... 207
6.1 Introduction ....................................................................................... 207
6.1.1 General Introduction to EPDM ............................................. 207
iii
6.1.2 EPDM Crosslinking ............................................................... 208

6.1.3 Studies into the Chemistry of Rubber Crosslinking ............... 209
6.1.4 Scope ..................................................................................... 210
6.2 Sulfur Vulcanisation ........................................................................... 211
6.2.1 Sulfur Vulcanisation of Polydiene Rubbers ............................ 211
6.3 Peroxide-curing .................................................................................. 225
6.3.1 General .................................................................................. 225
6.3.2 Polydiene Elastomers ............................................................. 226
6.3.3 EPDM .................................................................................... 227
6.4 Concluding Remarks and Future Outlook ......................................... 237
7 NMR Imaging of Elastomers ...................................................................... 247

7.1 NMR Imaging and Contrast .............................................................. 248
7.1.1 Principle of Fourier NMR ...................................................... 248
7.1.2 Spatial Resolution .................................................................. 251
7.1.3 Contrast ................................................................................. 253
7.2 Applications ....................................................................................... 264
7.2.1 Defects and Heterogeneities in Technical Elastomer
Products ................................................................................. 265
7.2.2 Covulcanisation ..................................................................... 269
7.2.3 Blending ................................................................................. 271
7.2.4 Crosslink Density ................................................................... 271
7.2.5 Vulcanisation Process ............................................................. 273
7.2.6 Ageing ................................................................................... 274
7.2.7 Sample Deformation .............................................................. 275
7.3 Spatially Resolved NMR .................................................................... 276
7.3.1 The NMR-MOUSE ................................................................ 277
7.3.2 Applications ........................................................................... 279
7.3.3 Imaging with the NMR-MOUSE ........................................... 282
7.4 Summary ............................................................................................ 282
iv
Contents
8 NMR in Soft Polymeric Matter: Nanometer Scale Probe ............................ 291

8.1 Introduction ....................................................................................... 291
8.2 Polymeric Networks .......................................................................... 292
8.2.1 Molten High Polymers ........................................................... 292
8.2.2 Crosslinked Chains ................................................................ 292
8.2.3 Semi-crystalline Polymers....................................................... 292
8.2.4 Block Copolymers .................................................................. 293
8.2.5 Loaded Polymers ................................................................... 293
8.2.6 Aggregated Polymers ............................................................. 294
8.2.7 Network Distribution Function ............................................. 294
8.3 Basis of the NMR Approach .............................................................. 295
8.3.1 Chain Elongation ................................................................... 296
8.3.2 NMR Evidence for Networks: Pseudo-solid Spin-echoes ....... 298
8.4 Crosslinked Chains ............................................................................ 299
8.4.1 End-linked Calibrated Chains ................................................ 299
8.4.2 Characteristic NMR Rates ..................................................... 299
8.4.3 Strand Length Dependence .................................................... 301
8.4.4 Randomly Crosslinked Chains ............................................... 302
8.4.5 Latex Suspensions .................................................................. 306
8.4.6 Kinetics of Gelation ............................................................... 307
8.5 Polymeric Crystallisation ................................................................... 307
8.5.1 Crystallisation-NMR Relationship ........................................ 307
8.5.2 Kinetics of Crystallisation ...................................................... 308
8.6 Entangled High Polymers ................................................................... 309
8.6.1 Temporary Networks ............................................................. 310
8.6.2 Short Chain Dynamics ........................................................... 312
8.6.3 Long Chain Dynamics ........................................................... 313
8.7 Adsorption on Mineral Aggregates .................................................... 315
8.8 Conclusion ......................................................................................... 317
v
9 Chemical Characterisation of Vulcanisates by High-Resolution

Solid State NMR ......................................................................................... 321
9.1 Introduction ....................................................................................... 321
9.2 Sulfur Vulcanisation Mechanism ........................................................ 322
9.3 The NMR Methods for Assigning Resonances to
Chemical Structure ............................................................................ 324
9.4 Unaccelerated Sulfur-vulcanisation of NR ......................................... 327
9.5 Accelerated Sulfur-vulcanisation of NR and IR .................................. 327
9.6 Sulfur-vulcanisation of BR ................................................................. 333
9.7 Sulfur-vulcanisation of SBR ............................................................... 338
9.8 Peroxide, Radiation, and High Pressure Vulcanisation ...................... 338
9.9 Vulcanisation of Other Elastomer Systems ......................................... 340
9.10 Effect of Carbon Black on Vulcanisation of Elastomers ..................... 341
9.11 Effect of Silica Filler on Vulcanisation Chemistry .............................. 347
9.12 Thermal-Oxidation of Network Structures ........................................ 347
9.13 Summary ............................................................................................ 348
10 Characterisation of Chemical and Physical Networks in Rubbery

Materials Using Proton NMR Magnetisation Relaxation ........................... 353
10.1 Introduction ....................................................................................... 353
10.2 Network Structure Analysis by Means of NMR
Transverse Magnetisation Relaxation ................................................ 355
10.3 Characterisation of Network Heterogeneity and
Network Defects ................................................................................ 360
10.4 Network Structure in Oil-Extended Rubbers
- Effect of Chain Entanglements......................................................... 366
10.5 Network Structure in Filled Rubbers - Rubber-Filler
Interface and the Structure of the Physical Network .......................... 368
vi
Contents
10.5.1 NMR Relaxation of Filled Rubbers ....................................... 368

10.5.2 Carbon-Black-Filled Rubbers ................................................. 369
10.5.3 Silica-Filled Silicon Rubbers ................................................... 374
10.5.4 Silica-Filled Conventional Rubbers ........................................ 378
10.6 Chains Grafted onto a Filler Surface .................................................. 379
10.7 Semi-Crystalline Elastomers ............................................................... 381
10.8 Ionic Viscoelastic Materials ............................................................... 383
10.9 Rubbery Phases in Blends and Emulsions .......................................... 384
10.10Real-Time NMR Experiments............................................................ 385
10.11Low-Field NMR Magnetisation Relaxation Experiments
for Quality Control Purposes ............................................................. 387
10.12Conclusions ....................................................................................... 388
11 High Resolution NMR of Elastomers ......................................................... 401

11.1 Introduction ....................................................................................... 401
11.2 Structural Feature of Elastomers ........................................................ 402
11.3 Analysis of Chemical Composition and Sequence Distribution .......... 404
11.3.1 Accuracy of NMR Measurements of Chemical
Composition .......................................................................... 404
11.3.2 Analysis of Chemical Composition Distribution
using SEC-NMR .................................................................... 410
11.3.3 Analysis of Sequence Distribution .......................................... 413
11.3.4 Analysis of Chemically Modified Structure
and Graft Polymers ................................................................ 419
11.4 Analysis of End-groups and Branching .............................................. 422
11.4.1 Assignment of Small Signals .................................................. 422
11.4.2 Functionality of Telechelic Diene Polymers ............................ 423
11.4.3 Structure of Terminal Groups ................................................ 424
11.4.4 Structure of Branch-points and Coupling Points .................... 426
vii
11.5 Structural of Naturally Occurring Polyisoprenes ............................... 429

11.5.1 Structure of Natural cis- and trans-Polyisoprenes .................. 429
11.5.2 Structure of Natural Rubber .................................................. 433
11.6 Application of High-Resolution NMR ............................................... 436
11.6.1 Multinuclear High-resolution NMR ...................................... 436
11.6.2 NMR of Swollen State and Latex .......................................... 443
11.7 Conclusion ......................................................................................... 448
129
12 Xe NMR of Elastomers in Blends and Composites.................................. 457
12.1 Introduction to 129Xe NMR Spectroscopy of Materials ...................... 457
129
12.1.1 Xe NMR Spectroscopy ....................................................... 457
12.1.2 NMR of 129Xe in the Gas Phase ............................................. 458
129
12.1.3 Xe NMR of Polymers ......................................................... 459
129
12.1.4 Xe Pulsed Field Gradient Echo (PFGE) Spectroscopy ......... 462
12.2 Experimental ..................................................................................... 463
129
12.3 Xe NMR of Ethylene/Propylene Copolymers in Blends
and Block-copolymers with Polypropylene ........................................ 465
129
12.4 Xe NMR of an Ethylene/Propylene/Diene Terpolymer
in Carbon Black Composites .............................................................. 474
12.5 The Self-diffusion Coefficient of Xe in Elastomers ............................. 480
12.5.1 General .................................................................................. 480
12.5.2 Xe Diffusion in a iPP/EPDM Blend ........................................ 482
12.5.3 Xe Diffusion as a Function of the Diffusion Time .................. 485
12.6 Conclusions ....................................................................................... 487
13 Swollen Rubbery Materials: Chemistry and Physical Properties

Studied by NMR Techniques ...................................................................... 491
13.1 Introduction - The Swelling of Crosslinked Rubbers ......................... 491
13.1.1 The Theory of Rubber Swelling ............................................. 491
13.1.2 Relationship with NMR Parameters ...................................... 492
viii
Contents
13.1.3 Reviews of NMR of Crosslinked Rubbers

and Polymers ......................................................................... 493
13.2 Motion of Polymer Chains in Polymer Solutions and
Swollen Networks .............................................................................. 494
13.2.1 Background ........................................................................... 494
13.2.2 Solutions and Bulk Rubbers ................................................... 494
13.3 Motion of Small Molecules in Swollen Rubbers ................................ 496
13.3.1 Self-Diffusion of Small Molecules in Rubbers ........................ 496
13.3.2 Self-Diffusion of Rubbers ...................................................... 498
13.4 NMR Imaging of Swollen Rubbers .................................................... 499
13.4.1 Macroscopic Diffusion of Small Molecules
in Swollen Rubbers ................................................................ 499
13.4.2 NMR Imaging of Swollen BR and Polyisoprene Rubbers ...... 500
13.5 Studies of Network Density in Swollen Rubbers and Blends .............. 502
13.5.1 Measurements of Transverse Relaxation Times ..................... 502
13.5.2 Estimation of Crosslink Density from NMR Linewidths ....... 506
13.5.3 High-Resolution 13C MAS NMR of Rubbers ......................... 510
13.6 Summary ............................................................................................ 511
14 Multidimensional NMR Techniques for the Characterisation

of Viscoelastic Materials ............................................................................. 519
14.1 Introduction ....................................................................................... 519
14.2 Basics of NMR in Viscoelastic Polymers ............................................ 520
14.2.1 Anisotropic Spin Interactions ................................................. 520
14.2.2 Manipulation of Spin Interactions ......................................... 522
14.2.3 Residual Couplings and Dynamic Order Parameters ............. 524
14.2.4 One-dimensional NMR Studies of Molecular
Motions and Dynamic Order ................................................. 526
14.3 Multidimensional NMR Spectroscopy of Viscoelastic Materials........ 529
14.3.1 Principle of Multidimensional NMR ..................................... 529
ix
14.3.2 Two-dimensional NMR Techniques and

Applications to Viscoelastic Polymers .................................... 530
14.5 Conclusion .......................................................................................... 550
15 Deuterium NMR in Rubbery Materials ...................................................... 557

15.1 Introduction ....................................................................................... 557
15.2 2H NMR Background ........................................................................ 559
15.3 Model, Labelled Rubber Systems ....................................................... 561
15.3.1 General Presentation: End-linked versus Randomly
Crosslinked Networks ........................................................... 561
15.3.2 Pseudosolid Behaviour ........................................................... 561
15.3.3 Dangling Chains .................................................................... 563
15.3.4 Time Scale of Motions ........................................................... 563
15.3.5 Uniaxially Deformed Model Networks .................................. 566
15.3.6 Physical Origin of the Induced Orientation............................ 573
15.3.7 Correlation to Elastic Properties ............................................ 575
15.4 Using Deuterated Probes .................................................................... 579
15.5 Filled, Composite and Thermoplastic Elastomers .............................. 582
15.5.1 Filled Elastomers .................................................................... 582
15.5.2 Semi-crystalline Polymers....................................................... 584
15.5.3 Thermoplastic Elastomers ...................................................... 585
15.6 Conclusion .......................................................................................... 588
Abbreviations .................................................................................................... 597
Contributors ...................................................................................................... 607
Index ................................................................................................................. 611
x
Preface
Preface
This book deals with the application of spectroscopic techniques for characterisation of
chemical and physical structures in viscoelastic materials, such as unvulcanised elastomers
and their vulcanisates, various rubbery materials and some plastics, which when blended
with particular additives (plasticisers) behave like rubbers. The common feature of all
these materials is a low glass transition temperature and a molecular network, which is
formed due to chemical and/or physical junctions between long chain molecules.
Depending on the applied force this network, neither liquid nor solid, reveals both solid-
like and liquid-like properties. This dual behaviour determines the application areas of
these materials. A distinguishing feature of rubbers is good elasticity, which is largely
determined by the network structure and phase/components composition of the material.
Despite the apparent simplicity, the network structure has a complex topology, which
can significantly affect the functional properties. Different types of network heterogeneity
are formed in the cured material, if no special precautions are taken to control the curing
chemistry and conditions. A difference in curing conditions through the sample volume,
such as the content of vulcanisation agents and temperature, results in spatial heterogeneity
of the network structure. Ageing and weathering can be other reasons for the spatial
heterogeneity. The structure of initial rubber and vulcanisation chemistry could cause
molecular scale heterogeneity of the network.
Analysis of the rubbery materials is complicated by the fact that rubbery products, such
as tyres, tubes, seals, V-belts and hoses, contain in the rubbery matrix a significant amount
of various compounds, i.e., fillers, vulcanising agents, antioxidants and plasticisers. Due
to the complex composition, no single technique can provide a good understanding of
the effect of chemical and physical structures on the functional properties of rubbery
materials. It is acknowledged that traditional methods, such as mechanical measurements
and equilibrium swelling, are not capable of giving reliable information about the network
structure and phase composition of rubbery materials, since these methods determine
volume average material properties. Local methods, which probe molecular properties,
are well suited for this purpose. Apparently, the most comprehensive information on
chemical and physical structures in the relation to material properties can be obtained by
using a combination of macroscopic techniques and methods that provide information
on the molecular level.
xi
Spectroscopy has become a powerful tool for the determination of polymer structures.
The major part of the book is devoted to techniques that are the most frequently used for
analysis of rubbery materials, i.e., various methods of nuclear magnetic resonance (NMR)
and optical spectroscopy. One chapter is devoted to (multi) hyphenated
thermograviometric analysis (TGA) techniques, i.e., TGA combined with Fourier
transform infrared spectroscopy (FT-IR), mass spectroscopy, gas chromatography,
differential scanning calorimetry and differential thermal analysis. There are already
many excellent textbooks on the basic principles of these methods. Therefore, the main
objective of the present book is to discuss a wide range of applications of the spectroscopic
techniques for the analysis of rubbery materials. The contents of this book are of interest
to chemists, physicists, material scientists and technologists who seek a better
understanding of rubbery materials.
The book brings together the various spectroscopic techniques for obtaining the following
information:
Chemical structure of rubbery materials: Chapters 1, 2, 3, 4, 5, 6, 9 and 11, describe

applications of (multi) hyphenated TGA techniques, optical and high-resolution NMR
spectroscopes for the analysis of chain microstructures and conformations, chemical
composition of components, additives and volatiles in rubbery materials, vulcanisation
chemistry, functional groups analysis and chemical modification of rubbery materials.
Network structure analysis is discussed in Chapters 7, 8, 10 and 13. These chapters deal
with the characterisation of the structure of chemical and physical networks, rubber-
filler physical network, network defects and its heterogeneity using NMR relaxation
techniques and NMR imaging.
Heterogeneity of rubbery materials: Use of NMR imaging, NMR MOUSE (mobile

universal surface explorer) and photoacoustic FT-IR is reviewed in Chapters 2 and 7.
These methods allow determination of defects and heterogeneity in technical elastomer
products, surface and depth profiles.
Physical properties of rubbery materials are discussed in Chapters 5, 7, 8, 10, 12, 13, 14
and 15, which cover the following topics: phase structure and composition, molecular
mobility, strain induced phenomena, phase separation and interfaces in rubbery
materials.
Functional properties and stability of rubbery materials: Chapters 1, 3, 4, 7, 12 and 13,

give examples of applications of spectroscopic techniques for the characterisation of
thermal stability and degradation, kinetics of thermal decomposition, ageing, oxidation
and weathering, self-diffusion of small molecules in rubbery materials, adhesion of
rubbers to metals, fluid adsorption and swelling.
xii
Preface
Processing of rubbery materials: The effect of milling on the heterogeneity of rubbery

materials and changes in their chemical and physical structures is determined by means
of optical spectroscopy, NMR relaxation experiments and NMR imaging (Chapters
3, 7 and 10). Welding of crosslinked polyethylenes is discussed in Chapter 5.
Quality control: Applications of NMR relaxation experiments, NMR imaging, NMR

MOUSE and (multi) hyphenated TGA techniques for quality control are discussed in
Chapters 1, 7 and 10.
As with any work of this type, this book is a result of a collective effort of many people.
The editors thank all the authors for their excellent contributions. We are indebted to
Ms. Frances Powers, Commissioning Editor of Rapra Technology, for her support in the
course of the work during last three years. Many thanks go to Ms. Claire Griffiths and
Ms. Sandra Hall for transforming the manuscripts into book format, and to all members
of RAPRA editorial group. We would also like to thank Ms. Caroline Barlow who
prepared the Index. Lastly, thanks to the family members of the editors, without whose
help and understanding the book could not come to completion.

2002
xiii
xiv
1
Characterisation of Elastomers Using (Multi)
Hyphenated Thermogravimetric Analysis
Techniques
Jan C.J. Bart and Charles Raemaekers
1.1 Introduction
Thermal analysis is a group of techniques in which a physical property of a substance is

measured as a function of temperature when the sample is subjected to a controlled
temperature program. Single techniques, such as thermogravimetry (TG), differential
scanning calorimetry (DSC), dynamic mechanical analysis (DMA), dielectric thermal
analysis, etc., provide important information on the thermal behaviour of materials.
However, for polymer characterisation, for instance in case of degradation, further analysis
is required, particularly because all of the techniques listed above mainly describe materials
only from a physical point of view. A hyphenated thermal analyser is a powerful tool to
yield the much-needed additional chemical information. In this paper we will concentrate
on simultaneous thermogravimetric techniques.
Thermogravimetry measures total mass changes under precise heating conditions. Thermal
events bringing about a change in the mass of a sample are adsorption and absorption,
desorption (diffusion), outgassing, dehydration or desolvation, foaming, vapourisation,
sublimation, decomposition, oxidation, reduction and other chemical reactions. These
phenomena can therefore be studied by means of TG. Some factors influencing component
loss are: chemical stability, volatility, rate of diffusion, sample thickness, flow rate,
temperature, heating rate, etc. Apart from sample weight changes no other chemical or
physical information is gathered about the matrix. For example, phenomena such as
polymerisation and cross-linking cannot be detected by TG. Although thermogravimetric
analysis is a powerful tool for quantitative polymer analysis its scope is further limited as
also no information is obtained about the qualitative aspects of the evolved gases during
polymer degradation. Both simultaneous and sequential techniques have been developed
for the purpose of identification of these gases and volatiles. The great advantage of
simultaneous thermoanalytical methods is that they allow the results to be interpreted
and evaluated jointly.
As may be seen from Figure 1.1 [1-27] hyphenated and multihyphenated methods greatly
enlarge the scope of TG and are becoming important analytical tools for materials
1
Figure 1.1 (Multi) hyphenated thermogravimetric analysis techniques. The figures

refer to references
characterisation. Some thirty different experimental approaches may already be identified,

many with a proven record. The present trend towards ever increasing product quality
standards demands greater sensitivity, accuracy, reproducibility and therefore constitutes
a challenge to thermogravimetric analytical techniques.
The thermal characterisation of elastomers has recently been reviewed by Sircar [28]
from which it appears that DSC followed by TG/DTG are the most popular thermal
analysis techniques for elastomer applications. The TG/differential thermal gravimetry
(DTG) method remains the method of choice for compositional analysis of uncured and
cured elastomer compounds. Sircar’s comprehensive review [28] was based on single
thermal methods (TG, DSC, differential thermal analysis (DTA), thermomechanical
analysis (TMA), DMA) and excluded combined (TG-DSC, TG-DTA) and simultaneous
(TG-fourier transform infrared (TG-FTIR), TG-mass spectroscopy (TG-MS)) techniques.
In this chapter the emphasis is on those multiple and hyphenated thermogravimetric
analysis techniques which have had an impact on the characterisation of elastomers. The
review is based mainly on Chemical Abstracts records corresponding to the keywords
elastomers, thermogravimetry, differential scanning calorimetry, differential thermal
analysis, infrared and mass spectrometry over the period 1979-1999. Table 1.1 contains
the references to the various combined techniques.
2
Characterisation of Elastomers Using (Multi) Hyphenated Thermogravimetric…
Table 1.1 Selected references to (multi) hyphenated

thermogravimetric analyses of elastomers
Technique Reference number
TG-FTIR [29-48]
TG-MS [25, 48-54]
TG-DSC [55-62]
TG-DSC-MS [50, 63]
TG-DTA [57, 64-84, 149]
TG-DTA-FTIR [39]
TG-DTA-MS [59-62, 85]
TG-Tenax-TD-GC-FTIR-MS [86]
1.2 Instrumental Part
Thermogravimetric measurements require a thermobalance. There are many different

types of TG analysers, varying in design (furnace, atmosphere, pressure, temperature
range, sample quantity, sensitivity, resolution, etc.), degree of microcomputer control of
the hardware, capabilities of the software, etc., all with typical specifications. The basic
TG experiment consists of recording the weight of a sample as it is heated in a given
environment either isothermally or at a controlled heating rate. The experimental record
is a plot (thermal curve) of some form of weight change versus time or temperature of
the sample. The simple additional step of using the derivative of the primary weight
change (DTG) extends the capability and scope of the analysis. Heating of the sample
may be combined with measurement of any parameter which further characterises the
thermally activated transformation, such as heat flow (DTA, DSC), mass (DTG) or total
gas flow (evolved gas analysis by MS or FTIR).
The family of thermogravimetric techniques, as shown in Figure 1.1, can be divided into
simultaneous and non-simultaneous methods. Non-simultaneous methods imply no real-
time generation of thermogravimetric and other physico-chemical information. A
continuous analysis of a gas stream from a thermal analyser is not readily possible by gas
chromatography (GC); usually one has to use a gas-sampling loop for intermittent
operation. Non-simultaneous methods usually involve cold traps (CT) or absorbent
charcoal (Tenax). Because of the intermittent function of the analysis the coupling of GC
to a thermogravimetric analyser is therefore still not so common. The separation function
3
by time in GC is, to a small extent, also given by thermogravimetry, as the gases are
‘temperature separated’ during analysis. Non-simultaneous TGA techniques, e.g., TG-
GC-MS, TG-Tenax-FTIR-MS, add surplus value and understanding when non-continuous
data are sufficient. Due to the experimental set-up, these methods give inherently less
insight in degradation behaviour than continuous methods. In particular, there is no
continuous monitoring of evolved gases. In fact, only information about a specific point
of the TG curve is collected at a given time. Discontinuous TG-GC, TG-GC-MS and TG-
Tenax-TD-GC-MS are obviously inherently slower than TG-MS. The possibilities are
improving though by using a fast column and an accurate fast sampling valve. However,
only the fragments that are non-reactive, thermally stable, and volatile can be analysed
by GC, whereas MS has none of these limitations.
Despite the utility of TG-FTIR and TG-MS, a distinct disadvantage is that the presence
of components at very low concentrations may be masked by higher concentration
interferants. Thus, some researchers have incorporated the separation power of gas
chromatography by collecting products in a trap or on the head of a capillary column for
all or part of the TG run [87, 88]. However, these methods necessarily result in the loss
of the time/temperature evolution data for the products analysed.
McClennen and co-workers [20] have described TG-GC-MS and TG-GC-IR illustrating
the feasibility of using a high-speed ‘transfer line’ GC technique for near real-time process
monitoring applications using MS [89]. The approach is thus to transport products from
the TG to the GC where they are rapidly separated on a short capillary column.
Combination of a pulsed automated vapour sampling inlet and transfer line type GC
column permits high-speed GC identification of individual TG products while maintaining
sufficiently high temporal resolution with a ~ 1 minute sampling interval. By selecting
columns to provide short retention times (< 1 minute) consistent with time-resolved
profiles of the TG curve for comparison to the DTG curves the resulting TG-GC-MS and
TG-GC-IR configurations are capable of separating thermal decomposition products,
while still allowing characterisation of the evolution profile for each component. The
classic chromatographic trade-off between efficiency and analytical time can be balanced
in such a way to provide both real-time thermal evolution profiles of multiple components
and separation sufficient to allow a significant degree of component identification by
means of TG-GC-IR and TG-GC-MS.
Meuzelaar and co-workers [90] have described an on-line high pressure TG-GC-MS system
which requires only very small amounts (10-100 mg) of sample and can be operated at high
pressure under different atmospheres (N2, He, H2, etc.). The system has been used in recycling
lower grade post-consumer polymers, as coloured polyethylene and polystyrene or used rubber
tyres, by co-processing with coal. The main characteristics of non-simultaneous methods,
TG-GC and techniques requiring a cold trap, are given in Table 1.2.
4
Table 1.2 Characteristics of some non-simultaneous TG methods [26]

TG-GC TG-CT
• Roughly equivalent to PyGC plus • Intermitted sampling allows for discontinuous
weight change. analysis of evolved gases.
• Permits the analysis of components • Continuous monitoring of the weight change
of different volatilities. process and gas analysis is impossible.
• GC allows analysis only of those • Recombination between products prior to
fragments that are non-reactive, subsequent analysis is not to be excluded.
thermally stable and volatile
It is especially important to have simultaneous techniques incorporating the most

frequently used thermal methods. Early work allowed joint measurement of TG, DTG,
and DTA. Paulik and co-workers [91, 92] devised simultaneous TG/DTG-DTA (the
derivatograph). Later, also TG/DTG-DTA-MS has been described [93]. TG-DSC (or TG-
DTA) allows simultaneous measurements for the determination of the mass change (TG)
and the energetic changes (DSC/DTA) on one sample under identical test conditions. As
with this method all factors which influence the measurement signals, e.g., atmosphere,
sample structure, temperature gradient, diffusion paths and packing density, are the same,
TG and DSC/DTA results can be correlated and interpreted more easily. The
complementary information obtained allows differentiation between endothermic and
exothermic events which have no associated weight loss, e.g., melting and crystallisation,
and those which involve a weight loss, e.g., degradation. Simultaneous TGA-DTA results
are most often used to improve the interpretation of thermal events.
Redfern [94] summarises the advantages of single sample simultaneous TG-DSC (or TG-
DTA) as follows:
• Greater efficiency (sample preparation time, run set-up time, instrument time).
• Reduced influence of sample preparation (a single sectioning of the original material).
• Higher accuracy of TGA temperature calibration (typically 0.1 °C for DSC and only
2 °C for stand-alone TGA).
• Simplified interpretation for perturbation of results arising from thermal history,

orientation effects, heat treatment, pressure during cutting, etc.
• Combined evaluation assuring identical experimental and sampling conditions for

both measurements with elimination of sources of uncertainty.
5
• Correlation of observed effects.
• Weight measurements validating quantitative measurements from DSC experiments.
• Detection of moisture and determination of in-situ dry weight.
• Quantitative measurements (based on correct dry sample mass and accurate enthalpic
measurements).
The ability to obtain two thermal measurements on a single material is useful for rapidly
comparing different batches of product or raw material in quality control. Simultaneous
TG-DTA provides valuable information even in materials where no weight changes occur
over the temperature range studied.
A disadvantage of TG-DSC (or TG-DTA) is that the data obtained give no direct
information on the nature of the chemical species involved. A typical modern assembly
for simultaneous TG-DSC (or ‘STA’) is shown in Figure 1.2.
Coupled instruments each need to operate under optimum conditions. Key elements in
performance are the interface system and an integrated software package. Interface design
is critical in allowing rapid mass transfer while not degrading the species of interest
being transferred. No standardised thermogravimetric couplings are available.
Descriptions of interfaces and interface techniques in coupled instrumentation have been
reported:
TG-FTIR [29, 30, 96], TG-MS [25, 49, 50, 63, 96], TG-DSC [55, 56], TG-DSC-
MS [50, 63], TG-DTA [56], TG-DTA-MS [93], TG-GC-MS [96], TD-GC-FTIR
[97], TG-Tenax-TD-GC-FTIR-MS [86].
For identification purposes the classical techniques of infrared spectroscopy and MS are
highly suited in hyphenation to thermogravimetry. Table 1.3 shows the main characteristics
of TG-FTIR and TG-MS.
Kaisersberger and co-workers [96] discussed hyphenation of a thermo-balance to infrared

spectrometers. In this instrumental coupling, the infrared radiation must be brought into
an intensive contact with the gas flow including the evolved gases, which is achieved in
specially designed gas-measuring cells. FTIR spectrometers provide the high scan speeds
which are required for the on-line coupling with thermal analysers for continuous
monitoring of the gas composition. FTIR coupling is made by connection of the gas cell
with the shortest possible heated transfer lines to the gas flow system of the thermobalance.
As opposed to the coupling of mass spectrometers, the whole gas flow from the
thermobalance should pass through the gas cell of the infrared spectrometer. While
6
Figure 1.2 Simultaneous TG-DSC

Reproduced with kind permission of Netzsch-Gerätebau GmbH, Selb. Germany [95]
experimental results obtained from coupling of TG and FTIR spectroscopy were reported
in 1981 [19], dedicated TG-FTIR instruments were not commercially available until
1987 [98-100]. Up to that time most work on the identification of evolved gases from
TG had been in the TG-MS combination and reports on polymer studies using hyphenated
TG-FTIR were relatively scarce [101, 102]. Several approaches to coupling TG and FTIR
instruments have been reported [19, 29, 31, 55, 57, 103-108]. In conventional
commercially available TG-FTIR systems, the evolved gases are led from the TG system
to the spectrometer via a heated transfer line by a carrier gas flow [19, 103-105]. Another
approach is the use of helium carrier gas at high flow rates, leading to the formation of
an aerosol of the evolved components [31, 109]. The aerosol is then led to the spectrometer
without loss of high-molecular-mass components. This system performs quantitative
measurements and preserves and monitors very high molecular weight condensibles. In
an ‘on-the-spot’ TG-FTIR technique the radiation is brought to the TG system - as opposed
to bringing the evolved components to the spectrometer [29, 110]. The IR beam is led
directly to the TG system, where it is reflected by a mirror mounted inside the TG
equipment and is subsequently detected by the standard FTIR detector. Compared to
7
Table 1.3 Characteristics of TG-FTIR and TG-MS methods [26]

TG-FTIR
• Characterises and quantifies gaseous decomposition products
• Permits specific compound analysis: fingerprint identification based on vibrational
spectra
• Identifies classes of compounds
• Requires IR absorbency
• Makes use of reference spectra
• Is of limited sensitivity only
ª Has lower compound specificity than TG-MS (sub-μg levels)
TG-MS
• Allows real time analysis of weight loss and temporal MS identification of evolved
gases
• Has restricted identification power of the gaseous species emitted according to mass
• Is designed for high sensitivity (pg level) and versatility
• Permits industrial problem solving capability
• Is a real asset in the analysis world, yet by many people still much ignored
High Resolution TG-MS
• Designed for improved resolution and identification
• Allows for stepwise and quasi-isothermal modes
TG-MS/MS
• Permits complex compositional analysis of thermally activated processes
• Allows rapid separation and identification of co-evolving compounds
• Differentiates between co-evolving isobaric ions
• Stands for real time analysis
FTIR methods which use heatable gas cells, e.g., fast thermolysis FTIR [111], the on-the-
spot TG-FTIR technique, which avoids transfer lines, monitors the gaseous atmosphere
directly above the sample pan; spectral information is obtained which is directly correlated
to the recorded mass change as a function of time and temperature. At the end of a TG
8
experiment an Evolved Gas Profile (EGP) can be reconstructed from the stored
interferograms according to Gram-Schmidt [112]. Each point in this EGP corresponds
to an IR spectrum of the evolved components in the TG equipment. Specific Gas Profiles
(SGP) and Functional Group Profiles (FGP) can also be reconstructed from the stored
interferograms in selected wavenumber windows to detect components with specific group
frequencies. Comparison of EGP as a function of time with the DTG curve yields a direct
comparison of TGA and spectroscopic data. The detection limits are in the sub μg/s
range and dependent on the extinction coefficient of the evolved components [29].
The commercial availability of FTIR systems capable of highly sensitive detection and
completely automated sampling and data manipulation, have given TG-FTIR polymer
analysis all the advantages of classic IR spectroscopic interpretation. Advantages of FTIR
detection are: functional group identification and specific compound analysis on the
basis of vibrational spectra; simultaneous spectral information on many species;
continuous scanning of effluent from direct thermal processing; quantitative analysis
(± 10%) using proper calibration from well-known absorption coefficient information
for IR-absorbing compounds [113]; reference (vapour-phase) spectral libraries [114]. In
general, the detection limits for components in the condensed phase are a factor-of-ten
lower than those in the gas phase.
The technique is especially useful for smaller molecules where the high specificity of strong
IR absorption bands makes up for the relatively low sensitivity of IR detection. It is rather
difficult to use IR to analyse mixtures of compounds with similar functional groups or
mixtures of weak IR absorbers in the presence of strong absorbers. The limitation of FTIR
lies in detecting only non-symmetrical gas molecules. FTIR cannot detect gases without IR
absorbance, e.g., O2, N2, and does not readily distinguish hydrocarbons above C3H8. If the
calculated weight loss of observed gases is lower than that measured by TGA, then it can
be inferred that other gases are being evolved that are FTIR blind.
FTIR uses much lower excitation energy than MS and can therefore detect larger functional
groups in evolved gases from TG experiments, such as high-boiling oligomers and heavy
tar products which can be analysed as fine aerosols in a gas stream [20, 100]. TG-FTIR
also distinguishes structural isomers [115].
Spectral subtraction and spectral search aid the identification of evolved gases, which
are often a mixture of products. Nevertheless, for unambiguous identification of unknown
volatiles more powerful methods are required. Jansen and co-workers [86] have
incorporated a parallel mass spectrometer onto the FTIR stage of a thermogravimetry-
FTIR (TG-FTIR). The sample is thermally decomposed by TGA and the products collected
in a Tenax (absorbent charcoal) trap. After desorption, the products are separated by a
GC and the sample split, with 99% going to the IR spectrometer and 1% to the mass
spectrometer.
9
TG-FTIR has become a quite popular, versatile, cost-effective and informative instrument
for modern polymer analysts concerned with compositional analysis and degradation/
reaction mechanism studies. The growth rate of TG/FTIR instrumentation currently
exceeds that of TG-MS.
Simultaneous TG-MS is a very powerful hyphenated technique combining the direct

measurement of weight loss as a function of temperature with the use of a sensitive
spectrometric detector. The usefulness of MS to TG coupling was suggested already in
1965 [8, 10]. The coupling with MS (in particular time-of-flight MS (ToFMS) and
quadrupole mass spectrometry (QMS)) adds chemical analytical features allowing the
chemist to assign the evolved gases to the detected weight losses thereby correlating
chemical information with the thermal event. Mass spectrometry permits temporal
resolution of the gases that are evolved during the thermal or thermo-oxidative degradation
of a polymer in controlled atmospheric conditions.
It is obvious from the history of TG-MS [25] that the interface is of crucial importance
and poses several problems. Within the interface, conditions are converted from the high
temperature and (usually) atmospheric pressure of TG to the room temperature and
(usually) high vacuum conditions in the mass analyser. Kaisersberger and co-workers
[96] have given an excellent account of the basic requirements and practical aspects of
coupling of gas analytical methods with thermal analysis instruments. Bart and co-workers
[25] have critically discussed a broad variety of TG-MS instrumental design solutions,
which depend partly on the sample characteristics and the desired conditions of the
thermal degradation, particularly in relation to polymer characterisation. Advantages
and disadvantages of TG-MS with respect to other evolved gas analysis techniques are
outlined. Mass spectrometry coupling can be achieved by connecting a capillary at the
end of the gas-flow system of a thermobalance or by means of a skimmer coupling,
integrated into the furnace of the thermal analyser [96]. Although capillary systems have
been described as a great limitation of coupling thermal analysis instruments with the
mass spectrometer [49] because of presumed restrictions in heating of capillary couplings
to a maximum of 200 °C and of most transfer lines to about 250 °C, a transfer experiment
with the commercial UV-absorber Tinuvin 234 has shown sufficient sensitivity and
transient response of a capillary TG-MS interface for molecular masses up to m/z 450 [116].
Although most TG-MS couplings have concentrated on evolved gas analysis, a few reports
concern residue analysis. Wiedemann and co-workers [117] have described a thermo-
molecular beam analysis (TMBA) technique, which allows to continuously and
simultaneously follow the weight of a sample (TG), the composition of its gaseous reaction
products (MS) and solid reaction products {high temperature X-ray diffraction (XRD)}.
Raemaekers and co-workers [26] have advocated video-imaging (VI-TG-MS).
As many of the curves from TGA or DSC are in fact much more complex than might
appear at first sight, TG-DSC/DTA with evolved gas analysers (MS and FTIR) are useful
10
tools [118]. In a top-loading STA analyser combined with a quadrupole mass spectrometer
by means of a two-stage gas inlet system, as described by Kaisersberger and co-workers
[50, 63], evolved gases and vapours are collected just above the sample crucible by an
orifice. A skimmer is arranged in the compression zone behind this orifice to achieve a
parallel molecular beam into the electron impact ion source of the MS. The optimised
orifice system and short transfer path in this advanced solution effectively eliminate
many of the coupling problems observed in most other designs. The arrangement excels
by high sensitivity [49]. A simultaneous TG-DSC-MS capable of operation up to 2000 °C
has been reported [119]. As a further extension of the classical TG-MS commercial TG/
DTG/DTA-MS (STA-MS) has been developed (sample size up to 5 g) as well as a macro
STA/MS/GC-MS (sample size 500 g), which allows examination of heterogeneous
materials and trace analysis [120].
The strengths of MS and FTIR complement each other. MS requires special high vacuum
capabilities and more stringent operating conditions but exhibits detection levels which
are several orders of magnitude better than FTIR (pg and sub-μg ranges, respectively). In
some conditions, MS results can be misleading because of secondary products resulting
from ion fragmentation [121]. Yet, MS identifies each individual compound and not a
class of compounds of the same functional group characteristics. Both MS and FTIR
need the support of a full spectrum vapour phase library.
Despite the utility of these techniques, a distinct disadvantage is that very low
concentration components may be masked by higher concentration interferants; in such
cases, devices based on thermal desorption (absorbent packing, static headspace sampling,
cryogenic trapping) may prove useful.
More complex thermal decomposition processes may require identification instead of

verification, for which single-step hyphenated techniques are ill equipped.
Multihyphenated methods have been described which combine a number of functions,
such as separation and identification (in various modes). If a pattern of complex species
is evolved during heating (as is easily the case for polymers), it is advantageous to achieve
separation prior to entering the final phase of the mass spectrometer. This can be
accomplished through the use of trapping TG-GC-MS [86, 122 or direct TG-GC-MS
[23, 123]. As mentioned already, this analysis is lengthy: for each point on the TG curve
(time, temperature) a GC-MS analysis has to be performed. McClennen and co-workers
[20] have shown the enhanced capability of the doubly hyphenated thermogravimetry-
based analytical techniques TG-IR-MS and TG-GC-MS for detection of minor products.
Arii and co-workers [124] have used an integrated simultaneous TG-DTA/GC-MS system
in high resolution mode. Controlled rate thermal analysis enhances accuracy of
identification and quantification. The TG-DTA/GC-MS system can be operated in two
interface modes, namely continuous sampling (direct coupling mode) and intermittent
11
sampling (trap coupling mode). In the direct coupling mode the GC is bypassed.
Intermittent sampling is ‘off-line’.
Jansen and co-workers [97] have described on-line TD-GC-FTIR. In the TD-GC-FTIR
system a thermal desorption (TD) cold trap injector is used for the temperature-controlled
outgassing of the samples with a maximum temperature of 350 °C. The volatile
components are transferred to the cold trap by the carrier gas and preconcentrated.
After completion of the outgassing process the cold trap is heated very quickly, causing
on-column injection of the trapped components onto the gas chromatograph. The
technique has recently been extended to include an ion-trap MS.
On-line thermal desorption-GC-FTIR-MS can be operated both in parallel and tandem

FTIR-MS configuration, where parallel FTIR-MS operation is preferred for several
reasons. Compared to FTIR alone, the parallel configuration of FTIR and MS enhances
and facilitates the elucidation of the evolved species and furthermore lowers the detection
limits from ppm to (sub) ppb level [86].
With a thermal desorption unit capable of accommodating Tenax absorption cartridges

sampling with preconcentration of thermogravimetric off-gases in possible which may
be followed off-line by analysis of the trapped components via TD-GC-FTIR-MS [86]. A
major disadvantage in routine analysis by these techniques is the throughput of samples;
the analysis time is determined by the longest retained compound. Quite obviously, neither
TG-GC-MS [20, 22, 23, 87, 125-129] nor TG-GC-FTIR [20, 24] provide continuous
monitoring of evolved gases. Other significant disadvantages of multihyphenated systems
are complexity, cost, and the need for a trained operator. It is therefore not at all certain
that such techniques hold the future.
1.2.1 Experimental
1.2.1.1 Sample Requirements for Polymers
For TG work it is generally recommended to use as little sample as possible within the
limits of resolution of the microbalance. The homogeneity of a sample can sometimes
limit how little sample can be used, e.g., in case of polymeric materials. Powdered samples,
of small particulate size, have the ideal form for TG studies. However, in polymer science,
samples are often also in the form of films, fibres, sheets, pellets, granules or blocks. The
packing density should be as uniform as possible.
Multicomponent separation of a rubber material performed with TG is typically as follows:

rapid heating in inert atmosphere (nitrogen) up to 100 °C (for loss of volatile oils and
12
extenders), successive heating up to 600 °C (for decomposition of the rubber component),

heating in oxygen to 950 °C (for combustion of carbon black) and determination of the
residue (fillers).
For (multi) hyphenated thermogravimetric analysis techniques typically a few mg of

sample are sufficient. Sample requirements for TG-MS vary according to the amount of
volatile substances liberated and the sensitivity of the mass spectrometer for a given
component. Moreover, as the temperature that corresponds to the maximum evolution
of a fragment is particle-size dependent, there is a size requirement as well. When
determining low level additives it is often quite necessary to weigh out as much as 100 mg;
sample sizes of 3-6 mg are a usual acceptable compromise between efficient balance
operation and overloading the mass spectrometer. For sample sizes < 5 mg, problems
associated with heat transfer and surface collisions are minimised; these samples are
more amenable to maintaining a low system pressure and reducing bimolecular collisions.
The pressure limit condition for the mass spectrometer implies that there is an upper
limit for the product of the rate of heating, dT/dt, and the mass M of the compound to be
decomposed: dT/dt. M < C, where the value of the constant C is dependent upon
instrument and sample. This condition should especially be considered in case of
quantitative measurements [130].
1.2.1.2 Calibration and Interlaboratory Reproducibility
Accurate temperature calibration using the ASTM temperature standards [131, 132] is
common practice for DSC and DTA. Calibration of thermobalances is more cumbersome.
The key to proper use of TGA is to recognise that the decomposition temperatures
measured are procedural and dependent on both sample and instrument related parameters
[30]. Considerable experimental control must be exercised at all stages of the technique
to ensure adequate reproducibility on a comparative basis. For (intralaboratory)
standardisation purposes it is absolutely required to respect and report a number of
measurement variables. ICTA recommendations should be followed [133-135] and should
accompany the TG record. During the course of experiments the optimum conditions
should be standardised and maintained within a given series of samples. Affolter and co-
workers [136] have described interlaboratory tests on thermal analysis of polymers.
Courtault [137] has critically considered calibration of TG, MS and the TG-MS coupling.
The TG (-MS) apparatus is often calibrated with calcium oxalate monohydrate. Bart
and co-workers [25] have discussed critically the calibration and interlaboratory
reproducibility of TG-MS.
The major hurdle to be overcome before a technique gains acceptance within the analytical
community and in practice is the achievement of interlaboratory validation. There still
13
appears to be sufficient room for improvement in interlaboratory reproducibility for

most other multihyphenated techniques considered in this paper.
1.3 Applications of Hyphenated Thermogravimetric Analysis

Techniques
Thermogravimetric analysis can yield a considerable amount of information about the

composition of an elastomer system and is a common means of testing of EPDM rubber
compounds [51]. The DTG curve may serve as an identifier of the type of elastomer in a
compounded formulation.
Sircar [138] has reviewed the analysis of elastomer vulcanisate composition by TG/DTG
techniques. The classical ASTM method, D297-93 [139], is too lengthy to be of much
practical use on a routine basis, often requires preliminary identification of the polymer
and is costly. TG has gained itself wide acceptance as a method for quantitative compositional
analysis of vulcanisates: ASTM E1131 [140], is basically designed for the analysis of rubber
compounds [141]. Thermogravimetric analysis can be used to determine:
• water, solvents, monomers and other volatile substances
• oil, extenders, plasticisers and other high temperature volatile substances
• elastomer, organic fillers by pyrolysis
• carbon black and graphite by high temperature oxidation
• carbonates, e.g., CaCO3, by decarbonation at elevated temperatures
• ash, pigments and inert fillers as measured residue
• thermal stability, kinetics of decomposition and thermal endurance.
TGA can be used as a quality control method in the production of rubber compounds (ASTM
Task Group E 3701.09), and is a recognised troubleshooting tool in the rubber industry.
Classical analysis of rubber compounds for identification of ingredients such as accelerators,
antioxidants, and antiozonants requires a very difficult separation step of such additives
from the rubber compound prior to analysis. Vulcanised rubbers are even more difficult to
analyse than thermoplastic polymers due to the compounding ingredients being locked into
the matrix by carbon black and cross-linking of the rubber polymer. The classical method for
separating volatile additives from rubber is by solvent extraction. TGA measurements can
14
differentiate between the most volatile components in rubber compounds, for example
moisture, plasticiser, residual solvent, processing oils and extenders, or other low boiling
point (below 300 °C) materials and components of medium volatility including the elastomer
portion of the compound. Although TG/DTG is a potentially effective analytical technique
for the compositional analysis of compounded elastomers such analyses are not always
completely accurate, i.e., a 100% materials balance cannot always be obtained [28]. This
may be due to an overlap of low molecular weight volatile material, e.g., oils, plasticisers and
resins, with polymer decomposition products, the decomposition of polymer blend components
in a similar temperature region, the formation of char, which decomposes in the region
assigned to carbon black, or the carry-over of early stage decomposition products to the ash
region. Typically, sulphur, accelerators, antioxidants and antidegradants are not observed as
independent weight loss in TG curves. For many systems weight losses overlap for most
choices of method parameters. Consequently, high resolution or reduced pressure methods
are frequently used. Apart from incomplete analysis due to poor resolution and the inability
to separate oligomers of polymer and oil, desorption methods such as TGA may produce
volatile fragments of additives. The interferences are discussed by Sircar [138]. Thus, there
are practical limits to the kind and degree of information that can be extracted from TG/
DTG analyses of unknown vulcanisate compositions.
Table 1.4 shows that the most numerous applications of hyphenated thermogravimetric
techniques for the study of elastomeric material make use of TG-DTA, followed by TG-
FTIR, TG-DSC, TG-MS and TG-DTA-MS, with only occasional recourse to TG-DSC-
MS and TG-GC-MS. Table 1.5 indicates the general performance characteristics of the
thermogravimetric techniques in use for the study of elastomeric materials.
There is a clearly defined industrial need for reliable and fast methods to study the
stabilisation and characterisation of elastomeric materials, to obtain knowledge on their
properties and behaviour, the effects of modifying structure, additives and processes to
engineer the most suitable material for any given application. Also, there is an increasing
(environmental) need to study the effects of ageing, the thermal stability, the degradation
processes and the products of decomposition under a wide range of conditions.
Evolved gas analysis may be used in two modes:
• as a stand-alone technique, such as temperature-programmed reduction
• as a hyphenated technique (TG, DSC, or STA combined with FTIR or MS).
Hyphenated thermogravimetric techniques have been used in a wide variety of qualitative

and quantitative problem-solving cases for elastomeric materials:
15
Table 1.4 Hyphenated thermogravimetric analysis studies of elastomeric

materials
TG-FTIR Vulcanisation [32], ageing characterisation [39, 48],
sulphur components in rubber [31], polyurethanes
[37], polymer degradation mechanisms [30, 40, 41],
identification of base polymers [36, 43, 44], thermal
stability [46], grafted flame retardants [47],
differentiation of EVA rubbers [45] and AN-NBR
rubbers [36, 44], degradation of chlorinated natural
rubber [42].
TG-MS Kinetics of degradation [54], ageing characterisation
[48], quality control [51], controlled release [52],
additive analysis [88, 50, 142, 143, 144], outgassing
[48, 49, 50, 145, 146], thermal degradation [25, 53,
54, 147], product development [148].
TG-DSC Additive interactions [58], reactivity [59-62].
TG-DSC-MS EPDM and chlorinated rubbers [63].
TG-DTA Characterisation of carbon black [149], flammability
evaluation [64], polymer degradation studies [65],
ageing studies [70-72], product control [77, 81],
combustion performance [83], safety evaluation [83],
antioxidation activity [68], pyrolysis of rubbers [82],
thermal stability [67, 69, 76, 77], interfacial junctions
in viscoelastic composites [78], weathering [72],
vulcanisation [73], oxidative behaviour [79], materials
evaluation [80], failure analyses [81].
TG/DTG-DTA-FTIR Weathering [39].
TG-DTA-MS Combustion performance [59-62, 85], safety
evaluation [59-62, 85].
TG-GC-MS Co-processing of waste polymers [150].
TG-Tenax-TD-GC-FTIR-MS Outgassing [86, 175].
AN: acrylonitrile
16
Table 1.5 Comparison of hyphenated thermogravimetric techniques for

elastomer analysis
Methoda Thermal effects Evolved gas analysis
with mass withoutb mass quantitative qualitativee
change change
TG + - - -
TG-FTIR + - +c +
TG-MS + - +c +
TG-DSC + + - -
TG-DSC-MS + + +c +
TG-DTA + + - -
TG-DTA-MS + + +c +
TG-GC-MS + - +d +
a
All combined techniques measure various effects on the same sample
b
Energetics of phase transitions or reactions (melting, crystallisation, TG, cross-linking,
etc.)
c
With appropriate calibration the system is suitable for quantitative determination.
d
Can be made quantitative
e
Structural assignment on the basis of molecular ion mass, fragmentation pattern or
vibrational mode
a. Thermal stability and degradation studies
• ageing characterisation in thermal and thermo-oxidative stability studies

weathering
• identification of degradation products
• elucidation of polymer degradation mechanisms
• intercomparison of materials properties
• characterisation of insoluble polymeric materials.
b. Structural characterisation and chemical analysis
• identification of base polymers and components

• differentiation of materials
• fingerprint identification (classification of competitor products).
17
c. Reactivity and curing
• vulcanisation
• characterisation of cure processes.
d. Product development
• optimisation of production processes of elastomeric materials

• product quality and batch-to-batch control
• determination of end-use conditions
• performance behaviour
• failure analysis
• product recycling.
e. Analysis of evolved gases during synthesis, processing and recycling
• outgassing phenomena
• trace solvent, monomer inclusions or impurity analysis
• analysis of additives or processing agents
• additive interactions
• determination of the effect of stabilisers
• controlled release
• combustion performance
• health protection studies
• toxicology of combustion.
f. Kinetics
• kinetics of degradation
• life-time predictions (oxidative induction time; OIT).
g. Quantitative analysis
• determination of the composition of elastomers.
Quantification of sample components by conventional TG may be made only if a

stoichiometric relationship exists between the gaseous decomposition products and the
original material. TG-FTIR allows quantitative analysis to be performed (calibration
necessary) even when more than one component of interest pyrolyses during a single
weight loss [152].
Courtault [137] has described quantitative aspects of TG-MS coupling, which is still a
difficult technique. Dyszel [153] has determined the styrene content in styrene-butadiene
18
rubbers. The complexity of quantitative measurements of thermal analysis using mass

spectrometry has recently been discussed by Morelli [154]. The development of a Hi-Res
TG-MS [17] is beneficial to the ability to obtain more accurate, quantitative results.
Quantitative interpretation of TG-MS data requires calibration of the system, i.e., the
determination of the relationship between the observed intensities of the ion currents
and the amount of the analysed species. Quantitative aspects of MS couplings have recently
been treated very clearly by Maciejewski and co-workers [155, 156], also introducing a
new experimental technology, Pulse Thermal Analysis (PTA). PTA enables the introduction
of a well-defined amount of a gas to the system at any temperature (non-isothermal)
and/or time (isothermal mode). Injected pulses can be used as a reference for the
quantification of the signals originating from the evolution of gas(es) formed during the
decomposition of solids. A linear dependence between the amount of injected gas and
the intensity of mass spectrometry signals enables the quantification of mass spectroscopic
data. An accuracy for the evolved species below 0.01 wt.% has been quoted. Unless
results are made quantitative it is arduous to draw any definite conclusions about
mechanisms of decomposition.
Table 1.4 lists applications of various hyphenated thermogravimetric techniques for the
study of elastomeric materials.
1.3.1 Thermogravimetry - Fourier Transform Infrared Spectroscopy
The applications of simultaneous TG-FTIR to elastomeric materials have been reviewed

in the past. Manley [32] has described thermal methods of analysis of rubbers and plastics,
including TGA, DTA, DSC, TMA, Thermal volatilisation analysis (TVA), TG-FTIR and
TG-MS and has indicated vulcanisation as an important application. Carangelo and co-
workers [31] have reviewed the applications of the combination of TG and evolved gas
analysis by FTIR. The authors report TG-FTIR analysis of evolved products (CO2, NH3,
CH3COOH and olefins) from a polyethylene with rubber additive. The TG-FTIR system
performs quantitative measurements, and preserves and monitors very high molecular
weight condensibles. The technique has proven useful for many applications (Table 1.6).
Mittleman and co-workers [30] have addressed the role of TG-FTIR in the determination
of polymer degradation pathways.
Some general applications of TG-FTIR are evolved gas analysis, identification of polymeric
materials, additive analysis, determination of residual solvents, degradation of polymers,
sulphur components from oil shale and rubber, contaminants in catalysts, hydrocarbons
in source rock, nitrogen species from waste oil, aldehydes in wood and lignins, nicotine
in tobacco and related products, moisture in pharmaceuticals, characterisation of minerals
and coal, determination of kinetic parameters and solid fuel analysis.
19
Table 1.6 Polymer chemistry applications

for TG-FTIR after [33]
Solvent and water retention
Curing reactions
Vulcanisation reactions
Isothermal ageing
Product stability
Thermal degradation
Identification of processing aids
Plasticisers
Mould lubricants
Blowing agents
Antioxidants
Flame retardants
Safety concerns
Product safety
Product liability
Fire hazards
The potential applications of an integrated TG-FTIR system were discussed by various

authors [33, 34]. Schönherr [35, 36] has discussed the advantages of the method, in particular
for rubber analysis. TG-FTIR examination of a polybutadiene sample with a high proportion
of inorganic fillers and spectral subtraction procedures identified water and plasticiser at
200 °C; CO2, CO, H2O, methane, ethylene, n-butane, n-pentane and cyclic hydrocarbons
at 500 °C. The results indicate that a single sample weight loss may well correspond to a
very complex mixture of evolved gases [57]. Another synthetic rubber sample evolved CO,
CO2, CS2 and hydrocarbon traces (at about 250 °C), methane and short-chained alkenes
(between 375 °C and 540 °C) [33]. TG-FTIR of a gold-filled epoxy resulted in evolution
of 1-epoxy-7-oxabicyclo [4.1.0] heptane at 185 °C, carboxylic acid and CO2 at 330 °C,
CO2, phenol and carboxylic acid between 360 °C and 460 °C [33].
TG-FTIR of a silicone O-ring under nitrogen showed formaldehyde and

hexamethylcyclotrisiloxane at 450 °C, and a hexamethylcyclotrisiloxane
octamethylcyclotetrasiloxane mixture up to 550 °C [33]. Brame [37] used TG-FTIR for
20
the determination of minor components in polyurethanes, which could not be discerned

by FTIR without a thermal separation procedure.
The technique was also used for the study of the decomposition of polybutadiene [38]
and of weathered sealants on a silicone and polyurethane basis (in an STA-FTIR
configuration) [39], as well as for the study of the thermal degradation mechanism of
tetrafluoroethylene-propylene copolymers [40]. Although the exact mechanism for the
thermal degradation of Aflas FA-150C rubber could not be defined, comparative TG-
FTIR experiments (see Figure 1.3) demonstrated that the main degradation products
(see Figure 1.4) were mixed hydrocarbon/fluorocarbons similar to those observed in
Tefzel 200 and 280 (alternating copolymer) degradation rather than what would be
expected if degradation occurred by unzipping to monomers. Suzuki and co-workers
[41] have studied the thermal degradation of acrylonitrile-butadiene-styrene (ABS)
terpolymer by TG-FTIR (see Figure 1.5) and compared with that of polystyrene (PS),
butadiene rubber (BR), styrene-acrylonitrile copolymer (SAN) and polyacrylonitrile. TG-
FTIR has also been used for a study of the thermal-oxidative and thermal degradation
processes of chlorinated natural rubber (CNR) [42]. Thermal degradation of CNR is a
one-step reaction with dehydrochlorination taking place at 160-190 °C, whereas
thermooxidative degradation of CNR is a two-step process with production of HCl and
CO2 at 160-390 °C and of CO2 at 390-575 °C.
Figure 1.3 Evolved gas profiles (EPG) for PTFE, Aflas and PP upon heating at 10 °C /
min. Ordinate axis values are arbitrary and simply relative
Reproduced with permission from H.G. Schild, Journal of Polymer Science A, 1993, 31, 6,
1629. Copyright 1993, John Wiley and Sons
21
Figure 1.4 Comparison of Aflas degradation product spectra at moderate (531 °C)
and high (726 °C) temperatures
Reproduced with permission from H.G. Schild, Journal of Polymer Science A, 1993, 31,
6, 1629. Copyright 1993, John Wiley and Sons
Figure 1.5 Infra red spectrum of the gases evolved from ABS terpolymer at 400 °C
Reproduced with permission from M. Suzuki and C.A. Wilkie, Polymer Degradation and
Stability, 1995, 47, 217. Copyright 1995, Elsevier Science Ltd
22
Schönherr [43] has described the combination of decomposition in a thermogravimetry

oven and FTIR spectroscopy for the identification of base polymers in elastomers, as
exemplified for nitrile rubber, and has presented infrared spectra for decomposition
products of various rubbers. The same author [36] studied use of the integrated TG-
FTIR system for the identification of sixteen vulcanised rubbers in mechanical goods
reporting the characteristic infrared spectra of the degradation products at temperatures
ranging from 334 °C to 635 °C.
As is well known, the properties of technical rubber goods are determined to a great extent
by the polymer type and comonomer ratios. Identification of the base polymer is therefore
of considerable interest, in particular in damage cases, raw material and product quality
control. Schönherr [44] used TG-FTIR to analyse an acrylonitrile(AN)-butadiene rubber
(NBR). TG-FTIR of NBR shows CS2 at 270 °C, HCN, NH3 and hydrocarbons at 450 °C.
The method was also applied to distinguish various uncured NBR (Perbunan NT and
Krynac) types with different AN content (18 to 48%). EGP and three FGP (characterising
HCN at 700-730 cm-1, NH3 at 800-1160 cm-1 and hydrocarbons at 2800-3000 cm-1) were
reported. NBR rubbers with different AN content are best distinguished on the basis of the
absorption intensities of HCN (714 cm-1) and CH compounds (2900-3000 cm-1). It thus
appears that the technically most important elastomer types can be discriminated by IR
spectra of their gaseous decomposition products [36, 44, 157]. The integrated TG-FTIR
system has been similarly used to analyse uncured and vulcanised ethylene vinyl acetate
copolymers (EVA) with different vinyl acetate (VAC) contents (40-70%) [45].
Polyesterurethanes, polycarbonate and silicone rubbers have been studied by TG-Tenax-

FTIR/MS. The degradation of polyesterurethanes yields CO2, water, tetrahydrofurans,
cyclopentanone, dicarbonic acid, and aliphatic diols and esters. The thermal decomposition
of silicone rubbers leads to the formation of polychlorinated biphenyls which are produced
in small amounts and can be observed in the mass spectrometer [86].
Post and co-workers [49] have used TG-FTIR to study the outgassing of a plasticiser
(type and amount) from an ethylene-propylene-diene terpolymer (EPDM) compound.
Figure 1.6 shows the thermogravimetric decomposition behaviour of the EPDM
compound. The plasticiser emerges in the first mass-loss step, which was identified as
adipic acid diisobutylester by on-line infrared analysis.
Melissaris and co-workers [46] have studied various novel p-ethynyl-substituted rigid
rod monomers by DSC, TG and TG-FTIR and identified the main decomposition products.
Void-free neat resin mouldings were made by compression moulding the monomers
followed by heating.
Wilkie and co-workers [47, 158] have used TG and TG-FTIR to study grafting of vinyl
monomers, such as methacrylic acid, onto styrene-butadiene block rubber (SBS) or ABS
23
Figure 1.6 Thermogravimetric decomposition behaviour of an EPDM compound

Reproduced with kind permission from E. Post, S. Rahner and F. Giblin, Proceedings of
Antec ‘97, Toronto, Canada, Volume 2, 2300, Figure 7. Copyright 1997, Society of
Plastic Engineers [49]
to achieve flame retardancy. Little difference was observed in the degradation pathway
between the grafted and virgin polymers. On-the-spot TG-FTIR of poly(butylene
terephthalate)(PBT)/octabromodiphenylether (molecular weight (MW) 801 Da) revealed
the brominated diphenylether at 275 °C and terephthalic acid (the starting monomer of
PBT) at 425 °C [110]. Similar high molecular species have never been reported in TG-
MS experiments; the flame retardant was also missed out in off-line TG-GC-FTIR-MS
analysis. The same authors [29] used the technique for the study of a carbon black-filled
styrene-butadiene rubber (SBR) providing information about the composition, including
organic additives, polymers, carbon black and inorganic fillers (see Figure 1.7). At 250 °C,
water, CS2 and morpholine were detected. The latter two components are degradation
products of the accelerator used, 2-(morpholinothio) benzothiazole. In an examination
of an ABS/polycarbonate (PC) blend with 8% triphenylphosphate (TPP), in addition to
the EGP, the SGP for the specific wavenumber windows of TPP (900-1200 cm-1), aromatic
compounds (3000-3100 cm-1), and carbon oxides originating from PC (2200-2300 cm-1)
were obtained. TPP evolving first was detected at about 150 °C (detection limit 0.5 μg/
s) [29]. With the direct IR detection of the on-the-spot technique, there is no loss of
evolved components by cold spots or discrimination of high-molecular-mass components.
In addition to the chemical composition of the evolved components, the technique also
provides information on the sequence and kinetics of the mass-loss process, which may
not show up in the mass-loss curve.
TG-IR has also been used to examine the thermally induced decomposition products of
polyvinyl chloride (PVC), polyacrylamide, tetrafluoroethylene-propylene, styrene-
24
Figure 1.7 IR spectrum of the degradation products of the accelerator of carbon black-
filled SBR at about 250 °C
Reproduced with kind permission from J.A.J. Jansen, J.H. v.d. Maas, A. Posthuma De Boer,
Applied Spectroscopy, 1992, 46, 1, 88. Copyright 1992, the Society for Applied Spectroscopy
butadiene composites and ethylene vinyl acetate (E/VAc) copolymers [159]. Because of
the mixture of products generated, identification of the constituents could only be made
when small molecules with simple infrared spectra, such as CO, CO2 and HCl were
generated. Alternatively, the effluent of TG-IR is captured on a trap constructed from a
GC capillary injector liner with Tenax solid phase adsorbent and analysed by GC-IR.
Other reported TG-FTIR applications concern: polyesterurethane [160], silicone rubber

[86, 33], ABS [161], synthetic rubbers used in the automotive industry [38] and the
determination of butadiene and styrene in unknown composites [162].
1.3.2 Thermogravimetry-Mass Spectrometry
Bart and co-workers [25] and others [34, 101, 163] have reviewed the application of TG-
MS for the study of polymeric materials, thermoplastics, thermosets and elastomers. This
thermoanalytical technique is used for the structural characterisation of homopolymers,
copolymers, polymeric blends and composites and finds application in the detection of
monomeric residuals, solvents, additives, (toxic) degradation products, etc. Information is
25
also obtained on the mechanism of solid-state reactions, chemical reactivity and curing;
TG-MS is also beneficial in matters of product formulation and development.
TG-MS is an ideal technique for identifying residual volatiles in polymers. The detection
of residual volatiles (and of other impurities) can often yield clues as to manufacturing
processes. In many cases, such as in the determination of highly volatile materials, of
residual solvents or plasticisers, use of TG-MS is requested. Specifically, there are reports
on the entrapment of curing volatiles in bismaleimide laminates [145] and elastomers
[48], on the detection of a curing agent (dicumylperoxide) in EPDM rubbers and of
bromine flame retardants in electronic waste [50], of plasticisers such as bambuterol
hydrochloride [142] or TPP and diethylterephthalate in cellulose acetate [143], on solvent
extraction and formaldehyde loss in phenolic resins [164], and on the evolution of toxic
compounds from PVC and polyurethane foams [146].
Kaisersberger and co-workers [50] have reported detection of nitrosamine precursory

compounds during rubber vulcanisation (originating from vulcanisation agents) and the
determination of toxic or environmentally damaging exhaust gases during technical
burning processes (polycyclic aromatic compounds, PCB, etc). TG-MS couplings are
increasingly used by the rubber industry especially since aromatic plasticisers are
toxicologically suspect. Post and co-workers [49] used the skimmer-MS coupling in TG-
MS measurements to study the outgassing of a plasticiser from an EPDM compound.
The sensitivity needed for identification of additives is, in general, greater than that
needed for identification of volatile pyrolysates.
In other applications the pattern of evolution of styrene, butadiene and acrylonitrile as a

function of temperature has provided a unique way for classifying different types of
ABS. The loss of the antioxidant butylated hydroxytoluene (BHT) was also detected by
MS preceding EVA copolymer degradation [165]; BHT was identified at a concentration
level of 20 ppm. Lehrle and co-workers [52] have described a successful controlled release
system for the stabilisation of rubber by encapsulating efficient but rather mobile
antioxidants to prevent loss from the host polymer. The performance of the controlled-
release of the antioxidant BHT from alginate matrix particles was studied by means of
DSC, TG and TG-MS. Polyisoprene rubber is more resistant to oxidation when protected
in this way than by the equivalent concentration of unencapsulated antioxidant.
TG-MS is ideally suited to reveal differences in pressure behaviour during thermal

decomposition of materials. This has been illustrated by Mol [144] in TG-MS analysis of
toluene diisocyanate (TDI) and methylene bis-4-phenol isocyanate (MDI)-based
polyurethanes, where the observed greater increase in pressure for the TDI polyurethane
than for the MDI derivative indicates a higher loss of low molecular weight fragments.
This is not possible to deduct from the TG curves alone. Such indications are of great
26
importance when evaluating kinetics or considering mechanically similar materials for

heat sensitive applications.
The presence of impurities and additives such as PVC as a fire retardant in polyurethanes
has also been detected by means of TG-MS analysis [144]. PU foam containing the flame
retardant tetrakis (2-chloroethyl) ethylenediphosphate decomposes in an oxidative
atmosphere at standard pressure in one rapid reaction whereby several highly toxic species
are formed; the TG-MS detection limit of this flame retardant was determined [166].
The most striking fact is the very early generation of vinylchloride from the flame retardant,
while HCN is produced at much higher temperatures.
Bart and co-workers [25] have reported the degradation of a complex, urethane-based,
thermoplastic elastomer and have identified the volatile products (mainly CO2, 1,3-
butadiene, tetrahydrofuran (THF) and cyclopentanone) by TG-MS and GC-MS. While
CO2, THF and HCN were observed by TG-MS a much richer evolved gas composition
with m/z = 44, 54, 70, 72, 84, 90, 200 (all components identified) was detected with TG-
CT-GC-MS. The main degradation mechanisms were identified as thermal degradation
of ester bonds of adipic acid and urethane moieties. This very complex polymer material
was chosen deliberately in an attempt to illustrate the limitations of the TG-MS technique.
As expected, not all the gases were readily identified. Several of the weight loss transitions
are caused by complex decomposition processes resulting in multiple materials being
evolved. Particularly if some of the components are present in minor amounts, the
separation becomes an integral part of the analysis. Recourse to GC-MS is an obvious
choice in case of complex mixtures in the TG atmosphere for which simple MS
identification of the component organic substances is not possible [88, 167]. Also several
EPDM products were studied by means of TG-MS [25]. TG/DTG of an EPDM without
filler and plasticiser shows that during the maximum weight loss phenomenon ENB (m/
z = 66, 91, 105 Da), aliphatics (m/z = 43, 56, 69 Da) and olefines (ethene: m/z = 26,
27 Da; propene: m/z = 40, 41, 42 Da) are detected. The dynamic DTG and MS curves in
inert atmosphere of an EPDM compound charged with oil, filler and carbon black,
indicates loss of oil (maximum at 336 °C), thermal stability of the polymer up to about
420 °C (maximum decomposition at 485 °C), and decarboxylation of the filler at 730 °C
(CO2: m/z = 12, 44 Da); finally, above 900 °C in O2 atmosphere carbon black is detected.
The same authors [25] have reported a TG-MS study of EPDM-SBR blends.
Griffiths and co-workers [53] have described the effect of the binder Alloprene (a chlorinated
rubber) on the ignition characteristics of some pyrotechnic compositions. Sklemin and co-
workers [54] have studied the kinetics of thermal degradation and evolution of volatile
products during thermolysis of ethylene-propylene rubber, synthetic fibre and carbon-fibre
reinforced phenol-formaldehyde copolymers. The temperature of initial degradation (5
wt.% loss) shifted to higher temperatures with increasing heating rate. Degradation rate,
27
rate constants of degradation and activation energies of degradation were determined.

Other applications of TG-MS in kinetic studies of polymers concern the (oxidative) thermal
degradation of polymethyl methacrylate (PMMA), polyurethanes and acrylonitrile (ACN)/
styrene copolymer [147]. Möhler and co-workers [51] have reviewed the suitability and
possibilities for thermal analysis in accordance with DIN 51005 [168] for quality control
(batch-to-batch control and optimisation) in the rubber industry, with special reference to
heat flux DSC, TG, TG-MS and thermomechanical methods. Results are reported for
various rubber compounds and blends.
TG-ToFMS has been used in product development by Kleineberg and co-workers [148]
as a means for the evaluation of the toxicity potential in normal use and catastrophic
situations of some 300 fire retardant materials applied in interiors of passenger and
cargo aircraft. ToFMS with its inherent high speed scanning capability enables the
toxicologist to relate conventional TGA information to the unequivocal identification of
potentially toxic thermal decomposition products. TG-ToFMS of a carboxynitroso rubber
showed abrupt, complete decomposition at 292 °C. As shown in Table 1.7 the mass data
indicated two primary decomposition products, namely carbonyl fluoride (m/z 66, 47,
50, 31, 19), perfluoro-N-methyl methylenimine) (m/z 133, 69, 114, 31, 50, 45, 26, 57,
64, 12, 19) and the secondary reaction products CO 2 (m/z 44) and
trifluoromethylisocyanate (m/z 111) and corrosion products (HF, SiF4, (CF3)2NH; m/z
85, etc.). Quantitative determination was achieved through correlation of the mass
spectrometric and the thermogravimetric data.
TG-MS and TG-FTIR have been used for ageing characterisation of the methylphenyl
silicone elastomers GE 566 (containing ferric oxide and silica filler) and GE 567 with the
Table 1.7 Decomposition products from a carboxynitroso rubber

m/z values
Carbonylfluoride COF2 66, 47, 50, 31, 19
Perfluoro-N-methylmethylenimine CF2=N-CF3 133, 69, 114, 31, 50,
45, 26, 57, 64, 12, 19
Trifluoromethylisocyanate CF3-N=C=O 111
Carbon dioxide CO2 44
Hydrofluoric acid HF
Silicon tetrafluoride SiF4 85
Perfluorodimethylamine (CF3)2NH
28
primary object of identification and quantitation of volatiles in these materials [48].

Cure catalysts were either 0.1 wt.% dibutyl tin dioleate or 0.5 wt.% dibutyl tin dilaurate.
TG-MS identified the trimethyl silyl fragment ion, benzene, styrene and cyclic oligomers
of dimethyl siloxane. The major difference between GE 566 and GE 567 found by TG-
MS was the much greater amount of weight loss and the presence of styrene for the
latter, irrespective of the catalyst used.
Other reported TG-MS applications concern polybutadiene [153], styrene-butadiene rubbers

[153], gums [14], polyisoprenes [52], polyurethanes [144, 146, 147, 166], ABS [144],
chlorosulphonated polyethylene elastomer [169, 170] and elastomer blends (NBR/SBR/
BR) [13]. Table 1.5 summarises the use of advanced TG-MS systems in elastomer analysis.
1.3.3 Thermogravimetry - Differential Scanning Calorimetry
Thermal analysis is a useful tool in the quality control of many incoming routine materials,
which can be tested against a reference standard developed internally by analysing a
large number of samples of known performance criteria to ensure that the quality of
supplies is maintained. Solid elastomers can be identified by glass transition temperature
(Tg) [70]. The rubber industry uses thousands of different raw materials, and this number
is steadily increasing. These materials are listed in [172].
Hyphenated TG-DSC provides simultaneous measurement of transition temperatures/

heat flows and weight changes in materials, thereby simplifying interpretation, increasing
productivity and assuring identical experimental sampling. TG-DSC can be used to study
thermal stability, decomposition behaviour, drying and firing processes, reaction rates,
sample composition, kinetics of reaction, transition and reaction temperatures, melting
and crystallisation processes as well as reaction mechanisms. Although TG is an excellent
technique for the compositional analysis of compounded elastomers, it does not reveal
the extent of cure. For that purpose DSC is required. Redfern [57] has reported TG-DSC
of an uncured polyimide resin in which a more accurate determination of the quantitative
measurement of the heat of cure is made possible by the simultaneous technique.
Kodama and co-workers [58] have reported TG-DSC curves for the analysis of the
interaction between vulcanisation accelerators (tetramethylthiuram disulphide,
dibenzothiazolyl disulphide, diphenylguanidine and N-cyclohexyl-2-benzothiazolyl-
sulphenamide) and fillers (carbon black, white carbon, hard clay and CaCO3). The initial
melting point (MP) of the accelerators was largely influenced by the fillers. The higher
the surface activity of the filler is, the lower and wider the melting range becomes.
Emmott and co-workers [59] have investigated the complex reaction between Sr(NO3)2
and the binder Alloprene (a pyrotechnic system) at about 300 °C by simultaneous TG-
29
DSC and TG-DTA-MS. The same techniques were used to examine the Ti-NaNO3-Alloprene
system [60, 61]. The NaNO3 reacts with HCl from Alloprene decomposition and the
carbonaceous residue of the Alloprene is oxidised by the O and NOx formed in the first
stage. Simultaneous TG-DSC and TG-DTA-MS were also used to investigate the first reaction
stage in the related Mg-NaNO3-Alloprene (chlorinated rubber) pyrotechnic system at about
300 °C [62]. The reduction in overall weight losses and the amount of main gaseous products
evolved (CO2, NO, H2O) suggested interaction of Mg with the gaseous products of the
reaction of HCl and NaNO3; this contrasts with the action of Ti [61].
The TG-DSC technique has recently been reviewed [56]. Redfern [57] has reviewed single
sample simultaneous thermal analysis, i.e., TG-DSC and TG-DTA studies of polymers.
1.3.4 Thermogravimetry - Differential Scanning Calorimetry - Mass

Spectrometry
Whereas Redfern [57] has pointed out the advantages of simultaneous thermal analysis
techniques (particularly TG-DSC and TG-DTA) over techniques conducted singly, an
even more complete thermal profile is provided when a thermal analyser is coupled to
some form of gas analyser (MS or FTIR). Möhler and co-workers [51] have reported
TG-DSC-MS of the thermal decomposition of the vulcanisation accelerator tetramethyl
thiuram disulphide (TMTD) in rubber; degradation of TMTD starts at about 155 °C, as
evidenced by m/z 76 (CS2) and 44 (radical of the secondary dimethylamine).
The high sensitivity of the instrumental combination was demonstrated by Kaisersberger

and co-workers [63] who published TG-DSC-MS data for EPDM showing cumyloxy
radicals (m/z 135, 136) from the dicumylperoxide system (see Figure 1.8). Without the
MS data, the mass loss in the range from 240 °C to 400 °C would only have been attributed
to the plasticiser content. Hyphenation prevents both a misinterpretation of the results
and permits optimisation of the process by adjusting the amount of DCP added to the
elastomer prior to vulcanisation. TG-DSC-MS was also used to recognise epoxy resin
fragments (m/z 58, 92, 135) in electronic scrap from the automobile industry and bromine
flame retardants in electronic waste [50, 63].
1.3.5 Thermogravimetry - Differential Thermal Analysis
Since TG and DTA complement each other, it is an obvious move to attempt both
investigations simultaneously [173]. TG-DTA measures mass and energy changes as a
function of temperature or time. Depending on the atmospheric conditions (vacuum,
inert or air conditions) thermal or oxidative stability is measured. Typical TG-DTA
30
Figure 1.8 EPDM rubber with peroxide curing agent, detection of cumyloxy fragments
Reproduced with permission from J. Janoschek, E. Kaisersberger and E. Post, Polymer
Materials Science Engineering, 1993, 69, 458. Copyright 1993, the American Chemical
Society [63]
applications are thermal and oxidative stability, determination of relative components,

decomposition temperatures and thermal decay reactions, action of heat stabilisers,
thermal ageing. The main use of DTA is to detect the initial temperatures of thermal
processes and qualitatively characterise them as endothermic or exothermic, reversible
or irreversible, etc. Ideally, the area under the DTA peak should be proportional to the
heat of the process that gave rise to the peak. TG/DTG-DTA instruments are commercially
available.
Negri and co-workers [149] have applied TG-DTA to the characterisation of different
types of carbon black in natural rubber (NR) vulcanisates. The method allowed the
determination of the overall carbon black content, but where combinations of different
blacks were present it was not possible to determine the proportion of each type, as
instead is possible by means of pyrolysis gas chromatography FID (PyGC-FID) [174].
TG-DTA has also been used to correlate thermogravimetric analysis in airflow and N2
gas flow and some other micro-scale flammability tests, i.e., the oxygen index, hot-plate
ignition and drum friction tests, on covers of different flame-resistant and non flame-
resistant rubber conveyer belts [64]. The minimum temperatures at which rapid weight
loss of each sample began to appear were determined and compared with the results
from the micro-scale flammability tests.
31
Lee and co-workers [65] have studied the thermal degradation of polyetherurethane elastomers
based on polytetramethyleneglycol (PTMG, MW 1000 or 2000), 4,4´-
diphenylmethanediisocyanate and 4,4´-diaminodiphenylmethane, m-phenylenediamine or
p-phenylenediamine by means of TG-DTA. Degradation is a two stage process; hard segments
degrade first, soft polymer segments next. Also the thermal decomposition of adhesive tapes
on the basis of acrylic rubber, acrylic-ethylene rubber, and acrylonitrile-butadiene rubber has
been studied by TG-DTA [66]. Similarly, new block copolyurethane elastomers consisting of
1,2-ethylene bis (4-phenylisocyanate), poly(etherglycol) and ethylenediamine, were
characterised by DSC, TG-DTA and FT-IR to determine the morphological structures and
thermal stability. PTMG-based copolyurethane shows superior thermal stability due to its
more cohesive hard domains and a better mutual-stabilisation effect between hard and soft
segments [67]. TG-DTA results have also shown that the presence of SnO2 and/or γ-Fe2O3
could effectively inhibit the oxidative decomposition of residual Si-OH groups in an addition
type silicone rubber and improve the thermal stability [68]. TG-DTA measurements have
further been used to indicate that elastic polyaniline has excellent thermal stability [69].
The combination of TG-DTA with DSC, DMA and TMA is useful in monitoring and
characterising the influence of artificial ageing on the chemical and physical properties
of EPDM roofing materials. Significant changes in Tg, the coefficient of thermal expansion
and the chemical composition were observed due to heat ageing [70]. Thermo-analytical
techniques and infrared have been applied to investigate the effect on the molecular level
of heat ageing, humidity and UV radiation over 8000 hours on several silicone-based
sealants [71]. TG-DTA confirmed the susceptibility to loss of polydimethylsiloxane
(PDMS) as well as other sealant components at elevated temperatures. Weathering of
thermoplastic polyester elastomers (TPEE) has been studied by TG-DTA, FTIR, 1H nuclear
magnetic resonance (NMR) and gel permeation chromatography (GPC) [72]. Ether parts
of the soft segments in the polymer are degraded selectively and ester bonds are formed.
There was a clear difference between tendencies of degradation by outdoor exposure
and accelerated weathering.
A TG-DTA study of the thermochemical processes occurring at vulcanisation temperatures

with N-oxydiethylene-2-benzthiazyl sulphenamide and N-cyclohexyl-2-benzthiazyl
sulphenamide and their mixtures with sulphur showed the formation of high molecular
weight polysulphides [73]. The influence of metallic oxides (Fe2O3, SnO2) on hot air
ageing of one-pack room temperature vulcanised fluorosilicone rubber has been studied
by means of TG-DTA [74, 75]. TG-DTA and TG were both applied to study the thermal
characteristics of room temperature vulcanised silicone rubber [76].
TG-DTA and DSC are suitable for product quality control as exemplified by OIT
measurements for polyethylene (PE) and quantitative analysis of the rubber phase in
ABS and of a polymer/ softener/soot/mineral filler mixture [77].
32
Khan [78] has examined the thermal response of various multiphase engineering materials
(such as modified PVC/EPDM composites) in relation to the integrity and morphology
of the interfacial junction. TG-DTA data displayed enhanced thermal stability by the
EPDM insulation in the advanced viscoelastic composite that compares well with
competitor advanced silicone composites. Riga and co-workers [79] have examined the
oxidative behaviour of commercial engineering plastics, polyolefins and elastomers by
simultaneous TG-DTA and pressure differential scanning calorimetry (PDSC). Riga [80]
has also reported illustrative case studies showing how PDSC, modulated differential
scanning calorimetry (MDSC), TMA, TG-DTA and DETA are used to select polymers,
elastomeric automotive seals and lubricant additives. Elastomeric parts must be flexible
and impervious to hostile engine environments, including hot-stressed motor oils possibly
containing additives. Thermal analysis can be used to evaluate seals for composition
(TGA), mechanical strength and glass transition temperature (DMA and/or TMA) and
oxidative stability (PDSC and TG-DTA) [81].
Yang and co-workers [82] have reported a mechanistic study carried out based on the
simultaneous measurement of enthalpy and weight loss by TG-DTA on tyre rubbers:
NR, SBR and BR. The hyphenated technique allows to attribute the observed exothermic
peak to chemical reactions and the endothermic peak to evaporation of pyrolysis products.
The evaluation of combustion performance and safety of Mg/polytetrafluroethylene

(PTFE) pyrotechnic compositions by means of TG-DTA has been reported by Miyake
and co-workers [83]. Similar work on different pyrotechnic systems containing a
chlorinated rubber binder has made use of TG-DSC and TG-DTA-MS [59-62]. The
hyphenated technique TG-DTA (DSC) has recently been reviewed [56].
1.3.6 Thermogravimetry - Differential Thermal Analysis - Mass Spectrometry
TG/DTG-DTA-MS equipment has been available since 1979 [11]. As mentioned before,
simultaneous TG-DTA-MS was used to investigate the exothermic reactions that take
place at approximately 300 °C between Alloprene binder and NaNO3 and Sr(NO3)2 in
various pyrotechnic compositions, including the effect of titanium [85]. For both binary
nitrate - Alloprene mixtures, reaction is initiated by Alloprene decomposition; another
characteristic reaction is that of HCl (from the chlorinated rubber decomposition) with
the nitrate to yield the metal chloride.
Manley [32] examined a cured phenolic formaldehyde (PF) resin by means of TG-DTA-
MS observing a lower sensitivity of TG relative to DTA. However, new TGA instrumental
developments have been reported since. The TG curve shows loss of phenol (MS evidence:
m/z 94); DTA observes water (MS: m/z 18), ammonia (MS: m/z 17) and formaldehyde
33
(MS: m/z 29), indicating disrupture of cross-links greatly effecting the mechanical
properties of PF moulded compounds. The MS traces show catastrophic deterioration of
PF resins at 200 °C. The DTA trace also signals a change around 200 °C. DTA is thus a
useful indicator of temperatures at which engineering properties may change but MS
shows clearly why these changes occur.
1.3.7 Thermogravimetry - Gas Chromatography - Mass Spectrometry
Volatile additives for vulcanised or unvulcanised rubbers can be accurately identified by

TG or by controlled heating of a test sample in a sealed vial equipped with an overhead
collecting headspace, transferring the heated volatile substances to a chromatographic
column and analysing the separated volatile components emerging from the
chromatograph column by various selective analytical detectors. Several illustrative
examples were mentioned before.
Although it is possible to obtain separate responses for several molecular species by TG-
MS, pyrolysis of samples such as polymers may give many products. It is an advantage if
these can be separated between the thermal analysis unit and the mass spectrometer by
gas chromatography. Gorman [175] has proposed a controlled thermal desorption and
concentration method (essentially headspace) for separating volatile additives from
vulcanisable rubber, in a TD-GC-MS configuration without the need for prior sample
preparation such as milling, extraction or pyrolysis. Meuzelaar and co-workers [150]
used high pressure TG-GC-MS to simulate solvent-free thermal and catalytic liquefaction
reactions for coprocessing of waste polymers (non-vulcanised SBR alone or mixtures of
waste plastics composed of PE, PS and waste rubber tyres) with coal.
1.3.8 Thermal Desorption - Gas Chromatography - Fourier Transform Infra-

red Spectroscopy/Mass Spectrometry
The coupling of thermal desorption and identification techniques constitutes a powerful

means for the detailed characterisation of outgassing processes with many potential
applications in the field of rubbers.
Gorman [175] has described a thermal desorption process for separating volatile
components such as accelerator fragments, antioxidants and other organic additives from
2-10 g (un)vulcanised rubber samples for the purpose of analysing the separated volatile
compounds. The process comprises sealing the test sample in a glass vial containing a
controlled atmosphere and overhead headspace, followed by heating for complete
desorption. The volatilised gases are then swept through a capillary GC column and
analysed by MS, FTIR and FID using column splitters.
34
Jansen and co-workers [86] have evaluated temperature-controlled outgassing processes

of plastics and rubbers using both off-line and on-line TD-GC-FTIR-MS. Decomposition
of polyesterurethanes by means of TG-Tenax off-line sampling followed by TD-GC-
FTIR-MS revealed CO2, H2O, tetrahydrofurane, cyclopentanone, dicarbonic acid,
aliphatic diols and esters [86]. The same authors have also described the detection of
polychlorinated biphenyls (PCB) in 2,4-dichlorobenzoylperoxide cured silicone rubbers
after outgassing products of a rubber silicone part obtained after desorption for 10 minutes
at 200 °C in the thermal desorption cold-trap and subsequent analysis by means of TD-
GC-MS. Using a mass range of 290-294 Da the MS can be used as a selective detector for
these substances.
1.4 Future Prospects of Hyphenated Thermogravimetric Techniques in

Elastomer Characterisation
As has been seen, competitive analytical systems correlating thermal and chemical
behaviour are TG-MS, TG-FTIR, TG-GC, TG-GC-MS, etc. This list of techniques may
easily be extended with various pyrolysis methods.
It may be noticed that in particular combinations of thermal analytical instrumentation

with MS provide a powerful means of analysing and characterising polymeric materials,
providing both qualitative data and giving quantitative applications. TG-MS is not the
solution to all problems in the study of polymer structure and degradation; other
techniques will continue to give valuable contributions. However, in many instances this
technique is a very valuable first experiment in order to identify reaction products and to
define the underlying chemistry. It is therefore expected that TG-MS will keep a key
position in problem solving for polymer analysts. Clearly though, routine application of
TG-MS requires skilled and dedicated operators. Reliable TG-MS equipment is now
commercially available. A limit of about 500 Da (bench-top range) is likely to be set by
interface design requirements.
The current interest of the material in events at lower temperatures (with evolution of
complex mixtures of components at much lower concentration at the onset of the reaction)
requires the need for quantification of components at very low concentrations and
demands high sensitivity, accuracy, and reproducibility of the analytical technique; at the
same time, the identification capability of the method requires strengthening.
Today’s challenge is no longer faster collection of mass spectral data over broader (m/z)
ranges, but more rapid identification of co-evolved species. In this respect several
approaches for improvement are possible, such as high resolution TG-MS experiments
(allowing for better experimental resolution of closely spaced thermal events), soft
35
ionisation techniques (increasing the identification power of the mass spectrometric event)
and principle component analysis (PCA) of TG-MS data (for more reliable interpretation).
Indeed, Alders and co-workers [151] have suggested that high resolution thermogravimetry
– electron impact/soft ionisation quadrupole mass spectrometry (HRTG-EI/SI QMS)
extended with PCA is currently the most advanced design for TG-MS coupling. PCA
analysis of TG-MS data has recently successfully been carried out by Tas and co-workers
[176]. The future will probably see a renaissance of TG-ToFMS.
A major challenge in TG/DTG-based analysis of elastomer vulcanisates is to demarcate

oil/plasticiser and elastomer regions, which often show overlapping TG events. Most of
these materials have volatilisation ranges rather than discrete volatilisation points because
they are chemical blends of components of various molecular weights and volatilities.
Deconvolution of the overlapping oil/plasticiser and oil/elastomer TG curves is expected
to be feasible with HRTG-MS-PCA which would substitute the dated methods for
graphical resolution of oil and polymer weight loss [177].
The increasing need to correlate thermal behaviour simultaneously with the underlying
chemistry (outgassing, thermostability, degradation) and the accompanying physical
phenomena (blooming, cracking, foaming, migration, change in colour) determines the
usefulness of a variety of endoscopic, audiometric and magnetometric extensions. For
example, VI-TG is a direct visualisation of the morphological and textural changes in
the substrate during thermal processing. VI-TG-MS combines chemical and physical
information with visual observations [26] and is expected to gain rapidly in popularity
on account of its close observancy of chemical and physical phenomena and for
documentation purposes.
Method standardisation of multihyphenated techniques remains an important issue.
1.5 Summary
The experimentalist disposes of a variety of powerful thermogravimetric tools for the

characterisation and study of the properties of elastomeric materials. Interpretation of
TG data is facilitated on the basis of spectroscopic and mass spectrometry information
which ease interpretation and quantification. The study of polymeric materials by a
combination of thermal analysis and chemical identification tools provides the chemist
with a wealth of information. In the case of TG-FTIR and TG-MS molecular weight
information is collected concerning the evolved gases which are responsible for the detected
weight losses. The chemical information obtained allows separation of concurrent or
overlapping reactions. New ways have been indicated to allow for a complete identification
of co-evolving species using simultaneous experimentation. Alternatively, it is
36
advantageous to separate the components in a gas chromatograph before analysis in the

mass spectrometer.
References
1. H.G. Wiedemann, Chemical Engineering and Technology, 1964, 36, 1105.
2. G.M. Stanton and E.M. Barrall, ACS Division of Petroleum Chemistry Preprints,
1996, 14, A59.
3. T.L. Chang and T.E. Mead, Analytical Chemistry, 1971, 43, 534.
4. J. Chiu, Analytical Chemistry, 1968, 40, 1516.
5. J.A.J. Jansen, University of Utrecht, 1992. [Ph.D. Thesis]
6. J.A.J. Jansen in Calorimetry and Thermal Analysis of Polymers, Ed., V.B.F.

Mathot, Hanser Publishers, Munich, Germany, 1994, 335.
7. W.R. Holdiness, Thermochimica Acta, 1984, 75, 361.
8. W.W. Wendlandt and T.M. Southern, Analytica Chimica Acta, 1965, 32, 405.
9. R.S. Gohlke and H.G. Langer, Analytical Chemistry, 1965, 37, 10, 25A.
10. K.W. Smalldon, R.E. Ardrey and L.R. Mullings, Analytica Chimica Acta, 1979,
107, 327.
11. W.D. Emmerich and E. Kaisersberger, Journal of Thermal Analysis, 1979, 17, 197.
12. G. Giovanoli and H.G. Wiedemann, Helvetica Chimica Acta, 1968, 51, 1134.
13. B. Shushan, B. Davidson and R.B. Prime, Analytical Calorimetry, 1984, 5, 105.
14. S.M. Dyszel, Thermochimica Acta, 1983, 61, 169.
15. F. Zitomer, Analytical Chemistry, 1968, 40, 1091.
16. P.S. Gill, S.R. Saurbrunn and B.S. Crowe, Journal of Thermal Analysis, 1992, 38,
255.
17. T.J. Lever and A. Sutkowski, Journal of Thermal Analysis, 1993, 40, 257.
37
18. E. Baumgartner and E. Nachbaur, Thermochimica Acta, 1997, 19, 3. 19. C.A.
Cody, L. DiCarlo and B.K. Faulseit, American Laboratory, 1981, 13, 93.
20. W.H. McClennen, R.M. Buchanan, N.S. Arnold, J.P. Dworzanski and H.L.C.
Meuzelaar, Analytical Chemistry, 1993, 65, 2819.
21. J. Mullens, R. Carleer, G. Reggers, M. Ruysen, J. Yperman and L.C. van Poucke,
Bulletin de la Société Chimique de Belgique, 1992, 101, 267.
22. S. Morisaki, Thermochimica Acta, 1974, 9, 157.
23. L.F. Whiting and P.W. Langvardt, Analytical Chemistry, 1984, 56, 1755.
24. P. Cukor and E.W. Lanning, Journal of Chromatographic Science, 1971, 9, 487.
25. K.G.H. Raemaekers and J.C.J. Bart, Thermochimica Acta, 1997, 295, 1-2, 1.
26. K.G.H. Raemaekers and J.C.J. Bart, Proceedings of the SPE ANTEC ’99, New
York, NY, USA, 1999, Volume 2, 2599.
27. A.N. Matzakos and K. Zygourakis, Review of Scientific Instruments, 1993, 64,
6, 1541.
28. A.K. Sircar in Thermal Characterization of Polymeric Materials, Ed., E.A. Turi,
Academic Press, San Diego, CA, USA, 1997, 888.
29. J. Jansen, J.H. v.d. Maas, A. Posthuma De Boer, Applied Spectroscopy, 1992, 46,
1, 88.
30. M.L. Mittleman, D. Johnson and C.A. Wilkie, Trends in Polymer Science, 1994,
2, 11, 391.
31. R.M. Carangelo, P.R. Solomon, R. Bassilakis, D. Gravel, M. Baillargeon, F.

Baudais, G. Vail and J. Whelan, American Laboratory, 1990, 22, 6, 51.
32. T.R. Manley, Progress in Rubber and Plastics Technology, 1989, 5, 4, 283.
33. D.A.C. Compton, D.J. Johnson and M.L. Mittleman, Research-Development,

1989, 31, 4, 68.
34. J. Mullens in Handbook of Thermal Analysis and Calorimetry, Ed., M.E. Brown,
Elsevier Science, Amsterdam, 1998, Volume 1, Chapter 12.
35. R. Schönherr, GIT Labor-Fachzeitschrift, 1997, 41, 8, 828.
38
36 R. Schönherr, Kautschuk und Gummi Kunststoffe, 1996, 49, 5, 371.
37. E.G. Brame, Die Makromolekulare Chemie - Macromolecular Symposia, 1993,

72, 1.
38. J.P. Redfern and J. Powell, ACS Symposium Series, 1994, 581, 90.
39. R.M. Paroli and A.H. Delgado, Polymer Materials Science Engineering, 1993,
69, 139.
40. H.G. Schild, Journal of Polymer Science A, 1993, 31, 6, 1629.
41. M. Suzuki and C.A. Wilkie, Polymer Degradation and Stability, 1995, 47, 2, 217.
42. S. Li, Z. Peng, H. Yu, J. Zhong and Y. Wei, Guangpuxue Yu Guangpu Fenxi,
1998, 18, 4, 103.
43. R. Schönherr, LaborPraxis, 1995, 19, 9, 24.
44. R. Schönherr, Kautschuk und Gummi Kunststoffe, 1996, 49, 11, 737.
45. R. Schönherr, Kautschuk und Gummi Kunststoffe, 1997, 50, 7-8, 564.
46. A.P. Melissaris, J.K. Sutter, M.H. Litt, D.P. Scheiman and M.A. Schuerman,
Macromolecules, 1995, 28, 4, 860.
47. C.A. Wilkie, M. Suzuki, X. Dong, C. Deacon, J.A. Chandrasiri and T.J. Xue,
Polymer Degradation and Stability, 1996, 54, 2-3, 117.
48. A.R. Cooper, M.M. Steiner, G.M. McCauley, G.R. Kwiatkowski, J.G. Moncur
and E.H. Kawasaki, Proceedings of the SPE ANTEC ’89, New York City, NY,
USA, 1989, 1137.
49. E. Post, S. Rahner and F. Giblin, Proceedings of the SPE ANTEC ’97, Toronto,
Canada, 1997, Volume 2, 2300.
50. E. Kaisersberger, E. Post and J. Janoschek, ACS Symposium Series, 1994, 581,
74.
51. H. Möhler, A. Stegmayer and E. Kaisersberger, Kautschuk und Gummi

Kunststoffe, 1991, 44, 4, 369.
52. F.E. Keen, R.S. Lehrle, E. Jakab and T. Szekely, Polymer Degradation and
Stability, 1992, 38, 3, 219.
39
53. T.T. Griffiths, J. Queay, E.L. Charsley and S.B. Warrington, Proceedings of the
15th International Pyrotechnics Seminar, Boulder, CO, USA, 1990, 353.
54. N.K. Sklemin, V.V. Grishin and R.A. Khmel’nitskii, Zavodskaya Laboratoriya,
1985, 51, 6, 53.
55. J.P. Redfern, Proceedings of the Applied Polymer Symposium, Alexandria, Egypt,
1994, 55, 65.
56. J. van Humbeeck in Handbook of Thermal Analysis and Calorimetry, Ed., M.E.
Brown, Elsevier Science, Amsterdam, 1998, 1, Chapter 11.
57. J.P. Redfern, Polymer International, 1991, 26, 1, 51.
58. S. Kodama, H. Kawasaki, K. Itatani, F. Kusano and T. Nakatsuka, Okayama-ken

Kogyo Gijutsu Senta Hokoku, 1981, 7, 1.
59. P. Emmott, T.T. Griffiths, J. Queay, E.L. Charsley and S.B. Warrington,
Proceedings of the 16th International Pyrotechnics Seminar, Jönköping, Sweden,
1991, 937.
60. P. Emmott, T.T. Griffiths, J. Queay, E.L. Charsley and S.B. Warrington,
Proceedings of the 18th International Pyrotechnics Seminar, Breckenridge, CO,
USA, 1992, 221.
61. E.L. Charsley, S.B. Warrington, P. Emmott, T.T. Griffiths and J. Queay, Journal of
Thermal Analysis, 1992, 38, 4, 641.
62. T.T. Griffiths, J. Queay, E.L. Charsley and S.B. Warrington, Proceedings of the
19th International Pyrotechnics Seminar, Christchurch, New Zealand, 1994, 716.
63. J. Janoschek, E. Kaisersberger and E. Post, Polymer Materials Science

Engineering, 1993, 69, 458.
64. Y. Nakagawa and T. Komai, Journal of Fire Sciences, 1990, 8, 6, 455.
65. H.K. Lee and S.W. Ko, Journal of Applied Polymer Science, 1993, 50, 7, 1269.
66. A. Tada and T. Yoshida, inventors; Sumitomo Bakelite Co., assignee; Jpn. Kokai
Tokkyo Koho 10,095,958, 1998.
67. T-L. Wang and T-H. Hsieh, Polymer Journal (Tokyo), 1996, 28, 10, 839.
40
68. Z. Su, D. Pan, J. Zheng, Z. Huang and B. Cai, Hecheng Xiangjiao Gongye, 1996,
19, 2, 103.
69. A. Kitani, K. Yoshioka, S. Maitani and S. Ito, Synthetic Metals, 1997, 84, 1-3,
83.
70. J.J. Penn and R.M. Paroli, Thermochimica Acta, 1993, 226, 1-2, 77.
71. M.A. Lacasse and R.M. Paroli in Science and Technology of Building Seals,
Sealants, Glazing, and Waterproofing, 4th Edition, Ed., D.H. Nicastro, STP1243,
1995, ASTM, Conshohocken, PA, 29.
72. Y. Nagai, T. Ogawa, L.Y. Zhen, Y. Nishimoto and F. Ohishi, Polymer

Degradation and Stability, 1997, 56, 1, 115.
73. D.V. Tarasov, I.I. Vishnyakov and B.S. Grishin, International Polymer Science
and Technology, 1998, 25, 10, 23.
74. D. Pan, Z. Su, J. Zheng and B. Cai, China Rubber Industry, 1998, 45, 7, 402.
75. Z. Su, D. Pan, J. Zheng and B. Cai, Hecheng Xiangjiao Gongye, 1998, 21, 2, 96.
76. S. Kumagai, X. Wang and N. Yoshimura, IEEE Transactions on Dielectrics and

Electrical Insulation, 1998, 5, 2, 281.
77. J. Koch, LaborPraxis, 1994, 18, 10, 58.
78. M.B. Khan, Proceedings of the 5th International Symposium on Advanced

Materials, Ed., M.A. Khan, Rawalpindi, Pakistan, 1997, p.57.
79. A. Riga, R. Collins and G. Mlachak, Thermochimica Acta, 1998, 324, 1-2, 135.
80. A. Riga, Proceedings of the 26th North American Thermal Analysis Society, Ed.,
K.R. Williams, Omnipress, Madison, USA, 1998, 648.
81. A. Riga, Proceedings of the 26th North American Thermal Analysis Society, Ed.,
K.R. Williams, Omnipress, Madison, USA, 1998, 667.
82. J. Yang and C. Roy, Thermochimica Acta, 1996, 288, 1-2, 155.
83. A. Miyake, K. Kitoh, T. Ogawa, M. Watanabe, N. Kazama and S. Tsuji,

Proceedings of the 19th International Pyrotechnics Seminar, Christchurch, New
Zealand, 1994, 124.
41
84. C.Y. Park. E.H. Hwang and S.K. Min, Eylasutoma, 1998, 33, 3, 201.
85. P. Emmott, T.T. Griffiths and J. Queay, Proceedings of the 17th International
Pyrotechnics Seminar, Beijing, China, 1991, 25.
86. J.A.J. Jansen, W.E. Haas, H.G.M. Neutkens and A.J.H. Leenen, Thermochimica
Acta, 1988, 134, 307.
87. A.E. Pavlath, K.S. Gregorski and R. Young, Thermochimica Acta, 1985, 92, 383.
88. H.K. Yuen and G.W. Mappes, Thermochimica Acta, 1983, 70, 269.
89. N.S. Arnold, M-G. Kim, W.H. McClennen, J.P. Dworzanski and H.L.C.
Meuzelaar, Proceedings of the Workshop Ion Mobility Spectrometry, Mescalero,
NM, USA,1992, 11.
90. K. Liu, E. Jacab, W.H. McClennen and H.L.C. Meuzelaar, ACS Division of Fuel
Chemistry, Preprints, 1993, 38, 823.
91. F. Paulik, J. Paulik and L. Erdey, Zeitschrift für Analytical Chemistry, 1958, 160,
241.
92. F. Paulik, J. Paulik and L. Erdey, inventors; no assignee; HU Patent 145, 369,
1955.
93. I. Berecz, S. Bohátka, G. Langer and G. Szöor, International Journal of Mass

Spectrometry and Ion Physics, 1983, 47, 273.
94. J.P. Redfern, International Labmate, 1986, 11, 1, 19.
95. Technical Data Sheet STA 409, Netzsch Gerätebau, Selb, Germany.
96. E. Kaisersberger and E. Post, Thermochimica Acta, 1997, 295, 1-2, 75.
97. J.A.J. Jansen and W.E. Haas, Analytica Chimica Acta, 1987, 196, 69.
98. D.A.C. Compton, International Labmate, 1987, 12, 4, 37.
99. R.C. Wieboldt, G.E. Adams, S.R. Lowry and R.J. Rosenthal, American
Laboratory, 1988, 20, 1, 70.
100. P.R. Solomon, M.A. Serio, R.M. Carangelo, R. Bassilakis, Z.Z. Yu, S. Charpenay
and J. Whelan, Journal of Analytical Applied Pyrolysis, 1991, 19, 1.
42
101. J. Chiu in Applied Polymer Analysis and Characterisation: Recent Developments

in Techniques, Instrumentation, Problem Solving, Ed., J. Mitchell, Hanser
Publishers, Munich, Germany, 1987, 175.
102. D.J. McEwen, W.R. Lee and S.J. Swarin, Thermochimica Acta, 1985, 86, 251.
103. J.O. Lephart, Applied Spectroscopy Reviews, 1982, 18, 265.
104. P.B. Roush, J.M. Luce and G.A. Totten, American Laboratory, 1983, 15, 10, 90.
105. A.G. Nerheim in Fourier Transform Infrared Spectroscopy.

Applications to Chemical Systems, Ed., J.R. Ferraro and L.J. Basile, Academic
Press, New York, 1985, Chapter 4, 147.
106. M.L. Mittleman, D.A. Compton and P. Engle, Proceedings of the 13th Meeting of
the North American Thermal Analysis Society, Philadelphia, PA, USA, 1984, 410.
107. D.A.C. Compton, S.L. Hill, N.A. Wright, M.A. Druy, J. Piche, W.A. Stevenson
and D.W. Vidrine, Applied Spectroscopy, 1988, 42, 972.
108. E. Post, S. Rahner, H. Möhler and A. Rager, Thermochimica Acta, 1995, 263, 1.
109. J.K. Whelan, P.R. Solomon, G.V. Deshpande and R.M. Carangelo, Energy and
Fuels, 1988, 2, 65.
110. J.A.J. Jansen, J.H. van der Maas and A. Posthuma De Boer in Integration of
Fundamental Polymer Science and Technology-5, Ed., P.J. Lemstra, Elsevier
Applied Sciences, Oxford, 1991, 316.
111. T.B. Brill, Analytical Chemistry, 1989, 61, 897A.
112. W. Herres, HRGC-FTIR-Capillary Gas Chromatography - Fourier Transform

Infrared Spectroscopy: Theory and Applications, Hüthig Publishers, Heidelberg,
Germany, 1987.
113. M.L. Mittleman, Thermochimica Acta, 1990, 166, 301.
114. R.A. Nyquist, The Interpretation of Vapor-Phase Infrared Spectra, Volume 1:

Group Frequency Data, Sadtler Research Laboratories, Philadelphia, PA, 1984, 87.
115. L. Meublat and P. Le Parlouer, Spectra 2000, 1991, 161, 59.
116. W. Schwanebeck and H.W. Wenz, Fresenius’ Journal of Analytical Chemistry,

1988, 331, 61.
43
117. H.G. Wiedemann and G. Bayer, Fresenius’ Zeitschrift fur Analytical Chemistry,
1973, 266, 97.
118. J.B. Henderson, E. Post, E. Treser and B. Fidler, Proceedings of the SPE ANTEC
’99, New York, NY, USA, 1999, Volume 2, 2589.
119. F. Giblin, E. Post, W-D. Emmerich and G. Bräuer, Proceedings of the 23rd
Meeting of the North American Thermal Analysis Society, Toronto, Canada,
1994.
120. Technical Data Sheet Coupling Systems, Netzsch Gerätebau, Selb, Germany,
1996.
121. J.P. Redfern, P.H. Newbatt and P. Larcey, Polymer Materials Science Engineering,
1993, 69, 144.
122. J. Chiu, Thermochimica Acta, 1970, 1, 231.
123. P.A. Barnes, G. Stevenson and S.B. Warrington, Proceedings of the 2nd European
Symposium on Thermal Analysis, Heyden, London, 1981, 47.
124. T. Arii, T. Senda and N. Fuji, Thermochimica Acta, 1995, 267, 209.
125. P.A. Barnes and G. Stephenson, Analytical Proceedings, 1981, 12, 538.
126. T. Okino, Netsu Sokutei no Shinpo, 1987, 5, 63.
127. J. Chiu, Analytical Calorimetry, 1984, 5, 197.
128. F.G. Buttler, A. Giles, F. Harrison and S.R. Morgan, Journal of Thermal Analysis,
1976, 1, 13.
129. H.L. Chung and J.C. Aldridge, Analytical Instrumentation (NY), 1992, 20, 2-3,
123.
130. A. Bouwknegt, J. de Kok and J.A.W. de Kock, Thermochimica Acta, 1974, 9,

399.
131. ASTM E967-97, Standard Practice for Temperature Calibration of Differential

Scanning Calorimeters and Differential Thermal Analysers, 1997.
132. ASTM E968-99, Standard Practice for Heat Flow Calibration of Differential
Scanning Calorimeters, 1999.
44
133. H.G. McAdie, Analytical Chemistry, 1967, 39, 543;
134. H.G. McAdie, Analytical Chemistry, 1972, 44, 640.
135. H.G. McAdie, Analytical Chemistry, 1974, 46, 1146.
136 S. Affolter and M. Schmid in Developments in Polymer Analysis and

Characterisation, Rapra Technology Ltd., Shawbury, UK, 1999, Paper 5.
137. B. Courtault, Analusis, 1979, 7, 481.
138. A.K. Sircar, Rubber Chemistry and Technology, 1992, 65, 3, 503.
139. ASTM D297-93, Standard Test Methods for Rubber Products – Chemical
Analysis, 1998.
140. ASTM E1131-98, Standard Test Methods for Compositional Analysis by

Thermogravimetry, 1998.
141. D.W. Brazier, Rubber Chemistry and Technology, 1980, 53, 3, 437.
142. M. Wingfield in Fachtagung. Kopplungen der Instrumentellen Analytik (TA, IR,

MS) für die Kunststoff- und Kautschukindustrie, Ed., H. Möhler, Süddeutsches
Kunststoff Zentrum, Würzburg, 1991.
143. R.G. Beimer, ACS Division of Organic Coatings and Plastics Chemistry,
Preprints, 1975, 35, 428.
144. G.J. Mol, Thermochimica Acta, 1974, 10, 259.
145. M. Simpson, P.M. Jacobs and F.R. Jones, Composites, 1991, 22, 2, 105.
146. A. Kettrup and K.H. Ohrbach, Proceedings of the 8th International Thermal
Analysis Conference, Alfa, Bratislava, 1985, 2, 629.
147. G.A. Kleineberg, D.L. Geiger and W.T. Gormley, Makromolekulare Chemie,
1974, 175, 2, 483.
148. G.A. Kleineberg and D.L. Geiger, Proceedings of the 3rd International Confernce
on Thermal Analysis, Davos, Switzerland, 1971, 325.
149. M. Negri and F. Alarcon - Lorca, Revue Generale des Caoutchoucs et Plastiques,
1994, 61, 637/8, 55.
45
150. K. Liu, E. Jakab, W. Zmierczak, J. Shabtai and H.L.C. Meuzelaar, ACS Division
of Fuel Chemistry, Preprints, 1994, 39, 2, 576.
151. A. Alders, K.G.H. Raemaekers and J.C.J. Bart, Proceedings of the SPE ANTEC
’99, New York City, NY, USA, 1999, Volume 3, p.3761.
152. D.J. Johnson and D.A.C. Compton, American Laboratory, 1991, 23, 1, 37.
153. S.M. Dyszel, Analaytical Calorimetry, 1984, 5, 277.
154. J. Morelli, Journal of Analytical Applied Pyrolysis, 1990, 18, 1.
155. M. Maciejewski and A. Baiker, Thermochimica Acta, 1997, 295, 1-2, 95.
156. M. Maciejewski, C.A. Müller, R. Tschan, W.D. Emmerich and A. Baiker,

Thermochimica Acta, 1997, 295, 167.
157. R. Schönherr, TGA-FTIR Atlas Elastomere, Verlag W.K. Schönherr, Burgdorf,

Germany, 1996.
158. D.D. Jiang, Q. Yao, M.A. McKinney and C.A. Wilkie, Polymer Degradation and
Stability, 1999, 63, 423.
159. B.J. McGrattan, Applied Spectroscopy, 1994, 48, 12, 1472.
160. R.G. Davidson, Mikrochimica Acta, 1988, 1, 301.
161. J.W. Washall and T.P. Wampler, Spectroscopy, 1991, 6, 38.
162. B. Bowley, E.J. Hutchinson, P. Gu, M. Zhang, W.P. Pan and C. Nguyen,
Thermochimica Acta 1992, 20, 209.
163. E.G. Jones, D.L. Pedrick and I.J. Goldfarb, Polymer Engineering and Science,
1988, 28, 1046.
164. C. Chang and J.R. Tackett, Thermochimica Acta, 1991, 192, 181.
165. H.K. Yuen, G.W.Mappes and W.A. Grote, Thermochimica Acta, 1982, 52, 143.
166. K-H. Ohrbach and A. Kettrup, Polymer Degradation and Stability, 1985, 13, 99.
167. W. Holzapfel, in Fachtagung. Kopplungen der Instrumentellen Analytik (TA, IR,

MS) für die Kunststoff- und Kautschukindustrie, Ed., H. Möhler, Süddeutsches
Kunststoff Zentrum, Würzburg, 1991.
46
168. DIN 51005, Thermal Analysis Terms, 1993.
169. J. Chiu and C.S. McLaren, Thermochimica Acta, 1986, 101, 231.
170. J. Chiu in Analytical Calorimetry, Eds., J.F. Johnson and P.S. Gill, Plenum, New
York, NY, USA, 1989, 5, 197.
171. A.K. Sircar, Journal of Scientific and Industrial Research, 1982, 41, 536.
172. J. Lippincot, Rubber World Blue Book: Materials, Compounding Ingredients,

and Machinery for Rubber (Annual), Bell Communications, New York, NY, USA,
1993.
173. D.A. Powell, Journal of Scientific Instruments, 1957, 34, 225.
174. G.R. Cotten and L.J. Murphy, Kautschuk und Gummi Kunststoffe, 1988, 41, 1,
54.
175. W.B. Gorman, inventor; Bridgestone/Firestone Inc., assignee; US 5,191,211,

1993.
176. A. Tas, F. Wülfert, K.G.H. Raemaekers and J.C.J. Bart, Unpublished Results,
2001.
177. S.J. Swarin and A.M. Wims, Rubber Chemistry and Technology, 1974, 5, 47,
1193.
47
48
2
Photoacoustic Fourier Transform Infrared
Spectroscopy of Rubbers and Related Materials
James R. Parker
2.1 Introduction
Photoacoustic Fourier transform infrared spectroscopy (PA-FTIR) differs from most

infrared techniques in that it is an emission rather than an absorption technique. It is a
highly versatile technique requiring little sample preparation other than sizing the sample.
It can be used for both qualitative and quantitative purposes and for surface and depth
profiling analyses. It has an interesting history with the first crude photoacoustic
spectrometer being built in the 1880s by Alexander Graham Bell, the inventor of the
telephone. In this chapter the basic principles and typical uses of what is known as rapid
scan PA-FTIR are presented. The analyses are those that can be carried out with any
mid-level grade Fourier transform-infrared (FTIR) spectrometer and a photoacoustic
detector.1
2.2 History of Photoacoustic Spectroscopy
The photoacoustic effect was discovered by Alexander Graham Bell in 1880 [1]. It led to
the development in 1881 of what Bell termed the spectrophone [2]. It had been found
that solids, liquids, and gases when irradiated with modulated sunlight produced sounds
of varying intensities depending upon the material. Bell realised that there was a
relationship between the radiation absorbed and the sounds produced and that the
technique would be useful in studying absorption spectra in the infrared. There appear
to have been several versions of the spectrophone and the version described in a biography
is a spectrophotometer utilising a prism to isolate wavelengths [3]. Others confirmed
Bell’s results and postulated various theories to explain the results [4-8]. At this point it
was clear that the photoacoustic effect was most pronounced with gases. It was also well
understood that the sound produced was the result of pressure/volume changes of the
1
‘Permission for the publication herein of Sadtler Spectra has been granted, and all rights reserved, by
BIO-RAD Laboratories, Sadtler Division.’
49
gas. However, the mechanism involving solids and liquids was not well understood.
Practical instrumentation for the determination of gases began to be developed in the
late 1930s and early 1940s by Veingerov, Pfund, and Luft [9-12]. This instrumentation
using typical infrared sources and microphone detectors developed the capability of
determining parts per million of gases. Such gas detectors are still available. If lasers are
used as sources than the detection limits can be pushed to the parts per billion level
[13,14]. Satisfactory results with solids awaited the development of nonresonant
photoacoustic cells by Rosencwaig and Gersho [15-17]. It awaited the further development
and utilisation of FTIR spectrometers. Such work began to be reported in the late 1970s
[18-21]. The theoretical basis for the photoacoustic effect in condensed phases was begun
by Parker [22]. A general theory was developed by Rosencwaig and Gersho (the RG
theory) and expanded upon by others [23-27].
2.3 Theory of Photoacoustic Spectroscopy
In photoacoustic FTIR spectroscopy the sample in a sealed chamber (as illustrated in

Figure 2.1) absorbs modulated infrared radiation.
The excited molecules give up that excess energy as heat through collisional processes.
The heat diffuses to the surface of the sample where it creates a pressure wave (an acoustic
Figure 2.1 Schematic of photoacoustic (PA) detector
50
Photoacoustic Fouriers Transform Infrared Spectroscopy of Rubbers and Related Materials
wave) which is then detected by a microphone. It is readily seen that any theory will have
to be fairly complicated and involve a number of optical and thermal parameters. The
RG theory uses the following parameters:
D the thickness of the sample (cm)
Do the optical absorption length is the distance the infrared radiation can travel through
the sample (cm), equal to 1/β where β is the optical absorption coefficient
Dt the thermal diffusion length is the distance the thermal wave can travel through the
sample (cm)
The book by Rosencwaig can be consulted for a thorough discussion of the theory [27].
To simplify the theory, samples are categorised depending on the relative magnitude of
the following parameters:
Optically transparent Do > D, the optical absorption coefficient is small
Optically opaque Do < D, the optical absorption coefficient is large
Thermally thin Dt > D, thermal conductivity is large
Thermally thick Dt < D, thermal conductivity is small
The typical polymer or rubber sample would be classified as optically transparent or

opaque and thermally thick except possibly for the strongest bands. In this case the
signal intensity would be proportional to the product of the optical absorption coefficient
(β) and the thermal diffusion length and show a – 3/2 dependence on the modulation
frequency (ω). The angular modulation frequency is a product of the interferometer
mirror velocity and the wavenumber:
ω = 2π (wavenumber, cm-1) (OPD velocity, cm/s) (2.1)
where OPD is the optical path difference
= 4π (wavenumber, cm-1)(mirror velocity, cm/s) (2.2)
for a Michelson interferometer
Sometimes the mirror velocity must be calculated from what the instrument manufacturer
specifies as the He-Ne laser modulation frequency.
Mirror velocity, cm/s = [(laser modulation frequency, KHz )(1000)] ÷ [(2)(15798)] (2.3)
51
The thermal diffusion depth is also related to the modulation frequency:
Dt = (2α/ω)1/2 (2.4)
where α is the thermal diffusivity (cm2/s) of the sample.
The thermal diffusivity is calculated from other thermal parameters and the density.
α(cm2/s) = (thermal conductivy,cal/cm-s-°C) ÷ [(density, g/cm3) (specific heat, cal/g-°C)]

(2.5)
Since only the radiation absorbed within the thermal diffusion depth gives rise to the
signal, the depth being sampled can be varied by changing the modulation frequency.
The sample spectrum is ratioed against that of carbon black or a polymer or rubber
sample highly filled with carbon black. This type of sample is classified as optically
opaque and thermally thin. Here the signal intensity is independent of the optical
absorption coefficient, photoacoustic saturation has taken place, and shows a –1 power
dependence on the modulation frequency. This single beam spectrum resembles the one
obtained when a background spectrum is obtained using a deuterated triglycine sulfate
(DTGS) detector, i.e., a blackbody emission curve. When the sample spectrum is ratioed
against the carbon black reference spectrum the resulting spectrum has a signal intensity
with a –1/2 power dependence on the modulation frequency (-3/2 ÷ -1 = -1/2). Chalmers
and co-workers, used the spectrum of polyvinyl chloride to test this dependence and
found good agreement for all the bands tested (963, 1100, 1331, and 1431 cm-1) except
for the strongest band (1250 cm-1) where saturation may have occurred [28]. The signal
increased with decreasing wavenumber in the manner predicted.
2.4 Instrumentation for PA-FTIR Analyses
A wide variety of FTIR instruments have been used to obtain PA-FTIR spectra. A cursory
examination of the literature reveals some of the instruments used: JEOL JIR-5500; IBM
IR-95, IR-98, 9195; Bruker IFS 66, IFS 88, IFS 113V; Perkin Elmer 1750, 1760-X, 1800;
Bomen DA 3.02; Nicolet 20 SX, 20 DXB, SX-170, 7199, 740, 800; Mattson Cygnus
100, Sirius 100, Polaris; Digilab FTS-10, FTS-10M, FTS-11, FTS-15, FTS-20, FTS-20E,
FTS-60, FTS-65, FTS-6000; Laser Precision Analytical (Analect) RFX-75 and Analect
FX-260. The Laser Precision Analytical Instrument required extensive modification. The
list is only intended to indicate the wide variety of different instruments which have been
used and may not be complete. The normal adjustments for maximum performance
should be made, beam splitter alignment, etc. The source aperture, if any, should be
opened fully for maximum source intensity. The gain control and other adjustments for
best operation at each of the different mirror velocities to be used should be made. The
52
instrument should be set on a solid table to minimise vibration and purged with dry air
or nitrogen. Prior to use of the PA detector the bench should be initialised with the
DTGS detector to establish that it is operating normally. Some of the commercial
photoacoustic detectors listed in the literature are: Ames; EDT OAS 401; LPA FX-400;
Digilab; PAR 6003; Gilford; Nicolet; MTEC 100, 200, and 300. MTEC appears to be
the only commercial producer at the moment. It is not clear that the results from all these
detectors are equally reliable and comparable although reviewers tend to treat them as
such. Figure 2.1 shows a general schematic of a photoacoustic detector. The MTEC
detector has a BNC connector (a metal device used to connect a coaxial cable to something
else) which is usually interfaced with the external detector port of the FTIR. A second
cable is connected to the MTEC power supply which is used to power the photoacoustic
detector and to remotely control the gain of the preamplifier in the detector. It is this
gain control which is adjusted when obtaining spectra of samples for the appropriate
signal level. Usually the detector is purged with helium because of its good thermal
properties and the instrument is operated with a low mirror velocity, e.g., 0.1 cm/s, and
a resolution of 8 cm-1 over a range of 450-4000 cm-1 with 32 scans. After the detector
has been aligned and is ready for sample data collection, a background spectrum of the
carbon filled polymer standard supplied by MTEC is obtained. Alternatively carbon
black powder, pelletised carbon black, carbon filled rubber, pyrolytic graphite, and other
totally infrared absorbing materials have been used. A typical background spectrum is
shown in Figure 2.2 with and without the use of a helium purge.
Figure 2.2 Single beam PA-FTIR spectra of MTEC carbon-filled polymer
53
Helium gave a 2.3 X improvement in sensitivity which is similar to the 2.5 X factor
found by Wong [29]. Early workers attributed the high sensitivity obtained with helium
to be due to its high thermal conductivity and low heat capacity [29, 30]. The other
purge gas which has sometimes been used is xenon, a highly polarisable gas. Its use
enhances the intensities of groups perpendicular to the surface and has been useful in
studying the orientation of hydroxyl groups and silane coupling agents on silica surfaces
[31, 32]. The sample can be of any shape or form as long as it fits into the sample holder
which is 10 mm in diameter. The thickness can be up to 6 mm. The standard large
sample cup is 3 mm in depth. There may be some optimum distance between the sample
and the top of the cup for optimum generation of a signal in the gas layer. Carter and
Wright tested the hypothesis that the optimum distance would be the thickness of the
gaseous boundary layer (BLg) [33]. It is the expansion and contraction of this layer that
generates an acoustic signal which is detected by a microphone. The thickness of the
boundary layer is equal to 2πDg where Dg is the thermal diffusion length of the gas.
BLg = 2πDg = (2α) ÷ [(4π)(wavenumber, cm-1 )(mirror velocity, cm/s)] (2.6)
In this equation, α is the thermal diffusivity of the gas which is 1.51 cm2/s for helium. Therefore,
at a mirror velocity of 0.08 cm/s and a wavenumber of 500 cm-1, the thermal diffusion length
is 0.78 mm and the boundary layer thickness is 4.9 mm. Carter and Wright concluded that
the optimum distance was somewhat less than that of the boundary layer thickness and that
other factors were involved. McClelland and co-workers, have suggested that a distance of
1mm or the thermal diffusion length of the gas be used [34]. The sample used for analysis can
be in a wide variety of forms. These can include irregular shaped solids, pellets, powders,
films, fibres, or liquids and need not be transparent. The primary restrictions be that the
sample fits in the sample holder and that it not degas since the spectra of gases are stronger
than those of solids. For samples containing moisture this might require oven drying, using a
desiccant beneath the sample in the sample holder, or simply allowing the sample to equilibrate
in the instrument atmosphere prior to insertion in the photoacoustic detector. It does not
mean that hydrates cannot be analysed. Powders give spectra similar to those of obtained for
solid samples as shown for polystyrene in Figure 2.3.
In general increasing surface area increases band intensity. For example, the C-H band
intensity of hydrogenated diamond powders correlated with the Brunauer, Emmet and
Teller (BET) surface area over a range of 0.2 to 22 m2/g [35]. For some types of powders
other factors also appear to be involved with signal enhancement. Yang and Fateley
reported that some type of extra signal enhancement occurred for quartz powder when
the particle diameter became smaller than the thermal diffusion length for a given peak
[36]. A more general theory is that in addition to the thermal signal predicted by the RG
theory there is a pressure signal caused by the interstitial gas which depends upon the
porosity of the powder [37, 38]. Pandurangi and Seehra showed that the change in signal
54
Figure 2.3 Effect of sample shape on PA-FTIR spectrum of polystyrene
intensity of silica powders of varying porosity (ε) has an (ε)/(1-ε) dependence for strong
bands and an e dependence for weaker bands in accordance with this theory [39]. Hövel,
Grosse, and Theiss using a somewhat different theoretical approach also concluded that
the expansion of the interstitial gas in porous samples is responsible for a significant
portion of the photoacoustic signal [40]. Perhaps a final consideration when obtaining
of PA-FTIR spectra is the FTIR computer software. Most IR experiments involve
measuring transmitted radiation while the photoacoustic spectroscopy (PAS) experiment
involves measuring emitted radiation. On some newer instruments the software may
already be set up to take this into account. On others a bit of trickery is involved. The
Mattson Cygnus 25 (Rev. 8) with First DOS based software is an example of one of the
latter. When obtaining a PA-FTIR spectrum the instrument parameters in the software
are set up to obtain a transmittance (.ras file) spectrum. The resulting sample spectrum is
obtained and ratioed against the carbon black or other reference spectrum. The resulting
.ras file must now be converted to an absorbance file (.abs) not through spectral conversion
routines but by changing the data file extension to .abs. In First software this requires
writing a macro or batch file routine for this purpose. The Y-axis may then be relabelled
relative intensity and the scale may be changed if that aids in the utilisation of some
desired software routine (library searching, etc). In any case the Y-axis is treated as being
proportional to concentration. Typical PAS spectra for natural rubber (NR) and silica
are shown in Figures 2.4 and 2.5.
55
Figure 2.4 PA-FTIR spectra of natural rubber (NR)

Upper spectrum reproduced with permission from Bio-Rad Laboratories, Sadtler Division
Figure 2.5 PA-FTIR spectra of amorphous silica

Upper spectrum reproduced with permission from Bio-Rad Laboratories, Sadtler Division
56
In the spectrum of natural rubber, the bands of particular importance are the C=C
stretching band at 16663 cm-1, the C-H bending band of the methyl group at 1378 cm-1,
and the C-H out-of-plane bending band from the cis unsaturation configuration at 837 cm-1.
The spectrum of silica shows a broad band at 1100 cm-1 from the Si-O-Si antisymmetric
stretching vibration and a band at 803 cm-1 from the Si-O-Si symmetric stretching
vibration. There is a very broad band centred around 3376 cm-1 which represents the O-H
stretching band of bound and unbound water and the Si-OH stretching band of silanols
involved in various degrees of hydrogen bonding to water and other silanol groups. The
distinct shoulder band at 3735 cm-1 represents the silanol groups which are least hydrogen
bonded. There is a small Si-O-Si combination band at 1872 cm-1 and a water O-H bending
bond at 1636 cm-1 which is on top of some type of Si-O overtone or combination band.
The band at 945 cm-1 decreases in size as drying takes place and the major part of it is
believed to represent the hydrogen bonding of water to silanols.
These spectra were converted from .ras to .abs files with no change in the Y-axis scale
and then used to search Sadtler libraries of commercial material using Mattson First
software with a correlation coefficient search algorithm which takes into account baseline
drift and differences in scaling. As can be seen the sample spectrum and the library
spectrum are usually quite similar. The library spectra shown here were the best hits in
each of the searches with a search correlation coefficient of >0.9 with 1.0 representing a
perfect match.
2.5 Analysis of Carbon-Filled Rubbers
Carbon-filled rubber samples are quite common. Many different methods for the IR
analysis of these materials have been used including use of microtomed sections, attenuated
total reflectance (ATR), specular reflectance, potassium bromide pellets, solubilisation
with o-dichlorobenzene, and pyrolysis. None of these methods is totally satisfactory and
there is a constant search for others. Problems arise because carbon black is such a
strong absorber of IR radiation, scatters radiation, raises the refractive index of the
sample, and is used in high enough concentrations to simply reduce the amount of sample
available for examination. Photoacoustic spectroscopy is another method which has been
examined since it is relatively simple to use and has a wider wavelength range than most
ATR methods and some MCT detectors used in conjunction with other methods. Teramae
examined nitrile-butadiene rubbers (NBR) containing up to 29 wt.% carbon black [41,
42] and Carter and co-workers, examined natural rubbers (NR) containing up to 25 wt.%
carbon black [43]. Waddell and Parker give an example where the presence of NR can be
detected in a sample containing 38.4 wt.% carbon black [44]. Figure 2.6 shows examples
of PA-FTIR spectra of styrene-butadiene rubbers (SBR) with 0, 10, 20, and 30 wt.%
carbon black.
57
Figure 2.6 Effect of carbon black on PA-FTIR spectrum of styrene-butadiene rubber

(SBR)
The maximum concentration of carbon black that can be tolerated has not been
established. Several ways of improving detectability have been proposed. The simplest
of these is the one suggested by McClelland and involves obtaining a single beam spectrum
of the sample and creating a background spectrum through the use of smoothing routines
[45]. He reports that the resulting sample/background ratioed spectrum looks something
like a derivative spectrum. Another possible way of creating a background spectrum is
to calculate one from the sample spectrum as suggested by Coates [46]. He has illustrated
this technique by using it to correct the severe curvature of a transmission spectrum of a
carbon-filled ethylene-vinyl acetate (EVA) copolymer. Finally, there is the technique of
linearising photoacoustic spectra. This involves recording photoacoustic spectra at two
different mirror velocities and requires the use of some special software which makes use
of phase information. Carter has used this technique to show the improvements available
in spectra of formulated NR containing up to 25 wt.% carbon black [47].
PAS-FTIR spectra have been used to find out the interaction of chlorosulphonated
polyethylene (CSM) and carbon black N110 [48]. A number of bands in the 1800 cm-1-
1680 cm-1 region in the spectrum of N110 (Figure 2.7) confirm the presence of different
carbonyl functionalities, which may include carboxyl group, lactone and quinone. The
band at 1651 cm-1 is characteristic of aromatic double bonds in the carbon black. The
58
Figure 2.7 FTIR-photoacoustic spectra of pure CSM, b) N110 carbon black, c) CSM/
N110 carbon black compound
59
spectrum of the carbon black filled CSM compound retains all the characteristics bands
due to the carbon black and CSM. However, there occur several additional bonds in the
1100-1000 cm-1 region, which indicate the formation of an ether. The –OH on carbon
black reacts with –SO2Cl group of the CSM, which on heating eliminates SO2, forming an
ether group, C-O-C, indicating polymer filler interaction, according to following equation:
(2.7)
2.6 Quantitative Analysis of Polymers
There is not an extensive body of literature on the quantitative PA-FTIR analysis of

rubbers and polymers. Early in the development of the technique there were indications
that it might be possible. Many of the early studies involved the use of vinyl acetate
copolymers. Vinyl chloride-vinyl acetate copolymers were briefly examined by Chalmers
and co-workers, in 1981 by PA-FTIR with indications that quantitation might be possible
[28]. They were also studied by Kirkbright and Menon in 1982 using near-IR
photoacoustic spectroscopy [49]. In this case the intensity of the second overtone of the
carbonyl band (4651 cm-1) was ratioed to the intensity of the first overtone of the C-H
stretching vibration (5714 cm-1) and correlated with the vinyl acetate content. A
chemometric PA-FTIR approach was used in 1991 by McClelland and co-workers to
determine the percentage vinyl acetate in EVA over a range of 10%-50% [50]. They used
three spectral ranges to represent compositional variations:
(1) 2750-3120 cm-1, C-H stretching
(2) 1630-1900 cm-1, C=O stretching and
(3) 578-1490 cm-1, molecular vibration fingerprint region.
In addition the band area in the 2750-3120 cm-1 region was used to correct for
experimental variabilities. Cross-correlation analysis gave standard errors of prediction
of 0.64%. In 1996 Parker and Waddell used two approaches for the PA-FTIR
determination of vinyl acetate in EVA [51]. The intensity ratio of the 610 cm-1 acetate
group bending band to the 722 cm-1 methylene group in-phase rocking vibration was
used to determine 10%-50% vinyl acetate. For levels below 10% (3%-9%), the intensity
of the carbonyl stretching band at 1740 cm-1, was ratioed to the intensity of the methylene
bending vibrational band at 1466 cm-1. Least squares linear regression analyses were
used in each case with correlation coefficients of 0.99 and 0.97 being obtained, respectively.
60
% Vinyl acetate = 38.8 x (610/722 Band Ratio) + 3.42 (2.8)
correlation coefficient = 0.992
% Vinyl acetate = 9.25 x (1740/1466 Band Ratio) – 3.07 (2.9)
An example of the PA-FTIR analysis of a rubber is the determination of acrylonitrile

(ACN) in NBR. A PA-FTIR spectrum of a NBR is shown in Figure 2.8.
Parker and Waddell investigated three methods [51]. These involved determining the
%ACN as a function of:
(1) the intensity ratio of the 2238 cm-1 nitrile stretching band to the 1443 cm-1 methylene
deformation or scissoring band
(2) the intensity ratio of the 2238 cm-1 nitrile stretching band to the 969 cm-1 C-H out-
of-plane bending band of the trans-butadiene double bond and
(3) the absolute intensity of the 2238 cm-1 nitrile stretching band.
Figure 2.8 PA-FTIR spectrum of nitrile-butadiene rubber (NBR)

Reprinted with label modifications from reference [49] with permission from Technomic
Publishing Co., Inc., © Copyright 1999
61
Least squares linear regression analyses of the data over the range of 20%-50% ACN
gave the following equations:
%ACN = 60.5 x (2238/1443 Band Ratio) – 0.271 (2.10)
%ACN = 79.74 x (2238/969 Band Ratio) + 5.08 (2.11)
%ACN = 4.84 x (2238 Band Intensity) + 1.99 (2.12)
Another example of the PA-FTIR analysis of a rubber is the determination of % styrene

in SBR. A typical PA-FTIR spectrum of an emulsion SBR is shown in Figure 2.9.
Traditional non-PA methods might involve measuring the 700 cm-1 styrene band in carbon
disulfide solutions as carried out by Binder [52], transmission measurements on films
Figure 2.9 PA-FTIR spectrum of emulsion styrene-butadiene rubber (SBR)

62
deposited from solution as done by Panyszak and Kovar where the 970 cm-1 butadiene
band was ratioed to the 755 cm-1 styrene band [53], or an ATR method like that of Sloan
and Clements where the ratio of the 698 cm-1 styrene band to the 967 cm-1 trans-1,4-
butadiene band was used [54]. Parker and Waddell studied the PA-FTIR analyses of cold
process emulsion SBR [51]. The styrene content was determined by ratioing the intensity
of the 698 cm-1 styrene band to the 968 cm-1 trans-1,4-butadiene band as a function of
% styrene. The results can be expressed by a least squares linear regression equation
over the range of 5%-56% styrene in SBR.
% Styrene = 53.6 x (698/968 Band Ratio) – 13.54 (2.13)
An alternative method was also studied. This involved ratioing the intensity of the 698 cm-1
styrene band to the intensity of the 2921 cm-1 C-H stretching vibration. Since oils and
other additives would interfere with this approach they were extracted with acetone.
Vacuum oven drying was then necessary to remove all traces of acetone prior to PA
analysis. Otherwise the PA spectrum would be that of acetone rather than that of the
rubber since the gas phase spectrum of the acetone would overwhelm that of the solid
phase rubber. This technique allowed both solution and emulsion SBR to be analysed by
a common method. The results can be expressed by a least squares linear regression
equation over the range of 5%-40% styrene in SBR.
% Styrene = 53.0 x (698/2921 Band Ratio) – 4.52 (2.14)
Another example of rubber analysis is the determination of vinyl-butadiene in SBR. A

PA-FTIR spectrum of a high-vinyl content SBR is shown in Figure 2.10. This determination
by mid-IR methods invariably involves the use of the 910 cm-1 band due to the wagging
motion of the pendent vinyl group of 1,2-butadiene.
These methods have been reviewed by Silas, Yates and Thornton [55]. In a PA-FTIR
method Parker and Waddell also used the intensity of the 910 cm-1 butadiene band and
ratioed it to the intensity of the 1450 cm-1 C-H bending band as a function of the %
vinyl-butadiene [51]. The results can be expressed by a least squares linear regression
equation over a range of 10%-60% vinyl-butadiene.
% Vinyl-butadiene = 28.4 x (910/1450 Band Ratio) – 8.16 (2.15)
63
Figure 2.10 PA-FTIR spectrum of high vinyl styrene-butadiene rubber (SBR)

A final quantitative example is that of determining the % ethylene in ethylene-propylene

copolymers (EPM). The International Institute of Synthetic Rubber Producers’ (IIRP)
Technical/Operating Committee and the American Society for Testing and Materials
(ASTM) cooperated to produce an updated standard method [56]. They also produced a
set of standards with the ethylene content established by 13C NMR. Those standards
were utilised by Parker and Waddell to study the photoacoustic determination of %
ethylene [57]. Three of the four infrared band ratios specified by ASTM were measured,
1378/1462, 1378/722, and 1156/722. The fourth one, 1156/4255, was not used since it
involves measurements in the near-IR region. Baselines were drawn by the valley-to-
valley technique as illustrated in Figure 2.11.
The 1378 cm-1 band is from the CH3 symmetric bending and the 1156 cm-1 band is a
complex skeletal vibration involving the CH3 branch of propylene. The 722 cm-1 band
represents the CH2 rock and the 1462 cm-1 band is a combination of the CH2 scissor and
the asymmetric CH3 bend. In the photoacoustic spectra the 1378 and 1462 bands are
strong while the 1154 and 722 cm-1 bands are weak. Least squares linear regression
64
Figure 2.11 Baseline constructions for PA-FTIR analysis of ethylene-propylene

copolymers (EPM)
analysis of the three band ratios versus % ethylene shows the best correlation coefficient
(0.9926) for the 1378/1462 band ratio, the next best (0.9670) for the 1378/722 band
ratio, and the worst (0.9566) for the 1156/722 band ratio. These are as would be expected
from the relative intensities of the bands involved. The corresponding equations for
these correlations are as follows:
% Ethylene = (121.35) – (100.22)(1378/1462 Band Ratio) (2.16)
Quantitation of PA-FTIR data appears to be generally applicable in the analysis of rubbers

and polymers. The chances of success are best if the strongest, possibly saturated, infrared
bands are avoided. The quality of the results are improved if band ratioing is used to
compensate for experimental variables.
65
2.7 Surface Analysis and Depth Profiling
A discussion of surface analysis requires a review of the depth being sampled during PA-
FTIR spectroscopy. The depth being sampled during PA-FTIR analyses of rubbers is the
thermal diffusion depth (Dt). This is a function of the thermal diffusivity of the sample,
the wavenumber, and the mirror velocity.
Dt = (2α/ω)1/2 where ω = 4π(wavenumber)(mirror velocity) (2.19)
Dt = [(2α, cm2/sec)] ÷ [(4π)(wavenumber, cm-1)(mirror velocity, cm/sec)]
For illustrative purposes a thermal diffusivity of 1.3 x 10-3 cm2/s is often used as being
typical of rubber and polymers. Some values from the literature for various materials are
given in Table 2.1. Using the value of 1.3 x 10-3 it can be calculated that a depth of 3 to
11 μm is being sampled at 2000 cm-1 as indicated in Table 2.2. This is an order of
magnitude greater than that sampled by ATR techniques.
One type of sample to be considered is one which has a bloom or similar surface layer.
This was an area first explored by Carter and co-workers [60]. They reported finding a
bloom of dimorpholinyl thione and zinc stearate formed from ingredients used during
the vulcanisation of the NR sample. As the mirror velocity was increased from 0.316 cm/s to
1.216 cm/s the bands associated with the bloom increased. Figure 2.12 shows the PA-
FTIR spectrum of a zinc stearate bloom on an amber coloured silica-filled SBR.
Table 2.1 Thermal diffusivities of rubbers and polymers

Thermal Diffusivity
Rubber/Polymer Reference
(cm2/s x 10-3)
Polyethylene (PE), high density 2.3 [58]
Polyethylene, low density 1.4 [58]
Polypropylene (PP) 1.2 [58]
Polyethylene terephthalate 1.1 [58]
Polystyrene 1.1 [58]
Polytetrafluoroethylene 0.95 [58]
SBR 1.23 [59]
BR 1.0 [59]
Polyisoprene 0.86 [59]
Polychloroprene (Neoprene) 0.68 [59]
66
Table 2.2 Thermal diffusion depth as a function of mirror velocity (for

thermal diffusivity = 1.3 x 10-3)
Mirror Velocity LMF Thermal Diffusion Depth (μm)
(cm/s) (kHz) 4000 cm-1 2000 cm-1 500 cm-1
0.084 2.7 7.8 11.1 22.2
0.115 3.6 6.7 9.5 19.0
0.180 5.7 5.4 7.6 15.2
0.316 10 4.0 5.7 11.4
0.633 20 2.9 4. 0 8.4
1.264 40 2. 0 2.9 5.7
LMF: laser modulation frequency
Figure 2.12 PA-FTIR spectrum of bloom on rubber
67
The 1539 cm-1 zinc stearate band can be ratioed to the 966 cm-1 SBR band and plotted as
a function of mirror velocity (depth) to show that it is increasing in concentration as the
surface is approached. This is shown in Figure 2.13.
Another type of sample for surface analysis is one where some kind of reaction has taken
place on the surface. A common example of this is the appearance of carbonyl bands from
the surface oxidation of PE. A spectrum of ozone treated PE film is shown in Figure 2.14.
The intensity of the 1738 cm-1 carbonyl band can be ratioed against the 1466 cm-1 PE
band and plotted as a function of mirror velocity (depth) to see that the oxidation decreases
with depth (Figure 2.15).
Delprat and Gardette studied the photo-oxidation of PP with and without titanium oxide
being present [61]. The intensity of the carbonyl band at 1713 cm-1 was followed using
a mirror velocity of 0.16 cm/s. As they point out roughened surfaces do not lend themselves
well to ATR techniques and microtoming may damage surface layers. Gonon, Vasseur,
and Gardette studied the photo-oxidation of styrene-isoprene copolymers by monitoring
the ratio of the 1718 cm-1 carbonyl band to the 1600 cm-1 aromatic ring band. Mirror
velocities from 0.0158 to 0.6329 cm/s were used [62]. A somewhat different type of
Figure 2.13 Zinc stearate bloom concentration as a function of depth
68
Figure 2.14 PA-FTIR spectrum of ozone treated polyethylene
Figure 2.15 Polyethylene oxidation as a function of depth
69
profiling was carried out by de Oliveira, Pessoa, Vargas and Galembeck [63]. They used
near-infrared photoacoustic spectroscopy to determine the depth profile of –CH3, =CH2,
and –OH groups in low density polyethylene over the 11-56 μm range [63]. They
concluded that the surface is richer in –CH3 and =CH2 groups than the bulk and play a
role in poor adhesion to polyethylene surfaces.
Laminates have been extensively studied by PA-FTIR and are the subject of extensive
investigations by the new much more complex step-scan PA-FTIR techniques. Kapton
film is often used to illustrate the analysis of this type of sample. Kapton 200FN919 film
is a DuPont product which consists of a 25 μm layer of polyimide laminated between
12.7 μm layers of Teflon. If a high mirror velocity is used then a spectrum of only the
Teflon layer is obtained. On the other hand if a slow mirror velocity is used a composite
spectrum of the Teflon and polyimide layer is obtained. The difference spectrum (low
mirror velocity spectrum minus high mirror velocity spectrum) represents that of the
polyimide layer. As illustrated in Figure 2.16 the difference spectrum closely resembles
the PA FTIR spectrum of polyimide except for the regions where the strong C-F bands of
the Teflon are present.
Figure 2.16 PA-FTIR spectra of Kapton film
70
2.8 Determination of Orientation Function of Polymeric Materials
Photoacoustic spectroscopy with polarised light offers some potential for a new kind of
investigation. For polarisation of IR radiation a KRS-5 wire grid polariser is inserted
immediately in front of the cell. Without any special preparation polarised PAS can be
directly applied to polymer granules, blocks, strings, sheets and plates as received from
extrusion or injection moulding processing. In this way the preferred orientation given
by the machine axis or stretching direction can be correlated with the polymer structure
and processing conditions. Klaus-Jochen Eichhorn and co-workers have investigated the
orientation of polymer components in different blends of thermoplastic polyurethane
(TPU) with common and modified polypropylene (PP) after processing and in some cases
after subsequent mechanical stretching [64]. TPU and PP form an immiscible blend but
with maleic anhydride grafted PP (MSA), properties are improved. Polarised PAS is used
to study the orientational behavior of TPU in the reactive compatibilised blend. The
orientation of the polymers and their individual components in the blends has been
investigated by means of the TPU bands between 1650-1800 cm-1 (carbonyl stretching
vibrations), at 1316 cm-1 (C-N), 1115, 1081 cm-1 (C-O-C), respectively, and using the PP
band at 1379 cm-1 (methyl deformation vibration). The direction of the dipole transition
moments of the carbonyl, NH stretching and methyl deformation vibration is located in
the perpendicular direction (⊥) relative to polymer chains axis and the direction of the
others is located in the parallel direction (||), details are given in reference [64].
Hence polarised FTIR-PAS spectroscopy is suitable for detecting orientation effects in

polymers and polymer blends, during processing like extrusion and injection molding.
With polarised PAS, the difference in the orientation behavior of the hard and soft segments
in a thermoplastic poly(ether urethane) can be determined [64]. Grafting of PP with
maleic anhydride gives rise to a better orientation of the TPU hard segments in the blends.
2.9 Conclusion
PA-FTIR has been shown to be a very versatile IR technique. All the usual types of
studies can be carried out including qualitative and quantitative, surface and bulk analyses.
The spectra can be searched, subtracted, annotated and otherwise manipulated although
some software may require a change of the data file extension. The lack of necessity for
any sample preparation offers many advantages which include allowing the
uncontaminated sample to be analysed by other techniques. No grinding is required
which may cause changes in the sample. There is no mixing with other reagents as in the
preparation of potassium bromide pellets which can introduce water or as in the
preparation of Nujol mulls which causes a masking of portions of the spectrum. Further,
since no contact with the sample is required sample shape or surface roughness does not
71
affect the spectra as it does with ATR techniques. The wavelength range covered is limited
only by the FTIR instrumentation and the PA detector window used and not by the
limitations of ATR crystals. Quantitation is possible even though the theory of
photoacoustic spectroscopy is somewhat complex. The main requirement being that the
strongest bands not be used or at least checked for saturation. The quality of the results
is improved if band ratioing is used since this compensates for instrumental and sample
variations. As a surface analytical method the depth being examined can be varied by
changing the interferometer mirror velocity. The depth being examined is usually
somewhat deeper than that by ATR methods and the techniques are complementary in
that respect.
Acknowledgements
This work was carried out in support of the research efforts of PPG Industries, Inc.,
Monroeville Technical Center, Pennsylvania, USA. I wish to thank Denise Callahan and
Audrey Anderson of the library staff and Vicki Potter formerly of the secretarial staff for
their assistance and Dick Obrycki and Singh Manocha, Heads of the Analytical
Departments, and Dave McKeough, Research Director, for their support.
References
1. A.G. Bell, American Journal of Science, 1880, 20, 305.
2. A.G. Bell, Philosophical Magazine, 1881, 11, 510.
3. R.V. Bruce, Bell: Alexander Graham Bell and the Conquest of Solitude, Little,
Brown and Company, Boston, USA, 1973.
4. J. Tyndall, Proceedings of the Royal Society of London, 1881, 31, 307.
5. W.C. Röntgen, Philosophical Magazine, 1881, 11, 5, 308.
6. Lord Rayleigh, Nature, 1881, 23, 274.
7. W.H. Preece, Proceedings of the Royal Society of London, 1881, 31, 506.
8. M.E. Mercadier, Comptes Rendus Hebdomadaires des Seances de l’Academie des

Sciences, 1881, 92, 409.
9. M.L. Veingerov, Doklady Akademii Nauk SSSR, 1938, 19, 687.
72
10. M.L. Veingerov, Doklady Akademii Nauk SSSR, 1945, 46, 182.
11. A.H. Pfund, Science, 1939, 90, 326.
12. K.F. Luft, Zeitschrift für Technische Physik, 1943, 24, 5, 97.
13. L.B. Kreuzer, N.D. Kenyon and C.K.N. Patel, Science, 1972, 177, 347.
14. E.L. Kerr and J.G. Attwood, Applied Optics, 1968, 7, 5, 915.
15. A. Rosencwaig, Optics Communications, 1973, 7, 4, 305.
16. A. Rosencwaig, Science, 1973, 181, 657.
17. A. Rosencwaig and A. Gersho, Science, 1975, 190, 556.
18. M.M. Farrow, R.K. Burnham and E.M. Eyring, Applied Physics Letters, 1978,
33, 8, 735.
19. M.G. Rockley, Chemical Physics Letters, 1979, 68, 2-3, 455.
20. M.G. Rockley, Applied Spectroscopy, 1980, 34, 4, 405.
21. D.W. Vidrine, Applied Spectroscopy, 1980, 34, 3, 314.
22. J.G. Parker, Applied Optics, 1973, 12, 12, 2974.
23. A. Rosencwaig and A. Gersho, Journal of Applied Physics, 1976, 47, 1, 64.
24. H.S. Bennett and R.A. Forman, Applied Optics, 1976, 15, 10, 2405.
25. L.C. Aamodt, J.C. Murphy and J.G. Parker, Journal of Applied Physics, 1977,
48, 3, 927.
26. F.A. McDonald and G.C. Wetsel Jr., Journal of Applied Physics, 1978, 49, 4,
2313.
27. A. Rosencwaig, Photoacoustics and Photoacoustic Spectroscopy, John Wiley &

Sons, New York, NY, USA, 1980.
28. J.M. Chalmers, B.J. Stay, G.F. Kirkbright, D.E.M. Spillane and B.C. Beadle,
Analyst, 1981, 106, 1179.
29. K.Y. Wong, Journal of Applied Physics, 1978, 49, 6, 3033.
73
30. M.J. Adams, A.A. King and G.F. Kirkbright, Analyst, 1976, 101, 1199, 73.
31. M.W. Urban and J.L. Koenig, Applied Spectroscopy, 1985, 39, 6, 1051.
32. M.W. Urban and J.L. Koenig, Applied Spectroscopy, 1986, 40, 4, 513.
33. R.O. Carter III and S.L. Wright, Applied Spectroscopy, 1991, 45, 7, 1101.
34. J.F. McClelland, R.W. Jones, S. Luo and L.M. Seaverson in Practical Sampling
Techniques for Infrared Analysis, Ed., P.B. Coleman, CRC Press, Boca Raton, FL,
USA, 1993, 107.
35. T. Ando, S. Inoue, M. Ishii, M. Kamo, Y. Sato, O. Yamada and T. Nakano,

Journal of the Chemical Society, Faraday Transactions, 1993, 89, 4, 749.
36. C.Q. Yang and W.G. Fateley, Journal of Molecular Structure, 1986, 146, 25.
37. J. Monchalin, L. Bertrand, G. Rousset and F. Lepoutre, Journal of Applied

Physics, 1984, 56, 1, 190.
38. S.J. McGovern, B.S.H. Royce, J.B. Benziger, Journal of Applied Physics, 1985,
57, 5, 1710.
39. R.S. Pandurangi and M.S. Seehra, Analytical Chemistry, 1990, 62, 18, 1943.
40. H. Hövel, P. Grosse and W. Theiss, Journal of Non-Crystalline Solids, 1992, 145,
159.
41. N. Teramae, International Polymer Science and Technology, 1985, 12, 6, T/43.
42. N. Teramae, Nippon Gomu Kyokaisi, 1984, 57, 3, 141.
43. R.O. Carter III, M.C. Paputa Peck, M.A. Samus and P.C. Killgoar, Jr., Applied
Spectroscopy, 1989, 43, 8, 1350.
44. W.H. Waddell and J.R. Parker, Rubber Chemistry and Technology, 1992, 65,
836.
45. J. McClelland, MTEC Corporation, Ames, IA, USA, Private Communication,

1999.
46. J.P. Coates, Applied Spectroscopy Reviews, 1999, 34, 1-2, 121.
47. R.O. Carter III, Applied Spectroscopy, 1992, 46, 2, 219.
74
48. A. Roychoudhary and P.P. De, Journal of Applied Polymer Science, 1995, 55, 9.
49. G.F. Kirkbright and K.R. Menon, Analytica Chimica Acta, 1982, 136, 373.
50. J.F. McClelland, S. Luo, R.W. Jones and L.M. Seaverson, Proceedings of the SPIE
- International Society Optical Engineering, 8th International Conference of
Fourier Transform Spectroscopy, Lübeck, Germany, 1991, Volume 1575, 226.
51. J.R. Parker and W.H. Waddell, Journal of Elastomers and Plastics, 1996, 28, 140.
52. J.L. Binder, Analytical Chemistry, 1954, 26, 12, 1877.
53. M. Panyszak and J. Kovar, Canadian Journal of Spectroscopy, 1986, 31, 5, 130.
54. J.M. Sloan and J.P. Clements, US Army Report AMMRC TR 84-25, National
Technical Information Service (NTIS), US Department of Commerce, Springfield,
VA, USA, 1984.
55. R.S. Silas, J. Yates and V. Thornton, Analytical Chemistry, 1959, 31, 4, 529.
56. ASTM Method D3900-95 (2000) Standard Test Methods for Rubber-Raw
Determination of Ethylene Units in EPM (Ethylene-Propylene Copolymers) and
EPDM (Ethylene-Propylene-Diene Terpolymers).
57. J.R. Parker, Photoacoustic Analysis of Ethylene-Propylene (EPM) Copolymers,

Poster Presentation, ISPAC-10, 10th International Symposium on Polymer
Analysis and Characterisation, University of Toronto, Toronto, Canada, 1997.
58. E.V. Thompson in Encyclopedia of Polymer Science and Engineering, 2nd

Edition, Ed., J. Kroschwitz, John Wiley & Sons, New York, USA, 1985, Volume
16, 711.
59. I. Furuta, S-I. Kimura and M. Iwama in Polymer Handbook, 4th Edition, Ed., J.
Brandrup, E.H. Immergut, E.A. Grulke, A. Abbe and D.R. Bloch, John Wiley &
Sons, New York, NY, USA, 1999, Volume 1.
60. M.C. Paputa Peck, M.A. Samus, P.C. Kilgoar, Jr. and R.O. Carter III, Rubber
Chemistry and Technology, 1991, 64, 4, 610.
61. P. Delprat and J-L. Gardette, Polymer, 1993, 34, 5, 933.
62. L. Gonon, O.J. Vasseur and J-L. Gardette, Applied Spectroscopy, 1999, 53, 2,
157.
75
63. M.G. de Oliveira, O. Pessoa, H. Vargas and F. Galembeck, Journal of Applied

Polymer Science, 1988, 35, 7, 1791.
64. K-J. Eichhorn, I. Hopfe, P. Pötschke and P. Schimdt, Journal of Applied Polymer
Science, 2000, 75, 1194.
76
3
Infrared Spectroscopy of Rubbers
Prajna P. De
3.1 Introduction
Infrared spectroscopy is one of the oldest techniques for the molecular level
characterisation of materials and it has been extensively used to study polymers. Exactly
two hundred years ago Sir William Herschel discovered infrared radiation in 1800. It
took another hundred years of research to find a correlation between the rotational and
vibrational motion of organic molecules and infrared radiation. Coblentz [1-4] produced
a monumental research work, which was published by the Carnegi Institution of
Washington in 1905. His work encompassed instrument design, and the development of
experimental technique as well as the determination of absorption and reflection spectra
of a large number of compounds both organic and inorganic. Infrared (IR) spectrometry
was used more in the USA and Europe after World War II, thanks to the availability of
high quality spectrometers at reasonable prices, which could produce several spectra per
hour and could be operated even by non-experts. The Perkin-Elmer single beam instrument
was first on the market and it was soon followed by double beam spectrometer [5],
offered by Bechman, Hilger and Gaertner. The instrument was based on a source of
infrared radiation, collimating or mirror systems, gratings, detectors and recorders.
The absorption or emission of energy in an electromagnetic spectrum occurs in discrete

packets of photons. The relationship between the energy of a photon and the frequency
appropriate for the description of its propagation is given by the famous Bohr equation:
Δ E = hν (3.1)
where Δ E in ergs represents energy absorbed (or emitted), as a molecule passes from one
quantised state to another; ν represents frequency in cycles per second; and h is a universal
constant known as Planck’s constant (6.6256 x 10-27 erg-s). For some purposes, it is
more convenient to think of radiant energy as a continuous wave motion in which λ
represents the interval between the nodes in the wave pattern. The relationship between
wave length and frequency is given as:
77
ν = c/λ (3.2)
and
1/λ = ν (3.3)
where λ is the wave length in centimeters and c is the velocity of propagation of radiant
energy in a vacuum (speed of light in vacuum is 2.9979 x 1010 cm. s-1 ), ν is the wave
number in cm-1.
The intensity of a beam of radiation is characterised by its radiant power, which is

proportional to the number of photons per second that are propagated in the beam. The
various regions in the electromagnetic spectrum are displayed in Table 3.1, along with
the nature of changes brought about by radiation [6,7]. The infrared region extends
from 2.5 to 50 μ (4000 to 200 cm-1), the region from 12,500 to 4000 cm-1 is called near
infrared and the region from 667 to 50 cm-1 is called far infrared.
Molecules do not absorb infrared radiation when they have a completely symmetrical
charge distribution and there is no change in dipole moment during molecular vibration
with different amplitude or rotation at different rates.
There are two kinds of fundamental vibrations for molecules:
Table 3.1 Electromagnetic spectral regions useful for chemical analysis

Region Name Approximate Boundaries
Wave length Frequency, Hz
X-ray 10-2 – 10 nm 3x1019 – 3x1016
Vacuum ultraviolet 10 – 200 nm 3x1016 – 1.5x1015
Near ultraviolet 200 – 400 nm 1.5x1015 – 7.5x1014
Visible 400 – 800 nm 7.5x1014 – 3.8x1014
Near Infrared 0.8 – 2.5 μm 3.8x1014 – 1x1014
Fundamental Infrared 2.5 – 50 μm 1x1014 – 6x1012
Far Infrared 50 – 300 μm 6x1012 – 1x1012
Microwave 0.3 – 0.5 m 1x1012 – 6x108
Radiowave 0.5 – 300 m 6x108 – 1x106
78
• Stretching, in which the distance between two atoms increases or decreases but atoms
remain in the same bond axis, and the angle between a vibrating bond and a chemical
bond that is attached to one of the atoms involved in the vibration, is not altered by
stretching vibrations.
Stretching vibrations are of two types:
(i) Symmetrical stretching
(ii) Asymmetrical stretching
• Bending vibrations are nuclear motions that cause a change in the angle between two
vibrating bonds.
Bending vibrations are of four types:
(i) Scissoring
(ii) Rocking
(iii) Wagging
(iv) Twisting
Different kinds of vibrations [8,9] are shown schematically in Figure 3.1.
Figure 3.1 Vibration of a group of atoms
79
A non-linear molecule that contains n atoms has 3n-6 possible fundamental vibrational
modes that can be responsible for the absorption of IR light, whereas a linear molecule
has 3n-5 possible vibrations.
An approximate value for the stretching frequency (ν in cm-1) of a bond is related to the
masses of the two atoms (Mx and My in grams), the velocity of light (c) and the force
constant of the bond (k, in N/m):
1
1 ⎡ (M x + M y ) ⎤ 2
ν= ⎢k ⎥ (3.4)
2πC ⎢⎣ M x M y ⎥⎦
Single, double and triple bonds have force constants that are approximately 500, 1,000 and
1,500 N/m, respectively.
Stretching vibrations are found to occur in order of bond strength [9]. The triple bond is
stronger than the double bond, which in turn is stronger than the single bond [9]. That
is, absorption of the triple bond occurs at 2300-2000 cm-1, while absorption for the
double bond occurs at 1900-1500 cm-1 and for the single bond at 1300-800 cm-1. Thus
IR spectra are very useful for distinguishing the different functional groups depending
upon their mode of vibration. The conception of group frequencies [10,11] is very useful
for practical interpretation of IR spectra, and nowadays software is attached to all the
modern instruments, providing all the group frequencies.
3.2 Sample Preparation
The selection of appropriate sample preparation technique is very important for accurate
analysis. Techniques of sample preparation for raw rubber identifications are as follows:
i) Preparation of raw rubber film [12]
ii) Preparation of film on low density polyethylene (LDPE) matrix [13]
iii) Preparation of film prepared from chloroform solution [12]
iv) Preparation of solution in chloroform or toluene [12]
v) Pyrolysis of raw rubbers [12]
Very rarely neat rubbers are used in end-use applications. In general, rubbers are
compounded with other materials such as fillers, antioxidants, accelerators and sulfur in
80
order to optimise a set of specific properties, e.g., tensile strength, modulus, impact
strength, and resistance to oxidation. More often than not, economic factors have an
overriding consideration. Carbon black is compounded into a large percentage of all
elastomers due to the fact that it is a relatively cheap filler and generally improves the
physical properties of rubbers. Hence in order to characterise the compounded and
vulcanised rubber products, different nondestructive methods are adopted, other than
the pyrolysis of rubber vulcanisates.
3.3 Different Types of IR Spectroscopy
Infrared spectra of optically dense materials such as silica and carbon black filled rubbers
cannot be obtained in the dispersive infrared spectrometers. In these cases only the destructive
methods like pyrolysis [12] of rubber vulcanisates are applied. With the advent of Fourier
transform infrared (FTIR) spectroscopy, numerous problems in the field of rubber and
polymer characterisation [14-16] could be easily solved due to an increase in the signal to
noise ratio, higher energy throughout, data processing capability and rapid scanning where
an entire spectrum can be recorded in a matter of seconds. Theoretical considerations of
the FTIR spectroscopy are available in the excellent books by Griffiths [17,18] and Ferraro
and Basile [19]. Several new sampling techniques allow the polymers to be examined in its
fabricated state, i.e., as a powder or fibre (diffuse reflectance), film (reflectance – absorbance),
coating (specular reflectance) or bulk sample (photoacoustic). Attenuated total reflectance
(ATR) spectroscopy, developed by Harrick [20] and Farenfort [21] is an ideal technique
for analysing the surface of crosslinked elastomeric materials [22] and carbon black filled
rubber [23]. Diffuse Reflectance Infrared Fourier Transform Spectroscopy (DRIFTS) has
recently become a powerful technique for the analysis of powders and coarse solids. The
sensitivity of DRIFTS and its quantitative accuracy for powdered samples have been
documented by Fuller and Griffiths [24,25]. Similar to ATR spectroscopy, infrared
photoacoustic spectroscopy (PAS), originally proposed by Alexander Graham Bell [26]
and developed by Rosencwaig and Gersho [27], attached to FTIR has the ability to examine
any type of solid, or semisolid material, whether it be crystalline, powder, amorphous,
smear, or gel. The surface selective nature of the PAS technique is advantageous for the
understanding of surface phenomena. Infrared PAS has been used to study surface treatment
of silica [28], and reactions involving silica [29].
3.4 Quantitative Analysis
The polymer systems are often complex. For example, crystalline and amorphous regions
coexist in semi-crystalline polymers. Any physical or chemical treatment of a polymer
will induce structural changes, the knowledge of which is essential for a better
81
understanding of the polymer properties. Infrared spectra of neat polymer can be taken
in solution and quantitative analysis is possible using Beer-Lambert’s law, following the
relationship between the concentration of the polymer solution and the amount of
radiation absorbed:
n
Io
Log
I
= ∑a c d
i =1
i i (3.5)
where, Io is the intensity of incident radiation,
I is the intensity of radiation after passage through the sample,
ci is concentration of the component (g/l),
d is the thickness of the sample cell in centimetre
and ai is the constant called absorptivity or extinction coefficient. Log (Io/I) is called the
absorbance. Transmittance (I/Io) of a sample is related to absorbance by the equation:
A = log (1/T) = -log T = al ci d (3.6)
A calibration curve is prepared, using absorbance versus concentration plot, so that the
concentration of the unknown component can be determined. But quantitative analysis
for a complex system like vulcanised rubber or a blend of two or three components, is
not possible. The use of computers with the FTIR spectrometer, increases the rapid
scanning capability, data processing for analysis of chemical or physical structural changes
in polymers as a function of time over the entire mid-IR frequency.
Quantitative analysis of such complex systems by data processing of digitised spectra

has been recently developed by Koenig and co-workers [30]. Koenig has developed several
methods for quantitative analyses:
i) Factor group analysis [31]
ii) Ratio method [32,33]
iii) Least squares curve fitting [34]
3.5 Applications of IR Spectroscopy
Hampton [35] reviewed the application of infrared spectroscopy in the characterisation

of rubbers. He discussed the role of infrared spectroscopy in the determination of polymer
82
composition, study of molecular structure, tacticity and sequence determination of

crystallinity and thermal transitions.
Another review [36] discusses the determination of composition and structure, including
tacticity, branching and end groups for diene polymers with the help of IR spectroscopy.
The infrared spectra for common rubbers have been reported in ASTM D3677 [12].
Most of these spectra were done either in solvent casting film or by pyrolysis.
Roy and De [13] used LDPE as the matrix material in the sample preparation, in which
the raw rubber is mixed with LDPE at 120 °C in a Brabender Plasticorder and made into
a thin film, which is stiff. Peak assignments of some rubbers are given in Tables 3.2a and
3.2b. The peak positions are almost in the same position as reported in ASTM D3677.
The representative spectra [13] of natural rubber (NR) and chloroprene rubber (CR) are
given in Figures 3.2a and 3.2b. For raw rubber analysis, this method is useful, as it does
not need any solvent, and there is no question of destruction of functional groups which
happens during pyrolysis. Furthermore, it also characterises the hanging groups of nitrile
rubber (NBR) or styrene-butadiene rubber (SBR), unlike the IR-ATR method which deals
with only the surface.
Composition and microstructure determination of polybutadiene (BR) and natural rubber

(NR) can be done by infrared spectra. Three different base units are possible for linear
addition polymers of 1,3 butadiene: units with cis or trans internal double bands from
1,4 addition and units with side vinyl groups from 1,2 addition (see Scheme 3.1a).
The IR spectra of butadiene polymers have been the subject of considerable study and
discussion [37-41]. Early work on the chemical and IR analyses of BR was reviewed and
a method described [35,42] for the IR determination of cis, vinyl and trans units from
absorption measurements at 724, 911, and 967 cm-1, respectively, as though each sample
was a mixture of three isomerically pure cis, vinyl and trans homopolymers (Figure 3.3).
Since isomerically pure polymers were not available, three different kinds of BR, each
relatively high in one of the three kinds of base units were used as standards [35]. The
band near 1308 cm-1 was identified [38,39] with the cis isomer and used for analyses
[43]. The 1308 cm-1 band is weak and relatively broad, with the appearance of an
unresolved doublet (1306, 1311 cm-1). The cis band at 730 cm-1 is more frequently used
in spite of some difficulties. Relatively pure, crystalline stereoregular polymers have been
prepared and structures were determined by X-ray diffraction for cis [44], trans [45] and
syndiotactic vinyl [46] and isotactic vinyl [47]. Infrared spectra [48-50] have been
published for the four stereoregular polybutadienes, with detailed analyses of the spectra
and band assignments for cis [51], trans [51] and syndiotactic vinyl [51] polymers. For the
spectrum of isotactic vinyl BR, bands at 1232, 1225, 1109, 943, 876, 807 and 695 cm-1
83
Table 3.2a Peak assignments for some commonly used rubbers

Name of Absorption bandsa Absorption Corresponding Assignmentb
polymer of pure polymer bands in LDPE chemical group
films, cm-1 matrix cm-1
CH3
NR 833 (s) 836(s) cis C=C (C-H)op def
1370 (s) 1374 (s) -CH3 (C-H) def
1665 (m) 1661 (w) —C = C (C-C) str
3,4 addition products
885 (m) 889 (w)
of isoprene
BR 741 (vs) 745 (s) cis-CH = CH— (=C-H)op def
990 (m) 993 (m) cis-CH=CH- (=C-H)op def
962 (m) - trans -CH=CH- (=C-H)op def
909 (m) 909 (m) CH2 = CH- (CH2)op def
SBR 699 (vs), 700 (vs), Monosubstituted
(C-H)op def
758 (s) 760 (s) benzene
1490 (m) 1493 (m) C=C in benzene ring (C-C) str
2000-1660 (m) (4 2000-1660 diff benzene ring
(C-H)op def
peaks) (w) (4 peaks) substitution
909 (m) 911 (w) butadiene part
962 (s) 962 (m) butadiene part
990 (m) 990 (w) butadiene part
EPDM (HD) 760 (m) - - -
710 (m) - - -
1370 (s) 1365 (m) -CH3 (C-H) def
1083 (s) 1074 (m) -CH3-CH2-CH2 -do-
900 (m) 889 (w) hexadiene
1145 - C(CH3)2 (C-H) def
a: s, strong; m, medium, w, weak
b: str, stretching; tw, twisting; def, deformation; op, out of plane; asym, asymmetric, sym,
symmetric. VS: very strong HD: hexadiene
Reprinted in modified form from S. Roy and P.P. De, Polymer Testing, 1992, 11, 1, 3,
Table 1, with permission of Elsevier Science, 1992, [13]
84
Table 3.2b Peak Assignments for some polar rubbers

Name of Absorption bandsa Absorption bands Corresponding Assignmentb
polymer of pure polymer in LDPE matrix chemical group
films, cm-1 cm-1
CR 1665 (vs) 1660 (s) >C = C (C=C) str

1315 (m) 1304 (s) —CH=CH2 (C-H) def
1110 (s) 1120 (w) - -
820 (s) 824 (s) C-Cl (C-Cl) str
NBR 2220 (vs) 2223 (s) —C ≡ N (-C=N) str
1300 (m) 1304 (w) —CH=CH2 (-C=H) def
1450 (s) 1466 (s) —CH2 (C-H) def
962 (s) 971 (s) Butadiene part
909 (m) 909 (s) -do-
CSM 1370 (vs) 1373 (s) —SO2 Cl (SO2) asym, str
1265 (m) 1258 —CH2 (CH2), tw
1150 (s) 1160 (s) —SO2Cl (SO2) sym, str
submerged in
740-720 C-Cl (C-Cl), str
723 band on PE
XNBR 2220 (vs) 2223 (vs) -C ≡ N (-C=N) str
1710 (s) 1698 (s) -COOH (C-O) str
962 (s) 969 (s) Butadiene part
909 (m) 916 (w) -do-
ENR-25 870 (m) 870 (w) epoxide Ring vibration
1250 (m) 1244 (m) epoxide -do-
835 (s) 836 (m) C=C(CH3)H (C=H)op def
1370 (s) 1375 (s) -CH3 (C-H) def
1665 (m) 1662 (w) C=C (C=C), str
85
Table 3.2b Continued

Name of Absorption bandsa Absorption bands Corresponding Assignmentb
polymer of pure polymer in LDPE matrix chemical group
films, cm-1 cm-1
Silicone methyl 1373 (w) 1375 (w) C H2 (C-H) def
1019 (s) 1020 (m) Si-O (Si-O) str
1264 (s) 1265 (m) Si-CH3 (-C-H) sym def
800 (m) 799 (m) Si-C (Si-C) str
880 (w) 882 (m) -CH=CH2
a: s, strong; m, medium, w, weak
b: str, stretching; tw, twisting; def, deformation; op, out of plane; asym, asymmetric, sym,
symmetric.
Reprinted in modified form from S. Roy and P.P. De, Polymer Testing, 1992, 11, 1, 3,
Table 2, with permission of Elsevier Science, 1992, [13]
Figure 3.2 (a) Infrared spectra of NR, (b) Infrared spectra of CR

Reprinted from S. Roy and P.P. De, Polymer Testing, 1992, 11, 1, 3, Figures 1 and 2, with
permission of Elsevier Science, 1992, [13]
86
Scheme 3.1a Structure of polybutadiene base units
Scheme 3.1b Structure of polyisoprene base units
Figure 3.3 Infrared spectra of butadiene polymers

Reprinted with permission from R.R. Hampton, Rubber Chemistry and Technology, 1972,
45, 3, 546. Copyright 1972, Rubber Division, American Chemical Society, [35]
87
have been used to differentiate it from the syndiotactic form. Trans-BR exists in amorphous
and two different crystalline forms. Recently Rao and co-workers [52] synthesised 1,3
butadiene using a neodymium chloride-tripentanotate-triethyl aluminium catalyst. FTIR
spectrum showed absorption peaks at 740, 912, 965 cm-1 for cis 1,4, vinyl 1,2 and trans
1,4 butadiene but microstructure determination shows that the cis content is about 99%,
and thus a stereospecific BR with high cis 1,4 and low vinyl 1,2 can be obtained.
Another widely used rubber is NR, which can be present in four base units in linear
addition polymers of isoprene (shown in Scheme 3.1b). Infrared analysis of polyisoprene
is difficult. The out of plane hydrogen deformations are satisfactory for determining
3,4 units at 888 cm-1 and 1,2 units at 908 cm-1 but the bands at 839 cm-1 for cis 1,4 and
842 cm-1 for trans 1,4 are relatively weak (Figure 3.2a).
Richardson and Sacher [53] measured total 1,4 structure at 815 and 857 cm-1, where the
broad weak trans and the sharper cis bands show identical absorptivities. Cis and trans
proportions were calculated from the absorbance at 843 cm-1, where absorptivity for the
cis form is nearly twice as great as for the trans forms. On the basis of X-ray diffraction
and chemical evidence, it was thought that NR and balata were 100% cis 1,4 - and
100% trans-1,4-polyisoprene, respectively. Modern infrared and nuclear magnetic
resonance (NMR) data strongly support this belief, but there has been an interesting
controversy about the possible presence of a small amount of 3,4 isomer in the natural
polymers. Binder and Ransaw [54,55] reported the presence of 2.2% 3,4 isomer in NR
and 1.3% 3,4 isomer in balata, and no evidence was found for the presence of trans
forms in NR, or for cis forms in balata. Following extensive IR studies, Saunders and
Smith [56] concluded that gutta percha and balata were trans 1,4 polyisoprene, while
Hevea and guayule rubbers were cis 1,4 polyisoprene (spectra shown in Figure 3.2a).
Infrared methods for copolymer composition like SBR, NBR, ethylene-propylene

copolymer (EPM), ethylene propylene diene terpolymer (EPDM) are listed in Table 3.3.
The determination of percentage of styrene and butadiene isomer distribution in

copolymers is an extension of the methods for the analysis of polybutadiene. The styrene
band at 700 cm-1 is largely independent of the sequence distribution and therefore useful
in styrene content determination [76]. A series of bands in the IR spectrum of crystalline
isotactic polystyrene at 758, 783, 898, 920, 1053, 1084, 1194, 1261, 1297, 1312 cm-1
have been attributed to the helical structure [77]. The absorption bands for butadiene in
SBR are similar to BR structures (Table 3.2a).
Like styrene content determination, acrylonitrile content in nitrile rubber can be

determined with IR spectra following the quantitative analysis method. Butadiene units,
adding up to an acrylonitrile unit in the polymer chain, show a strong tendency to add
1,4 trans structure. The acrylonitrile NBR has the following peaks:
88
2920, 2851 cm-1 for ν (-C-H of CH3-), CH2-

2222 cm-1 for ν (-C≡N)
1460 cm-1 for δ(C-H) of -CH2-, 1304 cm-1 for γ(CH2) of -CH2 -
962 cm-1 for δ(=C-H)op of trans -CH=CH-
909 cm-1 for δ(CH2)op of CH2 = CH-
and 740 cm-1 for δ(=C-H) op of cis - CH = CH- (Table 3.2b).
Table 3.3 Example of quantitative infrared analysis from copolymers

Determination of Sample analysed Infrared Technique Reference
polymer/rubber
Cis, trans, vinyl BR, blends AT R 57
Styrene SBR Film 58
Styrene SBR Film 59
Acrylonitrile, styrene Copolymer Solution 60
Acrylonitrile, styrene Copolymer Film 61
Butadiene, methyl methacrylate Copolymer Solution 62
NR SBR-NR vulcnaisate Pyrolysis 63
NR, SBR, IIR Blends Pyrolysis 64
NR, SBR Oil blends Pyrolysis 65
SBR NR-SBR blend Pyrolysis 66
NR, NBR, SBR Binary blends Pyrolysis 67
IIR, Polypropylene Blends Pyrolysis 68
NR, SBR, BR Blends Pyrolysis 69
BR Tyre tread Film 70
Propylene EPM Solution 71, 72
Propylene EPM Film 73
Dicyclopentadiene EPDM Film 74, 75
Reprinted in modified from [35] with permission of Rubber Chemistry Division,
American Chemical Society, USA, 1972
89
An IR spectrum shows the butadiene units to be almost exclusively trans [78], (that is,
98% trans and 2% vinyl, and 0% cis).
Propylene content in EPM rubber can be determined with the help of IR spectra. A
propylene band near 1155 cm-1 has been widely used [79] for EPM analysis, frequently
in combination with the polyethylene band at 721 cm-1. Tacticity is important in EPM
rubber, and the bands at 1229 and 1252 cm-1 are characteristic of syndiotactic and isotactic
structures, respectively, (both bands are present in atactic polypropylene as well). Polymer
structure may vary in the relative tactic placement of adjacent head to tail propylene
units and in the sequence distribution of base units along the chain. Some of them can be
identified [80] by infrared spectra, such as isolated or head to tail propylene units:
—CH2 –CH (CH3) –CH2-, at 1155 cm-1;
head to head propylene units, —CH2 –CH (CH3)–CH (CH3) –CH2 at 1120 cm-1; and
side ethyl groups, –CH—(CH2 - CH3)- at 775 cm-1 can be used to identify crystalline
ethylene blocks. Depending on the conditions of polymerisation the distribution of
monomer units in ethylene – propylene copolymers may be random or may show a
preference for the formation of relatively long sequences (blocks) of the same base unit
[81], or may show a tendency for alternation [82] of ethylene and propylene units. The
distribution of base units can be calculated theoretically [83] from monomer reactivity
ratios and can be measured experimentally by IR spectroscopy. For random copolymers
with propylene content, most of the propylene units were found in head to tail sequences
and thought to be isotactic [84].
An EPDM rubber is produced by the terpolymerisation of ethylene and propylene with a

small amount (typically of the order of 5%) of an unconjugated di-olefin. The di-olefins
used, include dicyclopentadiene, 1,4-hexadiene, 5-methylene-2-norbornene, 5-ethylidene-
2-norbornene and methyl tetrahydroindene, 1,5 cyclo octadiene. A number of other
dienes [74,75] have been tried. Infrared spectroscopy [35] is used to find out the ter
monomer content. The characteristic peaks for the ter monomer are shown in Table 3.4.
In view of the relatively low concentrations, it is probable that ter monomer base units
are present largely as isolated units in EPDM but the distribution of propylene and
methylene sequences is of considerable interest.
CR is a polymer with relatively large monomer repeating units that consequently exhibit
little vibrational coupling between the chemical units along the chain [85]. Infrared analysis
[86] showed the following bands: cis 1652 cm-1, trans 1660 cm-1; 1,2, 925 cm-1; 3,4,
883 cm-1. For emulsion polymerisation addition is almost entirely 1,4 with no more than
about 2% 1,2- and 3,4-. Cis 1,4 – CR exhibits characteristic infrared absorption bands
at 847, 1652, 3025 and 3282 cm-1 (C = C overtone). The corresponding trans bands are
at 822, 1660, 3018 and 3295 cm-1. Typical CR has 78%-96% trans, 1,4; 18% cis 1,4,
90
Table 3.4 Characteristic olefin infrared frequencies for EPDM

Termonomer Hydrogen stretch Double bond cm-1 Hydrogen
cm-1 stretch cm-1
Dicyclopentadiene 3043 1611 600
1,4 hexadiene - - 965
5-methyl-2-norbornene 3069 1662 872
5-ethylidene-2-norbornene 3040 1688 809
Reprinted from [35] with permission of Rubber Chemistry Division, American Chemical
Society, USA, 1972
0.3-2% 1,2-; and 0.2-2% 3,4- compositions. Infrared band at 780 and 950 cm-1 are
associated with the crystallinity, normally present in chloroprene rubber at room
temperature [87]. Coleman and co-workers [88] determined the proportion of the
crystalline phase of CR using the subtraction method. This technique considerably
simplified the assignment of the individual IR bands arising from the crystalline and
amorphous components of the spectrum [88]. The crystallinity in CR is affected by the
temperature of polymerisation. Polymers prepared in the range of –20 °C to 40 °C revealed
the specific crystalline bands, which were sensitive to a number of structural irregularities.
It is well known that as the polymerisation temperature is increased, the concentration
of the following units increased [89]: head to head, trans 1,4; cis 1,4; 1,2; 3,4 units.
Coleman and Painter [90] had made detailed studies of structural changes, crystallinity
changes of CR with temperature. The infrared spectrum [13] of CR is given in Figure 3.2b
showing the absorption bands at 1660, 1304, 1120, 824, 720 cm-1, characteristics of
trans isomers (Table 3.2b).
Polyurethanes (PU) are a special class of polymers containing the urethane group,
(—NH-CO-O-), commonly prepared by a condensation reaction between a diisocyanate
and a diol, such as hydroxyl terminated polyether or polyester:
HO –R —OH + OCN –R′ –NCO
( O —R —O —CO –NH –R′ –NH –CO ) n~
The isocyanate (2270 cm-1) uretedinedione ring carbonyl (1780 cm-1) and urea carbonyl
(1660 cm-1) groups can usually be identified. Carbonyls from ester, urethane, allophanate,
isocyanuric acid ring and Biuret groups all absorb near 1730 cm-1 and are difficult to
distinguish. Hydrogen bonds which can function as physical crosslinks in PU have been
91
studied for model compounds and polymers [89,91]. The 3305 cm-1 NH band and
inflections at 1780 and 1724 cm-1 on the 1735 cm-1 urethane carbonyl band are attributed
to hydrogen bonding; the free NH band is at 3340 cm-1. The NH group may bond to the
urethane carbonyl or to the ester carbonyl of polyester urethanes or to the ether oxygen
of polyether urethanes [89].
3.5.1 Rubber Blends
Polymer blends have had considerable attention recently due to the ease with which the
polymer properties can be manipulated to achieve some special characteristics that cannot
be achieved by the single polymer systems, or by simple copolymerisation. Different
polymers are blended to prepare the various types of industrial products, ranging from
giant tyres to small seals. Polymer blends may be immiscible, miscible, compatibilised,
reactive or self crosslinking. Utracki [92] described the criteria for the formation of various
types of blends. To study polymer miscibility, IR spectroscopy has emerged as one of the
most useful analytical tools. The technique can elucidate specific interactions, such as
hydrogen bonding, dipolar interaction, etc., [93,94]. The use of IR spectra for
characterisation of the blends of common rubbers (NR, BR, SBR) are shown in Table 3.3.
The production of reactive polymer alloys and blends has become a frontier area of
research and developmental activity in polymer technology in the past two decades.
Chemical modification of the polymer backbone, grafting onto a polymer chain, interchain
reactions and formation of interpenetrating networks (IPN) are the subject of a number
of reviews [95,96]. Maleation [97,98] is one of the oldest and most widely used techniques
for modification of polymers. It is the grafting of maleic anhydride onto a polymer
backbone by means of radical initiators such as peroxide. Maleation also facilitates specific
interactions that improve polymer-polymer compatibility [99].
In order to understand the mechanism of grafting, dibutyl maleate was grafted on to

polyethylene (PE) and EPDM using 0.5 wt% to 0.2 wt% dicumyl peroxide (DCP) in a
Brabender Plasticorder in two steps at 110 ± 2 °C and 180 °C [100]. The grafted polymers
(Peg dibutyl maleate (DBM) and EPDMgDBM) were characterised by IR analysis. The
IR spectra [100] of PE, PEgDBM and EPDMgDBM are presented in Figure 3.4.
In the spectra of functionalised polymers, the two bands at 1738 cm-1 and 1164 cm-1
were attributed to C=O stretching and C—O stretching of the ester group, respectively.
The carboxyl band at 1738 cm-1 was used to determine the degree of grafting. It is of
considerable technological importance to make heat, oil and fire resistant (HOFR) cable
sheathing compounds by combining the flame and oil resistance of polychloroprene (PCP)
and heat and ageing resistance of EPDM through application of judicious techniques of
blending. But PCP is not compatible with EPDM. Sen and co-workers [101] prepared
92
Figure 3.4 Infrared spectra of (a) PE, (b) PEgDBM, (c) EPDM, (d) EPDMgDBM
Reprinted with permission from A.K. Sen, B. Mukherjee, A.S. Bhattacharyya, P.P. De and
A.K. Bhowmick, Die Angewandte Makromolekulare Chemistry, 1991, 191, 3206, 15.
Copyright 1991, Wiley-VCH [100] with permission of Huthig and Wepf Verlag Basel,
1991
low halogen fire resistant low smoke (FRLS) cable sheathing compounds from the blends
of polyvinyl chloride (PVC) and polyolefins using dibutyl maleate grafted polyolefins as
compatibiliser and characterised their properties. PCP/EPDM blends were prepared using
dibutyl maleate grafted ethylene-propylene-diene rubber [102]. The probable mechanism
of compatibilisation [102] between PCP and EPDMgDBM is shown in Scheme 3.2.
EPDMgDBM gives rise to an interaction, mainly through a hydrogen bond between the
carboxyl group of DBM and methine hydrogen of PCP. Besides this, dipole-dipole
interactions of type >C=O…..Cl–C may also occur [102]. Peroxide initiated graft
copolymerisation of vinyl trimethoxy silane (VTMO) and vinyl triethoxy silane (VTEO)
onto PE and ethylene propylene copolymer (EPR) was also studied by De and co-workers
[103]. The absorbance of —Si –O –Si – group stretching of the grafted silane at 1090 cm-
1
was used for characterisation of PEgVTMO and PEgVTEO. Actually, in the presence
of moisture and a condensation catalyst, the alkoxy group of the silane derivatives pendent
to the grafted co-polymer is converted into silanol groups which undergo a condensation
reaction to form a –Si – O – Si – type network [103]. Both the hydrolysis of alkoxy silane
to silanol and their condensation reaction occur almost instantaneously forming the
93
Scheme 3.2 Probable mechanism of compatibilisation between EPDMgDBM and PCP.

Reprinted with permission from A.K. Kalidaha, A.S. Bhattacharya, A.K. Sen and P.P. De,
Die Angewandte Makromolekulare Chemistry, 1993, 204, 3484, 19. Copyright 1993,
Wiley-VCH, [102]
network [103]. These types of reactive blends are generally used for making special
purpose products for example butyl rubber (IIR) is blended with polyamide 6,66,610
copolymer (LPA) in the presence of alkyl phenol formaldehyde resin (PR) to produce
IIR-PR-LPA co-polymer [104] which has a novel gas-barrier property, used for rubber
hose in an air conditioning system. From Figure 3.5, it is clear that following the reaction
of LPA with PR, the peaks at 1548 cm-1 (NH bending) and 1647 cm-1 (C=O) stretching
of LPA shifted by the reaction to 1542 and 1640 cm-1, respectively.
The peak intensity of the NH bending deformation of the amide group at 1548 and
3085 cm-1 relative to that of CH2 stretching at 2950 cm-1 decreased and a peak of CH2OH
of the PR-LPA copolymer at 3372 cm decreased a little from PR. This indicates that the
hydroxyl group in PR reacts with the amide group in LPA to produce the co-polymer. IR
spectra of the IIR-PR-LPA copolymer and IIR are shown in Figures 3.5e, and 3.5d. In
Figure 3.5e, the peak intensity of the double bond at 890 cm-1 and that of CH2OH at
3372 cm-1 decreased, and a new peak at 1750 cm-1 appears, which may be due to a
Chroman-ring type of structure indicating that a Diels-Alder type of reaction takes place
between PR-LPA copolymer with IIR.
94
Figure 3.5 IR spectra of (a) polyamide 6,66,610 (LPA), (b) alkyl phenol formaldehyde
resin (PR), (c) reaction product PR/LPA, (d) IIR, (e) IIR-PR-LPA copolymer
Reprinted with permission from S.I. Goto, K. Kimura, T. Yamamoto and S. Yamashita,
Journal of Applied Polymer Science, 1999, 74, 3548. Copyright 1999, John Wiley & Sons,
Inc., [104]
Another such example is EPDMgNVP (ENVP), a new graft co-polymer which was
synthesised to improve the dyeability of EPDM [105]. ENVP was synthesised by grafting
N-vinyl pyrrolidione (NVP) with benzoyl peroxide as an initiator in toluene. The IR
spectra of ENVP exhibited a characteristic absorption band at 1670 cm-1 (C=O stretching)
due to NVP and 1460 cm-1 (-CH2 bending) and 1375 cm-1 (-CH3 bending) due to EPDM.
From the literature, many examples of such grafted and compatibilised blends can be
found where IR spectroscopy has been used to characterise the blends. Examples of such
blends are as follows:
95
carboxyterminated polybutadiene-nylon [106], polydimethyl siloxane (PDMS) – ethylene

methyl acrylate [107,108], epoxidised NR (ENR) – PE co-acrylic acid [109], PU-polyimide
[110], ethylene vinyl acetate – maleic anhydride [111], acrylate rubber – polyamide [112],
polybutadiene – CR [113], PU – aminoethyl aminopropyl PDMS [114], chlorohydrin
rubber grafted with polybutyl acrylate [115].
3.5.2 Self-crosslinking Blends
It has been already mentioned that functionalisation of polymers results in new materials
with a wide spectrum of properties not available in the parent polymers. Chemical
modification of the polymer backbone, grafting onto a polymer chain, interchain reactions
and formation of IPN are the subject of a number of reviews [116]. De and co-workers
studied a series of self-crosslinkable polymer blends, which are mixtures of two or more
functionally reactive polymers that are capable of undergoing mutual crosslinking via
condensation or substitution reactions at high temperatures. Such self-crosslinkable
rubber-rubber blends include ENR and carboxylated nitrile rubber (XNBR) [117,118].
ENR was found to form self-crosslinkable blends with many chlorine containing polymers
like CR [119], PVC [120] and chlorosulfonated polyethylene (CSM) [121-124].
Mukhopadhyay and co-workers [124] reported that CSM reacts with XNBR forming a
self vulcanisable miscible blend through the formation of a sulfonyl ester group
(Scheme 3.3a). Roychoudhury and co-workers [125] made a detailed study of the reaction
with FTIR-ATR spectra and showed that crosslinking takes place in two ways: either
through allyl chloride [Scheme 3.3b (1,2)] or through allyl chloride and amide
[Scheme 3.3b (3,4)] produced by hydrolysis of a –C≡N group.
The IR spectra of a self-crosslinked blend of CSM/XNBR [125] following the reaction

scheme (Scheme 3.3b) are presented in Figure 3.6.
It is evident that absorption at 1740 cm-1 in the crosslinked blend (Figure 3.6b) is likely
to correspond to the C=O stretching of ester. The band at 1124 and 1171 cm-1 may be
ascribed to C-O-C asymmetric stretching vibration of an aliphatic ether, but formation
of the ester is much stronger, because the band in the region of 1100-1200 cm-1 can also
be attributed to C-O-C stretching vibration of an aliphatic ester. The bands at 1697,
968, 916 cm-1 are due to hydrogen bonded –COOH, trans –CH=CH- and CH2 = CH~
(out of plane bending). The absorption band at 2237 cm-1, which is attributed to a C≡N
stretching mode is strongly attenuated by self-crosslinking of the blend. This indicates a
significant loss or conversion of cyanide groups, due to cyanide/cyanide [126] or cyanide/
carboxyl [127] reactions or even hydrolysis of the cyanide group (shown in Scheme 3.3b).
The band at 1553 cm-1 is due to an amide II band and the band at 1651 cm-1 is due to
96
Scheme 3.3a Reaction between Hypalon and XNBR

Reprinted with permission from S. Mukhopadhyay, P.P. De and S.K. De, Journal of
Applied Polymer Science, 1991, 43, 347, Figure 3. Copyright 1991, John Wiley and Sons,
Inc., [124]
amide I, while the band at 968 cm-1 (trans –CH=CH-) is attenuated in the crosslinked
blend, and the band at 725 cm-1 is a composite band of cis 1,4 –CH=CH- and CH2
rocking at 740-720 cm-1. Hence cis-trans isomerisation is also exhibited during the
crosslinking of blend. Other examples of such self-crosslinkable blends are: ENR/XNBR
[128], ENR/CR [119], ENR/XNBR/CR [128,129], ENR/CSM [130], NBR/PVC [131],
HNBR/PVC [132], ENR/PVC [133], PVC/ENR/XNBR [134], PAA/ENR [135], XNBR/
CIIR [136]. The blends have been characterised by IR spectroscopy.
97
Scheme 3.3b Reaction between CSM and XNBR

Reprinted with permission from A. Roychoudhury and P.P. De, Journal of Applied
Polymer Science, 1997, 63, 1761. Copyright 1997, John Wiley and Sons, Inc.
3.5.3 Polyurethanes
The properties of PU depend on the proportions of hard and soft segments in the polymer
structure. Paul and co-workers synthesised segmented block copolymers of NR and 1,3
butanediol – toluene diisocyanate oligomers [137], NR and bisphenolA-toluene
diisocyanate oligomers [138] and characterised them by IR spectroscopy. The
98
Figure 3.6 FTIR-ATR spectra of (a) uncrosslinked 50/50 (w/w) CSM/XNBR blend, (b)
self-crosslinked 50/50 (w/w) CSM/XNBR blend, (c) (b-a) difference spectrum
Reprinted from A. Roychoudhury and P.P. De, Journal of Applied Polymer Science, 1997,
63, 1761. Copyright 1997, John Wiley and Sons, Inc., [125]
99
stoichiometric amounts of hydroxyterminated liquid natural rubber (HTNR), 1,3

butanediol (BDO), dibutyl tin dilaurate (DBTDL) and toluene diisocyanate (TDI) in
tetrahydrofuran (THF), were mixed following a one shot or two shot process to prepare
the block copolymer, which were analysed by IR spectroscopy.
Hydrogen bonding plays a critical role in determining the morphology and overall properties
of PU, PU-ureas, and other polymers which have pendent functional groups capable of
forming hydrogen bonding. The extent of hydrogen bonding in PU can be qualitatively
studied by determining the frequency shifts in hydrogen bonded (-N-H) and (-C=O) peaks
(that is, -N-H…O=C) relative to the free (-N-H) and (C=O) peaks. The infrared absorption
bands for such hydrogen bondings are 1703-1710 (C=O….. H-N), 1730-1740 (free C=O,
urethane) 1660-1670 (C=O….. H-N, urea disordered), 1630-1645 (C=O……H-N, urea
ordered), such peaks are observed in PDMS and polyether based urethane-urea copolymers
[139], PU based on hydroquinone – bis (β-hydroxyethy)ether [140], propylene oxide/
polybutadiene bi-soft segment urethane/urea prepared from polypropylene oxide based
ioscyanate – terminated triol prepolymer with polybutadiene diol [141]. The FTIR studies
were conducted on UV curable films of water borne polyurethane-acrylate (PUA) ionomer
[142] and cyclo aliphatic di epoxide-ENR-glycidyl methacrylate (GMA) [143].
The IR spectra of a cycloaliphatic epoxide-ENR-GMA system based coating (Figure 3.7a)

[143] exhibited the following characteristic absorption bands:
• Epoxy ring stretching at 1255 cm-1,
• Terminal oxiranes of GMA at 909 cm-1,
• Weak symmetry in plane ring deformation of internal oxirane at 814-815 cm-1,
• Cycloepoxy stretching at 789 cm-1, 1638 cm-1, 844-858 cm-1 due isoprene double
bonds of ENR-50,
• Acrylic double bonds of GMA at 1638 cm-1 and 815 cm-1.
A marked increase in OH stretching at 3550 cm-1 on UV exposure (Figure 3.7b) with

concurrent decrease in the absorption bands corresponding to 1255 cm-1 (epoxy, ring
stretching) 909 cm-1, 814-815 cm-1 and 789 cm-1 establishes the successful accomplishment
of ring opening polymerisation. The increase in 1089 cm-1 can be attributed to the
formation of an ether bond. The marked decrease in the absorption bands of isoprene
double bonds at 1638 and 844-858 cm-1 (isoprene C-H wagging) showed that free radicals
generated during the photolysis of the initiator cause simultaneous radical polymerisation
involving the isoprene double bonds of ENR and the acrylic double bonds of GMA. The
absorption band at 909 cm-1, corresponding to terminal oxiranes of GMA almost
100
Figure 3.7a FTIR spectrum of cycloaliphatic epoxide-ENR-GMA hybrid system
Figure 3.7b FTIR spectrum of cycloaliphatic epoxide-ENR-GMA hybrid system after UV

exposure and post cure
Reprinted with permission from R.N. Kumar, C.K. Woo and A. Abusamah, Journal of
Applied Polymer Science, 1999, 73, 1569. Copyright 1999, John Wiley and Sons Inc., [143]
101
disappears on UV irradiation (Figure 3.7b), suggesting that the epoxy ring participates
in polymerisation reactions [143]. The results showed that the acrylic double bonds as
well as the epoxy groups of GMA, the isoprene double bonds and epoxy groups of ENR
and epoxy groups of cycloaliphatic epoxide resin all participate in photoinitiated
polymerisation and cross linking reactions, producing an interpenetrating network (IPN).
3.5.4 Rubber-filler Interaction
The reinforcement of an elastomer by a filler is associated with a strong interaction

between the filler surface and the elastomer [144]. Although a complete knowledge of
the exact nature of elastomer/filler interaction is still lacking, the polymer/filler attachments
appear to be both physical and chemical, depending on the physicochemical character of
the filler surface and the chemical nature of the elastomer [145]. IR spectroscopy in its
various forms is an important and forceful technique, which can provide useful
information about surface functional groups of carbon black and silica filler. The first IR
transmission spectrum of a carbon black with a high oxygen content was reported by
Hallum and Drushel [146] using a Nujol mull technique. Donnet [147] and Papirer
[148] examined the surface groups on a series of carbon blacks, using direct transmission
and internal reflection spectroscopy and observed the presence of bands at 3200-3400,
1740, 1600, 1275-1195 cm-1 indicating hydroxyl, carboxyl, lactone, quinone, ether, phenol
groups. Koenig and co-workers [149] utilised IR spectra to find the reversion resistance
in NR curing when filled with different grades of carbon black. They found the trans-
methine (965 cm-1) content is higher at low cure times but less at longer times for the
carbon black filled samples. Roychoudhury [150] while studying the interaction of ENR
with carbon blacks (N121 and oxidised N121), used diffused reflectance IR spectra to
study the surface groups present on the carbon black surface. The spectrum corresponding
to the oxidised sample was characterised by more intense peaks in the O-H and the C=O
stretching regions, which were responsible for chemical interaction between ENR and
oxidised carbon black. De and co-workers [151] using FTIR spectra characterised the
bound rubber formation between CSM and non-oxidised and oxidised N110 carbon
black. PAS-FTIR spectra [151] (Figure 3.8) of pure CSM, the carbon black (N 110) and
CSM/N110 were taken under high vacuum.
The spectra showed that carbon black filled CSM compound retained almost all the
characteristic bands (1736, 1697, 1651, 1558, 1458 cm-1) due to carbon black, but the
strong absorption bands at 1161, and 1366 cm-1, characteristic of –SO2Cl groups of
CSM diminished (Figure 3.8c) forming a coupling bond between the CSM macroradicals
and quinone or phenoxy radical sites on oxidised carbon blacks, giving an ether type of
bond (Figure 3.8c), e.g., CB—)—O-CH~ (1061,1096 cm-1), according to the following
reaction:
102
Figure 3.8 FTIR photoacoustic spectra of (a) net CSM, (b) carbon black N110, (c)
CSM/N110 compound
Reprinted with permission from A. Roychoudhury and P.P. De, Journal of Applied
Polymer Science, 1995, 55, 9. Copyright 1995, John Wiley and Sons, Inc., [151]
103
CB —) OH + ~ CH~
|
SO2Cl→CB. —) –O—CH ~+ SO2 + HCl (3.7)
Hence surface oxidation of carbon black leads to a greater extent of bound rubber
formation in polar rubbers [152] like ENR.
Chemical interaction between furnace blacks and rubbers is responsible for the greater
extent of reinforcement for polar rubbers [153]. Lately, it has been reported that –OH
and –COOH groups of surface oxidised furnace blacks react chemically with
functionalised polymers like CSM [154], XNBR [155] and ENR [156], forming a
crosslinked vulcanisate in the absence of any crosslinking agent. The mechanism of
chemical interactions is supported by IR spectroscopy.
It is a well known fact that coupling agents cause improvement in rubber-filler interaction
in the case of mineral fillers. Recently Bandyopadhyay and co-workers [157] showed
that 3-amino propyl triethoxy silane acts as a promoter in bonding between XNBR and
surface oxidised carbon black. Manna and co-workers [158] made detailed studies on
the effects on the surface oxidation of filler and silane coupling agent, namely N-(4-vinyl
benzyl) —N′—[3-(trimethoxysilyl)propyl] ethane-1,2-diamine mono HCl salt (Trade name
Z6032; made by Dow Corning, USA) on the properties of ENR filled with intermediate
super abrasion furnace (ISAF) carbon black in the absence of conventional crosslinking
agents. The various techniques of IR spectra were used to find out the chemical interaction
between ENR and oxidised ISAF black [158] and they are shown in Figures 3.9a and
3.9b.
The DRIFTS of carbon black sample [158] (Figure 3.9a) show peaks at 1350, 1385 cm-1
due to symmetric and asymmetric C-O stretching of phenol [147], 1574, 1634 cm-1 for
tetrahydroanthroquinone and polycyclic quinone [148], 1750, 1800 cm-1 for unsaturated
lactones [148], 3400, 3640, 3760 cm-1 for free OH groups.
The main difference between two grades of carbon black is the presence of carboxylic
acid groups at 1717 cm-1 for oxidised carbon black.
The ATR-FTIR spectra of ENR and transmission spectra [158] of thin film of the coupling
agent are shown in Figure 3.9b, in which the band at 1085 cm-1, is characteristic of
Si—OCH3 stretching vibration, the band at 1602, 1627 cm-1 are characteristic of ~NH2+
deformation modes and C=C (vinyl benzene) stretching, respectively.
The peak at 820 cm-1 is due to the rocking vibration of ~NH2+~, whereas the peaks at
1193, 715, 450 cm-1 are characteristic of secondary aliphatic amines, and the peaks at 878,
840 cm-1 are characteristic of epoxy ring vibration, while the peaks at 1050, 1114 cm-1 are
104
Figure 3.9a DRIFTS of carbon black (a) ISAF black, (b) oxidised-ISAF black
Reprinted with permission from A.K. Manna, P.P. De, D.K. Tripathy, S.K. De and M.K.
Chatterjee, Rubber Chemistry and Technology, 1999, 72, 2, 398, Figure 1. Copyright
1997, Rubber Division, American Chemical Society, Akron, [158]
due to C=O stretching of ENR, and at 1377 cm-1 and 1648 cm-1 are due to C-CH3 and
C=C stretching of ENR. The IR spectra of ENR mixed with ISAF black, oxidised ISAF
black, with and without coupling agent are presented in [158], and representative spectra
are given in Figure 3.10, in which the spectra of ENR with oxidised ISAF black (60 phr)
and coupling agent, Trade Name Z6032 (4 phr) with different moulding times show that
the intensity of the peak 874 and 840 cm-1 due to epoxy ring vibration decrease.
105
Figure 3.9b Transmission FTIR spectra of (a) coupling agent, (b) ENR
1997, Rubber Division, American Chemical Society, Akron, [158]
The difference spectrum shows negative absorbance at 1720 cm-1 which is due to the
reaction between –COOH groups on the oxidised carbon black surface and to –NH2+—
and Si-OCH3 of silane coupling agent. This is further substantiated from the appearance
of a peak at 1635 cm -1 with a shoulder at 1620 cm -1, which is due to formation of
–(C=O)—N—(amide) and –Si-O-(C=O)- (silyl ester).
The surface of hydrated (precipitated) silica is highly polar and hydrophilic due to its
numerous silanol groups. The polar surface chemistry assists the development of better
106
Figure 3.10 (a) ATR-FTIR spectra of mix EOZ4 (after 1 h moulding) (b) ATR-FTIR
spectra of mix EOZ4 (after 1 min moulding) (c) Difference spectrum (a-b)
1999, the Rubber Division, American Chemical Society, Akron, [158]
107
bonding of the silicas with polar polymers such as NBR [159], XNBR [160] and CSM
[161]. Koenig and co-workers [162] studied the interactions of silica filler in zinc activated
sulfur vulcanised cis-1,4 polyisoprene using FTIR-ATR spectroscopy. They observed that
an increase in the silica level resulted in band broadening and a frequency shift to lower
wave numbers of the silica absorbance region (1250-1000 cm-1) due to a combination of
physical and chemical adsorption of the rubber and parts of the cure system on the silica
surface. The peaks that appeared near 1040 and 1017 cm-1 were attributed to physical
interactions between silica and NR. Manna and co-workers [163] studied the role of the
coupling agent Z6032 in bonding between precipitated silica and ENR during high
temperature moulding at 180 °C. The ATR-FTIR spectra of EHZ5 (ENR 100, silica 60,
coupling agent 5 phr) at different moulding times, are shown in Figure 3.11. From the
difference spectrum, it is evident that the peaks at 876 and 3700 cm-1 show negative
absorbance. The results signify that a portion of the epoxy groups react during moulding
(Figure 3.11c).
The new peak at 485 cm-1 is characteristic of C-N-C (tertiary amine) deformation, whereas
the positive absorbance at 1078, 1011, 796 cm-1 are characteristic of the Si-O-Si
(asymmetric stretch), -Si(OCH3)2 deformation, indicating that the epoxy group reacts
with amine functionality to form C-N bonds. The negative peak at 1647 cm-1 confirms
that C=C bond reacts with hydrogen chloride (HCl) liberated from the coupling agent
during high temperature moulding. The positive absorbance at 704 cm-1 and 1260 cm-1
confirms the formation of C-Cl bonds, and secondary alcohols. In presence of coupling
agent, coupling bonds [163] are formed through functional groups (amine and methoxy)
in addition to silyl ether linkages as shown in Scheme 3.4.
Carbon-silica dual phase fillers (CSDPF), a new generation of materials for rubber
reinforcement, have been analysed by Murphy and co-workers [164] with IR spectra
(potassium bromide plate method). The region above 3000 cm-1 contains information
about the silanols, but unfortunately for carbon black and the dual phase fillers, this
information is lost due to interference from a large water band. The two major bands are
located, one at about 1600 to 1625 cm-1 and the other at 1115 cm-1. The 1600-1625 cm-1
has been assigned to bulk carbon and adsorbed water, when the 1100 cm-1 region has
been ascribed to SiO2 networks, a small peak near 1200 to 1250 cm-1 is indicative of
carbon phase oxygen groups.
3.5.5 Milling
The milling of rubber is an important operation for further processing of rubbers. During
milling, rubber molecules break down, and the chemical changes which occur, are a
function of time, temperature, rate of shear, shear stress and viscosity of rubber. The mill
108
Figure 3.11 ATR-FTIR spectra of mix EHZ5. (a) Sample moulded for 60 minutes (b)
Sample moulded for 2 minutes (c) The difference spectrum (a-b)
Reprinted with permission from A.K. Manna, P.P. De, D.K. Tripathy, S.K. De and D.G.
Peiffer, Journal of Applied Polymer Science, 1999, 74, 389, Figure 7. Copyright 1999,
John Wiley and Sons, Inc., [163]
breakdown behaviour of NR [165,166] and ENR [167] have been reported earlier, but
Kumar and co-workers [168] used IR spectroscopy to detect the formation of functional
groups during the milling of NR, ENR, SBR, CR and acrylic rubber. The concentration
of carboxyl (>C=O) groups at 1700-1800 cm-1 increases with time of milling for all the
109
Scheme 3.4 Probable mechanism of bonding between ENR and precipitated silica in
the presence of silane coupling agent
Reprinted with permission from A.K. Manna, P.P. De, D.K. Tripathy, S.K. De and D.G.
Peiffer, Journal of Applied Polymer Science, 1999, 74, 389, Figure 9. Copyright 1999,
John Wiley and Sons, Inc., [163]
110
rubbers. The concentration of hydroxyl groups (—OH) in NR and ENR also increases
during milling, but the concentration of epoxide rings decreases and that of furan rings
increases with time of milling.
3.5.6 Adhesion
IR spectroscopy plays a great role in the study of the chemical changes which occur
during the adhesion of polymers with metal. For example, FTIR has been used to analyse
the interface of laminates [169] composed of ethylene copolymers and aluminium. The
spectra of the thin solution cast films of carboxylic acid copolymer showed two bands in
the carbonyl region that were assumed to be due to free and dimeric acid groups, with a
free acid group at 1740 cm-1. The increased amount of free acid at interface can be
attributed to the interfacial hydrogen bonds between the hydroxyl groups of the acid
and the surface AlOH groups or eventually Al-O-Al, thus setting some of the carbonyl
groups free [169].
Tack is the ability of two rubbery materials to resist separation after bringing their surfaces
into contact for a short time under light pressure. The effect of para-tert-octyl phenolic
(PTOP) resin on the tack of isoprene/N660 compound has been studied before and after
ageing with FTIR spectroscopy [170].
The blends of XNBR and chlorobutyl rubber (CIIR) and XNBR-epichlorohydrin rubber
(ECO) act as adhesive for aluminium-aluminium bonding [171,172]. In order to study
the adhesion between CIIR-XNBR blend on aluminium foil, Bhattacharya and co-workers
[173] used FTIR-ATR spectra [173], and they found that there was no change of spectra
of the blend on Al-foil, i.e., on the leached aluminium surface, as indicated by the intensity
of the carboxyl peak at 1736 cm-1 (for free COOH group or ester group) and 1696 cm-1
(for hydrogen bonded carbonyl). Manoj and co-workers [174] found that in addition to
molecular interaction and physical entanglements of the molecular chains across the
interface in PVC-NBR rubber joints, at high temperatures and long contact times,
interfacial chemical bonds may be formed which seem to couple the two adherends,
thereby resulting in cohesive failure of the rubber matrix on peeling. IR spectroscopic
[174] studies of the PVC/NBR blend reveal formation of chemical bonds at the contact
temperatures studied. The changes in the absorbance in the regions 3500-2900 cm-1 and
1750-1500 cm-1 correspond to the chemical interaction between the functional groups
in the system are shown in Figure 3.12. The broadening of the peak at 3500-2900 cm-1 is
assigned to N-H and O-H bonds formed by the hydrolysis of acrylonitrile (C≡N) group
(2222 cm-1) in the presence of hydrogen chloride liberated during thermal degradation
of PVC [131].
111
Figure 3.12 Infrared spectra of 50/50 (w/w) PVC-NBR blends moulded at 150 °C for 2
and 60 minutes
Reprinted with permission N.R. Manoj and P.P. De, Journal of Adhesion, 1993, 43, 199.
Copyright 1993, Gordon and Breach Science Publishers, [174]
112
The increase in absorbance at 1730 cm-1 may be attributed to the formation of amide, acid
or ester groups in the system. The amide and ester crosslinks are formed via the reaction
between allylic chlorines in PVC and amide, acid groups in NBR [131]. The reduction in
the peak at 1532 cm-1 may be due to the ring opening of the triazine derivatives in NBR,
formed by the cyclisation of nitrile groups. The different amide bands could not be
distinguished from the C=C stretching vibrations. The existence of an interfacial reaction
between polyamide-6 and acrylate rubber (ACM) was confirmed by Jha and co-workers
[112] with the help of IR spectra, in which joints made of ACM rubber and polyamide-6
annealed at 220 °C for different times indicated a reaction between the amine (-NH2) and
carboxyl (-COOH) end groups of Nylon 6 with the reactive epoxy groups of ACM chains.
3.5.7 Degradation
On storage, the stability of polymers changes due to their ageing as a result of the reaction
of double bonds and other reactive groups. The degradation of ENR was studied with IR
spectroscopy [175], which showed that during thermal ageing of ENR, carbonyl, alcohols,
THF and ether, crosslinks are formed most probably due to the presence of a very small
amount of acid. The thermal degradation of NR [176] in air for 1, 2 or 3 hours at 145 °C
show the absorbance intensity of sample (37% gel content) decreases at 835, 1375, 1450,
1660 cm-1 with a prolonged heating whereas the absorbance at 1715 cm-1 and 3500 cm-1
increases, indicating the formation of aldehyde, ketone and carboxylic acid. FTIR is also
applied for the development of biodegradable polymers suitable for biomedical application
[177]. Two types of PU were made containing PEO (polyethylene oxide) and
polycaprolactone (PCL), an amino acid based diester chain extender was used to confer
degradability [177]. FTIR showed a large carbonyl peak (~1735 to 1724 cm-1) resulting
from the ester groups of PCL-PU whereas the PEO polyurethanes exhibit a distinct ether
peak at ª1100 cm-1. For PCL-PU the carbonyl stretching (1650-1800 cm-1) is dominated by
intense soft-segment ester band, located at 1724 to 1735 cm-1 depending on the presence
or absence of soft segment crystallinity [177].
3.6 Reverse Engineering

Rubber industries produce various types of complicated products like tyre, cable, belt,
seal, bearings, engine mounts, etc. The products are composed of rubber, plastics, fibre,
metal, fillers and many other additives like antioxidants, accelerators, etc. Reverse
engineering is a technique by which a rubber technologist can reconstruct the composition
of the products based on thermal analysis. Dormagen [178] and Baranwal [179]
reconstructed the formulation of a tyre, based on analyses of FTIR, spectra, thermal
analysis and high performance liquid chromatography.
113
3.7 Conclusion
It is concluded that IR spectroscopy provides information on qualitative as well

quantitative analyses of rubbery materials, apart from their microstructures (that is,
whether cis or trans, syndiotactic, atactic or isotactic). Different types of rubber blends
(compatibilised or self-crosslinked) can be identified by the infrared spectroscopy.
Synthesis, and degradation of polymers can also be followed by IR spectra. Mechanism
of interaction between rubbers and fillers, can also be studied by IR-spectra. Different
types of chemical reactions like the milling behaviour of rubbers, mechanism of adhesion
and degradation can also be studied with the help of IR spectroscopy. The technique
plays a great role in the product analysis under reverse engineering.
Acknowledgement
Thanks are due to Mr. Rajeev R.S. and Mr. Shambhu Bhattacharyya for assistance in
preparing the manuscript.
References
1. W.W. Coblentz, Applied Spectroscopy, 1953, 7, 109.
2. W.W. Coblentz, Investigations of Infrared Spectra, Carnegi Institution of

Washington, 1905, Paper No.35.
3 W.W. Coblentz, Investigations of Infrared Spectra, Carnegi Institution of

Washington, 1906, No.65.
4 W.W. Coblentz, Investigations of Infrared Spectra, Carnegi Institution of

Washington, 1908, No.97.
5. R.A.C. Isbell, Hilger Journal, 1955, 2, 1, 3.
6. I. Fleming and D.H. Williams in Spectroscopic Methods in Organic Chemistry,

5th Edition, Eds., D.H. Williams and I. Fleming, McGraw-Hill Publishing
Company Ltd., New Delhi, 1994.
7. R.G. White, Handbook of Industrial Infrared Analysis, Plenum Press, New York,
1964, Chapter 1.
8. R.D. Bruan, Introduction to Instrumental Analysis, McGraw Hill International

Editions, Singapore, 1987, Chapter 12.
114
9. J. R. Dyer, Applications of Absorption Spectroscopy of Organic Compounds,

Prentice Hall of India Pvt. Ltd., New Delhi, 1991, Chapter 3.
10. M.N. Bikales in Characterisation of Polymers, Wiley Interscience, New York,

1971, 129.
11. G. Socrates, Infrared Characteristic Group Frequencies, John Wiley and Sons,
New York, 1980.
12. ASTM D3677-00, Standard Test Methods for Rubber – Identification by Infrared
Spectrophotometry, 2000.
13. S. Roy and P.P. De, Polymer Testing, 1992, 11, 1, 3.
14. J.L. Koenig, Applied Spectroscopy, 1975, 29, 293.
15. M.M. Coleman and P.C. Painter in Applications of Polymer Spectroscopy, Ed.,
E.G. Brame, Academic Press, New York, 1978, 135.
16. L. D’Esposito and J.L. Koenig in Fourier Transform Infrared Spectroscopy,

Applications to Chemical Systems – 1, Eds., J.R. Ferraro and L.J. Basile,
Academic Press, New York, 1978, 61.
17. P.G. Griffiths, Chemical Infrared Fourier Transform Spectroscopy, Wiley, New
York, 1975.
18. P.G. Griffiths and J.A. de Haseth, Fourier Transform Infrared Spectroscopy,
Wiley, New York, 1987.
19. Fourier Transform Infrared Spectroscopy, Applications to Chemical Systems, Eds,

J.R. Ferraro and L.J. Basile, Academic Press, New York, 1978, Volumes 1 and 2.
20. N.J. Harrick, Journal of Physical Chemistry, 1960, 64, 1110.
21. J. Farenfort, Spectrochimica Acta, 1962, 18, 1103.
22. H. Ishida, Rubber Chemistry and Technology, 1987, 60, 497.
23. W.W. Hart, P.C. Painter, J.L. Koenig and M.M. Coleman, Applied Spectroscopy,
1977, 31, 3, 220.
24. M.P. Fuller and P.G. Griffiths, Analytical Chemistry, 1978, 50, 1906.
25. M.P. Fuller and P.G. Griffiths, Applied Spectroscopy, 1980, 34, 533.
115
26. A.G. Bell, Philosophical Magazine, 1881, 11, 510.
27. A. Rosencwaig and A. Gersho, Journal of Applied Physics, 1976, 47, 1, 64.
28. R.J. Bell, Introductory Fourier Transform Spectroscopy, Academic Press, New
York, 1972.
29. J.B. Kinny and R.H. Staley, Analytical Chemistry, 1983, 55, 343.
30. J.L. Koenig in Analytical Applications of FT-IR to Molecular and Biological

Systems, Ed., J.R. Durig, D. Reidel Publishing Company, Dordrecht, The
Netherlands, 1979, 80.
31. M.K. Antoon, L. D’Esposito and J.L. Koenig, Applied Spectroscopy, 1979, 33, 4,
351.
32. J.L. Koenig, L. D’Esposito and M.K. Antoon, Applied Spectroscopy, 1977, 31, 4,
292.
33. J.L. Koenig and D. Kormos, Applied Spectroscopy, 1979, 33, 349.
34. M.K. Antoon, J.H. Koenig and J.L. Koenig, Applied Spectroscopy, 1977, 31, 6,
518.
35. R.R. Hampton, Rubber Chemistry and Technology, 1972, 45, 3, 546.
36. Y. Tanaka, Nippon Gomu Kyokaishi, 1970, 43, 966.
37. J.L. Binder in Encyclopedia of Spectroscopy, Ed., G.L. Clark, Reinhold

Publishing Company, New York, 1960, 533.
38. J.L. Binder, Rubber Chemistry and Technology, 1962, 35, 57.
39. J.L. Binder, Journal of Polymer Science, 1963, A1, 47.
40. J.L. Binder, Journal of Polymer Science, 1965, A3, 1587.
41. J.L. Binder, Rubber Chemistry and Technology, 1966, 39, 945.
42. R.R. Hampton, Analytical Chemistry, 1949, 21, 923.
43. J.L. Binder, Applied Spectroscopy, 1969, 23, 17.
44. G. Natta and P. Corradini, Angewandte Chemistry, 1956, 68, 615.
116
45. G. Natta, L. Porri, P. Corradini and D. Morero, Chimica el Industria (Milan),

1958, 40, 362
46. G. Natta and P. Corradini, Journal of Polymer Science, 1956, 20, 251.
47. G. Natta, L. Porri, G. Zanini and A. Palvarini, Chimica el Industria (Milan),

1959, 41, 1163.
48. G. Kraus, J.N. Short and V. Thornton, Rubber Plast. Age, 1957, 38, 880.
49. G. Krauss, J.N. Short and V. Thornton, Rubber Chemistry and Technology, 1957,
30, 1118.
50. G. Natta, Chimica el Industria (Milan), 1957, 39, 653
51. D. Morero, F. Ciampelli and E. Mantica in Advances in Molecular Spectroscopy,

Ed., A. Mangini, Pergamon Press, Oxford, 1962, Volume 2, 898.
52. G.S.S. Rao, V.K. Upadhyay and R.C. Jain, Journal of Applied Polymer Science,
1999, 71, 4, 595.
53. W.S. Richardson and A. Sacher, Journal of Polymer Science, 1953, 10, 353.
54. J.L. Binder and H.C. Ransaw, Analytical Chemistry, 1957, 29, 503.
55. F.W. Stavely, J.L. Binder, H.C. Ransaw and co-workers, Industrial and
Engineering Chemistry, 1956, 48, 778.
56. R.A. Sanders and D.C. Smith, Journal of Applied Physics, 1949, 20, 953.
57. J. Hayashi, J. Furukawa and S. Yamashita, Kobunshi Kagaku, 1966, 23, 527.
58. A.S. Wexler, Analytical Chemistry, 1964, 36, 1829.
59. M.A. Post, Journal of Applied Chemistry (London), 1967, 17, 203.
60. R.T. Scheddel, Analytical Chemistry, 1958, 30, 1303.
61. T. Takeuchi, S. Tsuge and Y. Sugimura, Journal of Polymer Science, 1968, A1, 6,
3415.
62. R.M.B. Small, Analytical Chemistry, 1959, 31, 478.
63. M. Tryon, E. Horowitz and J. Mandel, Journal of Research of the National

Bureau of Standards, 1955, 55, 219.
117
64. T. Tanaka, H. Karino and H. Higashi, Nippon Gomu Kyokaishi, 1957, 30, 762.
65. T. Tanaka and H. Higashi, Nippon Gomu Kyokaishi, 1960, 33, 518.
66. H. Feuerberg, D. Gross and A. Zimmer, Kautschuk und Gummi Kunststoffe,

1963, 16, 199.
67. T. Takeuchi and K. Murase, Kogyo Kagaku Zasshi, 1965, 68, 2505.
68. I. Kral, Plaste und Kautstschuk, 1967, 14, 88.
69. D.A. MacKillop, Analytical Chemistry, 1968, 40, 607.
70. J.H. Schult, Kautschuk und Gummi Kunststoffe, 1964, 17, 707.
71. P.J. Corish and M.E. Tunnicliffe, Journal of Polymer Science, 1964, C7, 187.
72. P.J. Corish and M.E. Tunnicliffe, Rubber Chemistry and Technology, 1966, 39,
226.
73. F.J. Karol and W.L. Carrick, Journal of the American Chemical Society, 1961, 83,
2654.
74. R. Hank, Kautschuk und Gummi Kunststoffe, 1965, 18, 295.
75. R. Hank, Rubber Chemistry and Technology, 1967, 40, 936.
76. T.L. Ang and H.J. Harwood, Polymer Preprints of the American Chemical
Society, Division of Polymer Chemistry, 1964, 5, 1, 306.
77. M. Kobayashi, K. Akita and H. Takodoro, Die Makromolekulare Chemie, 1968,

118, 324.
78. J. Furukawa, Y. Iseda, K. Haga and N. Kataoka, Journal of Polymer Science,

1970, A-1, 8, 1147.
79. T. Gossl, Die Makromolekulare Chemie, 1968, 42, 1.
80. G. Natta, A. Valvassori, F. Ciampelli and G. Mazaanti, Journal of Polymer

Science, 1965, A3, 1.
81. H.J. Hagemeyer Jr., and M.B. Edwards, Journal of Polymer Science, 1964, C4,
731.
118
82. A. Zambelli, A. Lety, C. Tosi and I. Pasquon, Die Macromolekulare Chemie,

1968, 115, 73.
83. G. Natta, G. Mazzanti, A. Valvassori, G. Sartori and D. Morero, Chimica el

Industria (Milan), 1960, 42, 125.
84. G. Zerbi, M. Gussoni and F. Ciampelli, Spectrochimica Acta, 1967, A23, 301.
85. D.L. Tabb and J.L. Koenig, Journal of Polymer Science: Polymer Physics Edition,
1975, 13, 6, 1159.
86. J.T. Maynard and W.E. Mochel, Journal of Polymer Science, 1954, 13, 251.
87. J.T. Maynard and W.E. Mochel, Journal of Polymer Science, 1954, 13, 235.
88. M.M. Coleman, P.C. Painter, D.L. Tabb and J.L. Koenig, Journal of Polymer
Science: Polymer Letters Edition, 1974, 12, 10, 577.
89. R.W. Seymour, G.M. Estes and S.L. Cooper, Macromolecules, 1970, 3, 579.
90. M.M. Coleman and P.C. Painter in Applications of Polymer Spectroscopy, Ed.,
E.G. Brame, Academic Press, New York, 1978, Chapter 10.
91. T. Tanaka and T. Yokayama, Journal of Polymer Science, 1968, C23, 865.
92. Polymer Alloys and Blends, Thermodynamics and Rheology, Ed., L.A. Utracki,
Hanser Publishers, Munich, 1989.
93. C.M. Gomez and C.B. Bucknall, Polymer, 1993, 34, 10, 2111.
94. R. Santra, S. Roy and G.B. Nando, Die Angewandte Makromolekulare Chemie,
1993, 213, 7.
95. Functional Polymers, Eds., D.E. Bergbreiter and C.R. Martin, Plenum Press, New
York, 1989.
96. D.J. Burlett and J.T. Lindt, Rubber Chemistry and Technology, 1993, 66, 3, 411.
97. E.H. Farmer, Rubber Chemistry Technology, 1943, 16, 769.
98. L. D’Orazio, C. Mancarella and E. Martuscelli, Journal of Materials Science,

1988, 23, 1, 161.
119
99. A.Y. Coran and R. Patel, inventors; Monsanto Company, assignee; US Patent
4,338,413, 1982.
100. A.K. Sen, B. Mukherjee, A.S. Bhattacharyya, P.P. De and A.K. Bhowmick, Die
Angewandte Makromolekulare Chemistry, 1991, 191, 3206, 15.
101. A.K. Sen, B. Mukherjee, A.S. Bhattacharyya, L.K. Sanghi, A.K. Bhowmick and
P.P. De, Journal of Applied Polymer Science, 1991, 43, 1673.
102. A.K. Kalidaha, A.S. Bhattacharya, A.K. Sen and P.P. De, Die Angewandte
Makromolekulare Chemistry, 1993, 204, 3484, 19.
103. A.K. Sen, B. Mukherjee, A.S. Bhattacharyya, P.P. De and A.K. Bhowmick, Journal
of Applied Polymer Science, 1992, 44, 1153.
104. S.I. Goto, K. Kimura, T. Yamamoto and S. Yamashita, Journal of Applied

Polymer Science, 1999, 74, 3548.
105. J. Park, J.G. Park, C-S. Ha and W-J. Cho, Journal of Applied Polymer Science,
1999, 72, 1177.
106. A.Y. Coran and R. Patel, Rubber Chemistry and Technology, 1983, 56, 1045.
107. R.N. Santra, S. Roy and G.B. Nando, Polymer Plastics Technology and
Engineering, 1994, 33, 1, 23.
108. S. Mohanty, R.N. Santra and G.B. Nando, Advances in Polymer Technology,
1997, 16, 4, 323.
109. S. Mohanty, S. Roy, R.N. Santra and G.B. Nando, Journal of Applied Polymer
Science, 1995, 58, 1947.
110. M. Zuo and T. Takeichi, Polymer, 1999, 40, 5153.
111. B.G. Soares, S. Rodrigo and C. Colombarezzi, Journal of Applied Polymer

Science, 1999, 72, 1799.
112. A. Jha, A.K. Bhowmick, R. Fujitsuka and T. Inoue, Journal of Adhesion Science
and Technology, 1999, 13, 6, 649.
113. H. Mingyi, Z. Hua and L. Jianfeng, Journal of Applied Polymer Science, 1999,
71, 215.
120
114. Q. Fan, J. Fang, Q. Chen and X. Yu, Journal of Applied Polymer Science, 1999,
74, 2552.
115. G. Zhang, Z. Zhang, F. Xie, X.Q. Hu, X. Zuo and X. Chen, Journal of Applied
116. M.A. Dudley and D.A. Smith, Rubber Chemistry and Technology, 1967, 40, 445.
117. R. Alex, P.P. De and S.K. De, Polymer Communications, 1990, 31, 118.
118. R. Alex and P.P. De, Kautschuk und Gummi Kunststoffe, 1990, 43, 11, 1002.
119. R. Alex, P.P. De and S.K. De, Kautschuk und Gummi Kunststoffe, 1991, 44, 333.
120. P. Ramesh and S.K. De, Journal of Materials Science, 1991, 26, 2840.
121. S. Mukhopadhyay and S.K. De, Polymer, 1991, 32, 1223.
122. S. Mukhopadhyay and S.K. De, Journal of Applied Polymer Science, 1999, 42,
2773.
123. S. Mukhopadhyay and S.K. De, Journal of Materials Science, 1990, 25, 4027.
124. S. Mukhopadhyay, P.P. De and S.K. De, Journal of Applied Polymer Science,
1991, 43, 347.
125. A. Roychoudhury and P.P. De, Journal of Applied Polymer Science, 1997, 63,
1761.
126. R.J. Petcavich, P.C. Painter and M.M. Coleman, Journal of Polymer Science:
Polymer Physics Edition, 1979, 17, 165.
127. Y. Okamoto, Polymer Engineering Science, 1983, 23, 222.
128. S. Roy and P.P. De, Polymer Testing, 1994, 13, 419.
129. R. Alex, P.P. De and S.K. De, Polymer, 1991, 32, 2345.
130. A. Roychoudhury, P.P. De, N.K. Dutta, N. Roychoudhury, B. Haidar and A.

Vidal, Rubber Chemistry and Technology, 1993, 66, 230.
131. N.R. Manoj, P.P. De and S.K. De, Journal of Applied Polymer Science, 1993, 49,
133.
121
132. N.R. Manoj and P.P. De, Plastics and Rubber and Composites Processing and
Applications, 1995, 23, 2, 103.
133. P. Ramesh and S.K. De, Polymer, 1993, 34, 23, 4893.
134. P. Ramesh and S.K. De, Polymer Networks and Blends, 1992, 2, 4, 209.
135. A. Mallick, D.K. Tripathy and S.K. De, Polymer Networks and Blends, 1993, 3,
1, 51.
136. T. Bhattacharya and S.K. De, European Polymer Journal, 1991, 27, 1065.
137. C.J. Paul, M.R.G. Nair, N.R. Neelakantan, P. Koshy, B.B. Idage and A.A.
Bhelhekar, Polymer, 1998, 39, 26, 6861.
138. C.J. Paul, M.R.G. Nair, P. Koshy and B.B. Idage, Journal of Applied Polymer
Science, 1999, 74, 706.
139. E. Yilgar, E. Burgaz, E. Yurtsever and I. Yilgor, Polymer, 2000, 41, 849.
140. L. Zha, M. Wu and J. Yang, Journal of Applied Polymer Science, 1999, 73, 2895.
141. C-T. Zhao and M.N. de Pinho, Polymer, 1999, 40, 6089.
142. Z. Wang, D. Gao, J. Yang and Y. Chen, Journal of Applied Polymer Science,
1999, 73, 2869.
143. R.N. Kumar, C.K. Woo and A. Abusamah, Journal of Applied Polymer Science,
1999, 73, 1569.
144. J.B. Donnet and A. Vidal, Advances in Polymer Science, 1986, 76, 103.
145. E.M. Dannenberg, Rubber Chemistry and Technology, 1986, 59, 512.
146. J.V. Hallum and H.V. Drushel, Journal of Physical Chemistry, 1958, 62, 110.
147. J.B. Donnet, Carbon, 1982, 20, 266.
148. E. Papirer, E. Guyon and N. Perol, Carbon, 1978, 16, 133.
149. C.H. Chen, J.L. Koenig, J.R. Shelton and E.A. Collins, Rubber Chemistry and
Technology, 1982, 55, 103.
122
150. A. Roychoudhury, Chemical Interaction of Chlorosulfonated Polyethylene with

Functionalised Polymers and Surface Modified Fillers, IIT Kharagpur, India,
1994. [Ph.D. Thesis]
151. A. Roychoudhury and P.P. De, Journal of Applied Polymer Science, 1995, 55, 9.
152. A. Roychoudhury and P.P. De, Journal of Applied Polymer Science, 1993, 50,
181.
153. M.J. Wang, S. Wolf and J-B. Donnet, Rubber Chemistry and Technology, 1991,
64, 714.
154. A. Roychoudhury, S.K. De, P.P. De, J.A. Ayala and G.A. Joyce, Rubber Chemistry
and Technology, 1994, 67, 662.
155. S. Bandyopadhyay, P.P. De, D.K. Tripathy and S.K. De, Journal of Applied of
156. A.K. Manna, P.P. De, D.K. Tripathy and S.K. De, Rubber Chemistry and
Technology, 1997, 70, 624.
157. S. Bandyopadhyay, P.P. De, D.K. Tripathy and S.K. De, Journal of Applied
158. Ajoy K. Manna, P.P. De, D.K. Tripathy, S.K. De and M.K. Chatterjee, Rubber
159. M.J. Wang, S. Wolf and J-B. Donnet, Rubber Chemistry and Technology, 1991,
64, 559.
160. S. Bandyopadhyay, P.P. De, D.K. Tripathy and S.K. De, Plastics and Rubber and
Composites Processing and Applications, 1996, 25, 7, 327.
161. A. Roychoudhury, P.P. De, N. Roychoudhury and A. Vidal, Rubber Chemistry

162. M.L. Kralevich and J.L. Koenig, Rubber Chemistry and Technology, 1998, 71, 2,
300.
163. A.K. Manna, P.P. De, D.K. Tripathy, S.K. De and D.G. Peiffer, Journal of Applied
123
164. L.J. Murphy, M.J. Wang and K. Mahmud, Rubber Chemistry and Technology,
1998, 71, 998.
165. G.M. Bristow, Proceedings of the Natural Rubber Technology Seminar, Kuala
Lumpur, 1978, 163.
166. R.W. Keller and H.L. Stephens, Rubber Chemistry and Technology, 1982, 55,
161.
167. I.R. Gelling and N.J. Morrison, Rubber Chemistry and Technology, 1984, 58,
243.
168. N.R. Kumar, S. Roy, B.R. Gupta and A.K. Bhowmick, Journal of Applied
169. L. Ulren, T. Hgertberg and H. Ishida, Journal of Adhesion, 1990, 31, 117.
170. F.L. Magnus and G.R. Hamed, Rubber Chemistry and Technology, 1991, 64, 65.
171. T. Bhattacharya, B.K. Dhindaw and S.K. De, Journal of Adhesion, 1992, 34, 45.
172. T. Bhattacharya, B.K. Dhindaw and S.K. De, Journal of Adhesion, 1991, 39, 207.
173. T. Bhattacharya, D.K. Tripathy and S.K. De, Journal of Adhesion Science and
Technology, 1992, 6, 1165.
174. N.R. Manoj and P.P. De, Journal of Adhesion, 1993, 43, 199.
175. S. Roy, B.R. Gupta and T.K. Chaki, Kautschuk und Gummi Kunststoffe, 1993,
46, 293.
176. S-D. Li, H-P. Yu, Z. Ping, C-S. Zum and P-S. Li, Journal of Applied Polymer
Science, 1999, 75, 1339.
177. G.A. Skarja and K.A. Woodhouse, Journal of Applied Polymer Science, 2000, 75,
1522.
178. D.B. Dormagen, Kautschuk und Gummi Kunststoffe, 1986, 39, 12, 1165.
179. D. Coz and K. Baranwal, Rubber World, January, 1999, 30.
124
Application of Infrared Spectroscopy to Characterise Chemically Modified Rubbers …
4
Application of Infrared Spectroscopy to
Characterise Chemically Modified Rubbers and
Rubbery Materials
Nikhil K. Singha and Prajna P. De
4.1 Introduction
Chemical modification of polymers continues to be an active field of research [1-5]. It is

a common means of changing and optimising the physical, mechanical and technological
properties of polymers [5-7]. It is also a unique route to produce polymers with unusual
chemical structure and composition that are otherwise inaccessible or very difficult to
prepare by conventional polymerisation methods. For example, hydrogenated nitrile
rubber (HNBR) which has a structure which resembles that of the copolymer ethylene
and acrylonitrile, is very difficult to prepare by conventional copolymerisation of the
monomers. Polyvinyl alcohol can only be prepared by hydrolysis of polyvinyl acetate.
Most of the rubbers or rubbery materials have unsaturation in their main chain and/or
in their pendent groups. So these materials are very susceptible towards chemical reactions
compared to their saturated counterparts.
Infrared spectroscopy is one of the most important tools used to characterise the chemical
structure, composition and microstructure of different polymers [8-10]. In earlier chapters,
the principles and applications of infrared (IR) spectroscopy in the characterisation of
rubbers have been discussed. This chapter describes how IR spectroscopy can be used to
characterise different types of chemically modified elastomers.
4.2 The Infrared Spectra of Commonly Used Diene Rubbers
All diene rubbers may have few isomers depending on the addition of diene monomers.
Different modes of addition are shown in Scheme 4.1. ‘R’ may be different groups, e.g.,
for chloroprene rubber (CR), R = Cl; for polyisoprene or natural rubber (NR) R = CH3).
125
Scheme 4.1 Microstructures in the diene elastomers
Butadiene rubber (BR) is the simplest case with R = H. It shows strong bands at 1625,
1643 and 1666 cm-1 for C=C stretching (n), for pendent -CH=CH2 (vinyl), cis 1, 4 BR
and trans 1,4 BR, respectively. It also shows characteristic peaks at 967, 910 and 740-
732 cm-1 associated with wagging of the vinylic C-H bonds for trans 1,4, pendent and
cis 1,4 double bonds, respectively [11].
However, these values may change for other diene rubbers which have analogous structure
of BR, e.g., polyisoprene [12, 13] and CR [14] depending on the ‘R’ group in Scheme 4.1.
According to Urey-Bradley field theory it is assumed that the important forces in molecular
vibration act along chemical bonds and also includes interaction terms between adjacent
nonbonded atoms. Such a force field can thus take into account attraction and repulsion
between adjacent nonbonded atoms, which is known as field effects [15]. Slight deviation
in different frequencies in some elastomers from those of BR may be ascribed to the field
effect. The characteristic frequencies of different bands of various diene-based rubbers
are shown in Table 4.1. It is interesting to note how the position and intensity of different
absorption bands change during chemical modifications.
4.3 Hydrogenation
Hydrogenation is an important method of chemical modification of elastomers. Because

of the absence of carbon-carbon unsaturation, hydrogenated elastomers have good
resistance to oxidative and thermal degradation, improved weatherability and good
resistance towards chemicals and fluids [5-7]. Nitrile rubber (NBR) is a specialty rubber,
and because of its oil resistance properties, it has been used in oil-wells and the automotive
industry. Hydrogenation of NBR has been studied extensively because of its technological
importance [16-19].
It was observed that on hydrogenation there was no change in the position or intensity
of the peak at 2236 cm-1 (ν of -CN in acrylonitrile content) (Figure 4.1).
126
Table 4.1 Assignments of different characteristic bands in infrared

spectroscopy of elastomers
Elastomers Characteristic Assignments Ref.
Peaks (cm-1)
BR 1660 C=C stretching of cis —CH=CH— [11]
1640 C=C stretching of cis CH2=CH—
1470 CH2 deformation
1418 CH in-plane deformation of CH2=CH-
1408 CH in-plane deformation of CH=CH—
995 CH out-of-plane deformation
(wagging) of CH2=CH—
967 CH wagging of trans –CH=CH-
910 CH2 wagging of CH2=CH- unit
IR 1665 C=C stretching of -CH=C(CH3)- [12, 13]
1645 C=C stretching of CH2=C(CH3)-
1152 C-CH3 of trans -CH=C(CH3)-
1130 C-CH3 of cis -CH=C(CH3)-
890 CH wagging of CH2=C(CH3)-
845 CH wagging of trans CH=C(CH3)-
840 CH wagging of cis -CH=C(CH3)-
CR 1660 C=C stretching in trans CR [14]
1653 C=C stretching in cis CR
847 CH wagging in cis CR
822 CH (of =CH-) wagging in trans CR
720 —C-Cl stretching
NBR 2236 -CN stretching of acrylonitrile (ACN) [16-22]
970 CH wagging of trans -CH=CH- unit
750 CH wagging of cis -CH=CH- unit
SBR 1002 trans CH wagging of vinyl group [27-32]
980 CH wagging of trans -CH=CH- unit
750 CH wagging of cis -CH=CH- unit
700 Characteristic for polystyrene
127
Figure 4.1 Hydrogenation of NBR (34% acrylonitrile (ACN)) and hydrogenated NBR
with 22.7, 11.0 and 0.6% mol unhydrogenated butadiene units
Reprinted from [21] with permission from Huthig Publications, Heidelburg, Germany,
Copyright 1989
The peaks at 970, 920 and 750 cm-1 which are due to ω(CH2) of 1,4 trans, ω(CH2) of
pendent and ω(CH) of 1,4 cis content slowly disappear as the degree of hydrogenation
increases. A new peak at 723 cm-1 appears and its intensity increases with increasing
degree of hydrogenation. This peak is due to rocking of (CH2)n>4 in the hydrogenated
NBR. Marshall and co-workers [20] and Bruck [21,22] have studied the infrared spectrum
of HNBR (Figure 4.1).
The degree of hydrogenation in HNBR can be calculated using the following equation:
Degree of hydrogenation (mol%) = 100 - C(BR) x 100

C(BR) + C(HBR)
where C(BR) is the relative amount of carbon-carbon unsaturation remaining in HNBR

and C(HBR) is the relative amount of methylene group (CH 2) n formed during
hydrogenation of olefin group in NBR. The molar concentrations of different components
C(I) can be calculated by these equations,
128
1
C(ACN) =
ΣAi
A(970) 1
C(BR) = ⋅
K(970) ΣAi
A(723) 1
C(HBR) = ⋅
K(723) ΣAi
Where A is the normalised value of absorbances calculated using A(2236) as standard.

K is the calibration factors for the absorption at different absorbance maxima. Bruck
[21,22] calculated different ‘K’ values (K970, K723, K920, etc.), by using IR spectra as well
as nuclear magnetic resonance (NMR) of different NBR and HNBR with different ACN
content. Different spectral parameters of NBR and HNBR are shown in Table 4.2.
Bhattacharjee and co-workers [23] hydrogenated liquid carboxylated NBR (XNBR) using
homogeneous palladium and rhodium complexes. They observed that the absorption
band at 1730 cm-1 (>C=O of free -COOH) and 1705 cm-1 (H bonded -COOH) remained
unchanged when palladium catalysts were used. But the intensity of these peaks decreased
on hydrogenation using a rhodium catalyst, which implied that there were substantial
amounts of decarboxylation.
The simplest diene rubber for hydrogenation is BR. The structure of the hydrogenated
BR (HBR) depends on the different microstructures present in BR. Hydrogenation of BR
with high 1,4 structure converts this elastomeric polymer into a tough semicrystalline
Table 4.2 Assignments of different peaks in IR spectra of NBR and HNBR

Peak cm-1 Assignments Absorption Baseline drawn Reference
factor(k) between point
2236 -CN stretching of ACN 1.0 2280-2190 [21-23]
970 CH wagging of 1,4 trans 2.30 1030-930

(-CH=CH-)
920 CH2 wagging of 1,2 unit 2.24
(-CH=CH2-)
750 CH wagging of 1,4 cis (CH=CH) 0.36 860-660
723 (CH2)n rocking of HNBR (n >4) 0.25 840-680
129
polymer which resembles the properties of low density polyethylene (LDPE) [24].
Hydrogenation of BR having predominantly 1,2 content leads to a polymer with
thermoplastic elastomeric properties.
The IR spectrum of cis-1,4 BR (98% cis C=C) and HBR is shown in Figure 4.2.
The characteristic bands for the C=C unsaturation at 1640 cm-1 (C=C stretching for cis-
alkene), 1005 cm-1 (a small peak for trans CH wagging), 920 cm-1 (a weak peak for CH2
wagging for a vinyl group) and 750 cm-1 (a strong peak for CH wagging in cis-C=C)
disappear confirming quantitative hydrogenation. Mohammadi and Rempel [25]
hydrogenated BR containing 90% 1,2 units, which shows a strong peak at 910 cm-1
because of CH2 wagging of the 1,2 units. The disappearance of this peak and other
characteristic bands for unsaturation (1640, 992 and 970 cm-1) signify quantitative
hydrogenation in HBR. The study of hydrogenation using RhCl(PPh3)3 showed that 1,2
addition units were selectively hydrogenated over 1,4 addition units present in BR [24,25].
Hydrogenation of telechelic BR [26] was done either by diimide, a chemical generated
Figure 4.2 IR spectra of cis-1,4 BR and hydrogenated BR (HBR)

Reprinted from X. Guo and G.L. Rempel, Journal of Molecular Catalysis, 1990, 63, 279,
Figure 3, Copyright 1990, with permission from Elsevier Science
130
by a thermal decomposition of p-toluene sulfonyl hydrazide or in two steps via a partially

saturated intermediate prepared by prehydrogenation up to 95% using gaseous hydrogen
and a Ziegler-Natta catalytic system, followed by diimide hydrogenation up to 100%. It
can be seen that after 100% hydrogenation, all bonds pertaining to C=C bonds (910,
968, 995 and 1640 cm-1) disappeared.
Hydrogenation of the random and block copolymers of styrene and butadiene have been
reported in the literature [27-32]. De Sarkar and co-workers [30-32] characterised
hydrogenated styrene butadiene elastomer (HSBR) with IR spectroscopy. SBR shows
bands at 699 cm-1 due to an aromatic unit, at 767 cm-1 due to CH2 wagging in 1,4 cis,
unit at 909 cm-1 due to CH2 wagging in the 1,2 unit, and 967 cm-1 due to trans CH
wagging in the butadiene content (Figure 4.3). As the degree of hydrogenation increases,
the intensity of these peaks decreases and new peaks at 1450 cm-1 (-CH2 deformation in
HSBR) and 723 cm-1 (-CH2 rocking vibration (CH2)n>4 in HSBR) appear in the spectra of
HSBR (Figure 4.3a). The peaks due to aromatic styrene units are not changed, since the
styrene units are not hydrogenated.
So the peak at 699 cm-1 can be used for normalisation of other peaks in IR spectra of
HSBR or partially hydrogenated HSBR. De Sarkar and co-workers [32] hydrogenated
carboxylated SBR (XSBR containing 1-4% methacrylic acid) with diimide reduction
technique in the latex stage using hydrazine hydrate, H2O2 and Cu2+ as catalyst. They
observed selective hydrogenation of XSBR without affecting COOH groups. The infrared
spectra of XSBR show a peak at 1694 cm-1 due to >C=O stretching vibration in COOH
(Figure 4.3b).
The peak shape and the frequency of >C=O absorption suggested that most of the COOH
groups in XSBR were hydrogen bonded. The broad carbonyl peak for HXSBR can be
deconvoluted into three peaks, one at original 1694 cm-1 and the other two approximately
at 1720 and 1680 cm-1 (Figure 4.3b). The shift towards higher frequency was due to the
reduction of strength of hydrogen bond in HXSBR whereas the peak shift towards lower
frequency to 1680 cm-1 was ascribed to the ionic aggregates or clusters involving COOH
groups and Cu2+ ions [32].
Polyisoprene can have few isomers according to different mode of addition as shown in
Scheme 4.1. Cis 1,4 polyisoperene [12,13] shows a peak at 840 cm-1 whereas the trans
isomer shows a peak at 845 cm-1 and 3,4 polyisoprene shows a characteristic peak at
890 cm-1. NR is cis 1,4 polyisoprene. Singha and co-workers [33] hydrogenated NR and
observed that quantitative hydrogenation of NR led to an alternate copolymer of ethylene
and propylene. The peaks at 840 and 1663 cm-1 due to the double bond disappeared
[33] and a new peak at 735 cm-1 emerged due to (CH2)3 groups in hydrogenated NR [34]
(HNR). Burfield [35] carried out the solid state hydrogenation of natural rubber in order
131
Figure 4.3a IR spectra of SBR and HSBR (70% and 94%) hydrogenation
Reprinted from M. De Sarkar, P.P. De and A.K. Bhowmick, Journal of Applied Polymer
Science, 1997, 66, 1151, Figure 5, with permission of John Wiley and Sons, Copyright 1997
to increase the ageing properties. Like NR, epoxidised natural rubber (ENR) was
hydrogenated [36] in the presence of palladium acetate catalyst in a high pressure autoclave
at a temperature of 323 K with hydrogen gas under a pressure of 2.7 MPa according to
Scheme 4.2 [36]:
132
Figure 4.3b IR spectra of (a) XSBR and HXSBR with (b) 60% saturation (c) 80%
saturation
Reprinted from M. De Sarkar, P.P. De and A.K. Bhowmick, Polymer, 2000, 41, 907,
Figure 3, Copyright 2000, with permission from Elsevier Science
133
Scheme 4.2 Hydrogenation of ENR

Reprinted with permission from S. Roy, S. Bhattacharjee and B.R. Gupta, Journal of
Applied Polymer Science, 1993, 49, 375. Copyright 1993, John Wiley and Sons, Inc.
The peak due to the 1,4 >C = C moiety, i.e., 1665 cm-1, is for reduced hydrogenated ENR
(HENR). The peaks for 3,4 and 1,2 structures, at 1645 cm-1 and 1638 cm-1 disappeared
in the case of HENR, signifying that they are more prone towards hydrogenation. The
peak height of 1450 cm-1 corresponding to the >CH2 group increased, while the peak
height at 870 cm-1 (for epoxy group) remained unchanged.
CR may have four structural isomers, the trans 2-chloro-2-butenylidene-1,4, cis 2-chloro-
2-butenylidene-1,4 arising from 1,4 polymerisation and two other isomers arising from
1,2 and 3,4 polymerisation [14]. Cis and trans isomers show peaks at 1653 and 1660 cm-1
for C=C stretching, whereas the peaks due to CH wagging appear at 822 and 847 cm-1
for cis and trans isomers, respectively. Weak bands at 925 and 833 cm-1 appear due to
CH wagging from 1,2 and 3,4 addition units in polychloroprene [14]. Singha and co-
workers [37] hydrogenated CR using RhCl(PPh 3) 3 as a catalyst. They observed
dehydrochlorination and hydrogenation in hydrogenated CR (HCR). The characteristic
bands [14] at 1653, 1660, 925, 847, 833, 822 cm-1 disappeared and a new peak at
725 cm-1 emerged due to CH2 rocking which was similar to polyethylene. Hydrogenation
of CR using RhCl(PPh3)3 yielded a product with a predominantly linear polyethylene
sequence with minor proportions of vinyl chloride sequence [37].
134
4.4 Halogenation
Halogenation of elastomers profoundly alters its adhesion and cure characteristics.

Chlorination is carried out with either gaseous chlorine or sulfuryl chloride. The polymers
are dissolved in chloroform or dispersed in a swollen state in chloroform. Chlorine gas is
bubbled through the polymer solution or dispersion.
Cis and trans 1,4 BR have been chlorinated in chloroform using chlorine gas [38, 39].
Figure 4.4 shows the infrared spectra of chlorinated cis and trans BR.
Figure 4.4 IR spectra of chlorinated cis and trans 1,4 polybutadiene

Reprinted from N. Murayama and Y. Amagi, Journal of Polymer Science, Part B, 1966, 4,
119, Figure 2, Copyright 1966. Reprinted by permission of John Wiley & Sons, Inc.
135
On chlorination the strong band at 967 cm-1 in original trans BR which is due to C-H
out of plane vibration of –C=CH for 1,4 trans configuration, weakens and slowly
disappears as the chlorination proceeds. The band at 1090 cm-1 also disappears on
chlorination. As the chlorination proceeds, the intensity of the band at 1250 cm-1, which
corresponds to the band of polyvinyl chloride (PVC) at 1254 cm-1 of the C-H bending,
gradually increases (Figure 4.4b). On the other hand, the bands at 724 cm-1 and 1643 cm-1 in
cis BR, which are due to C-H wagging and C=C stretching, either decrease in intensity or
disappear. A weak band around 965 cm-1 appears again, which indicates the conversion
of cis isomer to trans isomer during chlorination [38, 39]. Chlorinated trans BR shows a
peak at 650 cm-1 and chlorinated cis BR shows peaks at 680 cm-1 and 590 cm-1 which are
due to the C-Cl stretching mode. According to Murayama and Amagi [38], during
chlorination of trans and cis BR, addition takes place first and then substitution occurs.
Chlorination of 1,4 BR leads to elastomers which resemble head to head PVC [38].
Bromination of BR can be carried out in n-heptane or chloroform with a solution of
bromine in an organic solvent. On bromination the intensity of characteristic olefinic
bands decrease, the peaks at 1235 cm-1 (due to CH2 bending in brominated BR) and at
1225 cm-1 (due to CHBr) increase [40]. Brominated BR shows bands at 785 and 550 cm-1
due to CBr stretching mode. Assignments of different peaks in chlorinated and brominated
BR have been described in Table 4.3.
Halogen can be introduced into elastomers by reactions with chlorocarbene. According

to Konietzny and Biethan [41], dichlorocarbene can be added to BR to form elastomers
with dicyclopropane rings. Dichlorocarbene was prepared in situ from chloroform and
aqueous caustic soda using phase transfer catalysis. The modified elastomer showed a
strong band at 810 cm-1 characteristic for CCl2 wagging vibration. Ramesan and Alex
[42] modified SBR with dichlorocarbene prepared by alkaline hydrolysis of chloroform
using cetyltrimethyl ammonium bromide as a phase transfer agent. Dicholoro carbene
modified SBR showed improved resistance towards heat, flame and solvents. The modified
SBR showed characteristic C-Cl absorption peaks at 806 and 1059 cm-1 which are due
to cyclopropane ring [43]. It was observed that CCl2 initially reacted with the cis double
bond. With the increase of reaction time, the intensity of absorption at 968 cm-1 (due to
the trans isomer) and at 1653 cm-1 (due to the cis isomer) slowly decreased. The peak at
698 cm-1 remained unaffected implying that aromatic ring did not take part in the reaction.
The absorption at 910 cm-1 (due to vinyl 1,2 content) remained almost unaffected during
modification. Pinazzi and Levesque [44] reported modification of 1,4 polyisoprene with
dichlorocarbene. As the reaction time increased, the intensity of the absorption band at
835 cm-1 decreased and a new peak emerged at 820 cm-1 due to C-Cl absorption and at
1025 cm-1 due to cyclopropane ring [43].
Okamoto and co-workers [45] reported γ-ray induced addition reaction of carbon
tetrachloride (CCl4) to the vinyl group of syndiotactic 1,2 BR. They observed that addition
136
Table 4.3 Different characteristic bands in various chemical modification

of elastomers
Modification Characteristic Assignments Ref.
Bands (cm-1)
Hydrogenation 725, 735 -CH2- rocking of (-CH2-)n (n>4) -(CH2) [16-37]
3-
in polyisoprene
Chlorination 1250, 650 C-H bending in chlorinated BR, C-Cl [38, 39]
680 and 590 stretching in trans BR, C-Cl stretching
in cis BR
Bromination 1235 1225 C-H bending in brominated BR, C-H [40]
785 and 550 bending in CHBr, C-Br stretching
Cyclisation 3040, 1175, CH stretching in trisubstituted double [51, 52]
820 bond due to cyclisation in BR, due to [57, 58]
terminal double bond in cyclised
polyisoprene
Hydrosilylation 1410 and Due to –SiCH2- [61]
1235
Hydroboration and 3340 O-H stretching in hydroxylated BR [64 ]
hydroxylation
1050 C-O stretching in hydroxylated BR
Hydroformylation 2700 C-H stretching in –CHO [65-67]
1725 >C=O stretching in -CHO
Oxidation 1260 and C-O-C ring vibration in epoxidised BR [68-70]
1380
1240 C-O-C ring vibration in epoxidised IR [71 ]
1720 >C=O in polyketones [72, 73]
1710 >C=O in –COOH
3500 O-H stretching
Phosphonylation 2525-2725 -OH(P=O)OH [74 ]
1150-1362 Free P=O
1087-1261 Hydrogen-bonded P=O
~ 800 (P-O-)-C
137
Table 4.3 Continued

Modification Characteristic Assignments Ref.
Bands (cm-1)
Sulfonation 1176 O=S=O stretching in –SO3H [75-82 ]
881 S-OH stretching
Ionomer formation 150 to 400 Depending on cations in sulfonated [83-85]
ionomers [87-90]
~1200 Asymmetric stretching of –SO3
1020 Symmetric stretching of –SO3
1550 >C=O stretching in carboxylate
ionomers (depends on nature cation and
ionic structure)
Weathering and ~3400 O-H stretching [109 -
degradation 115]
1700-1720 >C=O stretching
[115]
1175 Due to terminal >C= stretching in TPEE
1642, 1533 Due to –CONH2– of polyamide
of CCl4 to 1,2 BR is anti-Markovnikov and it leads to increase in molecular weight (Mn)

due to crosslinking. On addition of CCl4, vinyl bands at 3080, 1840, 1640, 1415, 995
and 910 cm-1 decrease, CH2 bands at 2990-2860 and 1460 cm-1 increase with increase in
molecular weight. Furthermore, intensity of the band at 680 cm-1 due to C-Cl stretching
increases as the addition of CCl4 proceeds.
Braun and co-workers [46] reported that reductive dechlorination of PVC using tri-n
butyl tinhydride (n-Bu3SnH) leads to vinylchloride-ethylene copolymers. Copolymers
were characterised by casting film from tetrahydrofuran (THF) solution in a potassium
bromide disk. It is thought that the IR absorption peak at 750 cm-1 is ascribed to the
(CH2)3 sequences and the peak at 720 cm-1 is due to the (CH2)n>5 sequences. As the
dechlorination starts, the peak at 750 cm-1 due to CH2 sequences appears and intensifies
with reaction time. The intensity of the peaks at 690 and 615 cm-1 due to the C-Cl
stretching vibration slowly decreases. If the reduced PVC contains more than 46 wt%
chlorine, only the absorption peak at 750 cm-1 appears in the IR spectra. If the chlorine
content is less than 46 wt%, the peak at 720 cm-1 weakens. In this case the (CH2)n>5
sequences become more prominent than the -(CH2)3 –sequences [46].
138
4.5 Isomerisation
In 1957, Golub [47] first reported the cis-trans isomerisation of an unsaturated

macromolecule. He showed the successful transformation of the cis 1,4 BR into
corresponding trans units by means of ultraviolet irradiation in the presence of a suitable
sensitiser, which may be any one of a wide variety of organic bromine or sulfur compounds.
In the photoisomerisation of BR or other diene elastomers, there is direct excitation of π-
electrons of the double bonds to an antibonding state in which free rotation and hence
geometrical interconversion can occur.
When BR is irradiated in the presence of some sensitisers, such as allyl bromide or phenyl
sulfide, the band at 710 cm-1 due to the cis isomer slowly weakens as the irradiation time
increases. On the other hand, the intensity of the peak at 970 cm-1 which is due to the
trans isomer slowly increases. From an initial cis/trans ratio of 57/43, BR was isomerised
with allyl bromide in nitrogen to a structure with a cis/trans ratio of 10/90. In this case
the band at 1450 cm-1 which is due to CH2 bending was taken as internal reference.
No side reactions like cyclisation or degradation were observed in this isomerisation

process. Under the similar reaction condition this isomerisation could not be induced in
NR [47]. The methyl group attached to the double bond appears to have a stabilising
influence on the given configuration. There are a variety of catalytic methods for cis-
trans isomerisation of 1,4 BR. They include heating the polymer with nitrogen dioxide
or thiol acid or sulfur dioxide or some metal catalyst based on rhodium and iron. In this
case the cis isomer forms a π-complex with the catalyst and subsequent release of the
catalyst regenerates the more stable trans form. Golub and Rosenburg [48] reported
photoinduced cis-trans isomerisation of the pendent 1,2 bonds in 1,2 – poly(cis-1,4
hexadiene) (CHD). On UV irradiation the bands at 700 cm-1 and 1666 cm-1 (due to
-CH=CH-) in CHD weaken and the intensity of the peak at 961 cm-1 due to the trans
isomer increases.
4.6 Cyclisation
Cyclisation of unsaturated elastomers has been the subject of interesting research.

Cyclisation yields hard resinous products which have commercial importance and are
designated as cyclised rubber [49, 50]. Cyclisation can be carried out with cationic,
radiation induced, photo-induced or by thermal methods. Among all these methods
cationic cyclisation has been extensively reported probably because there are less side
reactions. The generally accepted mechanism of cyclisation of 1,3 diene elastomers is
shown in Scheme 4.3a.
139
Scheme 4.3a Cyclisation of diene elastomers
Scheme 4.3b Monocyclic structure of 1,4 polyisoprene

Reprinted with permission from M. Stolka, J. Vodehnal and I. Kossler, Journal of Polymer
Science, Part A, 1964, 2, 3987. Copyright 1964, John Wiley & Sons, Inc.
The number of rings in the cyclised polymers depends on conditions of the cyclisation
and on the nature of the polymers.
Cis 1,4 BR has been cyclised with H2SO4 [51], TiCl4 [51] and alkyl aluminium halide in
combination with an organic halide [52]. Cationic cyclisation of BR is more difficult
with respect to polyisoprene rubber (IR) because of the lower stability of the secondary
carbonium ion in BR compared to the highly stable tertiary carbonium ion in IR. Presence
of the less stable carbonium ion in BR makes it difficult to control the side reactions and
hence either chain scission or crosslinking occurs. Analysis of the infrared spectrum
shows that the bands characteristic of cis 1,4 polybutadiene at 740, 1315 and 3008 cm-1
(cis 1,4 units), 910 and 990 cm-1 (1,2 units) and 967 cm-1 (trans 1,4 units) disappear and
two new peaks emerge at 820 and 3040 cm-1. The peak at 820 cm-1 can be ascribed to
out-of-plane deformation vibration and the peak at 3040 cm-1 is due to valency vibrations
of the tri-substituted double bond, C=CH-. These bonds are created in the ring when BR
undergoes intramolecular cyclisation. Figure 4.5 shows IR spectra in the 700-1100 cm-1
region of BR with different degrees of unsaturation. It shows that the intensity of the
band at 740 cm-1 (cis 1,4 BR) falls off sharply and a new peak appears at 820 cm-1 whose
intensity increases, as the degree of cyclisation increases.
140
Figure 4.5 IR spectra of cis 1,4-polybutadiene with (1) 97% unsaturation; (2) cyclised
polybutadiene with 58%; (3) 36% and (4) 30% unsaturation
Reprinted with permission from V.S. Shagov, A.I. Yakubchik and V.N. Podosok, Orskaya,
Polymer Science USSR, 1969, 10, 1092. Copyright Pergamon Press, UK, 1969
The intensity of the band at 967 cm-1 (due to the trans 1,4 unit) increases a little in the
initial stage and then it decreases. It indicates that in addition to cyclisation, cis-trans
isomerisation occurs at the initial reaction stage. Figure 4.5 shows a fully cyclised BR
(unsaturation 30%) though there is still some absorption at 740 cm-1 and 967 cm-1.
Although it is in fully cyclised BR, there is still a small proportion of segments with
double bonds, which cannot take part in further cyclisation because the neighbouring
double bonds have already reacted.
Information about the amount and the structure of the cyclic units present in cyclised BR
(CBR) can be obtained by different spectroscopic methods. C13 NMR yields the
quantitative information about unsaturation in CBR, i.e., linear and cyclic double bonds.
IR spectra provide information on the residual linear unsaturation. According to the
cyclisation mechanism it is assumed that one polycyclic sequence gives rise to one
unsaturation. The fraction of polycyclic units and the average number of cyclohexane
141
rings present in each polycyclic sequence can be calculated using the equation explained
in [53-56]. Priola and co-workers [52, 53] reported the average length of polycyclic
sequences to be between 2.5 and 4 depending on the extent of cyclisation in BR.
Quantitative IR and NMR spectra confirm the presence of the cyclic structures in some
polymers [54].
Cyclisation of cis-1,4 (Hevea), trans-1,4 (balata) and 3,4 polyisoprene has also been
reported using H2SO4, TiCl4 and p-toluenesulfonic acid as catalysts [57, 58]. Stolka and
co-workers [57] studied cyclisation of different isomers of polyisoprene using H2SO4
and characterised them with IR spectroscopy. The intensity of the bands at 1315 cm-1
due to C=C-H vibration in –C(CH3)=CH- groups in cis form and at 1130 cm-1 assigned
to the same group decreased in 1% H2SO4 and disappeared in concentrated H2SO4. The
absorption of CH2 groups near 1455 cm-1 shifted to 1465 cm-1.
When the cyclisation was complete, a very broad and strong band appeared at 1175 cm-1.
Cyclisation leads to monocyclic, bicyclic or polycyclic structures. The monocyclic structure
is shown in Scheme 4.3b. The peak at 1175 cm-1 is due to the terminal double bonds. It
has been confirmed by studying the IR spectra of model cyclopolymers [58, 59]. In the
completely cyclised state, new peaks appeared at 810, 885, 1040, 1175 and 1265 cm-1
together with a very weak band at 1200 cm-1. In the range of C-H stretching vibrations
around 3000 cm-1 the absorption intensity at 2750 and 3040 cm-1 decreased and new
bands appeared at 2685 and 3060 cm-1. During cyclisation the characteristic peaks of
trans 1,4 polyisoperene at 1330 and 1150 cm-1 slowly decreased and then disappeared in
the fully cyclised polymer. As in the case of cyclised cis isomer, a new peak emerged at
1175 cm-1. It was further observed that cyclised 1,4 cis and 1,4 trans isomers had identical
IR spectra in the whole range 400 to 4000 cm-1.
4.7 Hydrosilylation
Hydrosilylation is also a very useful chemical modification which leads to silane modified
polymers with special properties [60-62]. Silane modified polymers have improved
adhesion to fillers and better heat resistance. It also acts as a reactive substrate for grafting
or moisture catalysed room temperature vulcanisation. Guo and co-workers [61] carried
out catalytic hydrosilylation of BR using RhCl(PPh3)3 as the catalyst. Hydrosilylation
reactions followed anti-Markovnikov rule as shown in the Scheme 4.4.
The infrared spectra of BR and hydrosilylated BR are shown in Figure 4.6.
As the degree of hydrosilylation increases, the absorbance at 3100 and 910 cm-1 due to
the terminal C=C bonds slowly disappear. There are new absorbances at 1235 cm-1 due
to CH2 scissoring for the SiCH2 structure in the hydrosilylated polymer. Hydrosilylation
142
Scheme 4.4 Hydrosilylation in diene elastomers

Reprinted with permission from X. Guo and G.L. Rempel, Macromolecules, 1990, 23,
5047. Copyright 1990, American Chemical Society
Figure 4.6 (a) IR spectrum for PBD (A); (b) IR spectrum for PBD(A) + His (C2H5)3
(1) product
Reprinted with permission from X. Guo, R. Farwaha and G.L. Rempel, Macromolecules,
1990, 23, 5047, Figure 2. Copyright 1990, American Chemical Society
143
of SBR proceeds with greater difficulty than that of BR. Presence of the aromatic ring in
the copolymer chain creates a different electronic environment as well as the steric
environment for the reaction centre in the polymer chain. With hydrosilylation the peaks
at 915 cm-1 (=CH2 wagging) decreased and new peaks at 1220 cm-1 and 1005 cm-1 emerged
due to newly introduced –SiCH2- structure. Hydrosilylation of NBR leads to Markovnikov
addition products. Presence of -CN groups in NBR induces different electronic
environment [62].
4.8 Hydroboration
The hydroborated polymers are valuable intermediates that can be converted to a variety
of functional polymers [63, 64]. Chung and co-workers [64] carried out hydroboration
of 1,2 BR, 1,4 BR and 1,4 polyisoprene using 9-borabicyclo [3,3,1] nonane (9BBN) and
then hydroxylated the polymers with NaOH/H2O2. The relative reactivity of the various
polydiene is 1,2 BR>1,4 BR> 1,4 polyisoprene in the ratio 1000:167:125. As the degree
of hydroboration and subsequent hydroxylation increase, all absorption peaks for vinyl
groups at 3065, 1635, 990 and 905 cm-1 decrease. Two strong absorption bands emerge
at 3340 cm-1 and 1050 cm-1 due to ν(OH) and ν(CO), respectively, in hydroxylated BR.
4.9 Hydroformylation
A hydroformylation reaction in diene polymers introduces a formyl group which is an

extremely reactive functional group. Sibtain and Rempel [65] carried out hydroformylation
of SBR using HRh(CO)(PPh3)3 and reported anti-Markovnikov addition product.
Hydroformylation takes place preferentially in the 1,2 unit. As the degree of
hydroformylation increases new absorption bands appear at 1724 cm-1 due to ν(C=O)
and at 2700 cm-1 due to ν(C-H) in CHO. Bhattacharjee and co-workers [66] carried out
hydroformylation of NBR and observed new peaks at 1724 and 2700 cm-1 which are
characteristics of CHO groups.
Sanui and co-workers [67] carried out hydroformylation of polypentenamer (PPA). They
converted hydroformyl group of the modified PPA to the aldoxime and subsequently to
nitrile derivatives (Scheme 4.5). Then they carried out hydrogenation to convert the
amorphous, unsaturated nitrile derivative to the crystalline saturated polymer.
Figure 4.7 and Scheme 4.5 show the sequential chemical modification of PPA to nitrile
containing hydrogenated PPA. Disappearance of peaks at 1724 cm-1 and 2700 cm-1 in
PPA-CN indicates complete transformation of CHO, while emergence of the peak at
2240 cm-1 indicates presence of CN groups in PPA-CN.
144
Scheme 4.5 Chemical modification in polypentenamer (PPA)

Reprinted with permission from K. Sanui, W.J. MacKnight and R.W. Lenz.,
Macromolecules, 1974, 7, 952. Copyright 1974, American Chemical Society
Figure 4.7 IR spectra of PPA-CHO, PPA-CN and HyPPA-CN

Reprinted with permission from K. Sanui, W.J. MacKnight and R.W. Lenz,
Macromolecules, 1974, 7, 952, Figure 1. Copyright 1990, American Chemical Society
145
4.10 Oxidation
Oxidation induces polarity in the diene elastomers and hence increases resistance to
hydrocarbon solvent. Oxidation of diene elastomers with peracids leads to epoxidation
[68]. There are two types of epoxy rings depending on the chain microstructures. Epoxy
groups may be in the main chain (internal groups) or in the pendent groups (external
groups) [68]. Epoxidised cis-1,4 BR [69] shows characteristic peaks at 1260 and at
1380 cm-1 for ring stretching vibration (ν COC) and at 800 and 885 cm-1 for ring vibration.
Epoxidised trans 1,4 BR shows characteristic peaks at 1060 and 1365 cm-1 for ring
stretching vibration (ν COC) and at 710 and 880 cm-1 for ring vibration. Zuchowska
[68] showed that during epoxidation some secondary reactions like acid catalysed ring
opening take place. He observed a very intense peak at 1740 cm-1 due to hydroxyacetyl
group in the spectra of 1,2 BR. Epoxidised cis 1,4 polyisoperene [69] showed peaks at
1240 cm-1 for νCOC and other characteristic peaks at 680 and 1070 cm-1. With
epoxidation the intensity of the absorption bands due to carbon-carbon double bond
decreases. Roy and co-workers [70] studied the epoxidation of NR with performic acid
in the latex stage and their study reveals that the furanisation process (peak at 1068 cm-1)
predominates only at epoxy contents above 65%.
Diene elastomers can be oxidised to polyketones by a catalytic method [71]. Polyketones

have potential uses, as they are good precursors for the synthesis of many other polymers.
Polyketones are easily photodegradable. Iraqi and Cole-Hamilton [71] oxidised
polybutadiene to polyketones with t-BuOOH in the presence of [Pt (diphoe) CF3(CH2Cl2)]+
(diphoe stands for Ph2PCH =CHPh2). IR spectra of all the products showed strong
absorption at 1720 cm-1 indicating the induction of >C=O group in the polymer backbone.
They achieved up to 57% conversion of C=C to >C=O. The reactivity of different
microstructures follow the order [71], pendent BR > trans BR > cis BR.
Cole-Hamilton and co-workers [72, 73] carried out catalytic hydrocarboxylation of BR

using [PdCl2(PPh3)2]SnCl2 as the catalysts. Infrared spectra showed the disappearance or
diminution of peaks at 732-740, 910, 967, 1650 cm-1 which are characteristic of carbon-
carbon unsaturation. A strong peak at 1710 cm-1 (ν C=O) and a very broad new peak
(ν O-H) emerged due to the presence of the COOH group. Up to 93% of the unsaturation
can be converted to COOH groups. The hydrocarboxylated BR was found to be useful
in wood preservation as the polymer prevents the uptake of water [73].
4.11 Phosphonylation
It is difficult to introduce phosphorus into a linear polymer by conventional polymerisation

or co-polymerisation of phosphorus containing monomers. The polymers containing
146
phosphorus shows higher thermal stability and have better flame-retardant properties.
A phosphate salt containing polymer shows microphase separation of the salt groups to
form ionic domains. Azuma and Macknight [74] introduced dimethyl phosphonate ester
groups into PP using benzoylperoxide (BPO) as free radical initiator. They achieved 10%
phosphonylation of PP and hydrolysed using hydrochloric acid and then prepared its
caesium salt using CsOH. Phosphonylated PP shows a peak at 820 cm-1 which is absent
in its acid or salt form indicating it is due to P-O-C vibration. The different absorption
bands in phosphonylated PP and its assignment have been shown in the Table 4.3.
4.12 Sulfonation
Sulfonation is very useful chemical modification of polymer, as it induces high polarity
in the polymer changing its chemical as well as physical properties. Sulfonated polymers
are also important precursors for ionomer formation [75]. There are reports of sulfonation
of ethylene-propylene diene terpolymer (EPDM) [76, 77], polyarylene-ether-sulfone [78],
polyaromatic ether ketone [79], polyether ether ketone (PEEK) [80], styrene-ethylene-
butylene-styrene block copolymer, (SEBS) [81]. Poly [bis(3-methyl phenoxy) phosphozene]
[82]. Sulfonated polymers show a distinct peak at 1176 cm-1 due to stretching vibration
of O=S=O in the -SO3H group. Another peak appears at 881 cm-1 due to stretching
vibration of S-OH bond. However, the position of different vibrational bands due to
sulfonation depends on the nature of the cations as well as types of solvents [75, 76].
4.13 Ionomer Formation

Ionomers are polymers which contain up to about 10% mole percent of ionic group [83-
85]. Ionomers show interesting properties because of the presence of different interactions
which include hydrogen bonding, formation of charge transfer complexes and ion-ion
interactions. Polymers with carboxylic or sulfonic acid groups in its backbone on
neutralisation with zinc, sodium or other metal salts form ionic aggregates.
IR spectroscopy can be used to characterise the formation of ionomers by studying the

environment of the anions [85, 86]. Risen and co-workers [87, 88] used far-IR spectra
(150 to 400 cm-1) to demonstrate the sensitivity of low frequency vibrations to the anions
and cations and the degree of cluster formation in ionomers. For example, styrene sulfonic
acid ionomers with Na+ cation shows absorption bands at 220 cm-1, whereas the Cs+
cation shows bands at 100 cm-1.
Carboxylate ionomers have been characterised with Fourier transform-infrared (FT-IR)

in the region of antisymmetric stretching vibration of carboxylate anions. Figure 4.8
shows carboxylate ionomer [89] of ethylene methacrylic (4%) copolymer).
147
Figure 4.8 FT-IR spectra of carboxylate ionomers ionised with zinc salt in the range of
1900-1200 cm-1
Reprinted with permission from B.A. Brozoski, M.M. Coleman and P.C. Painter,
Macromolecules 1984, 17, 230, Figure 4. Copyright 1990, American Chemical Society
The absence of peaks at 1750-1700 cm-1 indicates complete ionisation. The nature of
band depends on the types of cations. The salts of potassium, caesium and zinc exhibit
sharp single bands whereas sodium, calcium, strontium and barium salts show distinct
doublet bands [90]. It is due to different ionic environment with various cations.
148
Sulfonated ionomers are also characterised by IR spectroscopy [75-77]. Agarwal and co-
workers [76] analysed the Zn+2 salt of sulfonated EPDM. The peak at about 1200 cm-1 is
due to the asymmetric stretching of the sulfonate group. The band at 1020 cm-1 is ascribed
to the symmetric stretching of the -SO3- group. The position and nature of the absorption
band depend on the nature of the cation [76]. The band at 610-615 cm-1 is due to C-S
stretching of the polymer -SO3- band.
Mixed anionic (sulfonated – carboxylated) ionomers [81] were prepared by sulfonation

of maleated block-copoly (styrene/ethylene-butylene/styrene) (m-SEBS) by acetyl sulfate,
followed by neutralisation of the sulfonated maleated product, leading to the formation
of a new block copolymer ionomer based on both carboxylate and sulfonate anions
according to Scheme 4.6. FT-IR spectra confirm the presence of both carboxylated and
sulfonate ions (Figure 4.9).
Figure 4.9 FT-IR spectra of m-SEBS, Na-m-SEBS and Na-s-m-SEBS

Reprinted from S.K. Ghosh, D. Khastgir, S.K. De, P.P. De, R.J. Albalak and R.E. Cohen,
Plastics, Rubber and Composite Processing and Applications, 1998, 27, 310, Figure 1,
with permission from the Institute of Materials, UK, Copyright 1998
149
Scheme 4.6 Sulfonation reaction of m-SEBS

Reprinted with permission from S.K. Ghosh, D. Khastgir, S.K. De, P.P. De, R.J. Albalak
and R.E. Cohen. Plastics, Rubber and Composite Processing and Applications, 1998, 27,
310. Copyright 1998,
150
The characteristic absorbance of m-SEBS occurring at 1716 cm-1 indicates the C=O
stretching frequency of a five membered maleic anhydride group in the hydrogen bonded
maleic acid form. The absorbance at 700 cm-1 is due to the C-H rocking vibration of the
polystyrene ring and the 1602 cm-1 band is due to phenyl ring vibration of polystyrene.
The carboxylated salt (Na-m-SEBS) shows the occurrence of two new bands at 1576 cm-1
and 1548 cm-1 as a doublet which is ascribed to the asymmetric carboxylate stretching
vibration [90]. The Na-s-m-SEBS shows an absorbance at 1048 cm-1, which is due to the
symmetric stretching of the SO3 - anion of the sodium sulfonate group [76]. Other
characteristic bands at 1581 and 1552 cm-1 are due to a doublet resulting from the
octahedral arrangement of the carboxylate anions. Ghosh and co-workers studied such
mixed anionic ionomers with zinc [91], magnesium [92] and barium [92].
4.14 Ionomeric Blends
The ionic aggregates present in an ionomer act as physical crosslinks and drastically
change the polymer properties. The blending of two ionomers enhances the compatibility
via ion-ion interaction. The compatibilisation of polymer blends by specific ion-dipole
and ion-ion interactions has recently received wide attention [93-96]. FT-IR spectroscopy
is a powerful technique for investigating such specific interactions [97-99] in an ionic
blend made from the acid form of sulfonated polystyrene and poly[(ethyl acrylate – CO
(4, vinyl pyridine)]. Datta and co-workers [98] characterised blends of zinc oxide-
neutralised maleated EPDM (m-EPDM) and zinc salt of an ethylene-methacrylic acid
copolymer (Zn-EMA), wherein Zn-EMA content does not exceed 50% by weight. The
blend behaves as an ionic thermoplastic elastomer (ITPE). Blends (Z0, Z5 and Z10) were
prepared according to the following formulations [98]:
Z0 m-EPDM 100, ZnO 10

Z5 Zn-EMA 50, m-EPDM 50, ZnO 10
Z10 Zn-EMA 100, ZnO 10
A typical infrared spectra of 50/50 blend of m-EPDM and Zn-EMA [98] is shown in
Figure 4.10.
The m-EPDM-ZnO system shows a broad diffused band at 1565 cm-1 which is believed to
be due to asymmetric stretching of bridging type carboxylate groups. The Zn-EMA-ZnO
system on the other hand, shows a high intensity band at 1586 cm-1 indicating a high
extent of metal carboxylate salt formation. The asymmetric stretching peak in the case of
mix Z10 occurs at higher frequency than that in mix Z0, due to the occurrence of stronger
ionic aggregates or clusters in mix Z10. In the absence of any interaction, the spectrum of
the blend is likely to be equivalent to the summation spectrum (Figure 4.10c) of the
151
Figure 4.10 IR spectra of (a) Z0 (m-EPDM); (b) Z10 (Zn-EMA); (c) Summation
averaged spectrum; (d) Z5 (blend 50/50 m-EPDM and Zn-EMA)
Reprinted with permission from S. Datta, P.P. De and S.K. De, Journal of Applied
Polymer Science, 1996, 61, 1839. Copyright 1996, John Wiley and Sons, Inc.
152
constituent polymers [98], indicating the absence of any interaction, but the blend
corresponding to mix Z5 (Figure 4.10d) shows only a sharp peak at 1586 cm-1. It is expected
that if the two ionic peaks combine, the intermediate band position should be around
1574 cm-1, but the band has been shifted to higher frequency side by 12 cm-1, which may
be due to intermolecular interaction. Recently De and his co-workers have characterised a
series of such ionomeric polyblends with infrared spectroscopy, as given in Table 4.4.
4.15 Weathering and Degradation of Polymers
It is possible to study the effect of weather on elastomers by IR spectroscopy. However,

IR bands of polymers are inherently broad and weak and it is difficult to detect minor
chemical changes on the polymer chain. But it is possible to recognise these differences
by using absorption subtractions of the original polymer from the reacted polymer. Koenig
[109] used the absorbance subtraction method to study the surface oxidation of BR at
30 °C. When BR was exposed to weather for 10 hours there was a change in the ratio of
cis and trans unsaturation and occurrence of a peak at 1065 cm-1 (for C-O) suggested
formation of an oxidised product. After longer exposure oxidation gave rise to absorptions
for OH (3300 cm-1) and C=O (1700 cm-1, 1720cm-1). Irradiation damage of polyethylene
was studied in the same way [110]. Chlorosulfonated polyethylene (CSM) is a specialty
elastomer because of its interesting properties. The changes in CSM on γ-irradiation
have already been studied [111]. The most important changes occur in the range of
3600-3200 cm-1 and 1800-1650 cm-1, where the accumulation of OOH/OH and C=O
groups occur. On γ-irradiation new peaks appear at 3400 cm-1 and 1725 cm-1 which are
due to OH stretching and >C=O stretching, respectively. The nature of chemical changes
depends on the intensity of dose of γ-radiation. Even after exposure at relatively small
dose of radiation there are modifications of the bands at 1370 and 1155 cm-1 due to
SO2Cl and SO2. The quantitative change of SO2Cl can be detected by the absorbance
ratio at the band 1370 and 1460 cm-1 (due to C-H bending in PE). On γ- irradiation
there is decrease in the intensity of the peak at 1370 and 1155 cm-1 due to SO2 evolution.
New peaks emerge at 1610 and 825 cm-1 due to C=C unsaturation generated during
modification. Weathering of acrylonitrile-butadiene-styrene terpolymer (ABS) has been
characterised by IR spectroscopy [112] and it was shown that the butadiene unit is
responsible for UV degradation.
Weatherability of thermoplastic polyester elastomers (TPEE) based on polybutylene

terephthalate and polytetramethylene glycol was studied using outdoor exposure [113].
The IR spectra of TPEE exposed to air for 6 months for the original TPEE were compared.
New peaks occurred in exposed TPEE at 1175 cm-1 (due to >C= stretching). The ratio of
the peaks at 1175 cm-1 to the absorbance at 1500 cm-1 was due to in-plane skeletal
vibration of the benzene ring which determines the extent of weathering in the sample.
153
Table 4.4 Characteristic bands in ionomeric polyblends

Constituents of Main peaks Peak assignment Ref.
polyblend in cm-1
1. Zn-s-EPDM and 1694 >C=O stretching [100 ]
Zn EMA
1584 bridging metal carboxylate
1259 coupled vibration of C-O stretch and OH
bending
2. Zn-m-EPDM 1562 asym. metal carboxylate stretching [101 ]
and
1462 -CH2- bending
Zn-m-HDPE
1366 -CH2-wagging
1538, 1563 asym. carboxylate stretching pair, octahedral
carboxylate stretching
1596 tetrahedral carboxylate vibration
3. Zn-m-HDPE 1596,1552 a doublet in Zn-m-HDPE, asym. stretching of [102]
and Zn-XNBR carboxylate
carboxylate
1587, 1541 asym. stretching of carboxylate in Zn-XNBR
720, 731 a doublet due to rocking vibration in Zn-m-
HDPE
4. Al-m-EPDM 1570, 1536 a doublet carboxylate asym. stretching in [103]
Al-CSM octahedral metal carboxylate in Al-m-EPDM
561 Al stearate in Al-CSM
1016 symm. stretching of sulfonate group
1156 asym. stretching of S=O
5. Zn-m-HDPE 1894, 1775 C=O stretching of five membered maleic [104]
and Zn-m- EPDM anhydride
1709 hydrogen bonded carboxylic acid pairs
1376 CH3 symm. Deformation
1562 assym. carboxylate stretching, of Ionomer
1588, asymm. carboxylate stretching
1542, 1532
154
Table 4.4 Continued

Constituents of Main peaks Peak assignment Ref.
polyblend in cm-1
6. Zn-m-EPDM 1709 Hydrogen bonded carboxy lic acid pairs [105]
and Zn-PEA
1464 —CH2 bending
1376 symm. Deformation
1560 asymm. carboxylate stretching in ionomer
1534 asymm. stretching of carboxylate ion
7. Zn-PPA and 1547, 1588 asymm. stretching of metal carboxylate [106]
Zn-XNBR
1458 —CH2 bending
1375 CH3 asymm. Deformation
1730 free carboxylic acid
1697 hydrogen bonded carboxylic acid
1448 coupled vibration of—CH2 – bending and
–CH3 asymm. deformation
8. Zn-m-EPDM 1714 hydrogen bonded carboxylic acid [107]
and Zn-PEA
1464 CH2 bending
1363 CH2 wagging
1560 assym. stretching of carboxylate
1630 acid salt complex
1595 asymm. tetrahedral carboxylate stretching
1564, 1541 asymm. octahedral zinc-carboxylate stretching
9. Zn-XNBR and 1665 C=C stretch [108]
Zn-PEA
1591, 1538 asymm. stretching of zinc carboxylate ion
1464 —CH2- bending
1714 hydrogden bonded carboxylic acid
1590 tetrahedral structure of zinc carboxylate ion
1562/1546 octahedral structure of zinc carboxylate
1620 acid salt complex
155
Bhattacharjee and co-workers [114] studied degradation of NBR and HNBR using IR
spectroscopy and X-ray photoelectron spectroscopy (XPS). They observed >C=O and
-COOR functionalities in HNBR and NBR after ageing. The extent of ageing was
quantified from the ratio of absorbances at 1735 cm-1 and 1463 cm-1 (due to CH2
deformation in HNBR) and 1732 cm-1 and 1446 cm-1 (due to CH2 deformation in NBR).
Ghosh and co-workers [115] studied the thermal degradation of segmented polyamides
at different temperatures with the help of IR spectra. In segmented polyamide [115], the
peak at 1727 cm-1 is indicative of the ester linkage connecting the soft polyether segment
and the polyamide hard block. The peaks occurring at 3311, 1642, 1553 cm-1 are due to
the presence of –CONH2- groups, whereas the peaks at 1036, 1116, 1185 and 1275 cm-1
denote the various types of deformations taking place in the –CH2-O-CH2- linkages of
the soft polyether segments. IR spectra of the degraded sample in air at 367 °C reveal the
presence of major peaks at 3311, 1642 and 1533 cm-1. The absence of peaks at 1727,
1275, 1185 and 1036 cm-1 suggests that the polyether soft blocks are adversely affected
by oxidative degradation.
4.16 Modification by Radiation
Modification of polymers by radiation is a potential method for the development of new

polymers and composites. Radiation crosslinking thus endows the polymers with special
properties. Datta and co-workers used IR spectroscopy to characterise modified ethylene
vinyl acetate (EVA) exposed to electron beam in the presence of trimethylolpropane
trimethacrylate (TMPTMA) [116] or triallyl cyanurate (TAC) [117]. The IR spectra of
pure EVA shows peaks at 1740 cm-1 (>C=0 of ester group), 1480 cm-1 (-CH2 blending),
1370 cm-1 (-C-H bending of CH3), 1260 cm-1 (C-O stretching) and 1030 cm-1 (-C-O-C,
ether linkage). The spectra of unirradiated blend of EVA and TMPTMA show a sharp
peak at 1640 cm-1 mainly resulting from >C=C stretching of trans-vinylene, present in
TMPTMA. When the blends are irradiated with 50 KGy doses of electron beam, the
peak at 1640 cm-1 disappears, due to grafting or crosslinking of TMPTMA with EVA
through unsaturation. Up to 150 KGy dose, various reactions like aerial oxidation,
crosslinking of EVA with TMPTMA, chain scission, disproportionation and cyclisation
reactions take place [116]. IR study showed that some residual unsaturations remained
in irradiated pure TMPTMA, while in blends all unsaturations were used up at a very
early stage of irradiation. Some ether linkages were formed during irradiation in pure
EVA and the blends [118], although in pure EVA the concentration of ether linkages
reached a maximum at the 2 Mrad dose and then decreased, while in blends it increased
with an increase in radiation dose [118]. Electron beam initiated grafting of TMPTA
onto EPDM has been carried out over a wide range of irradiation doses (0-200 kGy)
using a fixed concentration (10%) of TMPTA [119, 120]. IR studies indicate increased
peak absorbances at 1730, 1260, 1120 and 1019 cm-1 up to 50 kGy and hence increased
156
>C=O and C-O-C concentrations, which may cause crosslinking as well as chain scission
of the EPDM rubber. On the other hand, free radicals developed on trimethylol propane
triacrylate (TMPTA) participate in reactions like self-crosslinking, grafting, cyclisation
and copolymerisation. Similarly, the influence of polyfunctional monomers such as
tripropylene glycol diacrylate (TPGDA), TMPTA, TMPTMA, TAC on the structural
changes of fluorocarbon rubber (Viton B) exposed to electron beam has been investigated
with the help of attenuated total reflectance-IR spectroscopy [121]. The absorbance at
1397, 1021, 672, 504 cm-1 due to C-F group decrease on irradiation of the mixtures of
fluoroelastomer and TMPTA, indicating dehydrofluorination and scission. The
photopolymerisation of tetra chloroethyl acrylate, pentafluorophenyl acrylate,
pentafluorophenyl-methacrylate was examined by quantitative FT-IR [122].
4.17 Conclusion
IR spectroscopy can be used to characterise not only different rubbers, but also to
understand the structural changes due to the chemical modification of the rubbers. The
chemical methods normally used to modify rubbers include hydrogenation, halogenation,
hydrosilylation, phosphonylation and sulfonation. The effects of oxidation, weathering
and radiation on the polymer structure can be studied with the help of infrared
spectroscopy. Formation of ionic polymers and ionomeric polyblends behaving as
thermoplastic elastomers can be followed by this method. Infrared spectroscopy in
conjunction with other techniques is an important tool to characterise polymeric materials.
Acknowledgement
NKS and PPD acknowledge assistance from Dr. S. Sivaram, National Chemical Laboratory,
India, Dr. Olav M. Aagaard and Mr. Wil J. Belt, DSM Research, The Netherlands. Thanks
are also due to Ms. Rituparna Singha, Bhuwneesh Kumar and Mr. Shambhu Bhattacharya
for helping the authors to prepare the manuscript.
References
1. D.N. Schulz, S.R. Turner and M.A. Golub, Rubber Chemistry and Technology,
1982, 55, 807.
2. N.T. McManus and G.L. Rempel, Journal of Macromolecular Science C, 1995,

C35, 2, 239.
3. M.P. McGrath, E.D. Sall and S.J. Tremont, Chemical Reviews, 1995, 95, 381.
157
4. C. Pinazzi, J.C. Brosse, A. Pleurdeau and D. Reyx, Applied Symposium No 26,

1975, 73.
5. N.K. Singha, S. Bhattacharjee and S. Sivaram, Rubber Chemistry and

Technology, 1997, 70, 309.
6. J. Thoermer, J. Mirja and N. Shoen, Elastomerics, 1986, 118, 9, 28.
7. K. Hashimoto and Y. Todani in Handbook of Elastomers New Development

Technology, Eds., A.K. Bhowmick and H.L. Stephens, Marcel Dekker, New York,
NY, USA, 1998, Chapter 24.
8. J.L. Koenig, Spectroscopy of Polymers, American Chemical Society, Washington,

DC, USA, 1992.
9. Structural Studies of Macromolecules by Spectroscopic Methods, Ed., K.J. Ivin,

Wiley, New York, NY, USA, 1976.
10. Infrared Analysis of Polymer, Resins and Additives; an Atlas, Volume 1, Eds.,
D.O. Hummel and F.K. Scholl, Plastics, Elastomers, Fibers and Resins, Part 1:
Text, Part 2: Spectra Tables, Index, Wiley Interscience, New York, 1969. Volume
2, Additives and Processing Aids 1973.
11. J.L. Binder, Journal of Polymer Science, Part A, 1963, 1, 47.
12. J.L. Binder, Journal of Polymer Science, Part A, 1963, 1, 37.
13. Y. Tanaka, Y. Takeuchi, M. Kobayashi and H. Todakaro, Journal of Polymer

Science, Part A2, 1971, 9, 13.
14. J.C. Ferguson, Journal of Polymer Science, Part A, 1964, 2, 4735.
15. A.L. Smith, Applied Infrared Spectroscopy: Fundamentals, Techniques and

Analytical Problem-Solving, John Wiley & Sons, New York, NY, USA, 1979,
134.
16. N.A. Mohammadi and G.L. Rempel, Macromolecules, 1987, 20, 2362.
17. N.K. Singha, S. Sivaram and S.S. Talwar, Rubber Chmeistry and Technology,
1995, 68, 281.
18. S. Bhattacharjee, A.K. Bhowmick and B.N Awasthi, Industrial and Engineering
Chemistry Research, 1991, 30, 1086.
158
19. S. Bhattacharjee, A.K. Bhowmick and B.N Awasthi, Journal of Polymer Science,
Part A, Polymer Chemistry, 1992, 30, 471.
20. A.J. Marshall, I.R. Jobe, T. Dee and C. Taylor, Rubber Chemistry and
Technology, 1990, 63, 244.
21. D. Bruck, Kautschuk und Gummi Kunststoffe, 1989, 42, 107.
22. D. Bruck, Kautschuk und Gummi Kunststoffe, 1989, 42, 194.
23. S. Bhattacharjee, A.K. Bhowmick and B.N. Avasthi; Makromolekulare Chemie,

1992, 103, 659.
24. X. Guo and G.L. Rempel, Journal of Molecular Catalysis, 1990, 63, 279.
25. N. Mohammadi and G.L. Rempel. Journal of Molecular Catalysis, 1989, 50,
259.
26. J. Podesva, J. Spevacek and J. Dybal, Journal of Applied Polymer Science, 1997,
74, 3214.
27. X. Guo, P.J. Scott and G.L Rempel, Journal of Molecular Catalysis, 1992, 72,
193.
28. D.K. Parker, R.F. Roberts and H.W. Schiessl, Rubber Chemistry and Technology,
1992, 65, 245; D.K. Parker, R.F. Roberts and H.W. Schiessl, Rubber Chemistry
29. N.K. Singha and S.Sivaram, Polymer Bulletin, 1995, 35, 121.
30. M. De Sarkar, P.P. De and A.K. Bhowmick, Journal of Applied Polymer Science,
1997, 66, 1151.
31. M. De Sarkar, P.P. De and A.K. Bhowmick, Journal of Applied Polymer Science,
1999, 71, 1581.
32. M. De Sarkar, P.P. De and A.K. Bhowmick, Polymer, 2000, 41, 907.
33. N.K. Singha, P.P. De and S. Sivaram. Journal of Applied Polymer Science, 1997,
66, 1647.
34. J.C. Falk and R.J. Schott, Die Angewandte Macromolecular Chemie, 1972, 21,
17.
159
35. D.R. Burfield, Proceedings of the International Conference on Rubbers and

Rubber-like Materials, Jamshedpur, India, 1987, Paper No.G-1, p.419.
36. S. Roy, S. Bhattacharjee and B.R. Gupta, Journal of Applied Polymer Science,
1993, 49, 375.
37. N.K. Singha, S.S. Talwar, S. Sivaram, Macromolecules, 1994, 27, 6985.
38. N. Murayama and Y. Amagi, Journal of Polymer Science, Part B, 1966, 4, 119.
39. J.C. Bevington and L. Ratti, Polymer, 1975, 16, 66.
40. A. Marchetti and E. Martuscelli, Journal of Polymer Science, 1976, 14, 151.
41. A. Konietzny and U. Biethan, Die Angewandte Macromolecular Chemie, 1978, 74, 61.
42. M.T. Ramesan and R. Alex, Journal of Applied Polymer Science, 1998, 68, 153.
43. S.A. Liebman and B.J. Gudzzinowicz, Analytical Chemistry, 1961, 33, 931.
44. C. Pinazzi and G. Levesque, Journal of Polymer Science, Part C, 1969, 16, 4695.
45. H. Okamoto, S. Aachi and T. Iwai, Journal of Polymer Science, Polymer

Chemistry Edition, 1979, 17, 1267.
46. D. Braun, W. Mao, D. Bohringer and R.W. Garbella, Die Angewandte

Macromolecular Chemie, 1986, 141, 113.
47. M.A. Golub, Journal of Polymer Science, 1957, 25, 373.
48. M.A. Golub and M.L. Rosenberg, Journal of Polymer Science, Polymer
49. J.A. Brydson, Rubber Chemistry, Applied Science Publishers, London, 1978,
Chapter 7.
50. J.I. Cunnen and M. Porter in Encyclopedia of Polymer Science and Technology,
Volume 12, Eds., N.M. Bikales, N.G. Gaylord and H.F. Mark, John Wiley and
Sons, New York, NY, USA, 1970, 318.
51. V.S. Shagov, A.I. Yakubchik and V.N. Podosokorskaya, Polymer Science USSR,
1969, 10, 1092.
52. A. Priola, M. Bruzza, F. Mistrali and S. Cesca, Die Angewandte Macromolekulare

Chemie, 1980, 88, 1.
160
53. A. Priola, N. Passerini, M. Bruzzone and S. Cesca, Die Angewandte Chemie,

1980, 88, 21.
54. J.L. Binder, Journal of Polymer Science, Part B, 1966, 4, 19.
55. I. Kossler, J. Vodehnal, M. Stolka, J. Kalal and E. Hartlova, Journal of Polymer

Science, Part C, 1967, 16, 1311.
56. N.G. Gaylord, I. Kossler, M. Stolka and H. Vodehnal, Journal of Polymer

Science, Part A, 1964, 2, 3969.
57. M. Stolka, J. Vodehnal and I. Kossler, Journal of Polymer Science, Part A, 1964,
2, 3987.
58. R.K. Agnihotri, D. Falcon and E.C. Fredericks, Journal of Polymer Science, Part
A-1, 1972, 10, 1839.
59. D. Morero, E. Mantica and L. Porri, Nuovo Cimento. 1960, Supplement 15,
Series 10, 136.
60. G.G. Cameron and M.Y. Qureshi, Makromolekulare Chemie, Rapid

Communications, 1982, 2, 287.
61. X. Guo, R. Farwaha and G.L. Rempel, Macromolecules, 1990, 23, 5047.
62. X. Guo and G.L. Rempel, Macromolecules, 1992, 25, 883.
63. Modification of Polymers, Eds., C.E. Carraher and J.A Moore, Plenum, Oxford,
UK, 1982.
64. T.C. Chung, M. Raate, E Berluche and D.N. Schulz, Macromolecules, 1988, 21, 1903.
65. F. Sibtain and G.L. Rempel. Journal of Polymer Science, Part A, Polymer
Chemistry, 1991, 29, 629.
66. S. Bhattacharjee, A.K. Bhowmick and B.N. Avasthi, Makromolekulare Chemie,

1992, 193, 659.
67. K. Sanui, W.J. MacKnight and R.W. Lenz, Macromolecules, 1974, 7, 952.
68. D. Zuchowska, Polymer, 1980, 21, 514.
69. R.V. Gemmer and M.A. Golub, Journal of Polymer Science, Part A, Polymer
Chemistry, 1978, 16, 2985.
161
70. S. Roy, B.R. Gupta and B.R. Maiti, Elastomers and Plastics, 1990, 22, 280.
71. A. Iraqi and D.J. Cole-Hamilton, Polyhedron, 1991, 10, 993.
72. P. Narayanan, A. Iraqi and D.J. Cole-Hamilton, Journal of Materials Chemistry,

1992, 2, 1149.
73. P. Narayanan, B. Kaya and D.J. Cole-Hamilton, Journal of Materials Chemistry,

1993, 3, 19.
74. C. Azuma and W.J. MacKnight, Journal of Polymer Science, Polymer Chemistry
Edition, 1977, 15, 547.
75. J.J. Fitzgerald and R.A. Weiss, Journal of Macromolecular Science C, 1988, 28, 1,
99.
76. P.K. Agarwal, P.K. Datta and R.D. Lundberg, Polymer, 1987, 28, 1467.
77. P.K. Agarwal, H.S. Makowski and R.D. Lundberg, Macromolecules, 1980, 13,
1679.
78. B.C. Johnson, C. Tran, I. Yigor, M. Iqbal, J.P. Wightman, D.R. Lioyd and J.E.
McGrath, Polymer Preprints, 1983, 24, 31.
79. A. Noshay and L.M. Robeson, Journal of Applied Polymer Science, 1976, 20,
1885.
80. K. Jin, M.T. Bishop, T.S. Ellis and F.E. Karasz, European Polymer Journal, 1985,
21, 4.
81. S.K. Ghosh, D. Khastgir, S.K. De, P.P. De, R.J. Albalak and R.E. Cohen, Plastics,
Rubber and Composite Processing and Applications, 1998, 27, 310.
82. H. Tang, P.N. Pintauro, Q.H. Guo, S. O’Connor, Journal of Applied Polymer
Science, 1999, 71, 387.
83. W.J. MacKnight and T.R. Earnest, Journal of Polymer Science - Macromolecular
Reviews, 1981, 16, 41.
84. A. Eisenberg and M. King, Ion-containing Polymers, Physical Properties and

Structure, Academic Press, New York, 1997.
85. T.R. Earnest, Jr. and W.J. Macknight, Macromolecules 1980, 13, 844.
162
86. T. Kurian, P.P. De, D.K. Tripathy and S.K. De, Rubber World, 1995, 213, 41.
87. A.T. Tsatsas, J.W. Reed and W.M. Risen Jr., Chemical Physics, 1971, 55, 3260.
88. G.B. Rouse, A.T. Tsatsas, A.E isenberg and W.M. Risen Jr., Journal of Polymer
Science: Polymer Physics Edition, 1979, 17, 81.
89. B.A. Brozoski, M.M. Coleman and P.C. Painter, Macromolecules 1984, 17, 230.
90. B.A. Brozoski and M.M. Coleman, Journal of Polymer Science: Polymer Physics
Edition, 1983, 21, 301.
91. S. Ghosh, P.P. De, D. Khastgir and S.K. De, Journal of Applied Polymer Science,
2000, 78, 326, 743.
92. S. Ghosh, P.P. De, D. Khastgir and S.K. De, Journal of Applied Polymer Science,
2000, 77, 816.
93. A. Eisenberg, P. Smith and Z.L. Zhou, Polymer Engineering Science, 1982, 22,
455.
94. M. Rutkowska and A. Eisenberg, Journal of Applied Polymer Science, 1984, 29,
775.
95. M. Hora and A. Eisenberg, Macromolecules, 1984, 17, 1335.
96. X. Lu and R.A. Weiss, Macromolecules, 1992, 25, 6185.
97. Ka Zuo Sakurai, Elliot P. Douglas, Wiliam J. Macknight, Macromolecules, 1992,

25, 4506.
98. S. Datta, P.P. De and S.K. De, Journal of Applied Polymer Science, 1996, 61,
1839.
99. A. Garton, Infrared Spectroscopy of Polymer Blends, Composites and Surfaces,

Hanser Publishers, Munich, Germany, 1992, 183.
100. T. Kurian, S. Datta, D. Khastgir, P.P. De, D.K. Tripathy, S.K. De and D.G. Peiffer,
Polymer, 1996, 37, 4787.
101. P. Antony and S.K. De, Plastics, Rubber and Composite Processing and
Applications, 1997, 26, 7, 311.
102. P. Antony and S.K. De, Journal of Applied Polymer Science, 1998, 70, 483.
163
103. P. Antony, S. Datta and S.K. De, Plastics, Rubber and Composite Processing and
Applications, 1998, 27, 7, 303.
104. P. Antony and S.K. De, Polymer, 1999, 40, 1487.
105. P. Antony, A.K. Bhattacharya and S.K. De, Journal of Applied Polymer Science,
1999, 71, 1257.
106. P. Antony, S. Bandyopadhyay and S.K. De, Journal of Materials Science, 1999,
34, 2553.
108. P. Antony, S. Bandyopadhyay and S.K. De, Polymer, 2000, 41, 787.
109. J.L. Koenig, Applied Spectroscopy, 1975, 29, 293.
110. D.L. Tabb, J.J. Sevcik and J.L. Koenig, Journal of Polymer Science: Polymer
Physics Edition, 1975, 13, 815.
111. G. Ivan, M. Giurginca and T. Zaharescu, Macromolecular Symposia, 1998, 129,

163.
112. A. Davis and D. Gordon, Journal of Applied Polymer Science, 1974, 18, 1159.
113. Y. Nagai, T. Ogawa, L.Y. Zhen, Y. Nishimoto and F. Ohishi, Polymer

Degradation and Stability, 1987, 19, 113.
114. S. Bhattacharjee, A.K. Bhowmick and B.N. Awasthi, Polymer Degradation and
Stability, 1991, 31, 71.
115. S. Ghosh, D. Khastgir, Anil K. Bhowmick and P.G. Mukunda, Polymer

116. S.K. Datta, A.K. Bhowmick, T.K. Chaki, A.B. Majali and R.S. Despande,
Polymer, 1996, 37, 1, 45.
117. S.K. Datta, A.K. Bhowmick and T.K. Chaki, Radiation Physics and Chemistry,
1996, 47, 6, 913.
118. S.K. Datta, Anil K. Bhowmick, D.K. Triapthy and T.K. Chaki, Journal of Applied
164
119. P.S. Majumdar and A.K. Bhowmick, Radiation Physics and Chemistry, 1998, 53,
63.
120. P.S. Majumdar, A.K. Bhowmick, A.B. Majali and V.K. Tikku, Journal of Applied
121. I. Banik, S.K. Dutta, T.K. Chaki and A.K. Bhowmick, Polymer, 1999, 40, 447.
122. M. Johnck, L. Muller, A. Neyer and J.W. Hofstraat, Polymer, 1999, 40, 3631.
165
166
Infrared Spectroscopy of Rubbery Materials
5
Prajna P. De
5.1 Introduction
Linear amorphous polymers can behave as either Hookian elastic (glassy) materials, or
highly elastic (rubbery) substances or as viscous melts according to prevailing temperature
and time scale of experiments. The different transitions as shown schematically in Figure 5.1
are manifestations of viscoelastic deformations, which are time dependent [1].
Figure 5.1 Temperature dependence of log (shear) modulus in a polymer system

showing molecular mechanism of the deformations taking place at different point
Reprinted with permission from P. Ghosh, Polymer Science and Technology of Plastics and
Rubbers, Tata McGraw-Hill Publishing Company Limited, New Delhi, 1990, Chapter I.
Copyright 1990, McGraw Hill Publishing Company Ltd., New Delhi
167
The deformation in the transition regions is often mechanically reversible, but owing to
time dependency, thermodynamically irreversible. It is apparent from the diagram that
the useful plastic properties may be exhibited by polymers of wide molecular weight
range, while useful rubbery properties are exhibited by polymers which are essentially
highly polymeric in nature. To exhibit the rubberiness, the ambient conditions must be
far above the glass-to-rubber transition temperature (Tg) of polymer. Tg for elastomers
(or rubbers) occurs in the range of –20 to –100 °C, depending on the technique and test
conditions. This is manifested in polymers of low cohesive energy density and poor
molecular symmetry, which results in an amorphous state, at least in the unstrained
state, permitting enough freedom of molecular motion so that the deformation of high
magnitude takes place rapidly. Rubber has the ability to extend its length, several hundred
percent on application of stress, with virtually immediate and complete recovery on
release of the stress [2]. Rubbery behaviour is shown by many polymeric materials, and
the first to be exploited was natural rubber (NR). Discovery of Ziegler-Natta catalyst
systems, led to the availability of a new generation of synthetic rubbers with controlled
molecular architecture. In subsequent years the rapid growth and availability of
thermoplastics and in particular polyethylene and plasticised polyvinyl chloride (PVC),
led to the replacement of many rubbers in many such applications [2]. In the 1960s and
1970s some conventional rubbers were replaced by thermoplastic rubbers for a number
of applications where flexibility rather than high elasticity was the primary concern. The
common feature of the thermoplastic rubbers is that the chains are held together at
normal ambient temperatures by heat fugitive crosslinks or physical bonds which become
ineffective at elevated temperatures [3]. Such crosslinks may be obtained in a number of
ways of which the most important are ionic crosslinks, hydrogen bonds and formation
of block copolymers. The rubberiness of plasticised PVC is due to hydrogen bonding
between the polymer chains and ester plasticiser molecules [3], which by spacing the
polymer molecules depress the Tg of PVC. Low density polyethylenes (PE) are rubbery to
some extent and the rubberiness becomes more evident when ethylene is copolymerised
with a small amount of vinyl acetate.
Since infrared (IR) spectroscopy is one of the most widely used techniques for the
identification of materials at the molecular level, it has been extensively used to characterise
the rubbery materials. In this chapter the rubbery materials encompass PE, plasticised
PVC, thermoplastic elastomers and ionomers.
5.2 Polyethylenes
PE is a polymer having the simplest structure (CH2)n, with a certain amount of chain
branching [4]. As shown in Figure 5.2, the polymer shows strong bands at 2927, 1467 and
724 cm-1, assigned for C-H stretching, deformation and rocking modes of –CH2 groups.
168
Figure 5.2 Infrared spectra of low density polyethylene (LDPE)

Reprinted from [53]
The significant chain branching due to the presence of methyl groups gives a peak at
1370 cm-1 overlapped by a methylene band at 1365 cm-1. LDPE from high pressure
polymerisation [4] contains vinylidene groups, which absorb at 890 cm-1, while polymers
prepared by Ziegler type catalysts [4] show three types of unsaturation: a vinyl group at
910 cm-1, -CH=CH- trans at 965 cm-1, and defect structure of the occasional double
bond at 1640 cm-1.
At room temperature, PE is a semi-crystalline plastomer (a plastic which on stretching

shows elongation like an elastomer), but on heating crystallites melt and the polymer
passes through an elastomeric phase. Similarly, by hindering the crystallisation of PE
(that is, by incorporating new chain elements), amorphous curable rubbery materials
like ethylene propylene copolymer (EPM), ethylene propylene diene terpolymer (EPDM),
ethylene-vinyl acetate copolymer (EVA), chlorinated polyethylene (CM), and
chlorosulphonated polyethylene (CSM) can be prepared.
Studies on structural changes and orientation process occurring in PE during a stress-

strain experiment show that the changes in PE films during elongation are manifested in
the 730/720 cm-1 (CH2 rocking) absorption bands [5]. The transition moment vectors of
the corresponding vibrations are parallel to the a and b axes of the crystalline cell. The
samples used were blown PE films having an initial orientation in unit cell c and a axes,
169
which possess a preferred orientation perpendicular to the film plane and parallel to the
machine direction (which is identical to the stretching direction). As far as structural
changes during the stretching process are concerned, the orthorhombic structure of
crystalline PE is partially transformed into a monoclinic structure as observed by the
appearance of a characteristic absorption band at 716 cm-1. Faster stretching velocities
increase proportion of the monoclinic structure [6]. It is observed that the bands at 720,
731 cm-1, are resolved to 716, 731 cm-1 with faster stretching. The bands at 720 and
731 cm-1 are due to rocking [7] vibration giving a typical orientation. IR spectroscopy is
used to find out the dichroic ratio (R), which can be found out by stretching a PE film, in
which, direction of stretch is parallel to slit and then is made perpendicular to the stretching
direction. This provides two values of absorption, namely A-parallel and A-perpendicular.
The dichroic ratio (R) is defined as:
R= IR absorption for parallel radiation = A||

IR absorption for perpendicular radiation A⊥
The R values for PE at 731 and 720 cm-1 are 8.9 and 10.0, respectively.
Dynamic two-dimensional Fourier transform-IR (FT-IR) spectroscopy has been used to

study the nature of the interphase in LDPE [8]. A two-dimensional correlation analysis
of the dynamic spectra indicates that neat LDPE is comprised of three regions: an ordered
crystalline region, a disordered liquid-like region and a crystal/amorphous interfacial
region. The peak which occurs at ~1455 cm-1 has been assigned to methylene bending
vibrations in conformationally disordered groups [9], while the peak at ~1467 cm-1 is
assigned to bending vibrations in all trans methylene chain sequences outside the crystal
structures [10]. Figure 5.3a shows the –CH2 bending and rocking regions of the
conventional absorbance spectrum of neat LDPE, as received, whereas Figure 5.3b shows
the in-phase and quadrature dynamic step-scan spectra in the CH2 bending region.
The dynamic spectra contain only the changes in the absorption due to the application
of the mechanical perturbation. The in-phase spectrum in the CH2 bending region is
similar to the normal absorbance spectrum [8]. The in-phase and quadrature spectra (a
2-dimensional spectra, which contains just as much area of a certain square) can be used
to generate synchronous (Figure 5.3c) and asynchronous (Figure 5.3d) correlation plots
[8]. The synchronous plot shows strong auto-peaks at 1464 and 1472 cm-1, which typify
crystalline species in semicrystalline polymers. The composite nature of the CH2 bending
region is well demonstrated by the asynchronous plot, which shows that 1455 to 1475 cm-1
region splits into at least four independent components at 1458, 1462, 1467 and 1472 cm-1.
Analysis of crosspeaks shows that the two bands at 1462 and 1472 cm-1 are out-of-
phase with bands at 1458 and 1467 cm-1. This conclusion is in agreement with an earlier
work where the 1458 cm-1 band is assigned to the amorphous regions comprising
approximately 40% amorphous material, found in the core of the film; whereas skin is
170
a)
b)
c) d)
Figure 5.3 (a) Static IR absorbance spectrum of unoriented neat LDPE, (b) In-phase and
quadrature step-scan dynamic spectra of neat LDPE in CH2 bending region, (c) 2-
dimensional FT-IR synchronous correlation plot of neat LDPE in the CH2 bending, (d)
2-dimensional FT-IR asynchronous correlation plot of neat LDPE in CH2 bending region
Reprinted with permission from A. Singhal and L.J. Fina, Polymer, 1996, 37, 12, 2335.
Copyright 1996, Elsevier Science Ltd
171
crystalline [9] and the 1467 cm-1 band is assigned to methylene bending modes organised
in fully transplaner segments in the interphase region [10]. The addition of talc and
ethylene vinyl acetate (EVA) to LDPE does not change the dynamic deformation
mechanism of the LDPE but their morphological distribution in LDPE is found to be
very different [8]. Talc associates intimately with the crystallites and EVA associates with
all three of the morphological phases (crystalline, interphase and amorphous).
Khastgir and co-workers studied interactions between EVA copolymer and PE with the
help of IR spectroscopy [11, 12]. The spectrum of neat EVA shows two distinct carbonyl
peaks at 1746 cm-1 and 1729 cm-1. The 1746 cm-1 peak is assigned to the >CO stretching
vibration of vinyl acetate group and the peak at 1729 cm-1 is assigned to the keto carbonyl
stretching vibration. The formation of ketone [11] group in neat EVA during processing
is schematically represented by the reaction Scheme 5.1.
Scheme 5.1 Formulation of ketone in pure EVA during processing

Reprinted with permission from I. Ray, S. Roy and D. Khastgir, Polymer Bulletin, 1993,
30, 685. Copyright 1993, Springer-Verlag
But it is interesting to note that the 50/50 EVA/LDPE blend processed under the same
conditions shows a single well-resolved >C=O stretching band at 1740 cm-1 (Figure 5.4)
which indicates that there is no trace of keto carbonyl groups.
The absence of the keto group in the blend may be ascribed to the reaction of the
macroradical (formed during processing through abstraction of α-H atoms of EVA) with
vinylidene group of LDPE [13]. Hence it may be concluded from IR analysis (Figure 5.4)
that under the action of heat and mechanical shearing at the processing condition, LDPE-
g-EVA is formed [11] according to Scheme 5.2.
The grafting reaction takes place generally in the amorphous region of polymers [11].
Similarly, other polar groups like dibutyl maleate (DBM) [14, 15], vinyl trimethoxysilane
(VTMO) [16, 17], vinyl triethoxy silane (VTEO) [16, 17, 18] can be grafted onto PE, to
make cable sheathing compounds. Sen and co-workers [15] characterised low halogen
and non-halogen fire-resistant low smoke (FRLS) cable sheathing compounds from blends
172
Figure 5.4 IR spectra of EVA (———) LDPE (-o-o-) 50/50 EVA/LDPE blend (……)
processed at 170 °C, LDPE processed at 130 °C (-..-..-)
Scheme 5.2 Reaction of EVA with vinylidene group of LDPE

173
of functionalised polyolefins and PVC by IR spectroscopy. On grafting DBM onto PE,

the peak at 1738 cm-1 due to the carbonyl group of ester interacts with PVC through
hydrogen bonding, which is reflected by its shifting to lower wavelength [15]. Dipole-
dipole interactions of the type >C=0 .…. Cl-C- may also exist [15], as shown in Scheme 5.3.
Scheme 5.3 Mechanism of compatibilisation between PE grafted onto DBM

(PEgDBM) and PVC through hydrogen bonding and dipole-dipole interaction
Reprinted from [15] with permission of John Wiley and Sons, Inc, 1991
Vinyl silanes are grafted onto PE for the modification of properties such as adhesion and
dyeability, as well as for preparing compatibiliser for multicomponent polymer blends
[16, 17]. The VTMO and VTEO graft co-polymers of PE were prepared by reactive
processing in a Brabender extruder in the temperature range of 150-200 °C in the presence
of dicumyl peroxide (DCP) [16]. IR spectroscopy has been used to characterise the grafting
(Figure 5.5) showing new peaks in the range 800-1200 cm–1, such as 1190 cm-1 (-CH3
rocking in Si-O-CH3), 1094 cm-1 (the anti-symmetric Si-O-C stretching) and 800 cm-1
(the symmetric Si-O-C stretching). Moisture catalysed crosslinking of the silane-grafted
copolymer is believed to follow Scheme 5.4. The appearance of a shoulder at 1025 cm-1
due to Si-O-Si stretching indicates crosslinking [16].
Research on the application of radiation processing of polymers, particularly in the food

industry necessitates studies on the effects of radiation on various polymer films. This is
because different types of radiation may cause different structural changes, rearrangement
and fission or formation of chemical bonds in the polymeric film. As a result, the physical,
174
Scheme 5.4 Mechanism of hydrolysis and crosslink formation through the

condensation reaction of silanol groups
A.K. Bhowmick, Journal of Applied Polymer Science, 1992, 44, 1153. Copyright 1992,
John Wiley and Sons, Inc.
chemical, microbiological and mechanical properties of the polymer may suffer some
dramatic changes. Chaki and co-workers used electron beam irradiation crosslinking of
PE film for the cable sheathing materials [18]. PE (100 parts by weight) was mixed with
5 parts by weight of methyl methacrylate (MMA) at 120 °C and subsequently exposed to
electron radiation of different doses (up to 20 Mrad) to prepare PE/MMA graft copolymers.
Evidence of successful grafting was verified by IR spectroscopy. A sharp peak around
1030 cm-1 is ascribed to the formation of the methacrylate group in the irradiated crosslinked
sample. The grafting level was determined from the absorbance ratio of the peak at 1030 cm-1 to
175
Figure 5.5 IR spectra (a) PE, (b) PE-g-VTMO

A.K. Bhowmick, Journal of Applied Polymer Science, 1992, 44, 1153. Copyright 1992,
John Wiley and Sons, Inc.
that at 1450 cm-1 (due to –CH2 stretching). The methacrylate concentration of the irradiated
samples increased linearly with increase of MMA concentration.
Similarly Abou Zeid and co-workers made a detailed study on LDPE film, irradiated
with different types of radiation such as 60Co γ-rays, thermal and fast neutrons, and
electron beam irradiation [19]. The structural changes of PE films with thickness varying
from 8 to 24 μm, were characterised by FT-IR spectroscopy. The characteristic FT-IR
absorption bands of unirradiated, γ-irradiated and reactor irradiated LDPE film are
presented in Table 5.1.
The band at 1018 cm-1 is due to the presence of occulted oxygen in PE film (thickness,
22-24 μm). The bands at 452 and 456 cm-1 appear only for LDPE (thickness, 22-24 μm)
which are attributed to the β (C≡C-H). The disappearance of a small band at 888 cm-1 is
possibly due to >CH2 out of plane deformation. This may be explained on the basis of
evolution of hydrogen gas as a result of the degradation. The intensity of IR absorption
bands at 1017, 1113, 1131, 1175, 1241 cm-1 is shown to increase with increasing
irradiation dose. The intensity of the bands 1069 cm-1 (C-C stretching), 1287 cm-1 (OH
bending), 1377 cm-1 (symmetrical deformation of CH3) (which are initially present in IR
spectra of unirradiated films) increase with absorbed dose of radiation, whereas the
intensity of the sharp band at 1456 cm-1 is not affected at all. The small band at 965 cm-1
is attributed to C-H out of plane deformation of trans-vinylene groups (trans RCH=CHR).
176
Table 5.1 The Characteristic FTIR absorption bands of uniradiated, γ-

irradiated and reactor irradiated LDPE film
Original Position of FTIR band,
Observed Shifts, cm-1
cm-1
Bands Bands
Bands Intensity
initially induced
induced change Low γ Middle γ High γ
present in due to Literature
due to γ on γ dose dose dose
unirradiated reactor
radiation radiation
PE radiation
452 Constant
456 Constant
Dis- Dis- Dis-
888 -
appeared appeared appeared
965 967 965(+) + 965 965 963.98
1018 1017 1018 1068(+) + 1017 1018 1018
1069 1113 + 1114
1131 1131(+) + 1131 1131 1130
1175 + 1177 1177 1178
1241 1241
1287 +
1304 1306(+) + 1304 1302 1288
1377 1378(+) + 1378 1378 1377
1411 1410(+) + 1411 1411
1456 Constant 1456 1457 1457
1715 1714 1716(+) + 1714 1714 1701
3420 3420 +
+: increase
Reproduced from [19] with permission of John Wiley and Sons, 2000
The band at 1303 cm-1 is reported to be one of the bands characterising the amorphous
phase in PE. It follows that crosslinking causes a decrease in crystallinity. The band at
1411 cm-1 which is evidence for the occurrence of crosslinking due to the methylene
177
deformation influenced by an adjacent carbonyl group, shows a linear relationship over

a wide interval of γ-radiation absorbed doses. Similarly, the intensity of the band at
1716 cm-1, which is attributed to stretching vibration of ketonic carbonyl group, increases
remarkably with γ-radiation absorbed doses showing two stages of linear dependence.
The intensity of the band in the range (3370-3420 cm-1) increases with the absorbed
dose showing linear behaviour for LDPE of various thickness. Hence it can be concluded
that the influence of γ-radiation on the structure of PE shows degradation, crosslinking
and changes in both crystalline and amorphous regions.
Radiation induced graft copolymers with polyfunctionally substituted heterocyclic ring

derivatives comprise a very interesting class of polymers because of their significant
biological and pharmaceutical activity [20, 21]. The use of N-vinyl pyrrolidone (NVP)
as hydrophilic grafting monomer is well known in the radiation grafting on LDPE [22],
and the grafted polymer LDPE-g-PNVP is used as the starting material for the preparation
of biocompatible polymer surfaces. Here grafting of LDPE was done by 60Co gamma
source at a dose rate 1.60 Gy/s at 50 °C in a vacuum oven and the grafted polymer was
characterised by IR spectroscopy.
FT-IR spectroscopy has been used in the investigation of welding crosslinked polyethylene
pipes [23]. Three types of crosslinking systems were used namely, peroxide (PEXa), silane
(PEXb) and electron beam (PEXc). Scholten and co-workers [23] observed that only
PEXa pipes have a satisfactory electrofusion quality. The strength of electrofusion welds
of PEXb and PEXc pipes is not acceptable. The most likely explanation for the differences
in weld quality is related to the adhesion theory and more specifically to differences in
composition. Figure 5.6 shows the infrared spectra of medium density polyethylene
(MDPE), PEXa, PEXb and PEXc.
While PEXa exhibits an IR spectrum very similar to uncrosslinked PE, PEXb contains a
rather high concentration of silicon containing groups, as is obvious from the broad
band due to Si-O stretching and bending modes around 1080 cm-1. These groups are
residues of the crosslinking process. The large band at ~1740 cm-1 indicates that PEXc
contains EVA, which is supported by the occurrence of smaller bands at other positions
(not shown). These polar ‘impurities’ may have a deteriorating effect on the quality of
electrofusion welds produced with PEXb and PEXc.
In PE/aluminium laminates, which are used for packaging, the bond strength can be
improved by surface oxidation of the polymer. An alternative to surface oxidation is to
use copolymers of ethylene and monomers containing polar groups that can provide
stronger bonds to metal thus increasing adhesion [24, 25]. The effects of three functional
groups in ethylene copolymers, namely EVS [polyethylene-co-vinyl trimethoxy silane],
EBA [polyethylene-co-butyl acrylate], EAA [polyethylene-co-acrylic acid) on the adhesion
was studied [25]. The interface in polymer/metal laminates has been analysed by FT-IR
178
Figure 5.6 Part of IR spectra of PEXa, PEXb and PEXc are compared with that of
MDPE [bending vibrations at ~1450 cm-1 omitted]
Reprinted from [23] with permission of Elsevier Applied Science, on behalf of the Institute
of Materials, New York, 1998
spectroscopy. It is observed that the increased amount of free acid at the interface can be
attributed to interfacial hydrogen bonds between the hydroxyl groups of the acid and
the surface AlOH groups, forming Al-O-Al groups. Furthermore, the band at 1738 cm-1
represents the carbonyl group in an interfacial ester link between the polymer and the
metal originating from a reaction between surface AlOH groups and hydroxyls of acid
groups. The carbonyl of ester copolymer gives two absorptions in the IR spectrum. The
downshift of the carbonyl frequency at the interface can be attributed to interaction with
the aluminium oxide of the substrate. Ishida and co-workers reports that in the case of
EVS, the silane-related bands in the transmission spectrum are the same as that in the
case of silane modified polymer [25]. Examples are the bands at 1190 cm-1 (for –CH3
rocking in Si-O-CH3), 1094 cm-1 (anti-symmetric stretching of Si-O-C) and 800 cm-1 (the
symmetric stretching of Si-O-C). In addition the appearance of a shoulder at 1025 cm-1
representing the anti-symmetric Si-O-Si stretching, indicates formation of silanol groups
followed by condensation to siloxanes. Due to crosslinking via the silane groups, the
adhesive strength for EVS is higher than that of EBA and EAA.
179
The production of polymers with enhanced physical properties such as high modulus fibres
has led to detailed investigations on the polymer morphologies. In several cases PE single
crystals have been used as well characterised model starting materials for studies of
deformation. Recently the isotopic labelling technique has been used to establish the
conformation of individual polymer molecules in solution-grown single crystals of PE [26].
Okoroafor and co-workers have prepared mixed crystals [26] of normal polyethylene (PEH)
and fully deuterated polyethylene (PED) from dilute solution of mixed xylenes at 70 °C and
characterised them by using the mixed crystal IR technique. They have compared the CX2
bending vibration (X=H or D) for PEH and PED mixed crystal of PE and observed reduction
in the absorbance of the outer CD2 bending components, which provides evidence of lamellar
break up.
Characterisation of ethylene in ethylene-propylene copolymers (EP) and in isotactic

polypropylene (iPP)/EP systems by deconvolution of IR spectra was studied by Costa and co-
workers [27]. The method is based on the deconvolution of the spectrum in the CH2-rocking
range, i.e., 800-680 cm-1, where the bands of interest overlap. The overlapping bands at 720
and 730 cm-1 are due to ethylene crystallinity. Similarly a band for amorphous PE appears at
722 cm-1, a band for single ethylene group between two propylene groups, appears at
735 cm-1, a band at 752 cm-1 is assigned to –(CH2)2- between tertiary carbon atoms, for
head-to-head and a band at 770 cm-1 is assigned to pendent ethyl groups. The six bands
present in this region are signal averaged in position and width (Figure 5.7). The spectra are
then deconvoluted (curve fitted, Figure 5.7) assuming a Lorentzian shape for bands.
The band at 1167 cm-1 (with a shoulder at 1156 cm-1) corresponding to a CH3 vibration is
considered as an internal standard. The method is checked by varying some fitting parameters.
The amounts of total (-CH2-CH2) units, isolated ethylene and structural defaults in PP were
determined for different iPP/EP blends.
5.3 Polyvinyl Chloride
Like other polymers, the structure of PVC is identified by IR spectroscopy [4]. Figure 5.8
shows the spectra of PVC, indicating the presence of saturated C-H groups with CH2
deformation band at 1430 cm-1, which has moved about 30 cm-1, suggesting the presence of
polar substituents. The C-Cl stretching bands are observed in the range of 600-700 cm-1.
Usually two different polymers do not mix at the segmental level but a favourable interaction
between the two polymers can allow one to obtain homogeneous blends. FT-IR is a potential
tool for the investigation of the mutual compatibility of various polymers. The small spectral
changes due to these interactions can be detected by this method. If two polymers are
immiscible, one can synthesise a spectrum of the blend by co-adding, in the appropriate
180
Figure 5.7 Deconvolution of an iPP/EP sample in the range 800-680 cm-1

Reprinted from [27] with permission of John Wiley and Sons, Inc., 2000
Figure 5.8 Infrared spectra of PVC

Reprinted with permission from W.F. Maddams in Analysis of Polymer Systems, Eds., L.S.
Bark and N.S. Allen, Applied Science Publishers, London, 1982, Chapter 3, 59. Copyright
1982, Bark and Allen
181
proportions, the spectra of the two pure components, which is then compared with the
observed spectrum of the blend. On the other hand, when polymers are miscible, only
one phase exists and there should be spectral differences between the co-added spectra
of the pure components [5].
Compatible and incompatible polyester – PVC blends have been considered. Examples
include systems, poly-∑-caprolactone (PCL)/PVC which are compatible [5, 28], in the melt
and exhibit partial compatibility in solid state and poly-β-propiolactone (PPL)/PVC blends
which are known to be incompatible [5, 28]. Figure 5.9 shows the infrared spectra of the
carbonyl stretching vibration (in the range 1600-1800 cm-1) for the different blends.
Figure 5.9 (a) Spectra of PVC/PCL blends at room temperature for (A) neat PCL (B) 1:1 (C) 2:1
(D) 3:1 (E) 5:1 (F) 10:1, (b) Spectra of PPL and PPL/PVC (30/70) blend at room temperature
Reprinted with permission from Developments in Polymer Characterisation, Ed., J.V. Dawkins,
Applied Science Publishers, London, 1983, Chapter 3, 119. Copyright J.V. Dawkins, 1983
182
Specific interactions between PCL and PVC are clearly indicated. In the solid state
(Figure 5.9a) the spectrum of neat PCL indicates the presence of crystalline (1724 cm-1)
and amorphous (1737 cm-1) bands. At mole ratios up to 2:1 of PVC to PCL, the spectra
indicate that in the solid state the blends consist of crystalline and amorphous phases. As
the PVC concentration increases, a parallel increase of the intensity of the amorphous
band is observed. Moreover, the frequency shifts observed for both the crystalline and
amorphous bands as a function of the composition of the blend suggests that specific
interactions between the two polymers occur. No shift is observed in the carbonyl
stretching vibration of PPL/PVC blends, in the molten state or in the solid state over the
entire range of compositions and the two polymers are incompatible [28].
PVC and acrylonitrile butadiene rubber (NBR) blends undergo self-crosslinking [29-31]
during high temperature mixing in the absence of any external curing agents. Evidence
of the chemical interaction is confirmed by IR spectroscopic studies, which show that
the reaction takes place through the amide and acid groups formed by the hydrolysis of
the -C≡N groups, in the presence of hydrogen chloride liberated from PVC [32]. The
interaction [31] of PVC with nitrile rubber is shown in Scheme 5.5.
Scheme 5.5 Chemical Interactions of PVC and nitrile rubber during mixing
Reprinted with permission from N.R. Manoj and P.P. De, Polymer, 1998, 39, 3, 733.
183
The representative IR spectra for PVC and NBR blends [31] are shown in Figures 5.10
and 5.11 [31]. Neat PVC shows two sharp absorbance peaks at 1734 and 1590 cm-1
which are due to the carbonyl groups formed during the manufacturing stage [33] and
the polymer backbone defect [34].
Figure 5.10 IR spectra of PVC and NBR before mixing (a) PVC (c) NBR after mixing
for 60 minutes at 180 °C, (b) PVC (d) NBR
184
Figure 5.11 Infrared spectra of 50/50 PVC/NBR blends after mixing (a) for 5 minutes,
(b) for 60 minutes (stabilised with tribasic lead sulfate) at 180 °C, (c) at 160 °C, (d) at
180 °C, (e) at 200 °C for 60 minutes
During processing, PVC undergoes thermooxidative degradation leading to the formation

of conjugated polymers, which absorb at 1715-1660 cm-1 The broad peak at 3280 cm-1
may be due to the formation of –OH groups during degradation [31]. The difference in the
185
spectra of NBR before and after processing (Figure 5.10) is the slight broadening of the
band at 3510 cm-1 due to formation of –OH groups during degradation, and a new peak at
1575 cm-1 due to formation of triazine generated by cyclisation of adjacent nitrile groups
[31]. The absorption band at 2224 cm-1 is due to the stretching vibration of C≡N groups.
Figure 5.11 shows the spectra of 50/50 PVC/NBR blends at different temperatures. A
comparison of the spectra of the blends mixed for 5 and 60 minutes at 180 °C (Figures 5.11,
a and d) show the changes occurring during crosslinking. The broadening and the increased
absorption at 3500-3200 cm-1 show the formation of O-H and N-H groups in the system
due to hydrolysis of the nitrile groups. At 160 °C the peak is smaller and at 180 °C, there
is increased absorption indicating the temperature dependence of the reaction. At 200 °C,
the peak intensity decreases, indicating that some of the –OH groups react during mixing.
In spectra (Figure 5.11b) the blend is mixed with stabiliser, it shows the same three peaks
of PVC (without stabiliser) observed with almost equal intensity within 1715-1660 cm-1.
The intensity of these peaks change, when they are heated without stabiliser as seen in
Figures 5.11c, 5.11d, 5.11e. The peak centred at 1720 cm-1 represents the carbonyl groups
of aliphatic acids, esters and ketones. The carbonyl groups in amides absorb at 1630 cm-1
and carbon-carbon double bonds absorb at 1660 cm-1. During crosslinking reaction, the
concentration of carbonyl groups increase due to formation of amides, acids and esters
Figures 5.11 d and 5.11e and that of carbon-carbon double bonds increase due to the
simultaneous degradation (and e). This is manifested in the broadening of the peaks and
increased absorption as the processing temperature and time increases. The changes in the
concentration of nitrile group are not visible since the absorption is sharp and strong. But
in order to confirm any change of the intensity of the band for the acrylonitrile (-C≡N)
group, Manoj and co-workers [31] have mixed separately NBR, hydrogenated acrylonitrile
butadiene rubber (HNBR) and PVC/NBR (50/50 blend) and PVC/HNBR (50/50) blend at
180 °C for 60 minutes in the Brabender Plasticorder and IR spectra of each compound was
taken. The thermal treatment does not produce any perceptible change in the intensity of
the band at 2224 cm-1, but the decrease in —CN concentration in the blend is attributed to
partial hydrolysis of cyanide groups by the splitting of hydrogen chloride from PVC during
heating in the presence of moisture [31] (Scheme 5.5). De and co-workers studied chemical
changes in the hot air and fuel ageing of PVC/NBR and PVC/HNBR blends with the help
of infrared spectroscopy [35].
The thermal stability of PVC blended with polymers acting as impact modifiers depends
mainly on the type of polymer, blend morphology and degradation environment. The
effect of the type of epoxidised butadiene-styrene block copolymer [ESBS, linear butadiene/
styrene (B/S) or radial (E(B/S)n) containing 0%-27% of epoxy groups] on thermal
dehydrochlorination of PVC-ESBS blends has been investigated by Meissner and co-
workers [36] in the temperature range 170-180 °C under non-oxygen atmosphere.
Thermal stability of PVC-ESBS blends is estimated on the basis of induction time and
maximum rate of hydrochloride emission from the system. On the basis of the analysis
186
of the IR spectra of PVC and its mixtures with epoxidised co-polymers before and after
heating at 180 °C in the range 400-4000 cm-1, it is concluded that hydrogen chloride
liberated from PVC reacts with double bond and epoxy groups to form –CH2-CHCl-
and –CH(OH)-CHCl- as indicated by the bands at 3600 cm-1 (OH) and 698 cm-1 (C-Cl).
Similarly IR spectra has been used to find out the thermal stability of PVC-diocytyl
phthalate plastigels in presence of zinc stearate (ZnSt2) and epoxidised soyabean oil
(ESO) [37]. ZnSt2 added to PVC is characterised by an absorption band at 1540 cm-1
corresponding to the stretching vibration of carboxylate groups. The intensity of carboxyl
band is proportional to the concentration of ZnSt2, and it decreases with increasing
heating time, as it reacts with hydrogen chloride liberated from PVC.
The OH group of organic acid is characterised by an absorption band at 3400 cm-1,

corresponding to the stretching vibration. The best heat stability and gelation effects are obtained
with ESO for long heating times, which are confirmed by dehydrochlorination rate constants.
Like thermal ageing, photoageing of PVC has been studied by Verdu and co-workers
[38-41]. It is widely recognised that PVC photoageing is a very complex process due to
the co-existence of two (sharply interrelated) chain reactions, namely, oxidation and
hydrogen chloride elimination. Photoelimination creates chlorine (Cl) radicals able to
initiate oxidation chains, HCl able to catalyse hydroperoxide or peroxy decomposition
and polyenes able to play a photosensitising or photostabilising role depending essentially
on their length [39, 40]. Verdu and co-workers [39] have used FT-IR spectra to determine
the carbonyl concentration at 1720 cm-1 and have observed that a significant decrease of
number average molecular weight and the highest rate of carbonyl build up take place in
the superficial layer of about 50 μm thickness, whereas the subcutaneous layer of about
200-300 μm thickness is characterised by a predominant crosslinking and conjugated
polyene growth, when it becomes insoluble and highly discoloured. The accelerated
photoageing of non-photostabilised, non-pigmented bulk PVC samples have been studied
in photochemical reactors at different temperatures (40-70 °C) with sources of different
intensities starting from 7.2 x 1018 photons per cm2h-1. As was observed in earlier studies
[36, 37], they have observed the existence of two layers [40]: the superficial layer <100 μm)
rich in oxidation products like carbonyls and the subcutaneous layer (50 to 400 μm) rich
in conjugated polymers. Verdu and co-workers studied the effect of titanium dioxide
[41] on the photoageing of PVC, and have observed that pigments display a strong screen
effect in preventing discolouration as PVC discolouration occurs in a subcutaneous layer
because its precursors (presumably alkyl radicals) have a sufficient life time only in the
absence of oxygen. Titanium oxide pigments are able to absorb most of the incident light
in the layer where oxygen is available and scavenges most of alkyl radicals [41]. The
photooxidative degradation and stabilisation of PVC have been studied using dibutyl tin
maleate (DBTM) alone [42] and mixed with varying amounts of tris nitro (1,3, dihydroxyl-
187
2-hydroxyl methyl-2-nitropropane) compounds as stabilisers. The degradation reactions

have been monitored by IR.
Millan and co-workers studied the substitution reaction of PVC with sodium thiobenzoate
[43], sodium-2-mercaptobenzothiozolate [44] and the substitution reaction is confirmed
by IR spectra. The presence of bands at 757 and 727 cm-1 for 2-mercaptobenzothiazolate
and bands at 653, 690 and 775 cm-1 for thiobenzoate along with 615 and 637 cm-1 for
C-Cl indicate that substitution has taken place.
Grafting of 2-hydroxyl ethyl methacrylate (HEMA) onto the surface of PVC has been
studied by Sreenivasan [45] and grafting reaction is characterised by attenuated total
reflectance-IR (ATR-IR) spectrum. The hydroxy group of HEMA reacts with PVC forming
an ether linkage, as shown by a band at 1140 cm-1 whereas –OH group in the spectrum
is absent. The 1630 cm-1 peak of C=C of HEMA is also present indicating that C=C
bond remains intact during the coupling reaction.
5.4 Thermoplastic Elastomers

Thermoplastic elastomers (TPE) based on blends of crystalline thermoplastics and rubbers
are unique in the sense that they can be processed like thermoplastics but their physical
properties are similar to vulcanised rubbers and rubbery materials [46, 47]. A wide
spectrum of properties, from essentially plastic to predominantly elastomeric, can be
attained by varying the blend ratios. One of the advantages of thermoplastic elastomers
is that they can be recycled – indicating that interchain bonding in TPE consists of thermo-
reversible crosslinks arising out of physical interactions like hydrogen bonding, charge
transfer complexes, ion-dipole interactions and ion-ion interactions [48, 49].
The physical interactions in TPE can be characterised by IR spectroscopy. A few examples

of such studies are discussed here. Examples of PE based thermoplastic elastomers are
NR/PE blends [50, 52]. TPE [49] based on 50/50 NR/LDPE, forms co-continuous
morphological structure of both NR and LDPE. Thermal analysis shows that the blend
is immiscible and from IR spectra of the 50/50 NR/LDPE blends [53], it is observed, the
peaks of NR and PE exist almost in the same positions in the blend with a very little shift
(Figure 5.12). The absorption band at 833 cm-1 for cis >C = C in NR (Figure 5.12) is
shifted to 836 cm-1. Similarly the peak at 1370 cm-1 (C-H stretching of CH3 group) shifts
to 1373 cm-1, while the peak for C=C double bond shifts from 1660 cm-1 to 1658 cm-1,
and the band at 1467 cm-1 for –CH2 in LDPE (Figure 5.12) is shifted to 1462 cm-1. The
spectra thus confirm that there exist only physical interactions in NR-PE blend.
Polyurethane TPE are block copolymers consisting of alternating hard and soft segments,
separated into two phases [54] due to thermodynamic incompatibility.
188
Figure 5.12 Infrared spectra (a) LDPE, (b) NR, (c) NR/LDPE (50/50 blends)
Reprinted from [53]
Two series of polyether polyurethanes (PU) based on hydroquinone bis (β-hydroxyethyl)

ether (HQEE) or 1,4-butanediol (BDO) as a chain extender were prepared by the one
step bulk polymerisation process. By varying the mole ratio of poly tetra methylene
oxide (PTMO) extender (with Mn = 1000 and Mn = 2000) and 4,4´-diphenylene methane
diisocyanate (MDI) the two series of HQEE (PU1000H1, PU 1000H2, PU2000H1,
189
PU2000H2, PU2000H3) and BDO (PU1000B1, PU1000B 2, PU2000B1, PU2000B2,

PU2000B3) were prepared [55]. Evidence of hydrogen bond formation is obtained by
means of FT-IR spectroscopy [55]. The hard segment units derived from MDI react with
(a) HQEE and (b) BDO according to Scheme 5.6.
Scheme 5.6 Structure of HQEE-MDI and BDO-MDI

Reprinted with permission from L. Zha, M. Wu and J. Yang, Journal of Applied Polymer
Science, 1999, 73, 2895. Copyright 1999, John Wiley and Sons, Inc.
In both cases, the N-H group of the urethane serves as the proton donor, while the
acceptor includes the carbonyl oxygen in the urethane group. It is generally agreed that
C=O stretching region is favourable for the estimation of hydrogen bonding in PU,
qualitatively and quantitatively. From the spectra, it can be observed that the absorption
band at 3451 cm-1 associated with the free NH groups, seems to be absent in HQEE-
based PU. This indicates that approximately 100% N-H groups in HQEE-based PU are
hydrogen bonded. The carbonyl peak of urethanes splits into the following two peaks:
one at about 1702 cm-1 results from self-association of NH with C=O in the interior of
the hard domains; and the other peak at 1732 cm-1 is related to the free carbonyl groups
present in the mixed soft segment phase. To assist in probing the complex by hydrogen
bonding structure from the IR spectra of the PU, the technique of Fourier deconvolution
[55] has been utilised. The deconvoluted spectra for carbonyl stretching region are shown
in Figure 5.13.
As can be seen from Figure 5.13 the lowest frequency associated with ordered hydrogen
bonded carbonyl in crystallite for HQEE based PU is lower by 4 cm-1 than BDO-based
PU (1698 and 1702 cm-1) [56]. This indicates that the former shows a higher extent of
ordering and a stronger ordered hydrogen bond. Compared with the BDO-based hard
segment, the presence of benzene ring (-O-Ph-O-) in the HQEE based hard segment
190
Figure 5.13 Deconvoluted FTIR spectra in the C=O stretching region of (a) HQEE
based PU and (b) BDO based polyurethanes
191
contributes to the ordering due to chain interaction. Figure 5.13 shows that two types of
disordered hydrogen bonded carbonyls are observed at frequencies intermediate to the
free and ordered hydrogen-bonded carbonyls for HQEE-based PU, whereas only one is
observed for BDO-based PU.
As were assigned previously [56], for HQEE based PU and BDO based PU the band at
around 1715 cm-1 is associated with the hydrogen bond that is short-range in the
amorphous phase of the hard domains or the interfacial phase region, whereas the band
at about 1708 cm-1 corresponds to the long-range hydrogen bond in the amorphous
phase of the hard domains. With an increase in the hard segment content for HQEE-
based polyurethanes the absorption band at 1732 cm-1 associated with free carbonyl
diminishes, while the band at 1698 cm-1 increases dramatically among hydrogen-bonded
carbonyl stretching bands. Qualitatively, the spectra for hydrogen bonded carbonyl is
dominated by the band at 1708 cm-1 at a low hard segment content. As the hard segment
content is raised, the dominant band is replaced by 1698 cm-1 components. Hence it can
be concluded that compared with the BDO-based hard segment, the (-O-Ph-O-) linkage
in HQEE makes the hard segment more incompatible with polyether soft segment and
the cohesion in the hard domain enhances.
Hydrogen bonding plays a critical role in determining the morphology and properties of
polyamides and PU and polyurethane-ureas. Yilgor and co-workers studied the hydrogen
bonding in polydimethyl siloxane and polyether based urethane and urea type segmented
copolymers with IR spectroscopy [57]. The extent of hydrogen bonding in PU or polyureas
can be qualitatively studied by determining the frequency shifts in a hydrogen bonded (-N-
H-) and (-C=O) peaks, that is the peak due to (-N-H…. O=C) relative to the free (-N-H)
and (-C=O) peaks. Table 5.2 gives a detailed list of absorption frequencies for various
groups and their hydrogen bonded complexes. Comparative FTIR spectra for (-N-H) and
(-C=O) stretching regions of model urethane and urea compounds and the copolymers are
given in reference [57]. Similarly hydrogen bonding in crosslinked urethane/urea polymers
with two soft segments, polypropylene oxide and polybutadiene, have been studied by
Zhao [58] and co-workers with the help of IR spectra. Furthermore, IR analysis has been
used for characterisation of a series of segmented block copolymers of NR and 1,3 butene
diol-toluene diisocyanate [59] and bisphenol A-toluene diisocyanate oligomers [60].
Like PU TPE, blends of thermoplastic polyurethanes and polyamide-12 (PA-12) have

been studied by Polosmak and co-workers [61]. They have mixed two types of
thermoplastic polyurethane (TPU) based on oligoether (polytetramethylene oxide,
molecular weight, 1000) and oligoester (polyethylene butylene glycol adipate, molecular
weight, 2000) and PA 12 were characterised by IR spectra and thermal analysis. IR
spectra of TPU, PA-12 and their blends show that in amide one (A1) carbonyl absorbancy
is seen to split [55] into two main bands with maxima at 1705 and 1730 cm-1. At 1730 cm-1,
192
Table 5.2 Characteristic IR absorption frequencies for polyurethanes and

polyureas
Group Mode Frequency (cm-1)
N-H free 3445-345 0
N-H N-H … N-H 3315-334 0
N-H N-H … E (ether) 3260-329 0
C-O (urethane) free 1730-174 0
C=O (urethane) C=O .. H-N 1730-1710
C=O (urea) free 1690-170 0
C=O (urea) C=O… H-N 1660-1670 (disordered)
C=O C=O…H-N 1630-1645 (ordered )
NH-C=O Amide II 1540-1560
Reprinted from [57] with permission of Elsevier Science Ltd., London, 1999
absorption bands of free carbonyl groups of the urethane fragments are found along with
low energy hydrogen bands, e.g., dimer, trimers, and also hydrogen bonds developing
between NH group and oxygen in the oligoester block. The ratio of optical densities of
the bands at 1705 and 1730 cm-1 can also show the degree of self-association of urethane
fragments. Addition of PA-12 results in variation of the shoulder intensity ratio in A1
band and the extent of this influence depends on the nature of TPU, plastomer and
concentration of the latter. In ‘optical blends’ which are obtained by putting PA-12 and
TPU films on a surface of germanium crystal, the D1705/D1730 ratio (= optical density
of the bands at 1705 cm-1/optical density of the bands at 1730 cm-1) is constant being
independent on composition. Due to this, its variation in real blends results from the
influence of plastomer on hydrogen interactions in TPU. Addition of PA-12 lowers the
D1705/D1730 ratio for ester containing TPU2 (ester TPU), whereas for oligoether containing
TPU1 (ether TPU) it increases compared with the virgin thermoplastic elastomers. So in
the blends with PA-12, the TPU2 domain structure undergoes partial degradation whereas
for TPU1 the extent of micro segregation increases. For two TPU in blends with PA-12,
NH is usually shifted to lower frequency region, indicating formation of higher energy
hydrogen bonds than that in virgin components (PA-12 and TPU). It is observed that the
effect of thermoplastic polymer on hydrogen bonding in TPU depends not only on its
ability to form hydrogen bonds with PA-12 macromolecules, but also on the distribution
pattern of PA-12 in individual blocks [62]. Jha and co-workers [63, 64] prepared thermoplastic
193
elastomeric blends of Nylon 6 and acrylate rubber (ACM) and have studied the interaction
between the two components with IR spectroscopy [63], which shows reduction in the
intensities of the peaks corresponding to epoxy groups of ACM as well as carboxylic acid
and amine end groups of Nylon 6, suggesting that a chemical reaction between the above
groups takes place at the processing condition.
IR spectroscopy has been used to detect the chemical changes during the thermal degradation
of PU-epoxide resin (ER) foam [65] and photochemical degradation of aliphatic polyester
urethane [66]. According to Zhang and co-workers [65], the thermal stability of the rigid
neat polyurethane foam (PUF) is slightly improved when the ER is incorporated. With increase
of temperature, two stages of weight-loss process are observed for the rigid PU/ER
interpenetrating network foam (IPNF) and for the neat PUF. FT-IR analysis [66] suggests
that the first stage of weight loss is due to the degradation of the polyol-derived blocks of the
PU and the second weight loss stage is governed by the degradation of the MDI-derived
blocks of the PU, along with the degradation of the ER component. FT-IR analysis shows
that the ether bond (at 1072cm-1) disappears first at around 300 °C and the urethane group
(at 3323 and 1722 cm-1) is finally degraded at about 400 °C. Photooxidation of polyester-
urethane takes place through a free radical mechanism [66]. Conventional PU based on
aromatic diisocyanates such as MDI and toluene diisocyanate (TDI) are known to undergo
yellowing on exposure to UV radiation. It is generally agreed that the principal structural
species responsible for discolouration of MDI-based PU are mono and diquinone imide [66].
It is shown by FT-IR spectra that long wavelength irradiations provoke an induced oxidation
of urethane functions. This reaction is initiated by hydrogen atom abstraction on the methylene
groups in the α-position of nitrogen atoms. On irradiation at short wavelength the results
provide evidence for a dual mechanism of photooxidation and induced oxidation, which are
characterised by an appreciable loss of ester structure, as indicated by the decline of the
bands at 1731, 1263, 1167 and 1065 cm-1. Actually photo scissions involve the homolysis of
C-N and C-O bonds producing –COOH- and H2N-COO- groups, and with further irradiation,
decarboxylation takes place.
Introduction of ionic functional groups, which are pendent on the backbone chain, results in
formation of ionomers. In the case of polyblends, intermolecular ionic interactions facilitate
compatibilisation [67, 68]. The mechanical properties of the polyblends are dependent on
the specific interactions between the component polymers promoting interfacial adhesion
[49, 69]. There are several reports on compatible blends, wherein compatibility is induced by
ion-ion and ion-dipole interactions [70, 71]. Recently De and co-workers have reported
studies on thermoplastic elastomers based on ionomeric polyblends [72-81]. They have
characterised the blends with the help of infrared spectroscopy. For example, an ionic
thermoplastic elastomer (ITPE) [75] is prepared by melt blending zinc salts of carboxylated
nitrile rubber (Zn-XNBR) and maleated high density polyethylene (Zn-mHDPE) according
to formulations given in Table 5.3. The synergism in physical properties of the resulting
194
Table 5.3 Formulation of the mixes

Ingredient ph r
Mix number M0 M1 M2 M3 M4 M5 M6
XNBR 100 90 80 70 60 50 0
m-HDPE 0 10 20 30 40 50 10 0
ZnO 12 12 12 12 12 12 12
Stearic acid 1 1 1 1 1 1 1
phr: parts per hundred rubber
ITPE is due to the formation of strong intermolecular ionic crosslinks, which act as a
compatibiliser. The IR spectra [75] of Zn-mHDPE and Zn-XNBR in the range of 1750-
1250 cm-1 are shown in Figure 5.14. The absence of the band corresponding to hydrogen
bonded carboxylic acid pairs (1700-1720 cm-1) indicate almost complete neutralisation
of the acid groups in both polymers.
The spectrum of Zn-XNBR shows a weak band at 1665 cm-1, which is ascribed to the
–C=C— stretching mode and the asymmetric carboxylate stretching region shows a doublet
at 1587 and 1541 cm-1. These bands are strong and intense compared to the spectrum of
Zn-mHDPE, which shows a doublet at 1596 and 1552 cm-1. The stronger asymmetric
carboxylate stretching band observed for Zn-XNBR is due to the presence of a higher
proportion of carboxylate ions in Zn-XNBR than in Zn-mHDPE. The splitting of the
asymmetric carboxylate stretching band into a doublet is assigned to the different co-
ordinated structures of the zinc cation. The band, which is strong and intense at 1462 cm-1,
is due to a –CH2 bending vibration of Zn-mHDPE. The bands observed at 1445 and
1415 cm-1 may also have a contribution from the symmetric carboxylate stretching. A
weak band at 1358 cm-1 accounts for –CH2- wagging. IR spectra of ionomeric polyblends
(mixes M1, M3 and M5) in the range of 1750-1250 cm-1 (assigned for octahedral [82]
structure of zinc carboxylate ion) are shown in Figure 5.14. These spectra show a strong
and intense doublet at 1587 and 1541 cm-1 in asymmetric carboxylate stretching region.
The band at 1587 cm-1 is assigned to the tetrahedral structure of zinc-carboxylate and the
1541 cm-1 band accounts for octahedral structure. It is interesting to note that as Zn-
mHDPE content in the blend increases, the intensity of the 1541 cm-1 band decreases as
compared to that of 1587 cm-1 band. The intensity of the –C=C- stretching band at 1665 cm-1
decreased with increase in Zn-mHDPE content in the blend. The IR spectrum of M5 shows
a clearly resolved band at 1462 cm-1 indicating a –CH2- bending vibration. The difference
195
Figure 5.14 Infrared spectra of (a) Zn-mHDPE, (b) Zn-XNBR, (c) Mix M1, (d) Mix
M3, (e) Mix M5 in the range 1750-1250 cm-1
spectrum [75] in Figure 5.15 is obtained by subtracting the weighted addition spectra of
the neat polymers from the observed spectra of the corresponding blends. In the case of
incompatible blends, the spectrum of the blend should be similar to that obtained by the
summation spectra of the neat polymers [69]. But the compatible blends show marked
changes in the spectra. It is evident from the difference spectra that the ionic interaction is
stronger in the blends than that in neat polymers. This is supported by the positive absorption
band at 1586 cm-1, which is due to the tetrahedral asymmetric COO- stretching band in
the difference spectra. It is also noted that the 1552 cm-1 band intensity in the mixes M1
196
Figure 5.15 Difference spectra obtained by substracting the sum of The spectra of the
neat polymers from the blend spectra (a) Mix M1, (b) Mix M3, (c) Mix M5
Reprinted with permission from P. Antony and S.K. De, Journal of Applied Polymer
Science, 1998, 70, 483. Copyright 1998, John Wiley and Sons, Inc.
197
and M3 and the 1569 cm-1 band intensity in M5 decrease, indicating a change due to the
mutual interaction between the ionic groups present in the ionomers.
A series of sulfonated ionomers have been prepared and characterised by IR studies [83-
85]. Recently Ghosh and co-workers reported the preparation of an ionic thermoplastic
elastomer based on the zinc salt of sulfonated maleated EPDM rubber (Zn-s-m-EPDM)
[86-89]. They also studied incorporation of semi-reinforcing furnace (SRF) and
intermediate super abrasion furnace (ISAF) carbon black and precipitated silica on zinc
salt of sulfonated maleated EPDM rubber [88]. The rubber-filler bonding has been studied
by IR spectroscopy [89]. IR spectra of zinc stearate, precipitated silica, SRF and ISAF
carbon black are shown in Figure 5.16. In Figure 5.16a, zinc stearate shows the
characteristic absorbance at 1577 cm-1, which is due to asymmetric stretching of the
bridging type of carboxylate groups. The spectrum of precipitated silica is characterised
by the strong absorbance at 1096 cm-1, which is due to Si-O-Si asymmetric stretching.
The other characteristic peaks at 3760, 960, 798 cm-1 are due to -OH stretching of Si-
OH, Si-O stretching of Si-OH and Si-OH deformation, while the peak at 470 cm-1 is due
to Si-O-Si symmetric stretching [90]. The peaks at 798 and 470 cm-1 for precipitated
silica are not shown in Figure 5.16a. SRF carbon black shows characteristic absorbance
at 3730 cm-1, which is due to O-H groups in the substituted phenolic compounds [91].
The broad peak at 3162 cm-1 is attributed to composite absorption of hydrogen bonded
–OH groups from alcohols, phenols and enols. The peak at 1693 cm-1 is due to carbonyl
stretching. ISAF black shows bands at 1350 and 1385 cm-1 due to –OH bending in water
and C-O stretching in phenols [92]. The band at 1574 cm-1 confirms the presence of
tetra- hydroquinone and the peak around 1634 cm-1 is due to polycyclic quinines [91].
The absorbance bands at 1750 and 3400 cm-1 are due to lactones and hydrogen bonded
phenolic –O-H groups. The IR spectra of silica-filled compound is shown in Figure 5.16b,
where the peak at 3472 cm-1 is believed to be due to hydrogen-bonded structure in the
vicinal (adjacent) groups, whereas the peak at 3666 cm-1 is due to hydrogen bonded
structures. The peak in the asymmetric stretching region occurs at 1567, 1548, and
1532 cm-1 and the carboxylate symmetric stretching occurs at 1410 cm-1 The peak at
1189 cm-1 is due to the asymmetric stretching of the sulfonate groups. The split pattern
between 1000 and 1200 cm-1 indicates specific interaction of silica filler with sulfonate
groups. The ionomeric compositions filled with SRF and ISAF carbon blacks show a
broad band in the sulfonate stretching frequency region (800-1400 cm-1). The SRF carbon
black shows a broad band at 3225 cm-1 due to a composite absorption by hydrogen
bonded O-H groups present in both carbon black and the ionomer [91, 92]. The ISAF
carbon black-filled ionomer shows band splitting in the high-frequency region, a carbonyl
stretching frequency at 1656 cm-1 and hydrogen-bonded carboxylate anions at 1572 and
1536 cm-1. Similarly, IR spectrum of the zinc stearate filler-loaded ionomers shows peaks
at 1585, 1547, 1528 and 1223 cm-1 and thus indicates that an interaction occurs between
the active sites of the filler and the zinc stearate-filled ionomer.
198
Figure 5.16 (a) IR spectra of zinc stearate, SRF carbon black, silica, ISAF carbon
black, (b) Infrared spectra of Zn-s-m-EPDM containing zinc stearate, SRF carbon
black, silica and ISAF carbon black
Reprinted from [89] with permission of John Wiley and Sons, 2000
UV curable waterborne coatings made of PU-acrylate ionomers have been synthesised and
characterised by IR spectroscopy [93]. Similarly a series of novel PU ionomers with polydioxolane
(PDXL) as soft segment is prepared and identified by FT-IR spectroscopy [94].
199
5.5 Conclusion
Different types of interaction in rubbery materials, which include PE, plasticised PVC,
TPE, ionomeric polyblends, and blends of plastics and rubbers can be characterised by
different types of IR spectroscopic techniques.
Acknowledgements
The author thanks Professor S.K. De, Professor, Rubber Technology Centre, IIT Kharagpur
for his moral support and constant co-operation in proof reading. Thanks are due to Mr.
R.S. Rajeev, Mr. Shambhu Bhattacharyya and Mr. Bhuwneesh Kumar for their assistance
in preparing the manuscript.
References
1. P. Ghosh, Polymer Science and Technology of Plastics and Rubbers, Tata

McGraw-Hill Publishing Company Limited, New Delhi, 1990, Chapter 1.
2. J.A. Brydson, Rubbery Materials and Their Compounds, Elsevier Applied

Science, London, 1988, Chapter 1.
3. J.A. Brydson, Rubbery Materials and Their Compounds, Elsevier Applied

Science, London, 1988, Chapter 16.
4. W.F. Maddams in Analysis of Polymer Systems, Eds., L.S. Bark and N.S. Allen,
Applied Science Publishers, London, 1982, Chapter 3, 59.
5. B. Jasse in Developments in Polymer Characterisation, Ed., J. V. Dawkins,

Applied Science Publishers, London, 1983, Chapter 3, 119.
6. K. Holland-Mortiz, I. Holland-Mortiz and K. Van-Werden, Colloid and Polymer

Science, 1981, 259, 156.
7. N.M. Bikales, Characterisation of Polymers, Wiley-Interscience, New York,

1971, 144.
8. A. Singhal and L.J. Fina, Polymer, 1996, 37, 12, 2335.
9. G. Zerbi, G. Gallino, D.N. Fanti and L. Baini, Polymer, 1989, 30, 2324.
10. E. Agosti, G. Zerbi and M. I. Ward, Polymer, 1992, 33, 4219.
200
11. I. Ray, S. Roy and D. Khastgir, Polymer Bulletin, 1993, 30, 685.
12. I. Ray, S. Roy, T.K. Chaki and D. Khastgir, Journal of Elastomers and Plastics,
1994, 26, 168.
13. B. Sultan and E. Sorvik, Journal of Applied Polymer Science, 1991, 43, 1947.
14. A.K. Sen, B. Mukherjee, A.S. Bhattacharyya, P.P. De and A.K. Bhowmick, Die
Angewandte Makromolekulare Chemie, 1991, 191, 3206, 15.
15. A.K. Sen, B. Mukherjee, A.S. Bhattacharyya, L.K. Sanghi, P.P. De and A.K.
Bhowmick, Journal of Applied Polymer Science, 1991, 43, 1673.
16. A.K. Sen, B. Mukherjee, A.S. Bhattacharyya, P.P. De and A.K. Bhowmick, Journal
17. Y-T. Shieh and C-M. Liu, Journal of Applied Polymer Science, 1999, 74, 3404.
18. T.K. Chaki, R.S. Despande, A.B. Majali, V.K. Tikku and A.K. Bhowmick, Die
Angewandte Makromolekulare Chemie, 1994, 217, 3744, 61.
19. H.M. Abou Zeid, Z.I. Ali, T.M. Abdel Maksoud and R.M. Khafagy, Journal of
Applied Polymer Science, 2000, 75, 179.
20. Y.C. Lai, Journal of Applied Polymer Science, 1997, 66, 1475.
21. B.L.D. Silveria, European Polymer Journal, 1993, 29, 1095.
22. A-Z.A. Elassar, N.M. El-sawy and F.A. Aisagheer, Journal of Applied Polymer
Science, 1999, 74, 2963.
23. F.L. Scholten and M. Wolters, Plastics and Rubber and Composite Processing and
Applications, 1998, 27, 10, 465.
24. T. Hjertberg and J.E. Lakso, Journal of Applied Polymer Science, 1989, 37, 1287.
25. L. Ulren, T. Hjertberg and H. Ishida, Journal of Adhesion, 1990, 31, 117.
26. E.U. Okoroafor and S.J. Spells, Polymer, 1994, 35, 21, 4578.
27. D. Daoust, S. Bebelman, N. Chaupart, R. Legras, J. Devaux and J. Costa, Journal

28. M.M. Coheman and J. Zarian, Journal of Polymer Science: Polymer Physics
Edition, 1979, 17, 837.
201
29. N.R. Manoj, P.P. De and S.K. De, Journal of Applied Polymer Science, 1993, 49,
133.
30. N.R. Manoj and P.P. De, Plastics and Rubber and Composite Processing and
Applications, 1995, 23, 103.
31. N.R. Manoj and P.P. De, Polymer, 1998, 39, 3, 733.
32. G.B. Buttler, K.R. O’Driscoll and M. Shen, Reviews in Macromolecular

Chemistry, Marcel Dekker, New York, 1974, Volume 2.
33. R. Lukas, O. Pradova, J. Michalcova and V. Paleckova, Journal of Polymer

Science: Polymer Letters, 1985, 23, 85.
34. L.I. Nas, Encyclopedia of PVC, Marcel Dekker, New York, 1977.
35. N.R. Manoj and P.P. De, Polymer Degradation and Stability, 1994, 44, 43.
36. W. Meissner and D. Zuchowska, Polymer Degradation and Stability, 1998, 60, 415.
37. H. Baltacloglu, D. Balkose, Journal of Applied Polymer Science, 1999, 74, 2488.
38. C. Anton-Prinet, G. Mur, M. Gay, L. Audouin and J. Verdu, Polymer

Degradation and Stability, 1998, 60, 2-3, 265.
39. C. Anton-Prinet, J. Dubois, G. Mur, M. Gay, L. Audouin and J. Verdu, Polymer



42. M. Turoti, J.Y. Olayemi, J.B. Adeniyi and O. Peters, Polymer Degradation and
Stability, 1998, 61, 297.
43. N. Guarrotxena, G. Martinez and J. Millan, Polymer, 1999, 40, 629.
44. G. Martinez, C. Garcia, N. Gurrotxena and J. Millan, Polymer, 1999, 40, 1507.
45. K. Sreenivasan, Journal of Applied Polymer Science, 1999, 74, 113.
46. E.N. Kresgei, Polymer Blends, Eds., D.R. Paul and S. Newman, Academaic Press,
New York, 1978, Chapter 20.
202
47. A.Y. Coran and R. Patel, Rubber Chemistry and Technology, 1982, 54, 116.
48. M.M. Coleman, J.F. Graf and P.C. Painter, Specific Interactions and Miscibility of
Polymer Blends, Technomic Publishing Company, Lancaster, PA, 1991.
49. O. Olabisi, L.M. Robeson and M.T. Shaw, Polymer-Polymer Miscibility,

Academic Press, New York, 1979.
50. S. Akhtar, P.P. De and S.K. De, Journal of Materials Science, Letters, 1986, 5, 399.
51. S. Akhtar, B. Kuriakose, P.P. De and S.K. De, Plastics and Rubber and Composite
Processing and Applications, 1987, 7, 11.
52. N. Roychoudhury and A.K. Bhowmick, Journal of Materials Science, 1988, 23,
2187.
53. P.P. De, Private Communication.
54. S.L. Copper and A.V. Tobolsky, Journal of Applied Polymer Science, 1966, 10,
1837.
55. L. Zha, M. Wu and J. Yang, Journal of Applied Polymer Science, 1999, 73, 2895.
56. M. Harthcock, Polymer, 1989, 30, 1234.
57. E. Yilgor, E. Burgaz, E. Yurtsever and I. Yilgor, Polymer, 2000, 41, 849.
58. C-T. Zhao, M.N. de Pinho, Polymer, 1999, 40, 6089.
59. C.J. Paul, M.R. Gopinathan Nair, N.R. Neelkantan, P. Koshy, B.B. Idage and
A.A. Bhelhekar, Polymer, 1998, 39, 6861.
60. C.J. Paul, M.R.G. Nair, P. Koshy, B.B. Idage, Journal of Applied Polymer Science,
1999, 74, 706.
61. S.S. Pesetskii, V.D. Fedorov, B. Jurkowski and N.D. Polosmak, Journal of
Applied Polymer Science, 1999, 74, 1054.
62. S.S. Pesetskii, N.D. Polosmak, N.L. Malinin, N.V. Koval, Byelorussia Soviet
Socialistic Republic Academy of Science Reports (in Russian), 1990, 34, 616.
63. A. Jha and A.K. Bhowmick, Rubber Chemistry and Technology, 1997, 70, 798.
64. A. Jha, A.K. Bhowmick, R. Fujitsuka and T. Inoue, Journal of Adhesion Science
203
65. Y. Zhang, R.J. Heath and D.J. Hourston, Journal of Applied Polymer Science,
2000, 75, 406.
66. C. Wilhelm and J-L. Gardette, Polymer, 1997, 38, 16, 4019.
67. W.J. Macknight and R.D. Lundberg in Thermoplastic Elastomers, Eds., N.R.
Legge, G. Holden and H.E. Schroeder, Hanser, Munich, 1987, 245.
68. J.J. Fitzgerald and R.A. Weiss, J. Macromol. Sci., Chem. Rev. Macromol. Chem.
Phys., 1988, 28, 1, 99.
69. M.M. Coleman and P.C. Painter, Applied Spectroscopy Reviews, 1984, 20, 255.
70. Z.L. Zhou and A.E. Eisenberg, Journal of Polymer Science: Polymer Physics
Edition, 1983, 21, 595.
71. X. Lu and R.A. Weiss, Macromolecules, 1991, 24, 4381.
72. T. Kurian, S. Datta, D. Khastgir, P.P. De, D.K. Tripathy, S.K. De and D.G. Peiffer,
Polymer, 1996, 37, 4787.
73. S. Datta, P.P. De and S.K. De, Journal of Applied Polymer Science, 1996, 61,
1839.
74. P. Antony and S.K. De, Plastics and Rubber and Composite Processing and
Applications, 1997, 26, 311.
76. P. Antony and S.K. De, Polymer, 1999, 40, 1487.
77. S. Datta, S.K. De, G. Kontos and J.M. Wefer, Journal of Applied Polymer Science,
1996, 61, 177.
78. P. Antony, A.K. Bhattacharya and S.K. De, Journal of Applied Polymer Science,
1999, 71, 1257.
79. P. Antony, S. Datta and S.K. De, Plastics and Rubber and Composite Processing
and Applications, 1998, 27, 303.
81. P. Antony, S. Bandyopadhyay and S.K. De, Journal of Materials Science, 1999,
34, 2553.
204
82. M.M. Coleman, J.Y. Lee and P.C. Painter, Macromolecules, 1990, 23, 2339.
83. K. Sakurai, E.P. Douglas and W.J. Macknight, Macromolecules, 1992, 25, 4506.
84. N.R. Manoj, P.P. De, S.K. De and D.G. Peiffer, Journal of Applied Polymer
Science, 1994, 53, 361.
85. N.R. Manoj, P.P. De, S.K. De and D.G. Peiffer, Polymer, 1993, 34, 10, 2128.
86. S.K. Ghosh, D. Khastgir, S.K. De, P.P. De, R.J. Albalak and R.E. Cohen, Plastics
and Rubber and Composite Processing and Applications, 1998, 27, 7, 310.
87. S.K. Ghosh, P.P. De, D. Khastgir and S.K. De, Polymer Plastics Technology and
Engineering, 2000, 39, 1, 47.
88. S.K. Ghosh, P.P. De, D. Khastgir and S.K. De, Journal of Applied Polymer
Science, 2000, 78, 743.
89. S.K. Ghosh, P.P. De, D. Khastgir and S.K. De, Journal of Applied Polymer
Science, 2000, 78, 326.
90. D.H. Williams and I. Flemming, Spectroscopic Methods in Organic Chemistry,

McGraw-Hill, New York, 1987.
91. J-B. Donnet and A. Voet, Carbon Black, Marcel Dekker, New York, 1976.
92. G. Socrates, Infrared Characteristic Group Frequencies, Wiley-Interscience,

Chichester, 1980.
93. Z. Wang, D. Gao, J. Yang, Y. Chen, Journal of Applied Polymer Science, 1999,
73, 2869.
94. L. Wang, B. Wang, X-L. Wang, X-Z. Tang, Journal of Applied Polymer Science,
1999, 71, 1711.
205
206
6
Crosslinking of EPDM and Polydiene Rubbers
Studied by Optical Spectroscopy
Herman G. Dikland and Martin van Duin
6.1 Introduction
6.1.1 General Introduction to EPDM [1-3]
EPM is a copolymer, consisting of ethylene and propylene. As a result of the random

incorporation of the monomers, EPM with medium ethylene content are amorphous
elastomers. EPM with high ethylene contents have diffuse crystalline regions, yielding
some rigidity although retaining rubbery characteristics. As a result of the absence of
unsaturation, EPM cannot be vulcanised with sulfur and it is not very reactive to peroxide
curing. To increase the reactivity for crosslinking a diene, namely 5-ethylidene-2-
norbornene (ENB) or dicyclopentadiene (DCPD) is terpolymerised in combination with
ethylene and propylene. The resulting EPDM combines a saturated polymer backbone
with residual unsaturation in the side groups. Typical examples of such polymers are
shown in Figure 6.1.
EPM and EPDM, generally referred to as EPDM are classically produced via Ziegler-
Natta catalysis in solution or slurry processes. Over the last decade metallocene catalysis
[4, 5] and gas phase technology [6] have been developed. EPDM is produced on a
commercial scale in a variety of chemical compositions. Typically, the ethylene content
Figure 6.1 Chemical structure of EPDM with 2-ethylidene-5-norbornene (left) and

dicyclopentadiene (right) as diene
207
ranges from 40 to 80 wt.% and the diene content from 0 to 12 wt.%. The weight average
molecular weight of EPDM typically ranges from 100 to 1,000 kg/mole in combination
with molecular weight distributions of 2.0 to 10. The degree of long chain branching can
be controlled to a considerable extent [7]. As for all elastomers, the processing
characteristics of polymers in combination with the compound recipe are used to determine
the processing behaviour of the compound and as well as the mechanical, elastic and
dynamical properties of the final product after crosslinking.
As a result of its saturated polymer backbone, EPDM is more resistant to oxygen, ozone,
UV and heat than the low-cost commodity polydiene rubbers, such as natural rubber
(NR), polybutadiene rubber (BR) and styrene-butadiene rubber (SBR). Therefore, the main
use of EPD(M) is in outdoor applications, such as automotive sealing systems, window
seals and roof sheeting, and in under-the-hood applications, such as coolant hoses. The
main drawback of EPDM is its poor resistance to swelling in apolar fluids such as oil,
making it inferior to high-performance elastomers, such as fluoro, acrylate and silicone
elastomers in that respect. Over the last decade thermoplastic vulcanisates, produced via
dynamic vulcanisation of blends of polypropylene (PP) and EPDM, have been
commercialised, combining thermoplastic processability with rubber elasticity [8, 9].
EPDM was developed and commercialised in the late 1950s. With an annual production
capacity of more than 1,000 kt in 1998 [10]. EPDM is currently the fourth elastomer by
volume and has become more or less a commodity rubber. Actually, EPDM is the largest
non-tyre rubber. The annual growth rate is about 4%. DSM and Exxon are market
leaders with a combined market share of approximately 40%. PP/EPDM-based
thermoplastic vulcanisates which have currently the fastest growing rubber market
(8% per year).
6.1.2 EPDM Crosslinking [1-3]
Most EPDM applications require crosslinking except when used as an impact modifier
for PP, polystyrene (PS) and polyamides or as an oil additive, e.g., as viscosity index
improver or dispersant. Most commonly, accelerated sulfur vulcanisation is used for the
crosslinking of EPDM. As a result of the low amount of unsaturation in EPDM (< 1 mole/
kg versus NR - 15 mole/kg), sulfur vulcanisation of EPDM is rather slow and a relatively
large amount of accelerators is needed. Because of the low polarity of EPDM the solubility
of polar accelerators is limited, often resulting in low effectivity and/or blooming. Typically,
up to 5 different accelerators are used in EPDM formulations. As for other rubbers
environmental issues, such as nitrosamine formation and may be in the future the presence
of zinc, are prompting the development of new accelerator systems.
208
Crosslinking of EPDM and Polydiene Rubbers Studied by Optical Spectroscopy
Saturated EPM can be cured with peroxide/co-agent combinations, but EPDM is cured
even more efficiently due to the presence of the residual unsaturation. The share of peroxide-
cured EPDM is gradually increasing at the expense of sulfur-vulcanised EPDM [11] because
of the high thermal stability of the carbon-carbon crosslinks formed upon peroxide-curing
of EPDM in combination with the increasing high temperature demands in the market.
Peroxide-cured EPDM has a superior compression set at temperatures above 100 °C and it
is generally more heat-resistant than that of sulfur-vulcanised EPDM. In addition, peroxide-
curing is not affected by environmental issues, such as nitrosamine formation and use of
large amounts of the heavy metal zinc. However, with respect to ultimate and dynamic
properties sulfur-vulcanised EPDM is superior to peroxide-cured EPDM as a result of the
dynamic rearrangement of sulfidic crosslinks during sulfur vulcanisation, yielding a rubber
network which is relatively free of internal stresses [12, 13].
Resols (phenol-formaldehyde resins) are commercially used for effective crosslinking of

EPDM in the production of thermoplastic vulcanisates [8]. General studies on rubber
crosslinking for different diene rubbers are presented here.
6.1.3 Studies into the Chemistry of Rubber Crosslinking
The technology of sulfur vulcanisation of unsaturated elastomers has evolved since

Goodyear’s invention in 1839. Scientific studies into the chemistry of sulfur vulcanisation
started to appear in the late 1950s (for reviews see References [14-18]). Two experimental
approaches can be distinguished: the analysis of rubber vulcanisates themselves and the
so-called low-molecular-weight model studies.
Analysis of rubber vulcanisates is a direct approach and is, therefore, to be preferred.

Initially, it suffered from several drawbacks. The number of structures formed during
sulfur vulcanisation is rather large. For example crosslinks with variable sulfur bridge
length, crosslink precursors (pendent sulfur) and cyclic sulfides, and the corresponding
concentrations are small relative to the detection limits of classical spectroscopic
techniques, such as infra-red (IR) and Raman spectroscopy. In the case of Raman
spectroscopy, fluorescence is encountered as a result of the presence of the vulcanisation
chemicals [19]. The presence of fillers, especially carbon black, which are strongly light-
scattering and/or light-absorbing, results in a (strongly) decreased sensitivity for optical
spectroscopy methods in general. Finally, a rubber vulcanisate consists of an infinite
network of crosslinked polymer chains, rendering it insoluble and, thus, inaccessible to
chromatographic techniques and liquid-state nuclear magnetic resonance (NMR).
Therefore most progress towards the understanding of sulfur-vulcanisation chemistry

was originally made by ‘vulcanising’ low-molecular-weight model olefins. Numerous
209
model compound studies by the Tun Abdul Razak Laboratory (UK), the Dunlop Research
Centre (Canada), the Experimental Station of E.I. DuPont de Nemours & Co. (USA),
DSM Research (The Netherlands) and the Universities of Leiden (The Netherlands),
South Elisabeth (South Africa), and Cologne (Germany), among others, have been
published, but a comprehensive review of this work is still not fully available. Because of
their low-molecular weight, the corresponding model crosslinks are soluble and sometimes
even volatile, allowing chromatographic techniques, such as gas chromatography and
liquid chromatography, for separation and also liquid-state NMR for identification.
Clearly, the results of model olefin studies should be used with care. Several practical
pitfalls have been encountered in the past and a ‘translation’ of the results to the polymer
system may also be a source of errors [20, 21].
Over the last decade the development of advanced analytical techniques, such as Fourier
transform (FT) Raman and solid-state NMR spectroscopy, have been impressive,
resulting in a great deal of progress in the field of the sulfur vulcanisation of unsaturated
elastomers [22-25].
In general, most of the problems encountered in the study of the chemistry of the sulfur
vulcanisation of elastomers are also encountered in the study of peroxide-curing. In
comparison with sulfur vulcanisation only a limited number of spectroscopic studies on
peroxide-curing have been published.
6.1.4 Scope
The aim of this chapter is to review optical spectroscopy studies on sulfur and peroxide
crosslinking of polydiene rubbers, such as NR and BR (Sections 6.2.1 and 6.3.1, respectively),
and to discuss in detail recent FT-Raman and FT-IR spectroscopy studies into the sulfur
and peroxide crosslinking of EPDM (Sections 6.2.2 and 6.3.2, respectively). The results of
optical spectroscopy studies will also be discussed in the light of results obtained with
other techniques. Finally, the elucidation of the chemical structures of the crosslinks formed
will allow enhanced understanding of the mechanisms of crosslinking and some preliminary
insight into the structure/property relationships of crosslinked rubber.
Raman and IR spectroscopic studies dealing with the qualitative and/or quantitative
determination of rubber compounding ingredients, i.e., the elastomer itself [22, 26-31],
fillers [32, 33], vulcanisation chemicals and other additives [34-37], are not included
here. The same applies to studies dealing with the crosslinking of elastomers by means of
chemicals other than sulfur or peroxide [38-41], self-crosslinking of elastomers blends
[42-44], crystallisation (strain-induced) [45-48] and oxidation/ageing [49-53].
210
Raman and IR spectroscopy are similar techniques, since both are used to study oscillations
of structural moieties of molecules, namely stretching and bending vibrations [54]. Raman
and IR spectroscopy are different in the sense that strong Raman bands are associated
with vibrational motions of highly polarisable bonds and, so, Raman spectroscopy is
more sensitive for vibrations of apolar groups. In contrast, intense IR bands are primarily
due to motions of bonds with strong permanent dipole moments and thus, IR spectroscopy
is more sensitive for vibrations of polar groups. As a result, Raman and IR spectroscopy
are usually seen as complementary techniques which have proven their value in polymer
science and technology [55-56].
FT-Raman spectroscopy was used for studying the sulfur vulcanisation of EPDM, since
apolar moieties, such as the residual unsaturation of the third monomer of EPDM, and
sulfur crosslinks are involved (Section 6.2.2). FT-IR spectroscopy was used to study the
contribution of polar co-agents during peroxide-curing of EPDM (Section 6.3.2). The
FT technique was essential for both studies to obtain high-quality spectra with a sufficient
signal-to-noise ratio. For Raman spectroscopy the problem of fluorescence was overcome
by excitation of the Raman scattering at 1064 nm using a Nd:YAG laser.
6.2 Sulfur Vulcanisation
6.2.1 Sulfur Vulcanisation of Polydiene Rubbers
6.2.1.1 General
In the late 1960s optical spectroscopy was first mentioned for studies on the sulfur
vulcanisation of unfilled rubber. For instance, Stewart and Linnig showed, using IR
spectroscopy, that sulfur vulcanisation of NR in the presence of tetramethyl thiuram
disulfide (TMTD) and ZnO resulted in the formation of zinc dimethyl dithiocarbamate
(ZDMC) [57]. Some years later three series of optical spectroscopy studies were published
by the Koenig group [19, 58-63], which can be seen as a breakthrough, since they provided
the first structural data for sulfur vulcanisation of high-molecular-weight elastomers.
These studies allowed the verification of (intermediate) structures identified in the series
of low-molecular-weight studies by Porter and co-workers and the validation of the
mechanisms proposed [14, 16]. However, several experimental problems had to be
overcome before final conclusions could be drawn from Raman spectra. Firstly, the
problem of strong fluorescence due to the presence of accelerator impurities and/or
211
oxidation products, which resulted in low signal/noise spectra, had to be solved by

choosing optimum vulcanisation conditions, carefully extracting the vulcanisates and
laser irradiation of the samples for several hours before recording the spectra [19, 58].
Secondly, because of a strong overlap of the various peaks, the Raman spectra had to be
curve-resolved [58, 59]. Finally, tentative assignments had to be verified by using selective
extraction solvents and/or chemical probes for the selective conversion of disulfides and
polysulfides [59].
6.2.1.2 Sulfur Vulcanisation of cis-BR
Table 6.1 gives an overview of the structures that were finally identified by Koenig and
co-workers using Raman spectroscopy during the vulcanisation of cis-BR at 140 or 150 °C
[19, 58-59]. For sulfur-free vulcanisation of BR using only TMTD and ZnO not only the
formation of ZDMC but also the formation of alkenylsulfides and of monosulfidic,
disulfidic and polysulfidic crosslinks were demonstrated. In addition, main-chain
modifications were observed, such as cis-trans isomerisation and the formation of
conjugated trienes and cyclic unsaturated sulfides. For sulfur vulcanisation of cis-BR in
the presence of mercaptobenzothiazole (MBT), ZnO and lauric acid, similar structures
were identified with the exception of monosulfides. Sulfur/TMTD/ZnO/lauric acid
vulcanisation was studied as a function of time and the results were interpreted in
combination with data obtained from crosslink density measurements. As expected, similar
kinetics were found for the formation of alkenylsulfides and sulfidic crosslinks as well as
the development of the crosslink density, until a plateau was reached. Interestingly, the
triene formation followed a similar time profile. However, cis-trans isomerisation and
diene formation were shown to continue beyond the optimum cure time, indicating that
they were independent of the vulcanisation process.
Using IR the degree of cis-trans isomerisation of cis-BR during sulfur vulcanisation was
quantified. Bishop showed high degrees of isomerisation from 95% to about 35% cis if
large amounts of sulfur were used for vulcanisation [64], whereas Madge [65] and Blümel
[66] found only modest degrees of isomerisation from 95% to 88%, if much lower
sulfur concentrations were applied. These high levels of isomerisation again demonstrate
that isomerisation and vulcanisation are independent processes. Devlin showed that in
the presence of carbon black the cis-trans isomerisation was markedly reduced from
97% to 94% using ‘typical’ amounts of sulfur [67].
Coleman of the Koenig group identified the reaction products obtained after heating
various mixtures of sulfur, TMTD, tetramethyl thiurammonosulfide, ZDMC and/or ZnO
in the absence of elastomer or olefin using Raman spectroscopy. This enabled the
212
Table 6.1 Raman identification of structures formed upon

accelerated sulfur vulcanisation of cis-BR [19, 58, 59]
Raman band (cm-1) Structure
1575 to 1700 C=C stretching
1664 trans C=C
1650 cis C=C
1640 vinyl C=C
1633 C=C-C-S
a
1605 conjugated C=C-C=C
1623 conjugated C=C-C=C-C=C
1587 conjugated C=C-C=C-C=C
400 to 900 C-S and S-S stretching
744 alkenyl monosulfide C-S
720 alkenyl disulfide C-S
696 cyclic sulfide C-S
635 unsaturated 6-ring cyclic sulfide C-S
577 pendent dithiocarbamate
505 - 510 disulfide S-S
505 unsaturated 5-ring cyclic sulfide C-S
424 polysulfide S-S
a: tentative assignment
elucidation of the complex thiuram sulfide and dithiocarbamate chemistry [60, 61]. These
reactions will probably also occur during sulfur vulcanisation of unsaturated elastomer,
but it has to be realised that the reaction mechanism in the latter case is more intricate
and selectivities may be totally different. Intermediate free radicals, such as the thiuram
radicals and the thiurampersulfenyl radicals, were identified using electron spin resonance
spectroscopy. Sulfur and/or TMTD vulcanisation of BR proceed probably via free radical
mechanisms, whereas in the presence of ZnO or ZDMC contributions of ionic pathways
will dominate.
213
6.2.1.3 Sulfur Vulcanisation of NR
Chen and co-workers tentatively assigned new signals in the FT-IR spectra of accelerated
sulfur-vulcanised NR to the formation of C-S and S-S bonds corresponding to
monosulfides, disulfides and polysulfides [68]. The vulcanisation of NR was retarded
when clay was added to the NR compound.
Recently, it was shown using FT-IR that the decrease of vulcanisation rate and final
crosslink density of sulfur vulcanised NR upon increasing silica content may be related
to increased absorption of zinc stearate onto the silica surface [69].
In a third series of papers from the Koenig group, reversion of accelerated sulfur-vulcanised
NR was studied using IR spectroscopy in combination with crosslink density
measurements [62, 63]. Chen and co-workers showed that at 140 to 150 °C reversion of
sulfur-vulcanised unfilled NR occurs at high sulfur/accelerator ratios, which could be
related to the isomerisation of the cis-2-methyl-2-butene monomeric unit of NR into a
trans-2-methyl-3-butene unit and to the formation of conjugated trienes. It was shown
that reversion is a result of neither oxidation nor isomerisation of the cis-2-methyl-2-
butene unit into the corresponding trans unit. When NR compounds were extended
with carbon black, FT-IR spectra could still be recorded for carbon black levels as high
as 30 phr. In the presence of carbon black less cis-trans isomerisation was observed,
which was explained by the higher thermal stability of polysulfides as a result of polymer-
filler interactions. It was also shown that the reversion resistance was improved if more
reinforcing carbon black was used.
Schotman and co-workers tentatively assigned the new Raman peaks at 1625 and
1592 cm-1 observed during sulfur vulcanisation of squalene, to the formation of conjugated
dienes and trienes, respectively [70]. When vulcanisation was carried out in the presence
of 1,3-di(citraconimidomethyl)benzene, this resulted in a reduced intensity of these two
new peaks, corroborating that conjugated dienes and trienes, formed as a result of
reversion, react with the diimide. Obviously, the diimide is not an anti-reversion agent in
the sense that it prevents reversion, but it is in the sense that it repairs crosslinks when
reversion has occurred.
6.2.1.4 Mechanism of Sulfur Vulcanisation of Polydiene Rubbers
The results of the optical spectroscopy studies into sulfur vulcanisation of polydiene
rubbers correspond well with the results obtained via low molecular weight model olefin
studies and solid state 13C NMR studies. From all these studies the mechanism for
accelerated sulfur vulcanisation as shown in Figure 6.2 has emerged [14-18], which is
214
Figure 6.2 Simplified scheme of the mechanism of accelerated sulphur vulcanization of

polydiene rubbers [14-18]
now commonly accepted, although it should be realised that some details of the various
reaction steps still need to be refined.
First, sulfur, accelerator and activators react towards a soluble sulfurated zinc complex
(activated sulfur). This complex reacts with the unsaturated elastomer by substituting a
labile allylic hydrogen atom, which results in the attachment of accelerator residues to
the elastomer chain (pendent sulfur or crosslink precursor). This intermediate is converted
into a sulfur crosslink either via disproportionation with a second pendent sulfur structure
or by allylic substitution of the unsaturation of a second elastomer chain. Recently, it
was shown that the former route is probably most likely [71]. It is noted that during
accelerated sulfur vulcanisation dialkenylsulfides are predominantly formed, indicating
that the unsaturation is not consumed during reaction but that it activates the α-position,
whereas during unaccelerated sulfur vulcanisation mixtures of dialkenylsulfides,
dialkylsulfides and alkenylalkylsulfides are formed. Upon prolonged vulcanisation
215
network maturation reactions may occur, for example, the sulfidic crosslink may exude
sulfur, resulting in crosslink shortening, and/or it may be converted into cyclic sulfides,
which is usually associated with reversion.
6.2.2 Sulfur Vulcanisation of EPDM
6.2.2.1 Early Studies
The mechanism of the accelerated sulfur vulcanisation of EPDM is probably similar to

that of the highly unsaturated polydiene rubbers. The vulcanisation of EPDM has been
studied with emphasis on the cure behaviour and mechanical and elastic properties of
the crosslinked EPDM. Hardly any spectroscopic studies on the crosslinking chemistry
of EPDM have been published, not only because of the problems discussed in Section 6.1.3
but also because of the low amount of unsaturation of EPDM relative to the sensitivity
of the analytical techniques. For instance, high-temperature magic-angle spinning solid-
state 13C NMR spectroscopy of crosslinked EPDM just allows the identification of the
rubber type, but spectroscopic evidence for the presence of crosslinks is not found [72].
Only two spectroscopic studies on sulfur vulcanisation of EPDM by Fujimoto and co-
workers are available [73-74]. Using attenuated total reflectance (ATR) IR spectroscopy
they showed that during sulfur/TMTD/MBT/ZnO/stearic acid vulcanization, the C=C
bands at 3035, 966 and 870 cm-1 of the residual unsaturations of the EPDM third
monomers, DCPD, 1,4-hexadiene (HD) and 5-methylidene-2-norbornene (MNB),
respectively, decreased in intensity as a function of time at 140 and 150 °C. The relative
decrease in intensity was shown to correlate with the increase in crosslink density. In
Sections 6.2.2.2 and 6.2.2.3 it will be shown that this decrease of intensity should not be
interpreted as a loss of unsaturation during sulfur vulcanisation of EPDM.
A series of low-molecular-weight model studies using 2-ethylidene norbornane (ENBH)

by Duynstee and co-workers have provided a lot of detailed structural information on
the sulfur vulcanisation of EPDM [75, 76]. As for the polydiene rubbers, accelerated
sulfur vulcanisation of EPDM proceeds via allylic substitution, resulting in the formation
of dialkenylsulfides. The number of ENBH model crosslinks that has been separated by
preparative high-pressure liquid chromatography (HPLC) and identified with 1H NMR
amounts to about 40 as a result of the variation of the sulfur bridge length (1 to 5 sulfur
atoms), the Entgegen/Zusammen isomerism of the ENB ethylidene unit and the 3 different
allylic positions for sulfur attachment for both ENBH moieties (C3,exo, C3,endo and C9).
One of these ENBH-Sn-ENBH model crosslink structures is shown in Figure 6.3.
216
Figure 6.3 Accelerated sulphur vulcanization of EPDM as demonstrated by low-

molecular weight model olefin studies [75,76]
After these early studies an extensive FT-Raman study [77] was performed to bridge the
gap between the low-molecular-weight ENBH model vulcanisation studies and the
vulcanisation studies using high-molecular-weight EPDM. These studies will be presented
in detail. First, a series of low-molecular-weight dialkenylsulfides will be discussed in
order to determine the effect of sulfur vulcanisation on Raman spectra of olefins.
Subsequently, the attachment of the sulfur crosslinks at the allylic positions, the conversion
of ENB, the length of sulfur crosslinks and the network structure will be addressed for
unfilled sulfur vulcanisates of amorphous EPDM. Some preliminary network structure/
properties relationships will also be presented.
6.2.2.2 Dialkenylsulfides
It has been shown that during the accelerated sulfur vulcanisation of unsaturated elastomers
or model olefins allylic hydrogen atoms are substituted by sulfur atoms (Figures 6.2 and
6.3). So, the C=C vibrations at 1600 - 1800 cm-1, the S-S peaks at 300 - 700 cm-1 and the
C-S bands at 600 - 750 cm-1 provide the most relevant information for monitoring sulfur
vulcanisation using Raman spectroscopy. Table 6.2 gives FT-Raman data for a series of
alkenes and the corresponding dialkenylsulfides. The latter were either synthesised as pure
sulfides [78] or produced as a mixture via model olefin vulcanisation [79]. Sulfur substitution
of the allylic hydrogen atoms results in a downward shift of the wave number of the C=C
stretching vibration of 7 to 18 cm-1. For 3-methyl-1-pentene an upward shift of 20 cm-1 is
observed, but this is due to complete allylic rearrangement of the unsaturation towards a
more highly substituted unsaturation upon vulcanisation [80]. When compared with 3-
methyl-2-pentene a downward shift of 13 cm-1 is observed.
The S-S vibrations have been assigned using the spectra of pure di(2,3-dimethyl-2-
butenyl)disulfide, trisulfide and tetrasulfide. Literature data on dialkenylsulfides are
available only for diallyldisulfide and disqualenylsulfides (Table 6.2), but more data for
217
Table 6.2 T-Raman spectral data of alkenes and (mixtures of)

corresponding dialkenylsulfides
Dialkenylsulfide FT-Raman band (cm-1)
alkene sulfide alkene dialkenylsulfide
C=C C=C S-Sa
2,3-dimethyl-2-butene mono 1675 1666 -
di 1675 1663 510
tri 1675 1663 490
tetra 1675 1662 491/439
mixture 1675 1661 490/438/459
2-methyl-2-pentene mixture 1676 1663 475/460/494/439
3-methyl-1-pentene mixture 1640 (1673b) 1660 491/460/439
3-hexene mixture 1671 1660 459/437/503/475
ENBH mixture 1688 1678 492/441
DCDPH mixture 1612 1605 484/450/436
allene [19, 81] di - - 510
squalene [125] mixture 1667 1649
a: when two or more wave numbers are given, the first one corresponds to the Raman
peak with the largest intensity;
b: due to complete allylic rearrangement of 3-methyl-1-pentene upon sulfur
vulcanisation, the comparison should be made with 3-methyl-2-pentene
dialkylsulfides are available by Freeman [81]. For dialkyldisulfides and dialkyltrisulfides,

one peak each is observed in the ranges of 500 to 520 cm-1 and 480 to 500 cm-1,
respectively. For dialkyltetrasulfides two peaks are found at about 440 and 490 cm-1.
Dialkylpentasulfides are characterised by a peak at 490 cm-1 and two smaller peaks at
460 and 440 cm-1. The observation of more than one peak for the higher sulfides is due
to the presence of different S-S conformations. A comparison of the data for
dialkenylsulfides (Table 6.2) and the literature data for dialkylsulfides [81] shows that
the nature of the substituents on the sulfide (alkenyl versus alkyl) does not affect the
position of the Raman bands.
218
The FT-Raman spectra of the sulfur vulcanisates of the various model olefins do not
contain the characteristic disulfide signal at 510 cm-1, but do contain the typical higher
sulfide bands at 490, 460 and 440 cm-1 (Table 6.2). In addition, a new band at about
475 cm-1 is observed for the vulcanisates of 2-methyl-2-pentene and 3-hexene, which has
not yet been assigned (hexasulfide?). Results of HPLC analysis have shown that the
vulcanisate of 2,3-dimethyl-2-butene consists mainly of a mixture of disulfide to
pentasulfide with about 15 mole% of disulfide [79]. This illustrates that FT-Raman
spectroscopy is not very sensitive for the identification of disulfides. Because of an overlap
of signals, FT-Raman does not provide detailed, quantitative information on the presence
of the individual higher sulfides (S>2).
6.2.2.3 Unfilled Vulcanised EPDM
6.2.2.3.1 ENB Conversion
It has been shown that during sulfur vulcanisation of EPDM the C=C peak of the residual
ENB unsaturation at 1685 cm-1 seems to decrease in intensity in agreement with the
observations by Fujimoto and co-workers [73, 74] (see Section 6.2.2.1). However, in Section
6.2.2.2 it was shown that sulfur vulcanisation of the low-molecular-weight ENBH results
in a shift of the Raman C=C peak from 1688 to 1678 cm-1. Taking this into account a
closer inspection of the FT-Raman spectra reveals that the original C=C peak at 1690 cm-1
decreases in intensity, and a new peak is observed at 1681 cm-1. Actually, the C=C peak
broadens towards lower wave numbers, but in a first approximation the total area remains
constant. So, the sulfur substitution reaction of the allylic hydrogens is confirmed for the
polymer system. This corresponds to the observation by Koenig and co-workers, namely
that upon sulfur vulcanisation of cis-BR, the C=C peak at 1650 cm -1 decreases in intensity
and that of a new peak at 1633 cm-1 increases its intensity [19, 58].
The intensity changes in the FT-Raman spectra of EPDM upon vulcanisation have been
used to quantify the chemical conversion of this reaction. The C=C stretching region was
fitted for a series of spectra using a 3-dimensional fitting program, fitting identical band
positions and widths for all spectra, fitting the intensities of the bands individually for
each spectrum [77]. The two bands were set at 1690 cm-1 (unreacted ENB) and 1681 cm-1
(reacted ENB). The line width of the C=C vibration increases from 11 to 21 cm-1 upon
sulfur vulcanisation. This increase is larger than that for the low-molecular-weight model
ENBH (10 to 13 cm-1), which suggests that more lower sulfides may be formed for
EPDM. By using the 1450 cm-1 cluster as a reference it was determined that the Raman
scattering cross-section of the 1681 cm-1 band is 30% higher than that of the 1690 cm-1
band and so, the conversion of ENB can be calculated (Table 6.3; the NMR data in this
219
220
Table 6.3 Number of converted ENB units as determined by FT-Raman spectroscopy and
number of chemical crosslinks as determined by solid state 1H NMR relaxation studies of
unfilled EPDM vulcanisates
Sample FT-Raman NMR FT-Raman +
NMR
ENB Mooney sulfur ENB converted ENB molecular weight crosslink crosslink
content viscosity content conversion (mole/kg between crosslinks density stoichiometry
(wt.%) ML(1+4)125°C (phr) (%) EPDM) (kg/mole) (mole/kg) (mole/mole)
8.9 33 1.5a 41 0.30 3.4 0.15 2.0
8.9 46 1.5 48 0.36 3.0 0.17 2.1
8.6 63 1.5 43 0.31 3. 1 0.16 1. 9
4.6 63 1.5 63 0.24 3.1 0.16 1.5
10.0 65 0.6b 14 0.12 7.8 0.064 1.8

10.0 65 0.9 25 0.21 4.6 0.11 1, 9
10.0 65 1.2 34 0.28 3.4 0.15 1. 9
10.0 65 1.5 40 0.33 2.7 0.18 1.8
10.0 65 3.0 85 0.71 1. 6 0.31 2.3
a: EPDM compounds containing 5 phr ZnO, 1 phr stearic acid, 1 phr TMTD, 0.5 phr MBT and 1.5 phr sulfur were
vulcanised for 30 minutes at 160 °C;
b: for the compounds with different sulfur concentration the concentrations of ZnO, stearic acid, TMTD and MBT were
adjusted proportionally
table will be discussed in Section 6.2.2.3.3). The number of ENB molecules reacted with
sulfur can be calculated from the fraction of ENB converted, as determined by FT-Raman
spectroscopy, and the composition of the vulcanisation recipe.
From the data in Table 6.3 it follows that the relative ENB conversion ranges from 15%
to 85%, which is larger than often assumed. For the EPDM vulcanisates with 1.5 phr of
sulfur the absolute ENB conversion is about 300 to 350 mmole ENB/kg EPDM. This
conversion is not affected much by the Mooney viscosity of the original EPDM nor by
the original ENB content of the EPDM. The latter indicates that for these samples it is
not the ENB content but some other parameter is limiting the ENB conversion. When
the Raman data for the vulcanisates of the EPDM with 10.0 wt.% ENB and sulfur
contents varying from 0.6 to 3 phr are plotted against the sulfur content a straight line
going through the origin is obtained (Figure 6.4). Toluene extraction of these vulcanisates
followed by thin layer chromatography quantification showed that the amount of free
sulfur (S8) is less than 1% of the original content. Since the amount of sulfur present in
loop/cyclic structures is small (< 10%) (compare with Section 6.2.2.3.3), it is concluded
that the sulfur added is predominantly used for crosslinking. It can be calculated that the
average length of the sulfur crosslink is 2.7 sulfur atoms for the ISO 4097 [82] formulations
studied. For a given ratio of sulfur to accelerators the average length of the sulfur crosslink
seems to be constant, independent of the EPDM type used.
Figure 6.4 Absolute ENB conversion as determined with FT-Raman spectroscopy

versus sulphur content of EPDM compounds (data from Table 6.3)
221
6.2.2.3.2 Sulfur Crosslinks
The region between 300 and 700 cm-1 in FT-Raman spectra contains information regarding
the nature of the sulfur crosslinks. From the studies of low-molecular-weight sulfides it is
concluded that the location of bands assigned to disulfides, trisulfides and tetrasulfides
is at approximately 500-520, 485-500 and 490 + 440 cm-1, respectively [81]. In the FT-
Raman spectrum of sulfur vulcanised EPDM it can clearly be seen that the bands are
present at 570, 520, 490 and 440 cm-1. Acetone extraction experiments showed that the
peak at 570 cm-1 is due to the presence of ZDMC, formed from ZnO and TMTD. ZDMC
also has a major band at 440 cm-1, which coincides with the tetrasulfide band at the
same wave number. The 490 cm-1 band results from trisulfides and higher sulfides. The
520 cm-1 band is assigned to disulfides, which was not observed for the low-molecular-
weight model olefin vulcanisates (compare with Section 6.2.2.2). Although the average
length per sulfur bridge is only 2.7 sulfur atoms, the relative intensity of the tetrasulfide
bands is still considerable. However, one has to realise that the polarisability increases as
the sulfur rank of the crosslink increases. Consequently, the Raman scattering cross-
section rapidly increases with the length of the sulfur bridge.
Although detailed analysis of the 400-700 cm-1 region in the FT-Raman spectrum of
sulfur vulcanised EPDM is still a problem due to the poor signal:noise ratio, it appears
that the relative intensities of the bands in this region do not change with the conversion
of the third monomer or the type of EPDM. The trisulfide band at 490 cm-1 can be
observed without the problem of overlap. In the series of EPDM with 10 wt.% ENB,
with increasing amounts of vulcanisation chemicals the area of this band increases, which
is in agreement with the increase in ENB conversion observed. The fact that the relative
amounts of disulfide, trisulfide and tetrasulfide peaks are not affected can be explained
by the fact that for all samples, the relative amounts of the various vulcanisation ingredients
is kept constant.
Using chemical probe treatments [83] attempts were made to support the assignment of
the Raman sulfide bands for sulfur-vulcanised EPDM. Treatment of the vulcanisate based
on EPDM with 10 wt.% ENB and 1.5 phr sulfur with 1-hexanethiol/piperidine results in
the disappearance of all signals in the 300 to 600 cm-1 region. Treatment with 2-
propanethiol/piperidine/hexane results in the disappearance of the 490 cm-1 band, but
the 440 cm-1 peak remains and the 510 cm-1 signal actually increases in intensity. The
various reactions of sulfur crosslinks possibly occurring in the presence of alkylthiols are
depicted in Figure 6.5. Alkanethiols are not able to convert monosulfides (I), indicating
that C-S bonds are not affected [83]. Therefore, reactions II and VI will not occur either.
For low-molecular dialkenylsulfides it has been demonstrated that 1-hexanethiol degrades
all disulfidic and higher sulfidic species. According to reaction III, a disulfide crosslink is
converted into R-SH and R-S2-X. Since the applied molar ratio, 1-hexanethiol/sulfur
222
Figure 6.5 Possible reactions of sulphur crosslinks R-Sm-R (R = rubber) with thiols X-
SH (X= alkyl)
crosslink is larger than 100, R-S2-X will be converted according to reaction IV, which is
similar to reaction III. The combination of reactions III and IV will result in the ultimate
formation of R-SH and X-S2-X. R-SH rubber thiols do not contain S-S bonds and X-S2-
X is washed from the rubber, all sulfidic bands in the 300-600 cm-1 region have
disappeared.
Experiments with low-molecular-weight dialkenylsulfides have demonstrated that

treatment with 2-propylthiol results in the selective degradation of trisulfides and higher
sulfides probably for electronic reasons [83]. Reactions VI and VII are the first reactions
to occur. In a reaction sequence as discussed for 1-hexanethiol all trisulfides and higher
sulfides will finally be degraded to R-S2-X, R-S2H, R-SH and X-S2-X. The Raman spectrum
of the 2-propanethiol-treated EPDM vulcanisate indeed shows that the trisulfide and
higher sulfides peak at 490 cm-1 has disappeared. In addition, the intensity of the disulfide
peak at 510 cm-1 has increased. Since hexane is used as a solvent, ZDMC will probably
not dissolve and therefore, the observation of the band at 440 cm-1 is related to the
presence of this chemical. Tetrasulfides and higher sulfides do not contribute to this
signal, since the characteristic peak at 490 cm-1 is absent. So, it is concluded that the
chemical probe treatment has confirmed the assignment of the various sulfidic crosslinks
formed during EPDM vulcanisation.
6.2.2.3.3 Network Structure
NMR relaxation time measurements provide information on the mobility of polymer

chain segments and have been used for studying the effect of crosslinking, oil extension
223
and reinforcement of rubbers. Recently, a wideline solid-state 1H NMR method was

developed for determining the spin-spin 1H relaxation time, which can be converted into
the molecular weight between chemical crosslinks in rubbery networks (Mc) and, thus,
into the density of chemical crosslinks (1/2Mc) [84]. Using this approach the crosslink
density of the sulfur-vulcanised EPDM samples studied with FT-Raman spectroscopy
could be confirmed (Table 6.3). In agreement with the FT-Raman data, but with small
experimental error, the NMR data show that for all unfilled EPDM vulcanisates with
1.5 phr sulfur the number of chemical crosslinks is more or less constant independent of
the elastomer type, i.e., 160-180 mmole/kg.
For all samples, except the vulcanisate of EPDM with 4.6 wt.% ENB and 1.5 phr sulfur,
the ratio of the converted number of ENB moieties and the number of chemical crosslinks
is around 2.0 (± 0.2). The statistical spread in the ratio of the converted number of ENB
molecules and the number of chemical crosslinks is estimated to be 10-15% based on the
error in the Raman and NMR data. Furthermore, there might be a systematic error in
the crosslink density as determined by NMR, originating from the assumptions made in
calculating the number of chemical crosslinks from the total number of crosslinks
(chemical and entanglements).
The experimental ratios of 2 (± 0.2) agree with the expected value of 2.0 converted ENB
units for one chemical crosslink (ENB-Sn-ENB), indicating that crosslinks are formed
predominantly. Several explanations are proposed for deviations from the theoretical
value. Both pendent sulfur as in the crosslink precursor (ENB-Sn-X, X = accelerator
residue) and sulfur loops (sulfur bridge between two ENB units in the same EPDM chain)
will give rise to a ratio larger than 2.0. Formation of cyclic sulfide (addition of S to
unsaturation) will give rise to a ratio lower than 2.0. The fact that within experimental
error a ratio of 2.0 is found suggests that pendent sulfur, loops and cyclic sulfur hardly
occur (less than 10% of the total amount of reacted sulfur).
6.2.2.3.4 Network Structure/Properties Relationships
The analytical techniques discussed previously can be used to study the EPDM network
as such or its formation in time as well as to determine relationships between the network
structure and the properties of the vulcanisates. In a preliminary approach some typical
vulcanised EPDM properties, i.e., hardness, tensile strength, elongation at break and
tear strength, have been plotted as a function of chemical crosslink density (Figure 6.6).
The latter is either determined directly via 1H NMR relaxation time measurements or
calculated from the FT-Raman ENB conversion (Table 6.3). It is concluded that for these
unfilled, sulfur-vulcanised, amorphous EPDM, the chemical crosslink density is the main
parameter determining the vulcanisate properties. It is beyond the purpose of this review
to discuss these relationships in a more detailed and theoretical way.
224
Figure 6.6 Mechanical properties of unfilled sulphur vulcanized amorphous EPDM as

a function of crosslink density
6.3 Peroxide-curing
6.3.1 General
The mechanism of peroxide crosslinking of elastomers is much less intricate than that of
sulfur vulcanisation. Crosslinking is initiated by the thermal decomposition of a peroxide,
which is the overall cure rate determining step. Next, the active radicals thus formed
abstract hydrogen from elastomer chains to form macroradicals. Finally, crosslinking
results either from the combination of two macroradicals or from the addition of a
macroradical to an unsaturated moiety of another primary elastomer chain.
In view of the abundance of unsaturation in polydiene elastomers it may be expected

that this type of elastomers can be cured very efficiently with peroxides. This section will
show that a very high crosslinking efficiency (≡ moles of crosslinks formed per mole of
peroxide decomposed) can indeed be obtained. However, the substitution pattern of the
unsaturation plays a major role in the actual efficiency found.
The industrial relevance of peroxide-curing of elastomers is by far larger for main-chain

saturated elastomers, such as silicone elastomers, acrylic elastomers, fluoro elastomers
225
and particularly EPDM. Nonetheless, there has been little optical spectroscopical work
studying the curing kinetics and mechanism. The EPDM studies that have been published
indicate that detailed information on the curing chemistry can be obtained. The role of
the type and the amount of the third monomer in particular, as well as the role of co-
agents and scorch retarders, have been elucidated.
6.3.2 Polydiene Elastomers
6.3.2.1 Natural Rubber
In early communications it was reported that the efficiency of dicumylperoxide (DCP)

crosslinking of NR amounts to unity [85, 86], which means that one mole of chemical
crosslink is obtained as the result of the thermal decomposition of one mole of DCP. This
was explained by assuming that crosslinking results from the combination of polyisopropenyl
radicals only and crosslinking by macroradical addition does not take place.
It is not easy to support this crosslinking mechanism with spectroscopic evidence,

considering that the differences in peak intensity of the =C-H out-of-plane bending mode,
absorbing at 838 cm-1, before and after peroxide-curing, are difficult to quantify. Moreover,
the formation of crosslink structures is even more difficult to detect, because C-C stretching
modes have a very weak intensity. Nonetheless, data processing techniques, such as the
least squares curve-fitting procedure as developed for polymeric systems by Koenig and
co-workers [87], have proven their value in this respect. Via this method, the spectrum
of a mixture (in this case NR, DCP and the peroxide decomposition products) could be
fitted with the spectra of the pure components and the fractional amount of each
component was calculated, assuming that the absorptions of the various components
were additive in the mixture. For NR cured with an excessive amount of DCP (> 50 phr!)
it was shown that a decrease of double-bond concentration occurred at a rate comparable
with the formation of peroxide decomposition products. This indicates that the
crosslinking efficiency can exceed unity as opposed to the earlier reports [85, 86]
mentioned. FT-IR results did not provide evidence for cis-trans isomerisation, although
isomerisation was confirmed from solid-state NMR data that were generated on the
same samples. Interestingly, the formation of quaternary carbon-carbon bonds resulting
from crosslinking could be detected at 1320 cm-1, but quantification was impossible in
the light of the weak intensity and width of the peak.
In more recent studies from Gonzalez and co-workers [88-90] it was concluded from
dynamic mechanical analysis of peroxide-cured NR that a non-uniform crosslinked
network results if a large amount of peroxide is used. This result seems to be in line with
the optical spectroscopy studies discussed.
226
6.3.2.2 Polybutadienes
Patterson monitored the peroxide-curing behaviour of cis-BR by both FT-IR and FT-
NMR techniques, again using excessive amounts of DCP [87]. It was concluded that cis-
trans isomerisation occurred, based on a new absorption appearing in the IR spectrum
at 965 cm-1 (=C-H, out-of-plane). Moreover, it was concluded that crosslinking occurred
through radical addition reactions, because the peak intensity at 728 cm-1 (cis =C-H, out
of plane) decreased. Also in this case the formation of C-C bonds resulting from
crosslinking reactions was shown by a weak and broad absorption appearing at
approximately 1320 cm-1. Based on FT-IR data of both the double-bond consumption
and the formation of peroxide-decomposition products a crosslinking efficiency of
approximately 1.7 was calculated.
Loan used more practical amounts of peroxide and showed that the crosslinking efficiency
for cis-BR with 3% of 1,2-vinyl units is about 10, based on FT-IR data on the consumption
of the vinyl double bond at 911 cm-1 (=C-H, out-of-plane) [91]. Hummel and Kaiser [92]
and Van der Hoff [93] found similar values. By studying BR grades with different 1,2-
vinyl contents Van der Hoff showed that the crosslinking efficiency becomes as high as
50 for a product with 79% 1,2-vinyl units. FT-IR data also demonstrate that during
curing significant cis-trans isomerisation occurs.
Bellander and co-workers [94] showed that the crosslinking efficiency of BR also depends
on the pressure applied during curing. Using the change of the vinyl absorption at 911 cm-1
and the peak at 1435 cm-1 (stretching of main-chain CH2) as an internal reference, it was
concluded that the crosslinking efficiency for a BR grade with 11 wt% of 1,2-vinyl units
amounted to 8, if the unfilled compounded film was cured at 170 °C in a mould at a
pressure of 26 MPa. The crosslinking efficiency further increased to about 12, if the
pressure during curing was increased to 293 MPa. This effect was explained by assuming
that the rate of radical combination reactions is lowered as a result of the reduced mobility
of polymer chains under higher pressures.
6.3.3 EPDM
6.3.3.1 Effect of Third Monomer Type and Content
Fujimoto and co-workers reported in 1969 on the use of a new high-temperature ATR
apparatus that they constructed for studying, among other things, the role of the type of
third monomer on the peroxide curing-chemistry of EPDM [73, 74]. EPDM grades
containing either ENB, DCPD, HD or MNB as a third monomer were compounded with
227
DCP and uncured films were deposited on a multi-reflectance prism. Next, the ATR cell
was heated to 140 °C and spectra were recorded as a function of curing time, while the
angle of incidence was fixed at 45o, which minimised loss of radiation. The decrease of
the absorption relating to the third monomer pendent unsaturation was monitored at
1685, 3045, 966 and 870 cm-1 for EPDM polymers containing ENB, DCPD, HD and
MNB, respectively. The absorption was normalised with the methyl absorption at
1380 cm-1 which was not affected during curing.
In the case of MNB-EPDM the absorption of the pendent unsaturation rapidly decreased
within 25 minutes to about 20% of its initial value and then very slowly decreased to
about 15% of its initial value. Peroxide decomposition data [95] indicate that a significant
amount of peroxide was still present after 25 minutes, which means that apparently two
kinetic regimes exist. Although less obvious, the data presented for the consumption of
the pendent unsaturation in the case of the other EPDM grades suggest the same.
Although peroxide-decomposition data have to be used with great caution, it can be

concluded that double bonds are consumed by radical addition reactions. One can debate
whether the unsaturations are consumed by multiple radical addition reactions or via
consecutive radical addition/radical transfer sequences. The latter seems most likely,
considering the low tendency of alkyl radicals for addition to alkyl-substituted double
bonds under these relatively mild conditions. In radical addition reactions of this kind,
the stabilisation of radicals due to polar effects is negligible. Experimental studies show
that the reactivity is mainly controlled by steric effects [96]. The order of reactivity MNB
> DCPD ≈ ENB > HD towards radical addition reactions as found by Fujimoto and co-
workers [73, 74] is in line with these considerations.
Baldwin and co-workers have monitored the reactions of the unsaturation of MNB-EPDM
during peroxide-curing as well [97]. To this end thin films of unfilled compound were cured
with 0.6 phr of DCP at different time intervals at 170 °C and examined for IR spectroscopic
changes. It was shown that 38% of the pendent unsaturation (of 2.5 wt% MNB incorporated)
was consumed after complete peroxide decomposition. This means that the number of radical
addition reactions per radical species initiated amounts to about 8.
Baldwin and co-workers also studied the peroxide curing behaviour of other EPDMs with
a variety of third monomers [97]. As expected, they found that terminal olefins as present
in 5-vinyl-2-norbornene (VNB) and MNB containing EPDM provide a much higher peroxide
curing efficiency as compared to internal olefins as present in for instance ENB, DCPD and
HD containing EPDM. This again is in line with the earlier considerations mentioned [96].
Based on the literature reports discussed, a rough scheme for the peroxide-curing of
EPDM is proposed (Figure 6.7). This scheme is identical to the mechanism that was
proposed based on the results of rheometer studies [11].
228
Figure 6.7 Simplified scheme of the peroxide curing of EPDM with dicyclopentadiene
as third monomer
6.3.3.2 Co-agents in Peroxide-curing
6.3.3.2.1 General
Co-agents are multi-unsaturated compounds, which are used in the peroxide-curing of

elastomers. When classical co-agents, such as triallylcyanurate (TAC),
trimethylolpropanetrimethacrylate (TRIM) or diallylterephthalate (DATP), are added,
the crosslinking efficiency is enhanced [98-102]. Various mechanisms for the increase of
the crosslinking efficiency have been proposed. In all cases a fast reaction between the
229
co-agent and a tertiary macroradical is supposed to occur, thus suppressing unwanted

side reactions such as chain scission or disproportionation.
Apart from the effect on the crosslinking efficiency, the use of co-agents in peroxide-
curing also imparts the molecular structure of crosslinks. It has been reported that co-
agents with two or more unsaturated moieties can be incorporated as individual molecules
between two elastomer strands to form crosslinks [103-109]. In this way the crosslink
structure of peroxide-cured elastomers can be altered. Thus, apart from the expected
benefits, such as improved crosslinking efficiency, decreased compound viscosity and
faster cure, the use of co-agents may also provide a tool for manipulating mechanical
properties.
Considering the above, it is our opinion that conclusions on the mechanism of EPDM
peroxide curing in the presence of co-agents have been drawn based on a great deal of
speculation. Concepts have been generally accepted, although no solid evidence for the
proposed reaction pathways exists. In a number of studies the use of FT-IR spectroscopy
to elucidate the mechanism of EPDM peroxide-curing in the presence of co-agents has
been published. Based on spectroscopic evidence and supported by other analytical
techniques, it is concluded that the actual cure mechanism [110-117] seriously diverges
from the existing theories, as will be reviewed in the next paragraphs.
6.3.3.2.2 Suppression of Macroradical Side Reactions
If the effect of co-agents on crosslinking efficiency is just the suppression of macroradical

side reactions, such as chain scission and disproportionation, one should expect mono-
functional co-agents to be as effective as their multi-functional analogues (if compared
at the same molar level of unsaturation). This is definitely not the case, as will be
demonstrated.
In Figure 6.8 it is shown that 1-allyloxyoctane (AO) and 1,8-diallyloxyoctane (DAO)

react during the peroxide-curing of EPM with the same kinetics [118]. The allyl
absorptions measured directly after curing decrease to about 5% to 10% of the initial
values and the ether absorptions measured after curing and extensive acetone extraction
approach 100% of the initial values, indicating quantitative attachment of AO and DAO
to the EPM matrix. In spite of the fact that the cure kinetics are identical, the difference
compared to curing efficiency becomes apparent from the results of crosslink density
measurements, as shown in Figure 6.9.
The results clearly demonstrate that a diallyl functional co-agent such as DAO, provides
a marked improvement in the crosslinking efficiency, whereas its monoallyl functional
230
Figure 6.8 Change of IR absorptions during crosslinking of amorphous EPM at

170 °C with 2 phr of DCP and equimolar amounts of unsaturation present in 1-
allyloxyoctane (AO) or 1,8-diallyloxyoctane (DAO).
Figure 6.9 Effects of the addition of 1-allyloxyoctane (AO) and 1,8-diallyloxyoctane

(DAO) on crosslink density, estimated from equilibrium swelling measurements in decalin
231
equivalent, AO, does not enhance the crosslinking efficiency at all. Apparently the
improvement of the crosslinking efficiency by co-agents stems from a mechanism other
than the suppression of chain scission or disproportionation reactions.
In efforts to come to a better understanding of the action of co-agents, the reactions of

some classical co-agents in peroxide-curing of amorphous EPDM have been studied in
the absence of fillers and oils, using FT-IR spectroscopy as the major tool for the elucidation
of the cure mechanism.
6.3.3.2.3 Co-agent Studies
In order to study the role of DATP as co-agents in the peroxide-curing of EPM, various
compounds of elastomer, DCP and co-agent were cured at 170 °C in a press at different
time intervals and then FT-IR spectra were recorded. Furthermore, the vulcanisates were
extracted with acetone to remove the unreacted co-agent and again FT-IR spectra were
recorded. Typical results are illustrated in a study using DATP as a co-agent [110].
The absorptions of the reactive allyl groups (932 cm-1) and the inert ester groups (1730 cm-1) of
DATP were quantified using the CH2 vibration at 2722 cm-1 from EPM as a reference
peak. In Figures 6.10 and 6.11 results from FT-IR analysis are presented.
Based on the IR data of samples before extraction, it is concluded that the allyl groups
react rapidly to completion within about 2 minutes, whereas the ester absorption remains
constant. More allyl groups per unit of time react than peroxide radical fragments are
initiated, it can be concluded that the allyl groups react predominantly via radical addition
reactions, probably accompanied by radical transfer reactions. FT-IR analysis after
vulcanisate extraction indicates that the co-agent is covalently bound to the elastomer
matrix, as shown by the 100% recovery of the ester absorption after 2 minutes of curing.
In Figure 6.10 the relative allyl absorption is divided into three parts. Part A represents
the amount of co-agent attached to the elastomer matrix with one allyl group. Part B
represents the amount of unreacted co-agent (the extracted co-agent fraction) and the
remainder represents the amount of co-agent with which both allyl groups have reacted.
Thus, a distribution of co-agent molecules which have reacted with none, one or two
allyl groups, can be calculated as a function of cure time.
Comparing the distribution as obtained by experiments with a theoretical distribution,

in which it is assumed that all allyl groups have the same probability of reaction, a large
discrepancy was established. It was found that the number of molecules which had reacted
with only one allyl moiety is markedly lower as predicted according to this theoretical
distribution. A similar result was found by Kloosterboer in his studies on the mechanism
232
Figure 6.10 Relative allyl absorptions of EPM compounds cured with 2 phr of DCP
and 3 phr of DATP, measured before and after acetone extraction [110]
Figure 6.11 Relative ester absorptions of EPM compounds cured with 2 phr of DCP
and 3 phr of DATP, measured before and after acetone extraction [110]
233
and kinetics of bulk photocrosslinking of diacrylates [119]. By using a percolation

model, he explained this behaviour by showing that the reactions of diacrylates proceed
inhomogeneously. Since AFM measurements showed that DATP was not homogeneously
distributed in the EPM matrix (discrete spherical domains with an average droplet size
of about 200 nm were found), the conclusion was drawn that the reactions of DATP
during the peroxide-curing of EPM also proceed inhomogeneously. When a radical
fragment is initiated within a DATP domain or in its vicinity, a local oligomerisation of
allyl groups is expected to occur, resulting in the formation of a crosslinked spot within
a co-agent domain, thus accounting for the relatively small fraction of allyl groups
having reacted with only one allyl moiety.
Since all allyl groups have reacted during curing, and in radical addition reactions (or
radical transfer reactions) the active radical centre remains intact, it is proposed that
the crosslinking efficiency is increased by reactions of macroradicals from the elastomer
matrix with allyl groups from the co-agent domains (co-curing). This assumption was
validated by a series of curing experiments in which the amount of peroxide was set at
a fixed level of 2 phr and the amount of DATP was varied. A part of the cured samples
was immersed in a solution of tetrahydrofuran, ethanol and 10 M NaOH at room
temperature. It was demonstrated by FT-IR that after the alkali treatment, the aromatic
absorptions had completely vanished. Finally, the equilibrium swelling ratio in
decahydronaphthalene of the cured samples was determined both before and after
alkali treatment. As expected, untreated samples showed a decrease of equilibrium
swelling ratio with increasing DATP concentration. However, after alkali treatment,
the equilibrium swelling ratio of cured samples was more or less independent of the
amount of DATP added and the crosslink density of all samples was comparable to the
sample that contained only peroxide. Thus, results are in line with the proposed
mechanism that states that during peroxide-curing, bridges between co-agent domains
and the EPM matrix are formed and that these bridges are the key factor enhancing the
crosslinking efficiency. Moreover, it can be concluded that these bridges can be
considered as extra crosslinks; the density of ‘normal’ chemical crosslinks is determined
solely by the concentration of the peroxide.
Apart from DATP, the curing behaviour of two isomers of DATP, i.e. diallylphthalate
(DAP) and diallylisophthalate (DAIP), was also studied by FT-IR spectroscopy. Typical
absorptions for the allyl and ester groups for these isomers were found at similar wave
numbers, compared to DATP. Interestingly it was found that for DAP hardly any
structures, comparable with that of DATP representing co-agent molecules being
attached with only one allyl group could be identified. But, in spite of this fact both the
compound morphology obtained rom AFM and the rate of allyl consumption as a
function of cure time were similar. This behaviour was explained by assuming that
234
cyclopolymerisation takes place, analogous to the observations of Matsumoto and co-

workers [120, 121] and Holt and Simpson [122], for the bulk polymerisation of aromatic
diallylesters.
Because the occurrence of cyclopolymerisation might be the reason for the different
behaviour of the three isomers, attempts were made to find IR spectroscopic evidence
using the following kinetic considerations: if after the addition of an EPM radical to
an allyl moiety of the co-agent the next reaction was the addition of an allyl moiety of
another co-agent molecule (‘normal’ polymerisation), than it is expected that at very
short cure times (infinitely short) exactly one allyl group per co-agent molecule attached
to the EPM backbone would be converted. However, if cyclopolymerisation were
predominant, this would mean that the reaction of a co-agent molecule with an EPM
radical would consume both allyl moieties instantaneously. Thus, the percentage of
unreacted allyl groups per covalently bound co-agent molecule extrapolated to zero
cure time provides the ratio of ‘normal’ polymerisation to cyclopolymerisation for the
given compound formulation and reaction conditions. Treating the IR data accordingly,
it is concluded that DAP cyclopolymerises for about 71%, DAIP for 22% and DATP
for less then 5%. This result is reflected in the very low crosslinking efficiency obtained
with DAP.
Based on all of these observations a general scheme for the peroxide-curing of EPM in
the presence of co-agents is proposed (Figure 6.12).
In similar studies the kinetics and the mechanism of EPDM peroxide-curing in the
presence of TAC as a co-agent has been reported [112, 113]. Peak intensities for the
reactive allyl groups were monitored at 930 cm-1 and for the s-triazine ring at 822 cm-1
(out-of-plane) and 1566 cm-1 (in-plane). Again, spectra were recorded for different
compounds at a fixed number of cure time intervals, both before and after acetone
extraction of the cured samples. Results of these studies were in line with the previous
studies on the use of diallylphthalates as co-agents. Again it was found that the peroxide
crosslinking efficiency was markedly enhanced by the action of TAC, resulting from
the formation of additional crosslinks across the co-agent elastomer interface. The
latter observation was also supported by a study by Murgic and co-workers [113],
who used a chemical probe to selectively cleave the cyanurate ester bonds after curing.
Interestingly, it was found that TAC reacts via a cyclopolymerisation mechanism. It
was shown that the number of reacted allyl groups per bound co-agent molecule in the
initial stage of the curing process amounts to two, independent of the amount of peroxide
or TAC used.
Finally, mention is made to similar FT-IR studies carried out using different a,w-
diallyloxyalkanes, diallylpolyethyleneglycols and methacrylic esters of
235
Figure 6.12 Simplified scheme for the peroxide curing of EPM in the presence of a co-
agent
polyethyleneglycols [114]. Again the reactions of the reactive unsaturations were

monitored by IR spectroscopy before and after acetone extraction along with the inert
ether absorptions. In this case it was found that the reactions of the co-agents can be
described by a ‘normal’ polymerisation mechanism. For none of these co-agents evidence
for cyclopolymerisation was found.
6.3.3.2.4 Network Structure/Properties Relationships
In the previous section it was demonstrated that co-agents enhance the crosslinking
efficiency via co-curing of elastomer strands with co-agent domains. Because the use of
co-agents alters the crosslink structure, the elastic and mechanical properties of the cured
236
products obtained are also affected. However, it was shown that these effects are relatively
small [114]. As a general rule it can be stated that compact co-agent molecules having
aromatic structures, which vitrify effectively during peroxide curing, provide worse tensile
properties, because the difference in compliance between the elastomer matrix and the
embedded co-agent domains is large.
6.3.3.3 Scorch Retarders in Peroxide-curing
2,6-Di-tert-butylcresol (butylated hydroxytoluene or BHT) is the most commonly applied

scorch retarder in the peroxide curing of EPDM. In order to study the fate of BHT
during peroxide-curing a FT-IR spectroscopic study was conducted [123]. Phenolic
absorptions at 3649 and 1231 cm-1 and an aromatic absorption at 775 cm-1 were
monitored as a function of cure time, both before and after extraction of the cured
samples with acetone. The presence of hydroxyl and aromatic absorptions after acetone
extraction indicates covalent bonding of BHT fragments to the elastomer matrix, because
control experiments with uncured or sulfur vulcanised EPDM compounds showed
complete extraction of BHT.
The ratio of the retention of aromatic and hydroxyl absorptions after extensive acetone
extraction was similar for different compound compositions cured during different time
intervals. Thus, it was concluded that BHT fragments were not bound via EPDM
macroradical phenoxy radical combination reactions, leaving combination via EPDM
macroradicals and benzylic radicals as the most likely reaction path. In conclusion, it
was stated that about 25% to 30% of the BHT became chemically bound, which was
further supported by results of ageing experiments using cured samples that had been
extensively extracted.
FT-IR results also showed that one new (small) absorption at 1659 cm-1 appeared, which
could not be attributed to peroxide decomposition products. This absorption also appeared
when the peroxide-curing experiments were carried out using an amorphous EPM,
indicating that the absorption did not relate to rearrangement of the third monomer
moiety (ENB in this case). It is tentatively concluded that the absorption at 1659 cm-1 is
related to EPDM main-chain modifications, resulting from disproportionation reactions
of EPDM macroradicals with BHT radical fragments.
6.4 Concluding Remarks and Future Outlook
Optical spectroscopy (IR/NMR/Raman) has been extremely useful in the study of the
sulfur and peroxide crosslinking chemistry of elastomers, especially that of EPDM. The
237
development of FT, the use of lasers and the deconvolution of spectra has enlarged the
applicability of both Raman and IR spectroscopy. As a result of the high polarisability of
C-S and S-S bonds, Raman spectroscopy is especially suitable for studying the sulfur
vulcanisation of elastomers, whereas as a result of the high dipole moments of the polar
co-agents, IR spectroscopy has been of great value in the study of the co-agent-assisted
peroxide-curing of elastomers. Changes in the type of unsaturation, for instance resulting
from allylic substitution by sulfur or isomerisation, have been demonstrated both with
Raman and IR spectroscopy.
For sulfur vulcanisation of EPDM it was shown that the relative ENB conversion (20 to
60%) is higher than often assumed. The absolute ENB conversion was shown to be
governed by the vulcanisation recipe and to be independent of the EPDM type. For the
ISO 4097 [82] recipe the average length of the sulfur crosslinks is 2.7 sulfur atoms. The
number of converted ENB units per sulfur bridge is 2.0, indicating that crosslinks are
formed predominantly. In a preliminary study it was shown that the mechanical properties
of unfilled sulfur-vulcanised amorphous EPDM are determined by the chemical crosslink
density. Clearly, these studies should be extended to other vulcanisation recipes and
completely formulated compounds. Vulcanisation kinetics should be studied, preferably
at different temperatures.
It was shown that simple spectroscopic studies, which can be performed in any laboratory
and not requiring high spectroscopic skills, can provide valuable data that help to
understand the peroxide-curing behaviour of elastomers. Based on FT-IR studies and
supported by other analytical evidence it has been demonstrated that co-agents increase
the crosslinking efficiency in EPDM peroxide-curing via the formation of chemical links
between the elastomer matrix and co-agent domains. These chemical links should be
regarded as additional (multifunctional) crosslinks that are formed next to ‘normal’
carbon-carbon crosslinks, produced via combination of macroradicals. The latter is
influenced only by the amount of peroxide used. Suppression of macroradical side
reactions by co-agents, such as chain scission or disproportionation, hardly plays a role.
The effect of co-agents in the peroxide-curing of EPDM is very similar to the effect of
third monomers. It was concluded that the pendent unsaturation of the third monomer
acts as a co-agent, i.e., the amount of third monomer governs the amount of chemical
crosslinks formed by macroradical addition reactions via the unsaturated moiety of the
third monomer, whereas the amount of peroxide governs the amount of crosslinks formed
by macroradical combination reactions.
For IR and Raman spectroscopy studies there is a clear limitation when large amounts of
carbon black are applied, as is common practice in technical rubber goods. In that sense
the applicability of Raman and IR spectroscopy seems to be surpassed by solid-state 13C
238
NMR spectroscopy and solid-state 1H and 13C NMR relaxation-time experiments.

However, the sensitivity of solid-state 13C NMR is not as high as that of Raman and IR
spectroscopy. For instance, solid-state 13C NMR of sulfur-vulcanised EPDM could only
be performed when the ENB unsaturation of EPDM was fully isotopically enriched with
13
C NMR [124].
The ultimate aim of optical spectroscopy or similar studies is to elucidate the chemistry
of elastomer crosslinking with respect to both the structures formed and the reaction
kinetics in order to provide a sound basis for structure/property relationships. As a result,
it will be possible in the future to develop rubber products with less trial and error [18,
19, 55, 56, 125].
References
1. The Vanderbilt Rubber Handbook, Ed., R.F. Ohm, R.T. Vanderbilt Company
Inc., 13th Edition, Norwalk, 1990, 123.
2. W. Hofmann, Rubber Technology Handbook, Hanser Publishers, Munich, 1989,

Chapter 3.3.8.
3. J.A. Brydson, Rubbery Materials and their Compounds, Elsevier, London, 1988,
Chapter 7.
4. J.L. Laird, M.S. Edmondson and J.A. Riedel, Rubber World, 1997, 217, 42.
5. R.T. Sylvest, G.Lancester and S.R. Betso, Kautschuk und Gummi Kunststoffe,
1997, 50, 186.
6. E.T. Italiaander, Kautschuk und Gummi Kunststoffe, 1995, 48, 742.
7. H.J.H. Beelen, Kautschuk und Gummi Kunststoffe, 1999, 52, 406.
8. A.Y. Coran and R.P. Patel in Thermoplastic Elastomers, Eds., G. Holden, N.R.
Legge, R.P. Quirk and H.E. Schroeder, 2nd Edition, Hanser Publishers, Munich,
1996, Chapter 7.
9. D.J. Synnott, D.F. Sheridan and E.G. Kontos in Thermoplastic Elastomers from
Rubber-Plastic Blends, Eds., S.K. De and A.K. Bhowmick, Ellis Horwood,
Chichester, 1990, Chapter 5.
10. Worldwide Rubber Statistics, International Institute of Synthetic Rubber

Producers, Houston, TX, USA, 1998.
239
11. H.G. Dikland, Kautschuk und Gummi Kunststoffe, 1996, 49, 413.
12. A.V. Tobolsky and P.F. Lyons, Journal of Polymer Science, Part A2, 1969, 6, 1561.
13. J. Lal and K.W. Scott, Journal of Polymer Science, Part C, 1965, 9, 113.
14. L. Bateman, C.G. Moore, M. Porter and B. Saville in The Chemistry and Physics
of Rubber-like Substances, Ed., L. Bateman, MacLaren & Sons Ltd., London,
1963, Chapter 15.
15. M.M. Coleman, J.R. Shelton and J.L. Koenig, Industrial Engineering and
Chemistry, Product Research and Development, 1974, 13, 154.
16. A.V. Chapman and M. Porter in Natural Rubber Science and Technology, Ed.,
A.D. Roberts, Oxford University Press, Oxford, 1988, Chapter 12.
17. A.Y. Coran in Encyclopedia of Polymer Science and Engineering, 2nd Edition,
John Wiley & Sons, Inc., 1989, Volume 17.
18. M.R. Kresja and J.L. Koenig, Rubber Chemistry and Technology, 1993, 66, 376.
19. J.L. Koenig, M.M. Coleman, J.R. Shelton and P.H. Starmer, Rubber Chemistry
20. P.J. Nieuwenhuizen, Workshop Vulkanisation, Process-Technik-Wirkung,

Deutsches Institut für Kautschuktechnologie, Hannover, 1997, Paper No.5.
21. P.J. Nieuwenhuizen, J. Reedijk, M. van Duin and W.J. McGill, Rubber Chemistry
22. G. Ellis, P.J. Hendra, C.H. Jones, K.D.O. Jackson and M.J.R. Loadman,
Kautschuk und Gummi Kunststoffe, 1990, 43, 118.
23. P.J. Hendra and K.D.O. Jackson, Spectrochimica Acta, 1994, 50A, 1987.
24. M. Mori and J.L. Koenig in Annual Reports on NMR Spectroscopy, Eds., G.A.
Webb and J. Ando, Academic Press, San Diego, 1997, 231.
25. W. Gronski, U. Hoffmann, G. Simon, A. Wutzler and E. Straube, Rubber

Chemistry and Technology, 1992, 65, 63.
26. L. de Roo, Nederlandse Rubber Industrie, 1967, 28, 1.
27. P.L. Wancheck and L.E. Wolfram, Applied Spectroscopy, 1976, 30, 542, RR76.
240
28. D. Brück, Kautschuk und Gummi Kunststoffe, 1988, 41, 875.
29. K.D.O. Jackson, M.J.R. Loadman, C.H. Jones and G.Ellis, Spectrochimica Acta,
1990, 46A, 217.
30. E. Gomez, Revue Generale des Caoutchousc et Plastiques, 1991, 68, No.705, 45.
31. P.J. Hendra, C.H. Jones, P.J. Wallen, G. Ellis, B.J. Kip, M. van Duin, K.D.O.
Jackson and M.J.R. Loadman, Kautschuk und Gummi Kunststoffe, 1992, 45, 910.
32. K.D.O. Jackson, Journal of Natural Rubber Research, 1997, 12, 102.
33. M.L. Kralevich and J.L. Koenig, Rubber Chemistry and Technology, 1998, 71,
300.
34. P. Hendra, P. Wallen, A. Chapman, K. Jackson, J. Loadman, M. van Duin and B.

Kip, Kautschuk und Gummi Kunststoffe, 1993, 46, 694.
35. J. Schnetger, D. Schulze and T. Werner, Kautschuk und Gummi Kunststoffe,

1980, 33, 185.
36. M. Blanco, J. Coello, H. Iturriaga, S. Maspoch and E. Bertran, Analytical

Chimica Acta, 1997, 353, 351.
37. M. Blanco, J. Coello, H. Iturriaga, S. Maspoch and E. Bertran, Applied

Spectrosopy, 1995, 49, 747.
38. E.F. Devlin and T.L. Folk, Rubber Chemistry and Technology, 1984, 57, 1098.
39. S-Y. Lin, W-J. Tsay, Y-L. Chen and C-J. Lee, Journal of Controlled Release, 1994,
31, 288.
40. A.N. Theodore and R.O. Carter, Journal of Applied Polymer Science, 1993, 49,
1071.
41. G. Menges, L. Setiawan, C. Herschbach and H. Grün, Kautschuk und Gummi

Kunststoffe, 1988, 41, 1125.
42. A. Roychoudhury, P.P. De, N.K. Dutta, N. Roychoudhury, B. Haidar and A.

Vidal, Rubber Chemistry and Technology, 1993, 66, 230.
43. R. Alex and P.P. De, Kautschuk und Gummi Kunststoffe, 1990, 43, 1002.
44. N.R. Manoj, P.P. De, S.K. De and D.G. Peiffer, Journal of Applied Polymer
Science, 1994, 53, 361.
241
45. G. Schreier and G. Peitscher, Fresenius’ Journal of Analytical Chemistry, 1972,

258, 199.
46. M.M. Coleman, P.C. Painter and J.L. Koenig, Journal of Raman Spectroscopy,
1976, 5, 417.
47. H.W. Siesler, Die Makromolekulare Chemie, Macromolecular Symposia, 1986, 5,

151.
48. A.M. Healey, P.J. Hendra and Y.D. West, Polymer, 1996, 37, 4009.
49. B. Mattson, B. Stenberg, S. Persson and E. Oestman, Rubber Chemistry and

Technology, 1990, 63, 23.
50. D. Kiroski and D.E. Packham, Proceedings of IRC ’96, Manchester, UK, 1996,
Paper No.4.
51. K. Anandakumaran and D.J. Stonkus, Polymer Engineeering and Science, 1992,
32, 1386.
52. J. Lemaire, Revue Generale des Caoutchoucs et Plastiques, 1998, 771, 98.
53. M. Baba, J.L. Gardette and J. Lacoste, Polymer Degradation and Stability, 1998,
63, 121.
54. J.M. Chalmers and G. Dent, Industrial Analysis with Vibrational Spectroscopy,
Royal Society of Chemistry, Cambridge, 1997.
55. H.W. Siesler and K. Holland-Moritz, Infrared and Raman Spectroscopy of

Polymers, Marcel Dekker, Inc., New York, 1980.
56. P.C. Painter, M.M. Coleman and J.L. Koenig, The Theory of Vibrational
Spectroscopy and its Application to Polymer Materials, John Wiley & Sons, New
York, NY, USA, 1982.
57. J.E. Stewart and F.J. Linnig, Journal of Research of the National Bureau of
Standards, 1967, 71A, 19.
58. J.R. Shelton, J.L. Koenig and M.M. Coleman, Rubber Chemistry and Technology,
1971, 44, 904.
59. M.M. Coleman, J.R. Shelton and J.L. Koenig, Rubber Chemistry and Technology,
1972, 45, 173.
242
1973, 46, 938.
1973, 46, 957.
62. C.H. Chen, J.L. Koenig, J.R.Shelton and E.A. Collins, Rubber Chemistry and
Technology, 1981, 54, 734.
63. C.H. Chen, J.L. Koenig, J.R. Shelton and E.A. Collins, Rubber Chemistry and
Technology, 1982, 55, 103.
64. W.A. Bishop, Journal of Polymer Science, 1961, 55, 827.
65. E.W. Madge, Chemistry and Industry (London), 1962, 42, 1806.
66. H. Blümel, Kautschuk und Gummi Kunststoffe, 1963, 16, 571.
67. E.F. Devlin and A.L. Mengel, Journal of Polymer Science, Polymer Chemistry,
1984, 22, 843.
68. X. Chen, S. Zhang, X. Wang, X. Yao, J. Chen and C. Zhou, Journal of Applied
69. C. Hill, M. Kralevich and J.L. Koenig, Proceedings of the 152nd ACS Rubber
Division Meeting, Cleveland, OH, USA, Fall 1997, Paper No.69.
70. A.H.M. Schotman, P.J.C. van Haeren, A.J.M. Weber, F.G.H. van Wijk, J.W.
Hofstraat, A.G. Talma, A. Steenbergen and R.N. Datta, Rubber Chemistry and
Technology, 1996 , 69, 727.
71. P.J. Nieuwenhuizen, S. Timal, J.G. Haasnoot, A.L. Spek and J. Reedijk,
Chemistry - A European Journal, 1997, 3, 1846.
72. D. Gross and J. Kelm, Kautschuk und Gummi Kunststoffe, 1987, 40, 13.
73. K. Fujimoto and K. Wataya, Journal of Applied Polymer Science, 1969, 13, 2513.
74. K. Fujimoto and S. Nakade, Journal of Applied Polymer Science, 1969, 13, 1509.
75. J.H.M. van den Berg, J.W. Beulen, E.F.J. Duynstee and H.L. Nelissen, Rubber
76. E.F.J. Duynstee, Kautschuk und Gummi Kunststoffe, 1987, 40, 205.
243
77. M. van Duin and others, to be published.
78. P.J. Nieuwenhuizen, S. Timal, J.M. van der Veen, J.G. Haasnoot and J. Reedijk,
Rubber Chemistry and Technology, 1998, 71, 750.
79. P. Versloot, J.G. Haasnoot, J. Reedijk, M. van Duin, E.F.J. Dynstee and J. Put,
80. P. Versloot, J.G. Haasnoot, P.J. Nieuwenhuizen, J. Reedijk, M. van Duin and J.
Put, Rubber Chemistry and Technology, 1997, 70, 106.
81. S.K. Freeman, Applications of Raman Spectroscopy, Wiley Interscience, New

York, 1974, Chapter 8.
82. ISO 4097, Rubber, Ethylene-Propylene-Diene(EPDM)-Evaluation Procedure, 2000.
83. D.S. Campbell and B. Saville, Proceedings of the International Rubber

Conference, Brighton, UK, 1967.
84. V.M. Litvinov, W. Barendswaard and M. van Duin, Rubber Chemistry and
Technology, 1998, 71, 105.
85. K. Hummel, W. Scheele and K-H. Hillmer, Kautschuk und Gummi Kunststoffe,
1961, 14, 171.
86. O. Lorenz, Kautschuk und Gummi Kunststoffe, 1960, 13, 206.
87. D. Patterson and J.L. Koenig, Organic Coatings Applied Polymer Science
Proceedings, 1983, 48, 455.
88. L. Gonzalez, A. Rodriguez, A. Marcos and C. Chamorro, Rubber Chemistry and

Technology, 1996, 69, 203.
89. L. Gonzalez, A. Rodriguez, A. Marcos and C. Chamorro, Kautschuk und Gummi

Kunststoffe, 1994, 47, 715.
90. L. Gonzalez, A. Rodriguez, A. Marcos and C. Chamorro, Kautschuk und Gummi

91. L.D. Loan, Rubber Chemistry and Technology, 1967, 40, 149.
92. K. Hummel und G. Kaiser, Kautschuk und Gummi Kunststoffe, 1963, 16, 426.
93. B.M.E. van der Hoff, Applied Polymer Symposia, 1968, 7, 21.
244
94. M. Bellander, B. Stenberg and S. Persson, Kautschuk und Gummi Kunststoffe,

1999, 52, 265.
95. W.C. Endstra, Rubber Chemicals, Application Research Bulletin, Crosslinking

Agents, Akzo Nobel, 1985.
96. J. Fossey, D. Lefort and J. Sorba, Free Radicals in Organic Chemistry, John Wiley
& Sons, Chichester, 1995, 139.
97. F.P. Baldwin, P. Borzel, C.A. Cohen, H.S. Makoswki and J.F. van de Castle,
98. Wirtschaftsverband der Deutschen Kautschukindustrie, Grünes Buch Nr. 39:

Elastomere auf Basis Äthylen-Propylen, Frankfurt a/M, 1979.
99. W. Hofmann, Kautschuk und Gummi Kunststoffe, 1987, 40, 308.
100. W. Hofmann, Vulkanization und Vulkanizationshilfsmittel, Bayer AG,

Leverkusen, 1965.
101. R.C. Keller, Rubber Chemistry and Technology, 1988, 61, 238.
102. W.C. Endstra, Proceedings of the International Conference on Various Aspects of

Ethylene-Propylene Based Polymers, Leuven, Belgium, 1991.
103. L.D. Loan, Journal of Polymer Science, Polymer Letters, Part B, 1962, 2, 59.
104. R. Wiedenmann, Radiation Physics and Chemistry, 1977, 9, 701.
105. A. Zyball, Kunststoffe, 1977, 67, 461.
106. V.I. Dakin, Z.S. Egorova and V.L. Karpov, International Polymer Science and
Technology, 1977, 4, 11, T/34.
107. D. Simúnková, R. Rado and A. Saliga, Plaste Kautschuk, 1980, 27, 247.
108. P. Laurenson, E. Fanton, G. Roche and J. Lemaire, European Polymer Journal,

1981, 17, 989.
109. D. Apotheker, J.B. Finlay, P.J. Krusic and A.L. Logothetis, Rubber Chemistry and
Technology, 1982, 55, 1004.
110. H.G. Dikland, L. van der Does and A. Bantjes, Rubber Chemistry and
Technology, 1993, 66, 196.
245
111. H.G. Dikland, S.S. Sheiko, M. Möller, L. van der Does and A. Bantjes, Polymer,
1993, 34, 1773.
112. H.G. Dikland, R.J.M. Hulskotte, L. van der Does and A. Bantjes, Kautschuk und
Gummi Kunststoffe, 1993, 46, 608.
113. Z.H. Murgic, J. Jelencic and L. Murgic, Polymer Engineering and Science, 1998,
38, 689.
114. H.G. Dikland, T. Ruardy, L. van der Does and A. Bantjes, Rubber Chemistry and
Technology, 1993, 66, 491.
115. H.G. Dikland, R.J.M. Hulskotte, L. van der Does and A. Bantjes, Polymer
Bulletin, 1993, 30, 477.
116. H.G. Dikland, Workshop Vulkanisation, Process-Technik-Wirkung, Deutsches

Institut für Kautschuktechnologie, Hannover, 1997, Paper No.11.
117. H.G. Dikland, L. van der Does and A. Bantjes, Kunststoffe en Rubber, 1993, 46, 15.
118. H.G. Dikland, unpublished results.
119. J.G. Kloosterboer, Advances in Polymer Science, 1988, 84, 1.
120. A. Matsumoto, K. Iwanami and M. Oiwa, Journal of Polymer Science, Polymer

Letters, 1980, 18, 307.
121. A. Matsumoto, H. Sasaki and M. Oiwa, Die Makromoleculare Chemie, 1973,

166, 179.
122. T. Holt and W. Simpson, Proceedings of the Royal Society of London, 1956,
A238, 154.
123. H.G. Dikland, I.M. Leussink, L. van der Does and A. Bantjes, Kautschuk und
124. R. Winters, J. Lugtenburg, H.J.M. de Groot and M. van Duin, Macromolecules,

to be published.
125. Y.M. Tsai and F.J. Boerio, Journal of Adhesion, 1995, 55, 151.
246
7
NMR Imaging of Elastomers
Bernhard Blümich and Dan E. Demco
NMR imaging finds most of its applications in medical diagnostics of humans and
pathology studies of animal models [1, 2]. The success of the method is based on the
noninvasiveness of nuclear magnetic resonance [3–5] and the unsurpassed soft-matter
contrast, which is hard to achieve with competitive methods like x-ray or computer
tomography. The same advantages can be exploited in imaging studies related to materials
science [6–10]. Here an important class of soft-matter materials is given by synthetic
polymers above the glass transition temperature. Apart from semi-crystalline polymers
like poly(ethylene), poly(propylene), some poly(amides), and polymer melts, elastomers
constitute the most striking class of synthetic soft matter with a modulus similar to many
biological tissues [11]. Applications of imaging for which non-destructiveness or contrast
are essential, are competitive with other imaging techniques in the information gained
and the cost of the experiment. Such applications to elastomers concern distributions in
temperature, stress, cross-link density, modulus, and the dynamics of fluid absorption
and swelling.
The nucleus imaged most often is the proton. The reason is not only sensitivity but also
the weak dipolar couplings between protons in a chemical group and between different
chemical groups which dominate the signal decay by relaxation. These dipolar couplings
are motionally averaged by the often fast but nearly always anisotropic motion of
intercross-link chains. This motion and consequently the value of the residual dipolar
couplings is affected by chain stiffness, cross-link density, chain orientation, temperature,
additives etc. Given that the residual dipolar interactions are too strong to obtain chemical-
shift resolution without sample spinning or multi-pulse techniques, relaxation techniques
which probe different time regimes of molecular motion provide the primary access to
contrast in imaging of elastomers.
Relaxation can be probed in inhomogeneous fields, so that the homogeneous polarization

field B0 is not a necessity for successful applications of soft-matter imaging. Based on this
fact, small portable NMR sensors can be built, which provide NMR data from a single
volume element with the same specificity as the contrast in an NMR image [12]. Such
247
sensors are suitable for investigations of very large objects as well as for process and quality
control in an industrial setting. Their price is an order of magnitude lower than that of an
NMR imager, and the spatial resolution may be several millimeters compared to about
0.1 mm in typical applications of NMR imaging to elastomer materials.
The following text is divided into three parts. The first part gives a general introduction
to NMR imaging and contrast. The second part reviews illustrative examples of NMR
imaging to elastomer materials. The last part shows some examples of investigations of
elastomer materials with mobile NMR equipment.
7.1 NMR Imaging and Contrast
The power of modern NMR methods derives from the fact, that the phase of the transverse
magnetisation can be measured. By use of the Fourier transformation the phase
information can be converted into probability densities of resonance frequencies (Larmor
Frequencies) (multi-dimensional spectra) [13, 14], densities of position (NMR images)
[1–4], and probability densities of parameters like velocity and acceleration which
quantifiy translational motion [15–17]. The signal phase is the angle ϕ which the precessing
magnetisation M forms with the x axis of the rotating frame (Figure 7.1). In general, this
angle depends on time t, because the magnitude of the magnetic field can be changed
during the experiment and parts of the object can be in motion. The former is the case
when pulsed field gradients (PFG) are applied to the object during measurement; the
latter is the case, for example, when a polymer melt passes through an extruder or a
static mixer.
7.1.1 Principle of Fourier NMR
If ωL(t) is the time-dependent Larmor frequency, the precession phase (θ in Figure 7.1)
can be expressed in an inhomogeneous magnetic field according to:
t t
ϕ ( x,t ) = ∫ ω L ( x,t' ) dt' = − γ ∫ B ( x,t') dt'
z
0 0
(1)
⎡ t
∂Bz ( x,t' ) ⎤ ⎡ t ⎤
= − γ ⎢ B0 t + ∫ x( t' ) dt' + ... ⎥ = − γ ⎢ B0 t + ∫ Gx ( t' ) x( t' ) dt' + ... ⎥
⎢ ∂x ⎥ ⎢⎣ ⎥⎦
⎣ 0 x= 0 ⎦ 0
248
Figure 7.1 Most modern NMR techniques are based on the fact, that the phase ϕ of the
precessing transverse magnetisation M(t) kann be measured. By use of the Fourier
transformation the phase provides access to NMR spectra, images, and parameters of
translational motion like velocity v and acceleration a. Spectroscopic parameters as well
as components of translational velocity and acceleration can be used for generating
contrast in NMR imaging. In the drawing the magnetisation M(t) has been generated
from Mz by use of a 90° pulse of the B1 radio-frequency (rf) field in y direction
Here the inhomogeneous polarizing magnetic field points in z direction, and the spatial
inhomogeneity in x direction has been expanded into a Taylor series and truncated after
the second term. The first term of this series is the homogeneous field B0, the second one
is the field gradient Gx = (∂Bz/∂x)x=0 in x direction.
The quantity x(t) denotes the position of a magnetisation component on the x axis. For
flowing and moving objects the position is a function of time. For short times also x(t) is
expanded into a Taylor series and arbitrarily truncated after the third term,
1
x( t ) = x0 + v x0 t + a x0 t 2 + ... (2)
2
The expansion coefficients are initial position x0, initial velocity vx0, and initial acceleration
ax0 in x direction. Combination of equations (1) and (2) yields:
249
ϕ( t ) = ω 0 t
ϕ ( t ) = − γ B0 t
t
−γ ∫ G ( t') dt' x
x 0 + kx ( t ) x0
0
t
+ q x ( t ) v x0
−γ ∫ G ( t') t' dt' v
0
x x0 (3)
+ ε x ( t ) a x0
t
1
−γ ∫ G ( t') t'
2
x dt' a x0
2
0
− ..... + .....
The individual terms of this expansion refer to different forms of ?????. The first term
(γB0t) of Equation (3) concerns homogeneous magnetic fields and thus addresses NMR
spectroscopy, for example, the phase contribution from the chemical shift. The second
term describes the dependence of the signal phase on initial position in the presence of
magnetic field gradients. It forms the basis of NMR imaging. The higher order terms
describe the dependence of the phase on the transport parameters vx, ax, etc. In imaging
the first, the third, and higher order terms are exploited to define contrast in NMR
images. The experimental variables are the magnetic field gradient G and radio-frequency
(rf) pulses, whereby 180° pulse effects a sign change of the inhomogeneous magnetic
field in the rotating frame.
In NMR spectroscopy the gradient Gx(t) is zero so that the signal phase is given by ω0 t. In
pulsed NMR the free induction decay (FIND) s(t) is acquired as a function of time, and by
Fourier transformation of the FIND, the NMR spectrum S(ω), is obtained. Similarly, to
obtain an NMR image S(x0) the NMR signal s(kx) must be measured as a function of kx,
the Fourier-conjugate variable of x. By the same principle, distributions S(vx0) of velocity
and distributions S(ax0) of acceleration can be obtained by measuring the signals s(qx) and
s(εx) and subsequent Fourier transformation. According to Equation (3) kx, qx, and εx are
defined by the moments mn(t) of the gradient modulation function Gx(t),
t
mn ( t ) = ∫ G ( t') t'
x
n
dt' (4)
0
Thus by suitable modulation of Gx(t) the signal phase can be made dependent on position,
velocity, or acceleration. The effect of a distribution in frequency is handled either by
250
sufficiently large gradients to dominate the phase evolution or by indirect detection

schemes, where only the gradient amplitude and not the duration t of the gradient
modulation period is varied (see Section 7.1.2).
7.1.2 Spatial Resolution
In the absense of object motion and in a time-invariant field gradient Gx, Equation (3)
simplifies to:
t
ϕ ( t ) = ω L t = − γ B0 t − γ ∫ Gx (t' ) dt' x0 = ω 0 t − γ Gx t x0 = ω 0 t + kx (t ) x0 (5)

0
so that the frequency offset in the rotating frame is determined by the gradient only,
Figure 7.2 Generation of spatial resolution in NMR: (a) Conventional NMR imaging
with magnetic field gradients. A magnetic field gradient Gx in x0 direction (top)
converts the NMR spectrum (bottom) of an object (middle) into a projection of the
object, (b) Localization of NMR signals from a large object by use of a surface coil
251
Ω = ω L − ω 0 = − γ G x x0 (6)
As a consequence, the NMR spectrum of a narrow resonance provides a projection of
the object (Figure 7.2a). Each frequency Ω encodes another space position x0, and the
signal intensity is proportional to the number of nuclei at this coordinate. This number is
obtained by integration of the distribution of longitudinal magnetisation Mz(x, y, z) over
the other two space coordinates, where the magnetic field Bz is homogeneous. Such an
integral defines a projection in the mathematical sense, similar to a projection which is
measured by x-ray imaging. From many such projections measured with different gradient
orientations relative to the object an image of the object can be reconstructed. For reasons
Figure 7.3 Sampling principles in 2D k space: (a) Cylindrical coordinates. The angle of
the field-gradient direction with respect to the x axis is given by θ = arctan{Gy / Gx},
(b) Cartesian coordinates. For rectangular gradient pulse shapes ky = -g Gy t1 and kx =
-g Gx t2. Such sampling schemes are applicable to a slice which can be selected when
the rf pulse is applied selectively in the presence of a gradient Gz. The areas of k space
accessible by the pulse sequences shown are shaded in gray. TX: transmitter signal;
RX: receiver signal; Gx, Gy: gradient signals
252
that the FIND corresponding to the projections are measured in Fourier space, the image
data are sampled in in cylindrical k-space coordinates (Figure 7.3a), and the image
construction involves a conversion of the sampling grid to Cartesian k space and 2D
Fourier-transformation.
For a one-dimensional (1D) image of a static object like the projection in Figure 7.2a,
bottom, the signal measured in a frame rotating at frequency ω0, is given by:
(
s kx ( t ) =) ∫∫∫ M ( x , y , z )dz
z 0 0 0 0 { }
dy0 exp i kx ( t ) x0 dx0 (7)
where kx(t) x0 = -γ Gx t x0 is the phase from Equation (5). Relaxation is neglected in this
equation. Fourier transformation over kx yields the 1D image
S( x0 ) = ∫∫ Mz ( x0 , y0 , z0 ) dy0 dz0 (8)
Several trajectories for k-space sampling are known in imaging, most notably sampling
in Cartesian coordinates (Figure 7.3b) [1 – 5]. In this case two types of space encoding
are distinguished, phase-encoding and frequency encoding. Phase encoding is only
applicable to indirect detection referring to the t1 period. Here ky (cf. Equation 3, line 2)
is varied in repeating scans by changing the area under the gradient pulse at constant
pulse width. This procedure is more favorable than varying the gradient pulse length,
because it avoids changes in the relaxation weight during the gradient pulse and changes
in the signal phase from the evolution of the chemical shift or other spin interactions (cf.
Equation 3, first term). Frequency encoding is used for direct signal detection during t2.
Here the gradient amplitude is kept constant and as the time increases, kx increases, and
data are acquired. In this case the spatial resolution is limited by relaxation, i.e. the
linewidth and the width of the NMR spectrum, while for phase encoding, the spatial
resolution is limited only by the signal-to-noise ratio.
7.1.3 Contrast
Most conventional imaging techniques are variants of Fourier imaging in Cartesian k

space. The scheme in Figure 7.3b can be modified to enable sampling of positive and
negative halfs of k space in the sampling direction kx, if signal echoes are introduced in
terms of gradient echoes or Hahn echoes (Figure 7.4). Hahn echoes are generated by a
sequence of a 90° and a 180° rf pulse (Figure 7.4a). In the echo maximum (Rx) the
effects of background magnetic field gradients and chemical shift in the phase-encoded
dimension are eliminated, while for gradient echoes consisting of only one pulse
253
Figure 7.4 Pulse sequences used for imaging of elastomers. A slice in direction z is
defined by a frequency-selective pulse applied in a z gradient. Positive and negative
halves of k space are accessed in x direction by sampling the entire echo. In the y
direction the gradients are stepped through positive and negative values. The signal
echo appears at the echo time tE after the first pulse: (a) Hahn-echo imaging. Signal
dephasing in background gradients is refocused by application of a 180° pulse, (b)
Gradient-echo imaging. Fast repetition of the experiment is enabled by small flip-angle
excitation pulses α
(Figure 7.4b) they are not. Background gradients may arise from inhomogeneities of the
polarizing magnetic field B0 and from susceptibility differences at interfaces within the
sample. In the latter case the very sample heterogeneities to be investigated may the
source for magnetic field distortions, so that gradient-echo imaging produces increased
image contrast.
• Susceptibility Contrast
The differences in image contrast resulting from Hahn-echo and gradient echo imaging
are illustrated in Figure 7.5 by 2D images from six unvulcanized EPDM sheets separated
by PTFE layers [18]. While some sample heterogeneities are recognizable in the Hahn-
254
Figure 7.5 [15] Susceptibility contrast in EPDM samples at 363 K: (a) 2D Hahn-echo
image acquired with the sequence of Figure 7.4a (b) 2D gradient-echo image acquired
with the sequence of Figure 7.4b
echo image (a) the effects from processing by mixing and folding of layers are much
better visualized in the gradient echo image (b).
• Relaxation-Time Contrast
Although the spatial resolution is rarely better than 0.1 mm in NMR images of elastomer
materials, NMR imaging nevertheless is useful for their analysis, because features invisible
by other imaging techniques can be detected. Compared to other techniques, the number
of contrast parameters in NMR imaging is abundant. Most parameters accessible by
conventional NMR in homogeneous magnetic fields can be used for generation of contrast
[19–21]. However, the important contrast parameters are the relaxation times, in particular
those, which are sensitive to slow molecular motion. Examples are the transverse
relaxation time T2, the longitudinal relaxation time T1ρ in the rotating frame, and the
relaxation time T2e of the multi-solid echo decay. For example, when the space encoding
sequence is preceded by a T2 filter consisting of a Hahn-echo pulse sequence, the
magnetisation available for space encoding is given by:
255
⎧ t ⎫
Mz (r) = exp ⎨− E ⎬ M0 (r) (9)
⎩ T2 (r) ⎭
where M0 is the thermodynamic-equilibrium magnetisation. Mz(r) is then used for space

encoding following the principles outlined above. In Hahn-echo imaging, the T2 filter is
usually integrated into the space-encoding sequence (Figure 7.4a).
Parameter images are of interest in so far, as NMR parameters can be correlated with
material properties either by experimental calibration or by theoretical models. An
example for experimental parameter calibration is given in Figure 7.6 with a quantitative
Figure 7.6 NMR parameter image of a strained poly(dimethylsiloxane) rubber band

with a cut and calibration curves: (a) Experimental curve for T2 versus strain, (b)
Experimental stress-strain relationship, (c) Calibration curve for T2 versus strain
obtained from combination of curves a and b, (d) Stress image obtained by
recalibration of a T2 parameter image. The stress contours range from 0 to 2.4 MPa
256
stress image of a strained poly(dimethylsiloxane) (PDMS) band with a cut [22]. T2 has
been determined from the tail of the signal decay curve measured with Hahn echoes for
each picture cell or pixel and subsequently recalibrated with the help of the T2-versus-
strain curve and the stress-strain curve determined for the rubber band without spatial
resolution. The contrast variation is attributed to the strain distribution associated with
the cut and to the variance of the random distribution of active silicate filler which also
affects the chain mobility and thus T2. This simple calibration procedure neglects the
tensorial properties of stress but nevertheless results in an image which depicts basic
features of the stress distribution.
• Models of Transverse Relaxation in Crosslinked Elastomers
The transverse magnetisation decay and the longitudinal decay in the rotating frame
have been modelled in terms of the network topology [23–25]. The motion of chain
segments in inter-crosslink chains is usually fast but anisotropic. The restricted motion
leads to a non-exponential decay of the Hahn-echo maxima with solid-like contributions
from residual dipolar couplings in the chain. The magnitude of these couplings increases
with increasing cross-link density, decreasing temperature, and with other restrictions in
Figure 7.7 Fast but anisotropic segmental motion results in a solid-like contribution to
the NMR signal. This contribution is expressed in terms of a fractional contribution q
of the second moment M2 of the rigid lattice line of a single chain or residual dipolar
interactions between protons. The line splitting caused by the dipole-dipole interaction
depends on the orientation angle q of the internuclear vector of the coupling protons
in the magnetic field B0. The distribution of orientation angles changes with the
network deformation
257
chain mobility like strain and network deformation. In the latter the case transverse
magnetisation decay becomes macroscopically anisotropic, because the line splitting
associated with the dipole-dipole interaction depends on the orientation angle θ of the
internuclear vectors with respect to the direction of the applied magnetic field B0.
(Figure 7.7). In the former case the internuclear vector of the coupling protons are
isotropically distributed and no orientation dependence of the NMR signal is observed.
In the Gotlib-Fedotov-Schneider model [24,26–28] the solid-like contributions are

expressed in terms of a Gaussian with a fractional contribution q of the second moment
M2 of the rigid lattice line of a single chain. Further signal contributions from dangling
chains with approximately isotropic motional averaging are taken care of by an additional
exponential relaxation term. The simplified expression suitable for analysis of imaging
data is given by [24, 29] (and references therein):
⎧ t ⎫ ⎧ t ⎫
Mx (tE ) = A exp ⎨− E − q M2 tE2 ⎬ + B exp ⎨− E ⎬ (10)
⎩ T2 ⎭ ⎩ T2 ⎭
From a fit of Equation (10) to spatially resolved relaxation curves, images of the parameters
A, B, T2, q M2 have been obtained [3- – 32]. Here A/(A + B) can be interpreted as the
concentration of cross-links and B/(A + B) as the concentration of dangling chains. In
addition to A/(A + B) also q M2 is related to the cross-link density in this model. In
practice also T2 has been found to depend on cross-link density and subsequently strain,
an effect which has been exploited in calibration of the image in Figure 7.6. Interestingly,
carbon-black as an active filler has little effect on the relaxation times, but silicate filler
has. Consequently the chemical cross-link density of carbon-black filled elastomers can
be determined by NMR. The apparent insensitivity of NMR to the interaction of the
network chains with carbon black filler particles is explained with paramagnetic impurities
of carbon black, which lead to rapid relaxation of the NMR signal in the vicinity of the
filler particles.
In the Cohen-Addad-Sotta model [23, 25] the Hahn-echo decay is expressed without the
assumption of dangling chains by the expression:
−1
⎧⎪ ⎛ 2 i Δ tE ⎞ ⎛ i Δ tE ⎞ ⎫⎪
−1
Mx (tE ) = Re ⎨ M0 ⎜ 1 − ⎜1 + ⎟ ⎬ (11)
⎪⎩ ⎝ 3 Ne ⎟⎠ ⎝ 3 Ne ⎠ ⎪
⎭
where Δ is the average dipolar interaction on a crosslink chain and Ne is the effective
number of chain segments. At sufficiently high temperature above the glass transition
258
temperature Tg, Ne can be interpreted as the number of Kuhn segments in the freely
jointed chain. Thus the cross-link density is proportional to Ne-1, and the signal decay
curves for rubber samples with different crosslink densities collapse onto one master
curve, if Mx is plotted as a function of tE/Ne [25].
Although at very long echo times tE this approach [25] is less accurate than a refined
Gotlib-Fedotov-Schneider model [26 – 28] the accuracy of Equation (11) is sufficient to
analyze imaging data. A signal decay modeled by Equation (11) is depicted by a line of
diamonds in Figure 7.8. A biexponential fit (solid line) fails at short tE but provides excellent
agreement at long tE for the characterization of T2. At short tE a Gaussian curve (broken
line) provides the best fit. Here the solid-like contribution dominates the signal decay.
• Contrast from Multi-Quantum Filtering
A more detailed analysis of the NMR signal from elastomer samples, addresses finer
details such as the chemical structure of chain segments. In general a hierarchy of dipolar
interactions between protons exists instead of a single chain-averaged dipolar interaction
[34, 35]. For example, in cis-1,4-poly(isoprene) these different dipolar interactions can
Figure 7.8 The shape of the Hahn-echo decay in cross-linked elastomers is given by the
line of diamonds. In the short-time limit the curve can be approximated by a Gaussian
(broken line) and in the long time limit it follows an exponential (solid line) [30]
259
be discriminated even in static samples by multi-quantum NMR. CH groups give rise to

single-quantum signals only, CH2 groups give rise to single- and double-quantum signals,
and CH3 groups to single-, double-, and triple-quantum signals. In multi-quantum NMR
n-quantum signals arise from coherent motion of n protons which interact by the dipole-
dipole coupling. Protons in close contact and on chain segments undergoing slow or
highly restricted motion exhibit strong dipolar couplings, the more distant protons are
coupled only weakly. Thus it is not surprising, that the multi-quantum signals are mainly
produced by protons within one chemical group like CH2 and CH3 while the intergroup
couplings are less significant.
This is illustrated in Figure 7.9a by multi-quantum filtered 1H NMR spectra of a non-

spinning sample of poly(isoprene) [34, 35]. The conventional single-quantum spectrum
essentially does not provide chemical-shift resolution. However, the methylene and methyl
signals can be partially separated by double- and triple-quantum filtering. The double-
quantum filtered signal contains contributions mainly from the CH2 groups and a smaller
contribution from the CH3 groups. The number of coupling protons defines the maximum
order of the multi-quantum signal or coherence order which can arise, but any subset of
the coupling spins can contribute to multi-quantum signals of lower coherence orders.
For example, a double-quantum signal requires two coupling protons to flip
simultaneously by absorption of two rf photons or quanta (Figure 7.9b), but also each
proton can flip independent of its coupling partners in a single-quantum process.
Staying with the example of two coupling protons in a CH2 group, two types of double-
quantum processes can be distinguished. The double-flip and the flip-flop process. The
double flip is observed by way of filtering double-quantum coherence. The resultant
coherence order of the flip-flop process is zero. In contrast to established zero-quantum
spectroscopy [13] it involves no signal precession at all and refers to antiparallel spin
alignment of the coupling protons (Figure 7.9b). Nevertheless, it is noted that for an
isolated multi-quantum pair of spins the sequence can excite only dipolar encoded
longitudinal magnetisation and double-quantum coherences [34]. The dipolar encoded
longitudinal magnetisation apparently behaves like zero-quantum coherence.
The signals from different coherence orders can be separated by suitable phase-cycling
schemes [13] of the multi-quantum filter (Figure 7.9c). In such a filter multi-quantum
coherence is generated from longitudinal magnetisation by a preparation pulse sequence
of duration τp, allowed to evolve in a multi-quantum evolution period t1, and then
converted back to longitudinal magnetisation by a mixing pulse sequence of duration τm.
Usually τp = τm. For the purpose of imaging multi-quantum filtered signals, one of the
imaging schemes of Figure 7.4 follows the multi-quantum filter.
Depending on the strength of the dipole-dipole interaction, the multi-quantum filtered

signals are generated with different intensity (Figure 7.9d). Strongly coupled protons are
260
Figure 7.9 Multi-quantum NMR: (a) Multi-quantum filtered spectra of poly(isoprene).

In contrast to the single-quantum (1Q) spectrum the double- (2Q) triple-quantum
(3Q) spectra exhibit relative chemical-shift resolution. The double-quantum signal
mainly arises from CH2 groups. The triple-quantum signal derives from the CH3
groups, (b) Double-quantum excitation of a methylene group generates double-
quantum coherence by a double-flip process, (c) Pulse sequence for excitation and
detection of multi-quantum filtered signals. The excitation time is denoted by tp.
Multi-quantum signals of different orders are selected on the basis of the rf pulse
phases in combination with partial signal cancellation during averaging, (d) Build-up
curves of multi-quantum filtered transverse 1H magnetisation of polyisoprene: dipolar
encoded longitudinal magnetisation (left) and double-quantum filtered signals (right)
for τp = τm. Opposite contrast is obtained at short τp
261
observable at short tp, weakly coupled ones at long tp. Based on the different coupling
strengths of protons in different chemical groups multi-quantum signals of the same order
but of different origin can therefore be emphasized and attenuated by suitable variation of
τp. This fact is confirmed by the different chemical shifts observed for the double- and
triple-quantum filtered signals in Figure 7.9a. But the coupling strength also depends on
the modulus of the material, i. e. on cross-link density, chain extension by strain, and on
temperature. In the short time limit, strong multi-quantum signals are observed from chains
stiffened by high cross-link density, applied strain, and low temperature. In this regard
contrast obtained from multi-quantum filtering at short τp is inversely related to contrast
induced from T2 weighting. Application of the zero-quantum filter provides dipolar-encoded
longitudinal magnetisation. In images of elastomers acquired with such a filter the contrast
is similar to T2 contrast but more sensitive to material properties.
The different contrast achieved by conventional T2 weighted Hahn-echo imaging and by

multi-quantum filtered imaging for dipolar-encoded longitudinal magnetisation is
illustrated in Figure 7.10 by images of a section from a silicone-breast implant envelope
from poly(dimethylsiloxane) which had been implanted for 3 years [36]. The T2 weighted
image (a) shows little to no signs of material deterioration, whereas in the multi-quantum
filtered image (b) different islands of material deterioration are readily identified. The
explanted envelope was brittle and had lost elasticity compared to a new one.
Double-quantum filtered images of a strained rubber band with a cut are depicted in
Figure 7.11. The image (a) has been obtained by 1H NMR with weak smoothing of the
experimental data. The signal-to-noise ratio is considerably lower than for dipolar-encoded
longitudinal magnetisation (cf. Figure 7.10b), and the contrast is inverted. Regions with
Figure 7.10 Images of a section from a worn silicon-breast implant envelope: (a) T2-
weighted imaged, (b) image of dipolar encoded longitudinal magnetisation
262
Figure 7.11 Double-quantum images of a strained rubber band with a cut: (a) 1H
image, (b) 2H image of deuterated spy molecules incorporated into the rubber network
by swelling, (c) finite element simulation of the stress distribution
weak dipolar couplings appear bright in Figure 7.10b and dark in Figure 7.11a. Thus the
region of high strain in the center of the cut shows high signal intensity in the double-
quantum image. Similar information can be obtained from double-quantum filtered
imaging of deuterons [37]. From the point of NMR spectroscopy deuterons behave like
a pair of coupled protons. Deuterated butadiene oligomers have been incorporated into
a rubber band as spy molecules by swelling and subsequent solvent drying. The geometry
of molecular motion of the spy molecules probes the void geometry of the rubber network,
which becomes anisotropic upon straining the sample. The motionally averaged residual
quadrupolar coupling of the deuteron is sensitive to the anisotropy of molecular motion
in a way similar to the dipole-dipole interaction between two protons, except that it is
zero in isotropic elastomers, while the residual dipolar coupling of an intercross-link
chain is not. In either case multi-quantum NMR can be used to probe the residual dipolar
and quadrupolar interactions which are influenced by the state of the rubber network.
The stress distribution obtained for a strained rubber band by double-quantum deuteron
NMR (Figure 7.11b) [37, 38] agrees with the one obtained from proton NMR
(Figure 7.11a) [39]. The quality of the deuterium image appears better because of much
stronger data smoothing. The measured signal intensities match the stress distribution
simulated by finite element methods (Figure 7.11c). Like the stress image from T2 data
(cf. Figure 7.6) the double-quantum filtered images do not account for the orientation
dependence of the stress tensor. To this end multi-quantum filtered NMR images need to
be measured from different orientations of the strained sample in the magnetic field [40],
and the resultant data set should be analyzed for the principle components of the stress
263
tensor, that is, for the trace, the anisotropy, the main principle value, and the orientation
angles in the sample. This approach to imaging of anisotropic material properties has
been demonstrated in diffusion tensor imaging of ordered biological tissue [41 – 43].
7.2 Applications
Applications of NMR imaging outside medicine are relatively rare [6, 7]. Potentially
rather useful applications are in the elastomer industry. This section features some selected
examples which illustrate the type of information obtainable by imaging of elastomers.
Heterogeneities in technical elastomers arise in different stages of elastomer production

and product use. Even in a perfectly prepared homogeneous elastomer product, unavoidable
ageing processes induce space-dependent defects. Examples for sources of defects are:
• Mixing process: Technical rubbers are blends of up to about 30 different compounds

like natural rubber, styrene-butadiene rubber, silicate and carbon-black fillers, and
mobile components like oils and waxes. These components show a large variety of
physical, chemical, and NMR properties. Improper mixing leads to inhomogeneties
in the final product with corresponding variations in mechanical and thermal properties
(cf. Figure 7.4).
• Vulcanisation: The changes in segmental mobility from progress of the vulcanisation

reaction and from the associated sample temperature distribution can be monitored
directly by NMR imaging in situ [44]. Heterogeneous structures arise from effects of
thermal conductivity, which lead to space-dependent temperature profiles during the
vulcanisation process depending on the position of the heat source and on heat
dissipation. As a result inhomogeneous cross-link densities may be established [45]. In
the covulcanisation of blends and sheets from different formulations inhomogenities in
cross-link density may arise from differences in solubility and diffusion of the curatives
[46]. Different components of rubber blends can be mapped by exploiting the editing
capabilities of motionally narrowed, swollen rubber samples [47] and of proton-detected
13
C imaging, which are useful even in the case of unresolved lines [48].
• Ageing: Ageing processes are most often introduced by UV irradiation, exposure to

heat and oxygen, and by biological mechanisms (cf. Figure 7.10). Depending on the
load applied, different ageing processes are observed [49]. Thermal oxydative ageing
usually leads to the formation of hardened surface layers in natural rubber (NR) as
well as in synthetic rubber (SBR, styrene-co-butadiene rubber) [50 – 54]. Typically
these layers approach a thickness of 200 to 300 μm and inhibit the progress of the
ageing process further into the bulk of the sample. The material hardening is explained
264
by an increase in cross-link density. In the absence of oxygen, chain scission may

dominate at elevated temperatures with an associated increase in segmental mobility.
Other types of ageing involve aggressive fluids and gases. In this context a sample of
degraded rubber hose has been investigated [55], but also the degradation of
polyethylene pipes [56], and the enzymatic degradation of biologically synthesized
polymers [57] have been studied through NMR imaging. Related investigations have
been carried out on asphalts [58]. Ageing associated with swelling of the rubber
particles has been observed in crumb-rubber modified asphalts [59].
• Mechanical load: Static mechanical load by strain or compression leads to stretching

of random-coil polymer chains in the direction of sample elongation and chain
compression in the orthogonal directions (cf. Figure 7.7). Stress and strain effects can
be analyzed for instance by parameter maps of T2 [22] (cf. Figure 7.6), and by 1H [34,
35] and 2H [37, 38] multi-quantum imaging (cf. Figure 7.11). Dynamic mechanical
load leads to sample heating where the temperature distribution in dynamic equilibrium
is determined by the temperature-dependent loss-modulus and the thermal conductivity
of the sample. Because T2 scales with temperature for carbon black filled SBR, a T2
map provides a temperature map of the sample. Such temperature maps have been
measured for carbon-black filled SBR cylinders for different filler contents and
mechanical shear rates [60].
7.2.1 Defects and Heterogeneities in Technical Elastomer Products
A common source of sample heterogeneities arises from filler clusters. Hardly any technical
elastomers is free of it. Even the variance of the distributions statistics of active filler may
lead to heterogeneities detected as an average over a volume cell (or voxel) with the
spatial resolution of NMR imaging (cf. Figure 7.6d). Small filler clusters may be invisible
by Hahn-echo imaging but may give rise to enlarged imaged distortions from susceptibility
effects in gradient-echo imaging (cf. Figure 7.5).
An example for elastomer defects is shown in Figure 7.12 by images of a rubber gasket
which had been exposed to oil, heat, and pressure in an overload test. No signs of failure
can be detected by visual inspection of sections cut out of the gasket (a). However, in a
T2 weighted spin-echo image (b), filler defects can be identified (circular structure in the
center) as well as regions of severe sample hardening (dark strip on the right), and a
swollen periphery, where the technical oil has penetrated into the gasket (bright contour
lining the sample). In fact closer analysis reveals a chromatographic separation of the
swelling fronts associated with different components of the technical oil [61]. In the
image (c) from a different section of the gasket a crack is noticeable in the middle of the
image, and the strips of hardened material have grown larger.
265
The visibility of defects may be enhanced by swelling in a suitable solvent. This is illustrated
in Figure 7.13 by images of a filled and cured sample of the terpolymer isobutylene-p-
methylstyrene-p-bromoethylstyrene [47]. In the image (a) acquired with an echo time
tE = 2 ms a few large inhomogeneities identified as voids are recognizable. Much more
defects become oberservable by swelling in cyclohexane, which softens the matrix and
increases the transverse relaxation time for better signal-to-noise ratio. The position of
the voids remains unchanged, so that swelling preserves essential features of the elastomer
Figure 7.12 Images of a carbon-black filled rubber gasket: (a) Photograph, (b) T2-
weighted image from an undamaged region, (c) T2-weighted image from a region
damaged in an overload test from pressure applied in a bath from technical oil
Figure 7.13 1H Hahn-echo images of a terpolymer: (a) Unswollen sample, (b) sample
after swelling in cyclohexane. The images were taken from a 15 mm diameter cylinder
(adapted from [47] with permission of the authors)
266
matrix. This approach to imaging provides the opportunity to optimize contrast in a

sample-specific way by imaging the unswollen polymer, the swollen network, and different
solvents with chemical-shift-selective excitation.
Samples with very short relaxation times are difficult to image by Hahn-echo methods.
Images of such samples can be acquired by single-point imaging [62, 63], and its variant,
the SPRITE (single-point ramped imaging with T1 enhancement) method [64]. Basically
a small-flip-angle rf pulse is applied in the presence of a magnetic-field gradient, and a
single data point is acquired a short time after the pulse. Then the experiment is repeated
with a different value of the field gradient vector. In this way k space is scanned point by
point. At first sight the method appears to be inefficient in terms of acquisition time. But
Hahn echoes are avoided and small flip angles are used, so that repetition times are fast.
Moreover, the time delay between the pulse and data acquisition can be varied to introduce
chemical-shift modulation in the acquired signal. If a refocusing pulse is placed in the
middle of the acquisition delay, the chemical-shift evolution is refocused and just a T2
weight is introduced. 128 x 64 point SPRITE images of elastomers can be acquired in
about three minutes including signal averaging. An example of such an image is given in
Figure 7.14 for a tyre section consisting of carbon-black filled SBR layers and an NR
Figure 7.14 T2* weighted image of a car-tyre section showing layers of SBR (left) and
NR (right) which are separated by polymer fabric. The image was acquired by the
SPRITE technique in 200 s
267
layer separated by textile fabric. The acquisition delay of 5 ms results in a T2* weight
which enhances the signal from the softer NR component [65]. Spin density images are
obtained with this method for samples with long relaxation times compared to the
acquisition delay.
Two T2-weighted Hahn-echo image from sections of the tyre treads are compared in
Figure 7.15. The samples had to be cut from the steel belt in order to avoid severe
distortions of the polarizing magnetic field by the belt. In each image the tyre surface is
on the left (hard material, dark) and the base is on the right (soft material, bright). Next
to the base two layers of reinforcing polymer fibers (line of dots) are embedded in a hard
formulation in image (a) and in a soft formulation in image (b). Furthermore, in image
(b) the soft base material bulges out into the tread material, and filler clusters can be
identified. The performance of tyre (b) was inferior to that of tyre (a), where the base
material is hard. Overlaid to the images is a variation of T2 which had been measured by
the NMR-MOUSE (see Section 7.3 and Figure 7.22) at different depths starting from the
right and from the left of the sample. Extreme differences between the signals are found
at 6.5 mm depth in this case, which correlate with the hard and soft materials. This
example demonstrates, that the data measured by the NMR-MOUSE can deliver
information which correlates with that obtainable by NMR imaging. However, in
comparison with the imaging data, the signal-bearing volume is much larger and less
well defined. But the NMR-MOUSE measurements are nondestructive and can be
Figure 7.15 T2 weighted images of car-tyre sections showing in each image the tread (left)
and the base with polymer fibers (right): (a) Hard base, (b) Soft base. The overlaid curves
are T2 data measured as a function of depth by the NMR-MOUSE (see Section 7.3.1)
268
performed on tyres with steel belts, so that they can be carried out during tyre testing,
for example, on the race track.
7.2.2 Covulcanisation
During fabrication of elastomer products like car tyres different elastomer formulations
often need to be covulcanized. At the interface between an SBR and an NR layer, an
interfacial layer with a modulus higher than either SBR and NR had been detected in a
sample from a used car tyre [51]. Subsequently such interfaces were detected in other
samples as well. For example, in unfilled elastomer composites they can be seen even by
visual inspection [52]. In Hahn-echo images the interface is identified by a dark line
paralleled by a slightly brighter line (Figure 7.16a). The contours associated with spin
density and fast relaxing components are eliminated from the image by forming the ratio
between two Hahn-echo images acquired with different echo times, a short and a long
one. While the image in Figure 7.16a has been acquired with an echo time of tE = 3.3 ms,
the image Figure 7.16b has been obtained by normalizing the image acquired with
Figure 7.16 T2 weighted Hahn-echo image: (a) and quotient image (b) computed as
the quotient of two Hahn echo images acquired with different echo times of an
unfilled SBR/NR covulcanisate with dimensions 9 mm x 30 mm. The acquired signal
has been integrated over the sample thickness of 1 mm. Contrast in the quotient image
is determined mainly by transverse relaxation. Contrast in the Hahn-echo image is
formed by a mixture of spin density and relaxation
269
tE = 4.2 ms to an image acquired with tE = 3.4 ms. In the latter image the interface appears
abrupt and well defined.
The visual impression of the interface is similar to that observed in Hahn-echo images
(Figure 7.17a). In order to obtain further details about the interfacial region, 1D profiles
were acquired with single-point imaging and different contrast filters (Figure 7.17b) [65,
66]. If the acquisition delay inherent to single-point imaging is chosen appropriately
(0.35 ms), the signal decay is affected by the weak but nevertheless effective chemical-shift
modulation, so that the image contrast is dominated by chemical shift effects (right scale,
smooth line). This image shows a rapid transition between both components and a narrow
Figure 7.17 Interface between covulcanized sheets from unfilled SBR and NR: (a)
Photograph, (b) 1D NMR images acquired with single-point imaging and different
contrast filters. Smooth line: chemical contrast (right scale). Rugged line: relaxation
contrast (left scale)
270
chemical interface with a width of less than 0.1 mm. A 180° pulse in the middle of the
acquisition delay eliminates the chemical shift modulation and the signal decay is defined
only by transverse relaxation. The contrast in the quotient profile corresponding to an
image acquired with an intermediate echo time (acquisition delay) normalized to an image
acquired with a short echo time (rugged line, left scale) provides mainly relaxation contrast
and thus information about molecular mobility or sample hardness. This image of the
physical interface shows a broad transition from the hard SBR component to the soft NR
component with a width of the order of 0.5 mm. This is the space scale on which the
modulus changes. More complicated shapes of the interface can also be observed. The
shape and dimension of the interface are defined by the concentration differences in the
vulcanizing agents, which diffuse at elevated temperatures across the interface until their
diffusion is hampered by their role in the vulcanisation reaction. Thus the interface arises
from a delicate balance of diffusion, reaction, heat supply, and removal.
7.2.3 Blending
Interfaces similar to those encountered during covulcanisation may arise in blends with
incomplete mixing. In unvulcanized samples heterogeneities from blending can be detected
by gradient echo imaging, where the contrast is enhanced by differences in magnetic
susceptibility of the components (cf. Figure 7.5). However, fine structures are usually
homogenized during the vulcanisation process. Larger structures survive, in particular,
when one component had already been vulcanized. An example of such a case is illustrated
in Figure 7.18 by two orthogonal slices through a block of a vulcanized blend from a
soft (bright) and a previously vulcanized hard (dark) component. In slice (a) the
structures invoked by the mixing of components appear course but random, but slice
(b) reveals stream lines of material flow in the rolling mill.
7.2.4 Crosslink Density
Cross-link density and parameters relating to the network structure can be measured
by NMR by analysis of the transverse relaxation decay (cf. Section 1.3) and the
longitudinal relaxation in the rotating frame [67]. Combined with spatial resolution,
the model-based analysis of relaxation yields maps of cross-link density and related
parameters [68]. Often the statistical distribution of relaxation parameters over all
pixels provides a reduced data set with sufficient information for sample characterization
and discrimination [68].
Relaxation curves can be measured by simple NMR experiments. On the other hand,
information about cross-link density and chain stiffness can be retrieved by exploiting
271
Figure 7.18 T2 weighted orthogonal slices of a cabon-black filled rubber blend with a
soft (bright) and a hard (dark) component. Insufficient blending in a rolling mill is
visualized by the stream lines in slice (b)
residual dipolar couplings (cf. Section 1.3). Experiments to measure such couplings are
more sophisticated, but better image contrast can be obtained and more specific structural
information such as the dynamic order parameter of individual chemical groups within the
network [35]. Model-free data of cross-link density can be obtained by NMR when
calibrating NMR parameters against values of cross-link density determined by other
methods. However, NMR relaxation and residual dipolar couplings are sensitive not only
to cross-link density but also to restrictions in chain motion in general, so that effects from
temperature variation, strain, and cross-link density need to be separated based on a priori
information. Most important, comparative measurements on different samples near room
temperature need to be carried out at the same temperature or extrapolated to a reference
temperature based on previously determined temperature coefficients.
Conventionally, cross-link density is determined by measurements of the modulus, the

glass transition temperature Tg, and by solvent uptake in swelling experiments. In
these procedures, the chemical cross-link density cannot be discriminated from network-
272
filler interactions. If NMR experiments are performed over 90°C above Tg, physical
interactions imposed by chain entanglements exert no influence on transverse NMR
relaxation. Moreover, the presence of carbon-black filler usually does not affect the
NMR data in contrast to functionalized silicate filler. This is attributed to rapid signal
relaxation of network chain segments near carbon-black filler particles caused by
paramagnetic centers in the filler. Thus for the technologically important class of carbon-
black filled elastomers, NMR can provide the chemical cross-link density.
7.2.5 Vulcanisation Process
Variations in cross-link density may arise from spatial variations in the rubber formulation,
although short-scale variations are often smoothed by component diffusion during the
vulcanisation process. Differences on the mm scale can lead to interfacial structures like
those depicted in Figures 7.16 and 7.17. Another source of variations in cross-link density
on the mm scale is the curing process in combination with the sample geometry. Heat is
supplied to the sample for a certain time and after vulcanisation is removed from the
sample in a certain time. Near the heat source vulcanisation sets in first, and near the
heat sink it sets in last. Depending on how the heat is supplied to and withdrawn from
the object, complicated time-dependent temperature profiles are established in the sample.
The vulcanisation process has been followed across a simple 1.4 mm thick disc from
unfilled SBR by NMR imaging of the transverse signal decay [44]. The measurements
were carried out in a magnetic field slightly inhomogeneous from the construction of a
special vulcanisation probe, and the transverse relaxation time T2* in the inhomogeneous
magnetic field was measured. Measurement of T2 would have been too time-consuming
to follow the vulcanisation process. Despite accelerated signal decay from the field
inhomogeneities, T2* turned out to still be sensitive to chain mobility because of a
sufficiently strong contribution from T2. A somewhat nonlinear relationship between
the transverse relaxation rate 1/T2* and the reduced shear modulus was found.
Nevertheless, the relationship that low modulus corresponds to high T2* and high modulus
to low T2* can be used to interpret the imaging data across the SBR sheet measured
during the vulcanisation process (Figure 7.19). Immediately after sample heating (5 min.)
chain mobility is high and so is T2*. With the formation of cross links chain mobility
decreases and so does T2*. The vulcanisation front is defined by the center of change in
T2*. It can be seen to migrate through the thin sample on a time scale of about half an
hour with temperatures in the range between 140° and 170° C across the sample. With
increasing time the slope of the T2* curve flattens, indicating a broadening of the reactive
vulcanisation zone.
273
Figure 7.19 Time-resolved T2* parameter images across a 1.4 mm thick sheet of SBR
following the vulcanisation process
7.2.6 Ageing
Ageing of elastomers is a process which affects the mobility of intercross-link chains by

packing, chain scission, and formation of new cross-links. It can therefore be studied by
the same methods as cross-link density. Most conveniently ageing is investigated by
relaxation methods. Thermo-oxydative ageing of many elastomers including SBR and
NR leads to the formation of a brittle surface layer with reduced mobility which appears
dark in Hahn-Echo images. The aged region asymptotically approaches a thickness of
typically 0.3 mm with time [69]. Often this region is followed by a zone of brighter
image intensity, which can be attributed to chain scission or the accumulation of low
molecular-weight additives.
The time evolution of the thermo-oxydative ageing in SBR has been studied by NMR
imaging [52]. In order to detect the aged surface layer with reduced signal intensity in
Hahn-echo images, two aged sheets were stacked, and the stack was imaged along its
axis. In this way the signal from the soft core delineated the hardened surface layers
(Figure 7.20a). The signal loss from the hardened surface regions was specified in terms
of the area A1 in the image which appeared dark relative to the total area A1 + A2 of the
object in the image. This quantity defines the ageing parameter α. It was found to follow
a simple exponential law with time for a sample with and a sample without ageing
274
Figure 7.20 Analysis of thermal oxydative ageing by T2-weighted imaging (a) Two SBR
layers stacked after ageing. The aged layer suffers a signal loss marked by the area A1.
The relative signal decrease defines the ageing parameter a. The progress in ageing can
be followed by evaluating the ageing parameter as a function of ageing time
protectant (Figure 7.20b). For both samples the time constant is the same, so that the
ageing reaction proceeds with the same speed. However, the amplitudes are different, so
that the number of chain segments affected by the ageing reaction is different. Less
segments are exposed to the oxydative ageing attack in the sample with the ageing
protectant, confirming the efficiency of the ageing protectant as a radical scavenger.
7.2.7 Sample Deformation
Sample deformations modify the number of accessible conformations of intercross-link

chains (cf. Figure 7.7), so that they can be detected by analysis of relaxation and residual
dipolar couplings. This is illustrated for strained rubber bands with a cut in Figures 7.6
and 7.11.
Dynamic mechanical load on elastomer products is often exerted at small deformations

and low deformation rates but over extended time periods. Then part of the mechanical
energy is dissipated into heat depending on the value of the loss modulus. As a
consequence, a temperature profile is established within the sample. Then the modulus
275
varies across the sample depending on the temperature profile, and properties determined
for thick samples under dynamic load are averaged quantities.
The temperature profile associated with sample heating during weak dynamic shear
deformation of carbon-black filled SBR cylinders of 10 mm in diameter and 10 mm in
height has been imaged by NMR by use of a specially designed probe [70]. The transverse
relaxation time strongly depends on temperature (Figure 7.21a), so that temperature can
be mapped by parameter imaging of T2. Axial parameter projections have been acquired
in dynamic equilibrium at a shear rate of 10 Hz and a pixel resolution of 0.4 x 0.4 mm2
for carbon-black contents ranging from 10 to 70 phr. One-dimensional cross-sections
through those projections are depicted in Figure 7.21b. An increase of the temperature
in the center of the sample is observed with increasing carbon-black contents which
scales with the increasing loss modulus of the samples. The 70 phr sample is warmer by
over 10° C in the center of the sample than the 10 phr sample.
7.3 Spatially Resolved NMR

In a factory environment NMR instrumentation needs to be robust and possibly mobile
for quality and process control at different stages of product development, fabrication,
Figure 7.21 Temperature profiles from T2 parameter images of SBR cylinders with
different carbon-black filler contents undergoing oscillatory shear deformation: (a)
Temperature calibration curves, (b) Temperature profiles across the cylinders
276
and quality control. These demands are difficult to fulfill with sophisticated pulse
sequences and highly homogeneous magnetic fields B0. For this reason low-resolution
NMR is well established in industrial laboratories, for example in the food, the cosmetics,
and the polymer industry [71]. In low-resolution NMR signal amplitudes, relaxation
times, and diffusion constants are measured. These parameters are most important in
definition of contrast in imaging of elastomers, but can be measured also in inhomogeneous
magnetic fields. Therefore, permanent magnets can be employed at lower field strength,
where transverse relaxation is even more sensitive to slow molecular motion than at high
field. Low-resolution NMR in inhomogeneous magnetic fields has been pioneered in the
oil industry for well logging by single-sided inspection of rock formations [72, 73]. The
same principles of single-sided NMR [74] are applied by the NMR-MOUSE, a mobile
universal surface explorer which weighs one to three kilograms with which NMR
relaxation parameters can be acquired non-destructively from surface-near volume
elements of arbitrarily large objects [75 – 80]. The NMR-MOUSE provides interesting
applications for quality assessment of elastomer products. Because of the simplicity of
the device and the pulse sequences, it is suitable for use in a factory and can be transported
to the object of investigation for spatially resolved NMR of accessible sample regions.
7.3.1 The NMR-MOUSE
The NMR-MOUSE is a portable NMR sensor which works in highly inhomogeneous

magnetic fields. Because of field inhomogeneity NMR spectroscopy of the chemical shift
is not readily possible, but relaxation times and parameters of translational motion can
be measured by echo techniques. These are the most important NMR parameters which
are exploited for contrast in imaging. Unless fluids are investigated field inhomogeneities
are essentially no obstacle for relaxation analysis [80], because molecular motion by
translational diffusion is absent.
Basically the NMR-MOUSE consists of an u-shaped magnet with an rf coil in the magnet
gap (Figure 7.22a). In a portion of the accessible space above the gap, the field lines of
the polarizing magnetic field B0 and the rf field B1 possess orthogonal components. In
this region of space the NMR effect can be invoked. The device can be built rather small
(Figure 7.22b) and the field profile is not severely affected by magnetic components in
the vicinity of the device. For example, NMR measurements can be executed on intact
tyres with steel belts and on conveyor belts with steel cords (cf. Figure 7.25). Even the
laboratory weathering of a 0.5 mm thick poly(vinyl chloride) coating on a 1.5 mm thick
sheet of iron could be followed with the NMR-MOUSE [76]. The polarizing magnetic
field possesses field gradients of the order of 10 to 50 T/m [78]. In such large field
gradients, even a short rectangular pulse is a selective pulse and no free induction decay
signal can be observed with a minimum instrument deadtime of about 10 μs after an rf
277
Figure 7.22 The NMR-MOUSE: (a) Schematic. The NMR sensor consists of an u-shaped
permanent magnet with a solenoidal rf coil placed in the gap. (b) Photo of the NMR-
MOUSE testing a tyre. (c) Example of a train of successive Hahn echoes generated
according to Carr, Purcell, Meiboom and Gill (CPMG echo train) for carbon-black filled
SBR measured by the NMR-MOUSE. The time constant of the echo-envelope defines 2 T
pulse. Therefore echo techniques need to be employed, which eliminate the phase
differences from signal precession in different magnetic field strengths. Such echo
techniques are the Hahn echo and its multi-pulse version, the CPMG sequence. Other
techniques are the solid echo and its multi-pulse version, the OW4 sequence [80] (and
references therein). Figure 7.22c gives an example of a CPMG echo train from carbon-
black filled SBR. In theory, the echo maxima follow a decay, which has been discussed in
Section 1.3. In practice, the particular pulse sequence has an influence on the shape of
the echo decay function, because the signals are acquired in highly inhomogeneous
278
magnetic fields [81], and relaxation curves measured by the NMR-MOUSE are often
evaluated in terms of a mono- or biexponential decay. A mono-exponential decay function
is schematically indicated in Figure 7.22c. The time constant of this decay is the effective
transverse relaxation time T2. It can be calibrated against material properties.
7.3.2 Applications
Some selected applications of the NMR-MOUSE to elastomers are summarized in

Figure 7.23. In (a) T2 has been measured for a series of carbon-black filled NR samples
Figure 7.23 Applications of the NMR-MOUSE to elastomer materials: (a) T2 values

for a curing series of carbon-black filled NR. Comparison of values obtained at high
homogeneous field (DMX 300) and with the NMR-MOUSE, (b) cross-link series of
unfilled SBR with different sulfur content, (c) T2 versus glass-transition temperature Tg
of unfilled SBR by the CPMG and the steady-state CPMG methods, (d) normalized
Hahn-echo decay curves for poly(butadiene) latex samples. Different decay rates are
obtained for small medium and large cross-link densities
279
with different curing times. Normalized values measured at high and homogeneous
magnetic field (7 T, DMX 300 NMR Spectrometer) are compared to those measured
with the NMR-MOUSE (0.5 T). Although the values differ because of different B0 field
strengths, they closely follow the same trend. This confirms, that relaxation measurements
by the NMR-MOUSE are a valid alternative to relaxation measurements at homogeneous
magnetic fields.
In (b) relaxation times are shown for a series unfilled SBR samples with variations in
sulfur content. With increasing sulfur content the chemical cross-link density increases.
Relaxation times are given for measurements with Hahn echoes and with solid echoes.
The change in relaxation times and thus the sensitivity of the method is greatest for the
Hahn-echo measurements at small cross-link density and for the solid-echo measurements
at high cross-link density. Solid echoes reduce the signal attenuation from the dipole-
dipole interaction between two spins in addition to producing a Hahn echo. At high
cross-link density multi-center dipolar couplings become effective, and the solid echo
becomes less effective in reducing signal attenuation from dipolar interactions. The Hahn
echo does not affect the dipole-dipole interaction at all. This explains the difference in
contrast obtained with both methods.
A reduction in measurement time is gained, when the measurements are performed in

dynamic equilibrium between rf excitation and longitudinal relaxation [13]. NMR-
methods which operate in this regime are referred to as steady-state methods. They deliver
equivalent information compared to methods which demand complete relaxation to
thermo-dynamic equilibrium between scans [13]. T2 measurements performed with the
CPMG and the steady-state CPMG method are depicted in Figure 7.23 (c) and correlated
with the glass transition temperature Tg of unfilled SBR laboratory samples of different
cross-link density. The NMR measurements were done at room temperature, whereas Tg
had been determined by temperature-dependent measurements of the dynamic-mechanical
loss modulus [75,76]. The correlation between both measurements is not surprising,
because both probe the network dynamics. These data demonstrate, that in certain cases,
Tg can be determined locally and nondestructively on large samples at room temperature
by the NMR-MOUSE.
The large magnetic-field gradients give rise to rapid signal loss from molecules with
translational motion, so that signals from low molecular-weight fluids are suppressed at
echo times tE of the order of 1 ms and more. The signal detected at larger echo times is
from larger molecules or particles. This effect of solvent suppression is exploited in the
characterization of cross-link density in poly(butadiene) latex samples (see Figure 7.23 (d)).
The signal decay for weak cross-link density is slow compared to that for large cross-link
density. In summary, Figure 7.23 demonstrates, that the NMR-MOUSE is a suitable device
for determination of relative cross-link density in a number of different soft materials.
280
In elastomer samples with macroscopic segmental orientation, the residual dipolar

couplings are oriented as well, so that also the transverse relaxation decay depends on
orientation. Therefore, the relaxation rate 1/T2 of a strained rubber band exhibits an
orientation dependence, which is characteristic of the orientational distribution function
of the residual dipolar interactions in the network. For perfect order the orientation
dependence is determined by the square of the second Legendre polynomial [14]. Nearly
perfect molecular order has been observed in porcine tendon by the orientation dependence
of 1/T2 [77]. It can be concluded, that the NMR-MOUSE appears suitable to discriminate
effects of macroscopic molecular order from effects of temperature and cross-link density
by the orientation dependence of T2.
The NMR-MOUSE is simple mobile NMR sensor suitable for operation in an industrial
environment. Its use for quality control is demonstrated in Figure 7.24 by the T2 statistics
measured at both sides of a conveyor belt with steel cords. The B0 distortions from the
steel cords were minimized by suitable orientation of the NMR-MOUSE with respect to
the direction of the cords. The upper side of the belt exhibits a higher average T2 value
than the lower side, which indicates lower cross-link density. In addition to that, the
variance of the measured values is larger for the lower side, indicating a better quality
material on the upper side. Depending on the size of the rf coil and the size of the sensitive
volume, typical measurement times at room temperature range from ten seconds to some
ten minutes. Following suitable instrument optimization and adaptation to specific
processes, the NMR-MOUSE promises to be of use also in process control.
Figure 7.24 Quality control of a conveyor-belt section with steel cords: (a) Position of
measurement points. (b) T2 values for upper and lower sides determined with the
CPMG sequence
281
7.3.3 Imaging with the NMR-MOUSE
Images can be obtained with the NMR-MOUSE by measuring voxels individually by lateral
displacement of the device or a change of excitation frequency to shift the sensitive volume
in depth (cf. Figure 7.15). On the other hand, also pulsed magnetic field gradients can be
employed for phase encoding of the space information. Frequency encoding is hampered
by the nonlinear magnetic field profile.
In direction parallel to the magnet gap, the magnetic field gradient is weakest, and a
sensitive volume of a centimeter and more in width can be excited by a short rf pulse.
Solenoidal gradient coils can readily be incorporated into the gap and pulsed to produce
antiparallel magnetic fields (Figure 7.25b). Then a gradient field is established along the
gap, and spatial resolution can be introduced into the measurement by phase-encoding
techniques similar to single-point imaging. A suitable pulse sequence is shown in
Figure 7.25a. The effects of the background field inhomogeneity Bz – B0 and of chemical
shift on the magnetisation phase are balanced in the peak of the Hahn echo, and the only
phase evolution from the pulsed field gradient remains. By applying gradient pulses with
different amplitude, k space can be scanned in the direction along the gap, so that Fourier
transformation of the acquired signal produces a 1D image [66]. This measurement
protocol has been applied in a study of a rubber sheet with parallel textile fibers
(Figure 7.25c). The fibers do not contribute to the detected signal, so that their positions
can be located by the dips in the 1D profile.
Clearly, the basic imaging scheme can be extended to include relaxation-time contrast
for discrimination of variations in cross-link density and strain, and the 1D MRI-MOUSE
(magnetic resonance imaging MOUSE) can be extended with further gradient coils to
permit imaging in three dimensions. Numerous applications of the MRI-MOUSE can be
envisioned in soft matter analysis, in particular in those areas, where imaging with
conventional equipment has proven to be successful, and where smaller, less expensive,
and mobile devices are in need.
7.4 Summary
NMR imaging has been tested in a multitude of cases and many simple and sophisticated
methods have been developed. The most successful applications of the method, however,
are in the field of soft matter, because transverse relaxation times are compatible with
current hardware technology. In the non-medical field, NMR imaging of elastomers shows
great promise for genuine applications, and in addition to academic laboratories, the
method is used in industry for product development and control. While academic efforts
often strive towards the development more sophisticated measurement schemes and
282
Figure 7.25 One-dimensional imaging with the NMR-MOUSE: (a) Single-point

imaging sequence for phase-encoding of space, (b) drawing of the NMR-MOUSE
with coils for pulsed field gradients, (c) sample of an elastomer sheet with parallel
textile fibres and one-dimensional NMR image with the space direction
perpendicular to the fiber direction
283
hardware tools, the demands for routine industrial use are simplicity, precision, and
economy. In all three demands, conventional NMR usually does not score high: Most
modern NMR methods require many precise pulses and highly homogeneous magnetic
fields. The precision of the method is often limited by inherently low sensitivity of NMR.
Economy often sets preferences against NMR, because the method requires expensive
equipment and personnel. Nevertheless, for elastomers simple methods like Hahn-echoes
and gradient echoes are useful for materials characterization and imaging. Homogeneous
magnetic fields are not a prerequisite for imaging and relaxation measurements, and
inexpensive devices like mobile low-field spectrometers, the NMR-MOUSE, and mobile
imagers can be built for use near or in the production line and for operation by technicians.
For these reasons, NMR of elastomers including NMR imaging of elastomers appears to
address one of the industrially most relevant applications to NMR. Further developments
of NMR instruments and the growing understanding of NMR in inhomogeneous fields
are expected to be of beneficial impact for increased use of the method in engineering
and fabrication.
Acknowledgements
Continuous support of this work by Deutsche Forschungsgemeinschaft (DFG) and Fonds

der Chemischen Industrie (FCI) is gratefully acknowledged.
References
1. E.M. Haacke, R.W. Brown, M.R. Thompson and R. Venkatesan, Magnetic

Resonance Imaging in Physical Principles and Sequence Design, Wiley, New
York, 1999.
2. M.T. Vlaardingerbroek and J.A. den Boer, Magnetic Resonance Imaging, Theory
and Practice, Second Edition, Springer, Berlin, 1999.
3. P.T. Callaghan, Principles of Nuclear Magnetic Resonance Microscopy,

Clarendon Press, Oxford, 1991.
4. R. Kimmich, NMR Tomography, Diffusometry, Relaxometry, Springer, Berlin,

1997.
5. B. Blümich, NMR Imaging of Materials, Oxford University Press, Oxford, 2000.
6. Magnetic Resonance Microscopy, Eds., B. Blümich and W. Kuhn, VCH,

Weinheim, 1992.
284
7. Spatially Resolved Magnetic Resonance, Eds., P. Blümler, B. Blümich, R. Botto

and E. Fukushima, Wiley-VCH, Weinheim, 1998.
8. C. Chang and R.A. Komoroski in Solid State NMR of Polymers, Ed., L.J.
Mathias, Plenum Press, New York, NY, USA, 1991, Chapter 23.
9. N.S. Subhendra and R.A. Komoroski, Macromolecules, 1992, 25, 1420.
10. C. Chang and R.A. Komoroski, Macromolecules, 1992, 25, 600.
11. P. Blümler and B, Blümich, Rubber Chemistry and Technology, 1997, 70, 468.
12. B. Blümich, P. Blümler, A. Guthausen, R. Haken, U. Schmitz, K. Saito and

G. Zimmer, Magnetic Resonance Imaging, 1998, 16, 479.
13. R.R. Ernst, G. Bodenhausen and A. Wokaun, Principles of Nuclear Magnetic

Resonance in One and Two Dimensions, Clarendon Press, Oxford, 1987.
14. K. Schmidt-Rohr and H.W. Spiess, Multidimensional Solid-State NMR and

Polymers, Academic Press, London, 1994.
15. A. Caprihan and E. Fukushima, Physics Reports, 1990, 4, 195.
16. E. Fukushima, Annual Reviews of Fluid Mechanics, 1999, 31, 95.
17. P.T. Callaghan, Reports on Progress in Physics, 1999, 62, 599.
18. P. Blümler, V. Litvinov, H.G. Dikland and M. van Duin, Kautschuk und Gummi
19. B. Blümich, Concepts in Magnetic Resonance, 1998, 10, 19.
22. P. Blümler and B. Blümich, Acta Polymerica, 1993, 44, 125.
23. J-P. Cohen Addad, Progress in NMR Spectroscopy, 1994, 25, 1.
24. V.D. Fedotov and H. Schneider, Structure and Dynamics of Bulk Polymers by
NMR Methods, Volume 21, Springer, Berlin, 1989.
25. P. Sotta, C. Fülber, D.E. Demco, B. Blümich and H.W. Spiess, Macromolecules,
1996, 29, 6222.
285
26. J.J. Gotlib, M.J. Lifshitz, V.A. Shevelev, I.C. Lishansky and I.V. Balanina,
Vysokonolekulyarnye Soedineniya Seriya A, 1976, 18, 2299.
27. V.D. Fedotov, V.M. Thernov and T.N. Khasanovitsch, Vysokonolekulyarnye

Soedineniya Seriya A, 1978, 20, 919.
28. W. Gronski, U. Hoffman, G. Simon, A. Wutzel and E.R. Straube, Rubber

29. G. Simon, K. Baumann and W. Gronski, Macromolecules, 1992, 25, 3624.
30. W. Kuhn, P. Barth, S. Hafner, G. Simon and H. Schneider, Macromolecules, 1994,

27, 5773.
31. S. Hafner and P. Barth, Magnetic Resonance Imaging, 1995, 13, 739.
32. M. Knörgen, and U. Heuert in Spatially Resolved Magnetic Resonance, Eds., P.

Blümler, B. Blümich, R. Botto and E. Fukushima, Wiley-VCH, Weinheim, 1998, 211.
33. C. Fülber, NMR-Relaxation und Bildgebung an Kautschuknetzwerken,

Akademischer Verlag, München, 1996.
34. M. Schneider, L. Gasper, D.E. Demco and B. Blümich, Journal of Chemical

Physics, 1999, 111, 402.
35. M. Schneider, D.E. Demco and B. Blümich, Journal of Magnetic Resonance,

1999, 140, 432.
36. R. Giesen, C. Chwatinski, D.E. Demco, B. Blümich, A. Branderburg, U. Nietta,

B. Pfleiderer and A. Birkefeld, Proceedings of the 5th International Conference on
Magnetic Resonance Microscopy, Heidelberg, Germany, 1999, p.25.
37. M. Klinkenberg, P. Blümler and B. Blümich, Journal of Magnetic Resonance,

1996, A 119, 197.
38. M. Klinkenberg, P. Blümler and B. Blümich, Macromolecules, 1997, 30, 1038.
39. L. Gasper, M. Schneider, D.E. Demco and B. Blümich, Proceedings of the 5th
International Conference on Magnetic Resonance Microscopy, Heidelberg,
Germany, 1999, p.59.
40. M. Klinkenberg in Bildgebende 2H-NMR an Gedehnten Elastomeren, Shaker

Verlag, Aachen, 1996.
286
41. P.J. Basser, J. Mattiello and D. LeBihan, Biophysical Journal, 1994, 66, 259.
42. C. Pierpaoli, P. Jezzard, P.J. Basser, A. Barnett and G. Di Chiro, Radiology, 1996,
201, 637.
43. M.M. Bahn, Journal of Magnetic Resonance, 1999, 141, 68.
44. C. Fülber, K. Unseld, V. Herrmann, K.H. Jakob and B. Blümich, Colloid and
45. S.R. Smith and J.L. Koenig, Macromolecules, 1991, 24, 3496.
46. B. Klei and J.L. Koenig, Acta Polymerica, 1997, 48, 199.
47. P. Adriansens, A. Pollaris, D. Vanderzande, J. Gelan, J.L. White, A.J. Dias and M.
Kelchtermans, Macromolecules 1999, 32, 4692.
48. A. Spyros, N. Chandrachumar, M. Heidenreich and R. Kimmich,

Macromolecules, 1998, 31, 3021.
49. P. Denner, B. Walker and T. Willing, Macromolecular Symposia, 1997, 119, 339.
50. P. Blümler and B. Blümich, Macromolecules, 1991, 24, 2183.
51. P. Blümler, B. Blümich and H. Dumler, Kautschuk und Gummi Kunstoffe, 1992,
45, 699.
52. C. Fülber, B. Blümich, K. Unseld and V. Herrmann, Kautschuk und Gummi

Kunstoffe, 1995, 48, 254.
53. M. Knörgen, U. Heuert, H. Schneider, P. Barth and W. Kuhn, Polymer Bulletin,

1997, 38, 101.
54. M. Knörgen, U. Heuert, H. Menge and H. Schneider, Die Angewandte

Makromolekulare Chemie, 1998, 261/262, 123.
55. J.A. Chudek and G. Hunter, Journal of Materials Science, Letters, 1992, 11, 222.
56. M. Sardashti, B.A. Baldwin and D.J. O´Donnell, Journal of Polymer Science B:
Polymer Physics, 1995, 33, 571.
57. A. Spyros, R. Kimmich, B.H. Briese and D. Jenddrossek, Macromolecules, 1997,

30, 8218.
287
58. F.P. Miknis, A.T. Pauli, L.C. Michon and D.A. Netzel, Fuel, 1998, 77, 399.
59. F.P. Miknis and L.C. Michon, Fuel, 1998, 77, 393.
60. D. Hauck, P. Blümler and B. Blümich, Macromolecular Chemistry and Physics,

1997, 198, 2729.
61. A. Guthausen, Die NMR-Mouse: Methoden und Anwendungen zur

Charakterisicrung von Polymeren, RWTH-Aachen, 1998. [Ph. D. Thesis]
62. S. Emid and J.H.N. Creyghton, Physica B, 1985, 128, 81.
63. S. Gravina and D.G. Cory, Journal of Magnetic Resonance, 1994, B104, 53.
64. B.J. Balcom, R.P. MacGregor, S.D. Beyea, D.P. Green, R.L. Amstrong and T.W.
Bremner, Journal of Magnetic Resonance, 1996, A123, 131.
65. P. Prado, L. Gasper, G. Fink, B. Blümich, V. Herrmann, K. Unseld, H-B. Fuchs, H.

Möhler and M. Rühl, Macromolecular Materials and Engineering, 2000, 274, 13.
66. P. Prado, L. Gasper, G. Fink and B. Blümich, Applied Magnetic Resonance, 2000,
18, 1.
67. R. Kimmich and G.Z. Voigt, Zeitschrift fur Naturforschung A, 1978, 33, 1294.
68. P. Barth, S. Hafner and P. Denner, Macromolecules, 1996, 29, 1655.
69. P. Blümler and B. Blümich, Macromolecules, 1991, 24, 2183.
70. D. Hauck, P. Blümler and B. Blümich, Macromolecular Chemistry and Physics,

1997, 198, 2729.
71. P. Prado, B. Blümich and B.J. Balcom in Spectroscopy in Process Analysis, Ed.,
J.M. Chalmers, Sheffield Academic Press, Sheffield, 2000, Chapter 8.
72. R.L. Kleinberg in Encyclopedia of NMR, Eds., D.M. Grant and R.K. Harris,
Wiley, New York, NY, USA, 1996, p.4960.
73. R.L. Kleinberg, A. Sezginer, D.D. Grifin and M. Fukuhara, Journal of Magnetic
Resonance, 1992, 97, 466.
74. G.A. Matzkanin in Nondestructive Characterisation of Materials, Eds., P. Höller,

V. Hauk, G. Dobmann, C.O. Ruud and R.E. Green, Springer, Heidelberg, 1989.
288
75. G. Eidmann, R. Savelsberg, P. Blümler and B. Blümich, Journal of Magnetic

Resonance, 1996, A 122, 104.
76. G. Zimmer, A. Guthausen, U. Schmitz, K. Saito and B. Blümich, Advanced

Materials, 1997, 9, 987.
77. B. Blümich, P. Blümler, G. Eidman, A. Guthausen, R. Haken, U. Schmitz, K. Saito

and G. Zimmer, Magnetic Resonance Imaging, 1998, 16, 479.
78. A. Guthausen, G. Zimmer, P. Blümler and B. Blümich, Journal of Magnetic

Resonance, 1998, 130, 1.
79. G. Zimmer, A. Guthhausen and B. Blümich, Solid State Nuclear Magnetic

Resonance, 1998, 12, 183.
80. M. Mehring, Principles of High Resolution NMR Spectroscopy in Solids, 2nd

Edition, Springer, New York, 1983.
81. F. Balibanu, K. Hailu, R. Eymael, D.E. Demco and B. Blümich, Journal of

Magnetic Resonance, 2000, 145, 246.
289
290
8
NMR in Soft Polymeric Matter: Nanometer
Scale Probe
Jean Pierre Cohen Addad
8.1 Introduction
Thinking of plastics all around us, polymers are usually regarded as materials substituted
for wood, glass or metal even though macromolecules also go into soft matter such as
rubber, latexes, contact-lenses, thin films; polymeric substances may be substituted for
leather or may behave like putty. The essential feature about soft matters, generally
referred to as gels, is the existence of networks generated by chain couplings at widely
separated loci in space; consequently, physical inquiry into the behavior of the statistical
structures of polymeric networks must focus both on segmental properties corresponding
to a few nanometer scale in space and on local properties induced by the dynamics of ten
or less monomeric units. It is shown here, that in contrast to usual local nuclear magnetic
resonance (NMR) approaches [1, 2] the proton magnetic relaxation is also a suitable
technique for providing a nanometer scale characterisation of polymers observed above
the glass transition temperature (Tg) [3]. This Chapter is intended to provide a simple
guide for rapidly characterising soft polymeric matters; the characterisation relies on
low resolution proton magnetic relaxation which leads to the determination of standard
parameters unambiguously interpreted.
Considering investigations into properties of soft polymers, the specificity of these physical
systems, compared with ordinary liquids formed from small molecules, originates
essentially from the linkage of chemical units whatever their nature; resulting chains
comprise of thousands of bonds forming a backbone. Significant properties induced by
the effect of linkage are, on the one hand, the curvilinear orientational correlation of
units which determines a persistent length along one chain and gives rise to the stiffness
effect. On the other hand, there is an elastic component which characterises the response
to a stress and which originates from the presence of topological constraints: molecules
having their atoms bonded in concatened sequences exhibit a physical exclusion of volume
which precludes the passage of one chain directly through the backbone of another. The
macromolecules of the liquid do not have time to adjust their positions by viscous
movement over the interval during which a force is applied and the only possible motion
is that of deformation analogous to the behaviour of a solid. There is, in turn, a resistance
291
to shear forces; the higher the frequency of the stress the stronger the resistance. This
effect adds to a natural component of viscosity due to the friction of chemical units, i.e.,
to the diffusion of molecular momentum. Both the atomic connections and the topological
hindrance generate a frequency dependent effect of elasticity in macromolecular liquids;
this effect is enhanced when polymeric networks are formed [4]. Both the elasticity and
the viscosity components can be detected from the magnetic relaxation of nuclei attached
to the polymer.
8.2 Polymeric Networks
A wide variety of network structures are currently observed.
8.2.1 Molten High Polymers
It is known that any high polymer is a non-Newtonian liquid which exhibits unusual
flows: it climbs a rotating rod or it swells when it emerges from a tube [5]. Correspondingly,
any molten high polymer conceals a temporary network which originates from the
existence of an uncrossable backbone contour characterised by a well-defined mean length
[6]. Furthermore, it is now well-established that the dynamic responses of molten polymers
originate from the collective motions of segments along one chain; depending on the
chain length, the time evolution of internal chain fluctuations exhibits characteristic
exponents. The relevant space scale of description of chain dynamics is determined from
the specific backbone contour length which governs the component of elasticity associated
with the presence of a temporary network.
8.2.2 Crosslinked Chains
Statistically defined structures may also arise from the formation of crosslinks in a melt;
the resulting gels are described within a percolation framework which predicts the
existence of definite meshes [7, 8]. Contact-lenses, jellies or even jellyfish are common
examples of gels. Latex beads with specific functionalities attached, such as antigens, are
used in biodiagnostics.
8.2.3 Semi-crystalline Polymers
In semi-crystalline polymers, small ordered domains are connected to one another by

chain segments; thus they are in coexistence with amorphous domains and their random
292
NMR in Soft Polymeric Matter: Nanometer Scale Probe
distribution in space can give rise to three-dimensional networks which originate complex
textures [9] (Figure 8.1). The semi-crystallisation of polymers yields a soft matter which
is currently substituted for leather.
8.2.4 Block Copolymers
Similarly, a polymeric medium characterised by strong cohesion, is also obtained from

di- or tri-block copolymers made by linking two or three chemically homogeneous
sequences which are incompatible with one another; usually, the Tg of one of the two
sequences is above room temperature while it is below for the other sequence [10]. There
is a phase separation; glassy segments are connected to one another by amorphous
segments and they play the role of ordered domains formed in semi-crystalline polymers.
8.2.5 Loaded Polymers
The process of chain adsorption on mineral aggregates immersed in a molten polymer

generates loops and tails on the surface of the filler; a chain structure thus appears. In
Figure 8.1 Topological constraints and defects along the chains hinder the complete
crystallisation of polymers. Chain segments are ordered inside small domains; both the
free enthalpy of bulk crystallisation and the surface energy are involved in the
formation of domains which occur, consequently, at temperatures lower than the
melting temperature
293
addition to the adsorption effect, a process of connection of aggregates, bonded by the

polymer may occur, leading to a percolation effect [11, 12]. Networks resulting from
both the formation of sulfur crosslinks and the adsorption of polybutadiene chains on
carbon black play a crucial role in the viscoelastic behaviour of tires.
8.2.6 Aggregated Polymers
When considering structural aspects of polymeric systems, solutions wherein partial

polymer association occurs, must also be taken into consideration. In concentrated or
semi-dilute solutions, long polymer chains can form networks through the association of
short segments randomly distributed along the chains; the physical association may arise
from charge transfer or from hydrophobic interactions; networks may also result from
the presence of chains which both enter in the formation of small aggregates and connect
them to one another.
8.2.7 Network Distribution Function
Any polymeric matter, observed above the Tg, can be pictured as an ensemble of chain
segments issuing from coupling junctions, formed by entanglements, crosslinks, crystallites
or contact points of chains adsorbed on the surface of mineral fillers. The formation of
such junctions, also called nodes, gives rise to networks that are defined statistically in
space; their structures consist of meshes determined by chain segments embedded in a
viscous medium. The relevant linear space scale of description of the network structures
is not the size of one monomeric unit; it is defined by about the mean distance between
two consecutive nodes: ≈ 100 Å. The need for characterisations of network structures
defined on a nanometer scale clearly arises from the description of physical properties of
polymers observed above the Tg. Nodes are sensitive to macroscopic deformations of
networks; they undergo displacements when a network is strained. The density of nodes
is a crucial factor in determining the responses to macroscopic deformations, which may
result from the effect of swelling induced by a good solvent or from a stress applied to
the polymeric system.
It is considered that the understanding of the behaviour of networks relies on the

knowledge of the distribution of the ensemble of segments which connect nodes; let
< rij >
G( ) denote the probability distribution function in which <rij> is the mean vector
ζ ij
between two consecutive i and j nodes. The correlation length ζij depends on the number
294
of bonds nij in the segment that joins these two nodes; it reflects the statistical framework
applied to the description of the polymer matter [7]. The statistics are specific to the
state of the polymeric medium, which is observed. The existence of a distribution function
cannot be detected directly but its moments are readily determined according to specific
experimental procedures. It is shown, here, that the transverse magnetic relaxation curves
of protons attached to strands can be given a simple analysis that reveals properties
< rij >
specific to the distribution function G( ).
ζ ij
8.3 Basis of the NMR Approach
Any strong steady magnetic field applied to a spin-system induces naturally an axial
symmetry for magnetic properties of nuclei; this axial symmetry leads to well distinct
irreversible behaviours of the longitudinal and the transverse components of the
macroscopic magnetisation, respectively. The longitudinal relaxation implies a quasi-
resonant exchange of energy between the spin-system and the molecular thermal bath:
the frequency window of observation of random molecular motions is necessarily defined
around the Larmor frequency of nuclear spins (≈ 109 rad.s-1). The transverse relaxation
reflects mainly the loss of quantum phase coherence of spins and no displacements of
nuclei are necessarily involved in such a process [13]. The axial symmetry is of particular
interest for observing the broad spectrum of relaxation rates related to the hierarchy of
fluctuations, which affect any polymer chain in a melt (from about 109 down to less than
1 s-1). For protons attached to polymer chains, the irreversible dynamics of the component
parallel to the direction of the magnetic field is sensitive to properties generated by the
local viscosity, which governs the random rotations of monomeric units, in the polymeric
medium. With regard to the transverse magnetisation, the relaxation process cannot be
analysed without considering the time interval allotted for the full random rotations of
chemical units; it is close to the time interval (≈ 1 s or more) required for the full renewal
of the chain configurations. However, this is too long a process for inducing any magnetic
relaxation mechanism; consequently, the transverse component is sensitive to a part of
the hierarchy of chain fluctuations, only. In other words, random motions of units are
detected as non-isotropic rotations; the irreversible dynamics of the magnetisation is
thus governed by the non-zero average of spin-spin interactions that results from the
anisotropy of rotations of skeletal bonds, generated by topological constraints. A solid-
like behaviour of the transverse magnetisation is expected to be associated to the property
of elasticity, provided the time scale of the renewal of configurations is longer than the
NMR scale of observation. This effect is considerably enhanced when permanent networks
are observed.
295
8.3.1 Chain Elongation
The NMR approach relies on the sensitivity of this technique to the effect of reduction of
entropy, which accompanies the formation of networks. Considering first a liquid made
from small molecules, these undergo random rotations and translations, which manifest
large amplitudes. Then, the existence of connections between atoms in a macromolecule
restricts greatly the configurations these atoms may assume although angles of rotations
about bonds of the molecular backbone permit a diversity of configurations; bond
rotations are affected by internal potential energies of interactions between chemical
groups. Thus, the linkage of chemical units forming macromolecules represents the first
step in the entropy reduction. In any polymeric network observed above the T, the existence
of coupling junctions, whatever their exact nature, eliminates translations and prevents
the strands from completely reorientating; this in turn leads to the second step in the
reduction of entropy.
The natural tendency of a chain segment, comprised of nitrogen bonds, is to retract, at a

given temperature; for such a chain segment, the equation of elasticity is derived from
r
the contribution of the configurational entropy to the chain free energy: f = - 3kT
Nb 2
where - f is the force exerted to the chain segment ends for counterbalancing the retractive
force resulting from the small extension r (Figure 8.2).
Figure 8.2 Stretching vector of a chain segment. The schematic illustration shows that
configurational fluctuations pervade a large volume
296
N is the number of skeletal bonds in one chain and k is the Boltzmann constant while b
is the mean skeletal bond. The mean square end-to-end distance, Nb2, is also referred to
as the square of the Gaussian correlation length between the chain ends, σ(N); it reflects
the effect of linkage of chemical units. The previous relationship between the force f and
the extension r is extended to any real chain submitted to a small elongation provided
the correlation length, σ(N), includes the stiffness property of the polymer: σ(N)2 = λKNb2;
λK is referred to as a persistence length. The related reduction of entropy is expressed as:
r2
ΔS = -3k . The longer the stretching vector, r, the higher the reduction of entropy.
2σ 2
Turning our attention to NMR, one chain segment is considered; its supposed fixed ends
prevent the segment from completely reorientating: the freedom of internal bond rotations
is impeded by the hindrance generated by the fixed ends; this constraint gives rise to fast
but non-isotropic random rotations of skeletal bonds which induce, in turn, a non-zero
average of the energy of magnetic interactions between protons attached to the segment.
There are residual dipole-dipole interactions, hHR, expressed as:
(3 cos2 θ r − 1)r 2
hH R ≈ hH D Λ
N2 b2
the Hamiltonian, hHD, represents the strength of dipole-dipole interactions established

in the absence of any chain fluctuations; the chain stiffness is accounted for by the
parameter Λ and θr is the angle that the stretching vector, r, makes with the steady magnetic
field direction [3]. This key equation ensures the transfer from local (a few Angströms)
to semi-local (a few nanometers) NMR properties. The longer the stretching vector, the
higher the strength of residual spin-spin interactions. Considering the residual interaction,
hHR, the proton relaxation is thus governed by non-isotropic diffusional rotations of
monomeric units which induce a solid-like behaviour of the transverse magnetisation;
this is therefore sensitive in a specific way to the presence of chain junctions and
consequently to the architecture of the polymeric network. In the frame rotating at the
Larmor frequency of the spins (≈ 109 rad.s-1), the normalised transverse magnetisation
associated to one segment is expressed as:
mx(t) = Tr [exp( i HRt ) Mx exp(-i HRt) Mx )] /Tr [ Mx2)]
Mx is the quantum operator associated with the transverse magnetisation. With regard
to the proton magnetic relaxation, the probe determined by the end-to-end vector r, is
substituted for any chemical unit attached to the chain segment. The observation is thus
delocalised over the space scale defined by the distance r ≈ 5 nm. From the spectroscopy
point of view, HD represents a dispersion of non-coherent broadening frequencies and
297
the rate of fluctuations of this dispersion is that of internal bond rotations (≈ 109 rad.s-1):
it is much higher than the width HD (≈ 105 rad.s-1); consequently, the observed residual
interactions do not result from bond fluctuations which might be too slow, compared
with the NMR scale (≈ 10-5 s) but they arise mainly, from the anisotropy of bond rotations.
8.3.2 NMR Evidence for Networks: Pseudo-solid Spin-echoes
The NMR characterisation of polymeric systems requires first the search for the existence
of networks. The observation of a time reversal effect, specific to residual spin-spin
interactions, gives evidence for the presence of polymeric networks [3]. This property is
reflected by so-called pseudo-solid spin-echoes formed by applying a suitable radio-
frequency pulse sequence that results in a rotation of the spin operators (Figure 8.3).
The partial recovery of the quantum phase coherence of nuclear dipoles originates from
the non-commutative property of the Zeeman energy with the quantum operator which
represents the residual interaction after rotating the spins. This rotation has no effect on
the magnetisation dynamics when the residual interaction, hHR, is equal to zero. No
Figure 8.3 Normalised proton transverse relaxation curve recorded from end-linked
calibrated chains. Eight pseudo-solid spin-echoes are recorded to illustrate the time
reversal effect specific to the presence of any polymeric network
Reproduced with permission from J.P. Cohen Addad and H. Montes, Macromolecules,
1997, 30, 12, 3678, Figure 6A. Copyright 1997, American Chemical Society
298
echoes are observed on ordinary liquids or on short polymer chains. The pulse sequence
applied to the spin system for observing a pseudo-solid spin-echo is determined by:
90°/x - τ/2 - 180°/x - τ/2 - 90°/y - (t-τ)/2 - 180°/y - (t-τ/2).
8.4 Crosslinked Chains
Permanent networks formed from crosslinked chains are of particular interest for
illustrating the NMR approach. In this Section, it is shown how the proton transverse
relaxation observed on networks is characterised in a standard way and provides values
of relevant parameters.
8.4.1 End-linked Calibrated Chains
Calibrated gels are synthesised by using polymer precursors also called telechelic chains
because their ends bear chemical functions; they are characterised by a polydispersity index
close to one. These chains are end-linked by reacting with a suitable chemical reagent. The
gelation process is usually supposed to be near completion; correspondingly, the fraction
of polymer extracted by washing samples is smaller than 2%. Since the length of segments
is constant, the only relevant variable is the vector, r, between two consecutive nodes. The
transverse magnetisation observed over a whole dry gel, is written as:
1 3r 2 3
M xR (t) = ∫ Tr o [exp(iH R t )o M x exp( − iH R t)M x ] exp( − )d r
(2πσ 2 )3 / 2 2σ 2
where the probability distribution function of the vector r is Gaussian. Taking the
expression of HR into consideration, it is clearly seen that the reduced variable used to
carry out the above integral is actually r/σ. Consequently, the effect of the non-zero
average is accounted for by simply dividing HD by N, the number of bonds in any segment.
The description of the effect of linkage on NMR amounts to considering the renormalised
interaction defined by HD/N (≈ 103 rad.s-1). Correspondingly, the timescale of relaxation,
currently observed, is about 1 ms for N = 100.
8.4.2 Characteristic NMR Rates
Experimental transverse relaxation curves cannot be usually described from exponential

time functions but integral treatments of these curves yield standard parameters equivalent
299
to relaxation rates. Without entering into too many details, it may be worth noting that
the integral treatments amount to calculating several moments of the probability
distribution function of the end-to-end vector, r. It is more convenient to give the numerical
estimate of moments of the distribution function than to determine the exact expression
of the relaxation function M0m. The first and the third moments of the distribution
function, G, are obtained from the two following integral treatments of the experimental
relaxation curve:
φ1 = ∫ oM Tx (t)/ t dt
0
and
∞
dM Tx (t)
φ3 = ∫ / t dt
0
dt
with MTx(0) = 1. It is shown that the quantity called φ1 is proportional to the first moment
of the probability distribution function G(ρ)
∞
m1 = ∫ ρG( ρ )dρ
0
while the quantity called φ3 is proportional to the third moment of G(ρ)
∞
m 3 = ∫ ρ 3G( ρ )dρ
0
with ρ = <r>/σ(N). More precisely,
φ1 ≈ Λ−1 / 2 Δ−G1 / 2 < σ ( N ) / b > m1 /m 2 (8.1)
and
φ3 ≈ Λ1 / 2 Δ1 / 2;G < b / σ ( N ) > m 3 /m 2 (8.2)
The quantum average Δ1/2;G characterises spin-spin interactions established in the absence
of any chain fluctuations. The ratio defined by:
300
χ c = φ3 / φ1
is called NMR structural parameter; it is expressed according to Equations 8.1 and 8.2 as:
χc ≈ Λ1/2Δ1/2;G <b/σ(N)>m3/m2 (8.3)

−1 / 2
and Δ*G is given by the ratio ( Δ G / Δ G ). The parameter χc is defined in a standard way
1/ 2
and is like a relaxation rate; 1/χc is proportional to the correlation length of segments
between two consecutive nodes. In contrast to the two- or three-dimensional NMR
approaches, local information is sacrificed, here, in order to focus attention on the space
averages that characterise the network structure considered as a whole. Integrals convert
the time dimension of relaxation curves into simple numbers: it is a zero-dimension
NMR approach.
8.4.3 Strand Length Dependence
8.4.3.1 Thermal Behaviour of Dry Gels
Applying Equation 8.3 to calibrated dry gels, the correlation length between two
consecutive nodes is expressed as σ(N)2 = λKNb2; 1/χc is thus proportional to the number
of skeletal bonds, N, between two crosslinks. The thermal behaviour of the transverse
relaxation exhibits three domains, usually observed on heating gels from the Tg(Mn)
which is a function of the molecular weight of the polymer precursor, Mn. Chain
configurations start taking place over the range Tg to Tg + 40 K but they fluctuate too
slowly to be detected from NMR; the NMR response is thus insensitive to temperature
variations. Then, the progressive onset of chain fluctuations as detected from NMR is
observed over the range Tg + 40 K to Tg + 100 K. On raising the gel temperature, segmental
fluctuations become faster and faster and more and more configurations are detected;
there is a transient behaviour described by the empirical equation
1
= β [T − Tg (M n ) − T0 ][1 + γN ] (8.4)
χc
with β = 9 10 ms.K-1 and γ = 4.2 x 10-2; the quantity 1/χc is expressed in ms, in Equation 8.4.
The last factor in Equation 8.4 is a linear function of the number of skeletal bonds, N,
between two adjacent nodes; it reflects the effect of small elongation of network segments
due to the presence of fixed ends. Finally, for temperatures higher than Tg + 100 K,
configurational fluctuations are fully observed; they are only restricted by the small
301
stretching of chain segments connecting any two nodes throughout the gel. Considering
calibrated polypropylene-oxide gels, the relaxation rate has been shown to be simply
proportional to 1/N:
χc = γ p / N
with γp = 23 ms-1; the molecular weight of segments was varied over the range 200 to
4000 [14].
8.4.3.2 Swelling Effect
The dependence of NMR on the segmental length between crosslinks, is also conveniently
detected from gels swollen by a good solvent. Chain segments obey the excluded volume
statistics and the correlation length is written as σ(N)2 = N1.2 λKb2 [7]. The proton
relaxation rate, χc, is then expressed as a function of the swelling ratio, Qm, as:
χc ≈ Qm-1.5
Qm is the volume of the swollen gel divided by the volume of the dry gel: Qm ≈ N.0.8.
8.4.3.3 Linkage Effect
Considering calibrated gels, the effect of linkage of chemical units on NMR is easily
disclosed when chain ends are frozen by the formation of crosslinks. The linear molecular
weight dependence of χc-1 may serve as a reference for calibrating other gels made from
the same polymer species.
8.4.4 Randomly Crosslinked Chains
Network structures are still determined by nodes and strands when long chains are
crosslinked at random, but the segmental spacing between two consecutive crosslinks,
along one chain, is not uniform in these systems which are currently described within the
framework of bond percolation, considered within the mean field approximation. The
percolation process is supposed to be developed on a Cayley tree [15, 16]. Polymer
chains are considered as percolation units that will be linked to one another to form a
gel. Chains bear chemical functions that can react with functions located on crosslinkers.
The functionality of percolation units is determined by the mean number f of chemical
functions per chain and the gelation (percolation) threshold is given by pc = (f-1)-1. The
302
probability that two chains have been bridged is called p. The variable of gelation
(percolation) is then defined as the deviation from the gelation threshold ε = (p-pc)/pc.
The formation of meshes accompanies the vulcanisation process; their size is a function
of e and the resulting network structure is specific to the state of gelation associated to
this variable. Thus, the weight fraction of synthesised gel, the modulus of elasticity, E
and the swelling ratio, Qm are physical quantities which must depend on the variable of
gelation. The transverse relaxation of protons attached to the chain segments is sensitive
to the relevant space scale determined by the mesh size.
8.4.4.1 NMR-Vulcanisation Relationship
More precisely, the magnetic relaxation depends on the variable of gelation, i.e., the
density of crosslinks, and is closely related to the modulus of elasticity, E, on the one
hand and to the swelling ratio, Qm, on the other hand. Long polybutadiene chains are
currently randomly crosslinked, using sulfur; they can serve to illustrate the NMR
approach to the characterisation of vulcanised polymers. It has been shown that the
2γ sM W
variable of gelation is expressed as: ε = γ s is the sulfur concentration (w/w),
< ms >
<ms> is the mean weight of one sulfur bridge and Mw is the weight average polymer
molecular weight. The probability that two polybutadiene chains have been bridged by
sulfur is p=2γsMm/<ms> while the percolation threshold is defined as pc=Mm/Mw (Mm is
the molar weight of one monomeric unit). The above equation relies on the assumptions
that any monomeric unit can react with sulfur and that the sulfur is entirely involved in
the formation of bridges between chains [17]. The standard relaxation rate, χc, is shown
to be actually a function of ε when the weight average polymer molecular weight, Mw, is
varied over the range 70 x 103 to 180 x 103 while the sulfur concentration, γs, is varied
from 0.002 to 0.01 g/g. Similarly, the NMR-percolation relationship is also well illustrated
from a slightly modified silicone which is a copolymer composed of dimethyl monomeric
units (-O-Si(CH3)2) and of a small number of vinylmethyl units (-O-Si(CH3)(CH=CH2)),
randomly distributed along the chains [18]. One vinyl group can react with a neighbouring
methyl group, in the presence of a catalyst; the reaction occurs at 150 °C under pressure
and yields a link between two chain segments. The kinetics of formation of crosslinks is
observed by quenching polymer films in ice to stop the chemical reaction and to keep the
polymer network in the gelation state reached at a given time t. The number of segments
issuing from one node is 4. Let α(t) denote the fraction of vinyl groups that have reacted
with methyl groups at a time t. The gelation variable ε is thus written as:
ε(t) ≈ α(t)2 Cvi2 - 1,
303
Cvi is the number of vinyl groups per chain. The direct observation of the decrease in
amplitude of the Fourier Transform-Infrared spectroscopy (FT-IR) spectrum of vinyl
groups leads to a square root dependence of α(t) on time. The threshold of gelation
occurs at a time t0 such that ε(t0) is equal to zero. Again, the magnetic relaxation rate is
a function of the variable, ε (Figure 8.4).
Figure 8.4 Typical variation of the standard NMR parameter, χc, as a function of the
variable of percolation e; the different states of gelation were obtained by quenching the
polymer in ice, at different times during the kinetics of synthesis of the network which
occurs at 150 °C. The polymer is a slightly modified silicone chain which bears randomly
distributed vinyl groups as comonomeric units (the mean number of vinyl groups along
one chain is Cvi = 2 x 10-3). Links are created between vinyl and methyl groups; the three
symbols correspond to different catalyst concentrations (redrawn from [18])
8.4.4.2 Elasticity-NMR Relationship
One important feature of randomly crosslinked chains is that they can be highly strained
and/or swollen; they recover their initial shape when the stress is interrupted and/or after
deswelling. Neither the dependence of the modulus of elasticity, E, upon the variable ε
nor that of the swelling ratio, Qm, have been exactly predicted, until now. Nevertheless,
these two physical quantities as well as the rate χc are functions of the variable, ε.
304
Consequently, they are closely related to one another. The NMR-elasticity relationship is
illustrated from the linear dependence of the relaxation rate, χc, on the modulus of elasticity
of the gel, E (≈ 0.5 MPa):
χc = χ0c+ qE
with q ≈ 0.3 ms-1MPa-1, in Figure 8.5, corresponding to polydimethylsiloxane chains

crosslinked at random and already described in Section 4.4.1; the relaxation rate, χ0c ,
measured in the absence of any strain, reflects the elasticity of the highly entangled molten
polymer [19]. It is worth noting that NMR measurements are performed without any
stretching of the polymeric gel. Segmental fluctuations involved in the linear mechanical
response to a small strain are directly observed from the transverse magnetic relaxation,
in the absence of any strain.
Figure 8.5 Linear dependence of the NMR standard parameter, cc, on the modulus of
elasticity E measured on randomly crosslinked polydimethyl siloxane (PDMS) chains. NMR
measurements were performed without any sample deformation (redrawn from [19])
Adapted with permission from J.P. Cohen Addad, B. Phan Thanh and H. Montes,
Macromolecules, 1997, 30, 15, 4374, Figure 2. Copyright 1997, American Chemical
Society
305
8.4.4.3 Swelling Effect
Similarly, considering the swelling effect of vulcanised poly(butadiene), induced by a

good solvent, the relaxation rate has been shown to obey the simple equation:
χc ≈ Qm-2
the dependence upon the swelling ratio, Qm, reveals the Gaussian statistics which applies
to the description of the swollen system. It is contrasted with the Qm-1.5 dependence
which reflects the excluded volume statistics specific to the description of end-linked
calibrated chains swollen by a good solvent.
Considering calibrated gels or randomly crosslinked chains, the effect of linkage of

monomeric units on NMR is well detected in spite of the purely statistical definition of
the segmental spacing between crosslinks in vulcanised chains.
8.4.4.4 Stretched Gels
The anisotropy of random motions of monomeric units can be mechanically induced by

stretching gels or vulcanised chains. The resulting non-zero average of the rank 2 tensor
of space variables concerns the dipole-dipole interaction as well as the quadrupolar one.
The deuterium NMR spectroscopy has been extensively studied; in particular, the
quadrupolar splitting has been quantitatively related to the deformation ratio, λ, resulting
from the uniaxial stretching of a network [20, 22]. The transverse proton magnetisation
is also sensitive to the elongation of polymeric networks even though the interpretation
of experimental results requires a careful analysis of the relaxation curves [23, 17].
8.4.5 Latex Suspensions
The NMR approach applies to suspensions formed from latex beads. The micrometer
particles are subject to Brownian motion from the thermal fluctuations in the solvent;
however, the random rotation of beads is too long a process to induce any magnetic
relaxation process. Consequently, particles can be considered as fixed and the transverse
magnetic relaxation of protons attached to the latex reflects unambiguously network
properties inside polymeric beads. This approach holds as long as the mean size of particles,
D, is such that the rotational correlation time derived according to the Stokes equation
(≈ 6πD3η0/kT, the solvent viscosity, η0 is set equal to 0.1 Ps-s and T = 300 K) is longer
than the inverse of the residual proton-proton interactions: 10-3 s, i.e., D ≥10 nm.
306
8.4.6 Kinetics of Gelation
NMR is a suitable technique for characterising the kinetics of gelation of polymers, in situ.
The sensitivity of the detection is considerably enhanced by observing amorphous segments
rather than crosslinks. For example, the presence of only one crosslink is reflected by about
100 monomeric units located around it because they are submitted to additional topological
constraints generated by the formation of one bridge between two segments [24].
8.5 Polymeric Crystallisation
As one might anticipate, the transverse relaxation is greatly sensitive to the degree of
crystallinity of a polymer. It has proved to be a suitable tool for discriminating the
relaxation of protons embedded in solid domains from the relaxation of protons located
in amorphous parts, in a semi-crystallised polymer. More precisely, in the case of semi-
crystallised polymers, ends of amorphous segments are embedded at the interface of the
nuclei that form and which act like physical crosslinks. Fluctuations that occur along
segments, with fixed ends, are conveniently characterised from NMR and the effect of
confinement, exerted on amorphous segments by ordered domains can be examined. For
a given state of semi-crystallisation of polymers the crystallinity can be correlated to the
properties of proton relaxation in the amorphous state.
8.5.1 Crystallisation-NMR Relationship
A great deal of effort has been devoted to characterising the crystallisation of pure polymers
[25, 26]. Investigations have been extended to random copolymers, using x-ray scattering,
differential scanning calorimetry (DSC) and polarised light microscopy [27, 28]. In this
Section, attention is focused on the NMR characterisation of the state of semi-
crystallisation observed during the progressive annealing of propylene-ethylene
copolymers, quenched at room temperature, from the melt. It has been shown that
statistical copolymers with low ethylene contents crystallise from the melt in different
forms. The development of each form depends on the cooling rate from the melt while
the supermolecular structure is determined by the undercooling temperature of observation
[29]. Ignoring the crystalline forms, the degree of crystallinity, Xc, can be detected, step
by step, observing the relaxation of the transverse magnetisation of protons attached to
the chains. The sharp decay associated with protons located in ordered domains is
contrasted to the long decay corresponding to protons attached to amorphous segments.
The ethylene content in random copolymers, hereafter called C2, varied over the range
0 to 4.7 % (w/w). Let T0m denote the melting temperature of pure polypropylene, the
307
presence of co-monomeric units is expected to induce a depression of the melting

temperature of the copolymer Tm(C2). The empirical temperature dependence of the
degree of crystallinity, as detected from NMR, is found to obey a single curve; this curve
is translated along the temperature axis in accordance with the melting point depression
when the ethylene content varies: Xc (Tm(C2)-T). The degree of crystallinity is primarily
a function of the undercooling, whatever the ethylene content. Considering any state of
semi-crystallisation, a strong correlation between the degree of crystallinity and the
relaxation rate of protons, attached to amorphous segments, is established. It is seen
from Figure 8.6 that the NMR sensitivity to the crystallinity is considerably enhanced
when detected from amorphous properties [30].
8.5.2 Kinetics of Crystallisation
With regard to the crystallisation, polymers stand in sharp contrast to other materials
such as metals that crystallise completely at the melting temperature, Tm; the topological
Figure 8.6 Variations of the superposition factor s with the degree of crystallinity Xc; s
is the factor applied to the time scale to bring all relaxation curves into coincidence
with a given curve chosen as a reference. ■ : pure polypropylene; ethylene content, ● :
4.7 % , ❍ : 3.8 % , ▲ : 3.7 % and Δ : 3.0 %, w/w (redrawn from reference [30])
Reproduced from L. Dujourdy, J.P. Bazile and J.P. Cohen Addad, Polymer International,
1999, 48, 561. Copyright Society of the Chemical Industry. Reproduced with permission.
Permission is granted by John Wiley and Sons on behalf of the SCI
308
hindrance and the presence of chain microstructure defects prevent polymers from com-
pletely crystallising. Consequently, both the free enthalpy of bulk crystallisation per unit
volume - Dhf, and the surface energy, γc - are involved in the formation of small ordered
domains. This process occurs necessarily below the melting temperature and its kinetics
(up to a few months) depend strongly on the undercooling temperature interval, (Tm-T).
The curves of kinetics of isothermal crystallisation observed on most polymers can be
analysed according to the equation proposed by Avrami [31]:
Xc(t) =X ∞ [1- exp(-Ktn)]

c
Xc(t) is the crystallinity ratio measured at time t, while X ∞ is the maximum of the
c
crystallinity ratio measured at the temperature of observation, T; the exponent, n, de-
pends on the mean shape of the ordered domains that are formed parallelepipeds or
platelets. The analysis of the experimental curves of crystallisation kinetics leads to the
determination of the kinetics constant, K which depends on the ratio γcTm/Δhf(Tm-T); the
experimental value of γc is thus derived from K. This ratio is an estimate of the minimum
size that nuclei which form spontaneously must have for starting to grow and giving rise
to the polymer crystallisation. This analysis also applies to the crystallisation of statisti-
cal copolymers which takes place in solutions: the mean segmental spacing between two
consecutive comonomeric units along one chain, required for the formation of ordered
domains, must be longer than the length determined by the ratio γcTm/Δhf(Tm-T).
8.6 Entangled High Polymers
The main feature about molten high polymers (molecular weights higher than about
104) concerns the broadness of the relaxation spectrum that characterises the viscoelastic
response of these systems. This broad two-dispersion spectrum may spread over a range
of relaxation times going from about 10-9 up to several seconds [4]. It is well illustrated
from the modulus of relaxation observed after applying a sudden stress to the polymer;
the resulting sudden deformation of the sample is then kept constant and the applied
stress is released in order to avoid the flow of the polymer. For example, the release of
the constraint σxy(t) is expressed as a function of the shear modulus of relaxation Gxy(t):
∂γ xy
Gxy(t): σxy(t) = Gxy(t) ( t − > 0) dt
∂t
∂γ xy
where ( t − > 0) is the time derivative of the initial shear applied to the sample
∂t
during δt seconds. The time dependence of Gxy(t) is currently considered as reflecting the
309
actual dynamics of internal fluctuations of any chain, in the melt; it consists of three
parts. For short times, t ≤ 10-5 s, Gxy(t) is a decreasing function, independent of the chain
molecular weight and reflects the dispersion of relaxation times associated with segmental
fluctuations. There is then a plateau, which corresponds to the break in the relaxation
spectrum; it is characterised by a modulus of temporary elasticity G0N, about equal to
0.1 MPa. The plateau is involved in the storage of energy during any deformation, a
crucial property for polymeric materials currently pictured as temporary networks; by
analogy with the description of the property of elasticity of permanent gels, the equivalent
mean number Ne of skeletal bonds in strands is expressed as:
Ne = ρRT/G0NMb
where ρ is the polymer density and Mb is the molar weight of one skeletal bond [4].
Finally, the second dispersion of relaxation times is associated with the full renewal of
chain configurations which is achieved during the third part of the time dependence of
Gxy(t); this relaxation, is characterised by a terminal rate TR-1 which depends strongly on
the chain molecular weight, M: TR ≈ M3 [7]. The existence of a two-dispersion spectrum
is an effect resulting from the linkage of monomeric units.
8.6.1 Temporary Networks
It is known that rates of molecular fluctuations are easily detected from the proton
magnetic relaxation only if they are higher than the spin-spin interactions (in rad.s-1)
[13]. As a consequence of this general requirement, any molten high polymer, observed
from NMR, behaves like a gel when the rate of dissociation of the temporary network is
smaller than the residual dipole-dipole interaction of protons attached to the chains [32,
33]. The existence of a break in the relaxation spectrum implies that long-range chain
fluctuations cannot be detected while segmental fluctuations are easily studied. Most
observed transverse magnetic relaxation curves of protons exhibit, well shaped pseudo-
solid echoes which cannot be distinguished from echoes recorded from permanent gels;
their observation gives a clear evidence for the existence of a temporary network as
detected from NMR. In addition to this evidence, the maximum of the spin-lattice
relaxation rate of protons, observed on high polymers, occurs at about 100 K above the
Tg when the Larmor frequency is 109 rad.s-1; the maximum reveals, without any ambiguity,
the presence of random motions of monomeric units characterised by a correlation time
close to 10-9 s. The spin-lattice relaxation rate is usually independent of the chain molecular
weight, except for short chains; in other words, long range effects are screened when
observing local motions in strands determined by the mean number Ne of skeletal bonds.
These motions are non-isotropic and give rise to the residual dipole-dipole interactions,
which are observed in high polymers as well as in permanent gels. The pattern of transverse
310
Figure 8.7A: Semi-logarithmic plot of the proton transverse relaxation functions

recorded from 1,2 polybutadiene, at 348 K. Vinyl contents are: χ1,2 = 0.82 (curve a),
0.66 (curve b), 0.58 (curve c), 0.40 (curve d) and 0.22 (curve e)
8.7B: Time scale shifts are applied to the relaxation functions to illustrate the
property of superposition. Time shift factors are 1.05 (χ1,2 = 0.40), 1.16 (χ1,2 = 0.58),
1.30 (χ1,2 = 0.66), and 1.62 (χ1,2 = 0.82)
relaxation curves recorded from polybutadiene in the melt illustrates the existence of a
temporary network as detected from NMR (Figure 8.7). In this system, both the Tg and
the modulus of temporary elasticity G0N depend strongly on the concentration of
monomeric units in the vinyl 1-2 conformation, along one chain.
311
The pattern can be obtained from the polymer temperature or concentration variations
in addition to the change of G0N. The relaxation function may be too complicated a
mathematical expression ever to be calculated, nonetheless, it obeys a property of
invariance which allows the superposition of all normalised relaxation curves to one
another by adjusting a suitable factor to the time scale of each curve. The time shift
factor is found to obey the equation
Γ
s[T,Tg (φ ) − T0 ] ≈ [T − Tg (φ ) − T0 ][1 + ] (8.5)
G φα
0
N
where Γ is a constant; Tg(φ) is the Tg of the polymer solution and φ is the polymer
concentration [34, 35]. The temperature T0 corresponds to the onset of fluctuations as
detected from NMR; the exponent α is equal to 2.2. Equation 8.5 applies to polyisobutylene
chains in concentrated solutions or to polyethylene-oxide, too. It is worth noting that the
behaviour of the relaxation time scale, represented by Equation 8.5, extends to calibrated
gels when the number N is substituted for the ratio Γ/G0Nφα ; this behaviour actually
concerns protons attached to any chain segment embedded in a viscous medium.
8.6.2 Short Chain Dynamics
The NMR approach is two-fold. Along with the characterisation of the equivalent network
specific to any high polymer, this technique is a suitable tool for studying the dynamics
of short chains (molecular weight < 10,000), in the melt. The principle of the description
of the chain dynamics has been outlined 45 years ago; it is based on the assumption that
information about the short-range behaviour at high frequencies can be sacrificed in
order to reduce considerably the huge number of degrees of freedom, associated with
any chain [36]. Any polymer is represented as a flexible chain of Gaussian submolecules
and long range fluctuations are described as the collective motions of coupled
submolecules; each Gaussian segment is submitted both to the effect of viscosity and to
the property of elasticity which applies between its two ends. Any submolecule is
characterised both by its time dependent end-to-end vector and by the related residual
spin-spin interaction of attached nuclei which is in turn time dependent, too; coupled
equations of all end-to-end vectors lead to the quantitative description of configurational
fluctuations. The terminal chain relaxation time is proportional to the square of the
chain molecular weight. Correspondingly, the slow time dependence of the residual
interaction induces a transverse relaxation process specific to the long-range fluctuations
of the chain. It has been shown that the molecular weight dependence of the proton
relaxation rate is logarithmic [36].
312
8.6.3 Long Chain Dynamics
Considering the hierarchy of long-range fluctuations that occur within long chains, the
Rouse model applies also to the description of the collective motion of submolecules.
During the time scale associated with the Rouse model, long range fluctuations are not
sensitive to the back and forth displacement (reptation) of one chain supposed to move
within a tube representing the lateral hindrance induced by surrounding chains [7]; the
time scale associated with the reptation motion is much longer than the time scale
determined by the Rouse model. This model was used as a convenient numerical tool for
giving the quantitative interpretation of the proton transverse relaxation observed on
polyethylene-oxide as a polymer molecular weight function [38, 39, 40]. It is now well
established that the observed normalised relaxation function MTx(t) can be analysed
using the equation:
M Tx (t) = (M xR (t) + M xEnd (t))Φ R (t)
where ΦR(t) is the contribution to the relaxation induced by fast but non-isotropic rotations
of monomeric units while MxEnd (t) is represented by an exponential function associated
with the relaxation of so-called end-submolecules; finally, MxR(t) is the part of the
relaxation function sensitive to the chain dynamics hindered by entanglements, with:
τs t /τs u1
1n(M (t))/t = − Δ τ s ( ) ∫ du1 ∫ du 2 Γ ( u 2 )
R
x
2
t 0 0
This second order expression accounts for experimental results with:
1
Γ( u ) = [ ∑ exp(− u ) / τp)]2
Ns p
according to this model, one chain is arbitrarily divided into Ns submolecules and the
correlation time τp is defined as:
1 1 πρ
= sin 2 sm 2 (8.6)
τ p τs Ns
where τs is the correlation time associated with one submolecule comprised of n skeletal
bonds:
313
C∞d 2ξ0 n 2
τs = (8.7)
12 kT
Γ(|u|) is actually a function both of the single time constant τs and of the number of
submolecules Ns in one chain; C∞ is the characteristic ratio of one polyethylene oxide
(PEO) chain. The mean square length of one skeletal bond is d2 while ζ0 is the friction
coefficient associated with one skeletal bond. The parameter Δ2 includes both the quantum
and the spacial square averages of the spin-spin interactions resulting from local motions
which occur within any submolecule; its numerical value cannot be exactly calculated
but the order of magnitude of Δ, derived from the mean segmental spacing between
entanglements, is 103 rad.s-1. The very good accuracy of the numerical description of the
proton relaxation is illustrated in Figure 8.8 where the computed expression:
[ln( MxT(t) )]/t
is compared with the experimental curve recorded from polybutadiene observed at 318 K.
Figure 8.8 The computed expression (ln(MxT(t))/t is compared with the curve
calculated from the relaxation curve recorded from molten polybutadiene (1.16 x 105
g/mol) at 318 K; = 1.05 rad.s-1, = 0.42 ms; Ns = 17
314
The estimate of the numerical value of (Δ/Ns) amounts to dividing the dipole-dipole
interaction, |HD| by the total number of skeletal bonds in one chain: N=nNs. This estimate
(≈ 15 rad.s for a PEO chain molecular weight equal to 200 K) is independent of the
partition of the chain into submolecules. The good agreement between the theory and
experimental results provides the numerical value of the correlation time τs associated
with one submolecule; then, knowing τs, the value of the friction coefficient ζ0 associated
with one skeletal bond is obtained considering Equation 8.7. Furthermore, the NMR
approach reveals without any ambiguity the existence of end-submolecules which play a
crucial role in the reptation motion of one long chain in the melt ; the molecular weight
of one end-submolecule is 11000 g/mol and 4850 g/mol for polyethylene-oxide and
polybutadiene, respectively [39, 40]. The approach has been extended to PEO solutions.
8.7 Adsorption on Mineral Aggregates
The accurate description of the mechanisms of reinforcement of elastomers is still a

baffling problem which requires the characterisation of several phenomena involving
mainly the polymer adsorption on the filler surface and the bond percolation resulting
from the random polymeric connection of mineral aggregates. NMR applies to the
determination of the law of chain adsorption on the filler surface by observing the
relaxation of protons attached to monomeric units fixed on aggregates; knowing the
weight of polymer adsorbed per gram of filler, this observation provides the mean number
of contact points of one chain with the surface.
Fumed silica is usually mixed with polydimethylsiloxane; it consists of aggregates with a

fractal structure formed from sticked elementary beads (diameter ≈ 14 nm); the average
length of aggregates is about 200 nm while their average diameter is about 70 nm and
the specific surface area is 150 m2g-1. Silanol groups cover the silica surface (≈ 2 groups
per nm2) and can form hydrogen bonds with oxygen atoms of siloxane chains. The
mechanical mixing of the polymer with silica leads to a process of chain adsorption
involving loops and tails which correspond to a configurational entropy reduction detected
from NMR (Figure 8.9) ; both the initial amount of polymer in a mixture and the amount
of adsorbed polymer on silica are conveniently detected from NMR [41].
The density of adsorbed chains is νc and the density profile of loops comprised of n units,
S(n), has been predicted to vary as:
S(n) = νc/n3/2
(N is the number of chemical units in one chain) [42]; the resulting mean number of
contact points of one chain with the silica surface is N1/2 [11, 12]. The NMR sensitivity
315
Figure 8.9 Semi-logarithmic plot of the proton transverse relaxation curve recorded
from a silica-polydimethylsiloxane mixture. The sharp decrease is associated with the
proton magnetisation of adsorbed monomeric units while the long time dependence
corresponds to loops and trains of linked chains
to the loop density profile is particularly enhanced by swelling the network using a good
solvent; the corresponding transverse magnetic relaxation rate has been shown to vary
as N-1/2. Such a specific dependence results from a purely structural property of adsorption
and no chain diffusion process is involved in the proton relaxation process. A network is
obtained when the relevant percolation variable, N1/2φSi, determined from the silica density,
φSi, and N, has suitable values. The fraction of bound polymer, fB obeys the equation:
fB = βN1/2φSi (1- βN1/2φSi /4 )
where β accounts for the specific surface area of silica AT ;
β = AT Mm1/2/( σeA0 )
σe is the surface area associated with one silanol group, Mm is the molar weight associated
with one skeletal bond and AO is the Avogadro number [11, 43]. Finally, NMR is a
suitable tool for characterising the dynamics of short chains moving through the network
formed by aggregates connected to one another by polymer chains [44].
316
Similar approaches apply to carbon black - polybutadiene mixtures [45, 46]. A thorough
study of ethylene-propylene diene terpolymer (EPDM) - carbon black has been developed
using NMR; a clear evidence for adsorbed EPDM is given in addition to the presence of
polymeric strands connecting aggregates to one another [47].
8.8 Conclusion
Properties of soft polymers, neither solids nor true liquids, have been investigated, for
many years, from their mechanical behaviour. The description of their texture has been
mainly referred to their macroscopic responses to external stresses or strains whether
these were static or time dependent. For several years, light scattering, fluorescence
correlation, and neutron scattering experiments have been used to get a stronger
correlation between the textures of networks and their mechanical behaviour. In this
Chapter, attention was focused essentially on the simple and non-expensive low resolution
NMR approach to the characterisation of polymeric network structures whatever their
physical origin. Investigations are based on the observation of the transverse relaxation
of protons attached to chain segments. In addition to the conventional effect of viscosity
well observed from NMR in any liquid, the transverse relaxation is particularly sensitive
to reductions of configurational entropy induced by topological constraints resulting
from coupling junctions of chains. The effect of linkage of chemical units gives rise to a
specific effect that underlies the NMR approach; it provides us with observations
delocalised over more than 5 nm. Examples, chosen in this survey, highlight this
characteristic property. Nowadays, the quantitative interpretation of properties of the
proton transverse relaxation may lead to a sharp characterisation of complex rubbery
materials, provided calibrated samples associated to each polymer species have been
previously studied for determining standard NMR parameters. Here, the NMR technique
is a suitable tool for observing the mesh size of any network structure including adsorption
on mineral aggregates and the lifetime of the networks when molten polymers are
considered. The deuterium and 13C NMR spectroscopies provide useful additional insights
into the mechanisms of network deformations which are mechanically induced.
References
1. R.R. Ernst, G. Bodenhausen and A. Wokaun, Principles of NMR in One and Two
Dimensions, Clarendon Press, Oxford, UK, 1987.
2. K. Schmidt-Rohr and H.W. Spiess, Multidimensional Solid-state NMR and

Polymers, Academic Press, New York, USA, 1994.
317
3. J.P. Cohen Addad, NMR and Fractal Properties of Polymeric Liquids and Gels,
Progress in NMR Spectroscopy, Eds., J.W. Emsley, J. Feeney and L.H. Sutcliffe,
Pergamon Press, Oxford, England, 1993.
4. J.D. Ferry, Viscoelastic Properties of Polymers, 3rd Edition, John Wiley, New
York, USA, 1980.
5. R.B. Bird and C.F. Curtis, Physics Today, 1984, 37, 26.
6. W.W. Graessley, Advances in Polymer Science, 1974, 16, 1.
7. P.G. De Gennes, Scaling Concepts in Polymer Physics, Cornell University Press,

Ithaca, NY, USA, 1979.
8. D. Stauffer and A. Aharony, Introduction to Percolation Theory, 3rd Edition,

Taylor & Francis, UK, 1994.
9. L. Mandelkern, Crystallisation of Polymers, McGraw-Hill, New York, NY, USA,

1964.
10. F.S. Bates and G.H. Fredrickson, Physics Today, 1999, 52, 32.
11. J.P. Cohen Addad, Polymer, 1992, 33, 2762.
12. J.P. Cohen Addad, Polymer, 1989, 30, 1820.
13. A. Abragam, Principles of Nuclear Magnetism, Oxford University Press, Oxford,

UK, 1961.
14. J.P. Cohen Addad, L. Pellicioli and J.J.H. Nusselder, Polymer Gels and Networks,
1997, 5, 201.
15. W.H. Stockmayer, Journal of Chemical Physics, 1943, 11, 45.
16. P.J. Flory, Principles of Polymer Chemistry, Cornell University Press, Ithaca, NY,
USA, 1969.
17. J.P. Cohen Addad and E. Soyez, Macromolecules, 1992, 25, 6855.
18. J.P. Cohen Addad and H. Montes, Macromolecules, 1997, 30, 3678.
19. J.P. Cohen Addad, B. Phan Thanh and H. Montes, Macromolecules, 1997, 30,
4374.
318
20. B. Deloche and E.T. Samulski, Macromolecules, 1981, 14, 575.
21. P. Sotta and B. Deloche, Macromolecules, 1990, 23, 1999.
22. A.I. Nakatani, M.D. Poliks and E.T. Samulski, Macromolecules, 1990, 23, 2686.
23. P.T. Callaghan and E.T. Samulski, Macromolecules, 1997, 30, 113.
24. J.P. Cohen Addad, E. Soyez and A. Viallat, Macromolecules, 1992, 25, 1259.
25. B. Wunderlich, Macromolecular Physics, Volume 3, Academic Press, New York,

NY, USA, 1980.
26. B. Lotz, J.C. Wittman and A.J. Lovinger, Polymer, 1996, 37, 4979.
27. S. Laihonen, U.W. Gedde, P.E. Werner and J. Martinez-Salazar, Polymer, 1997,
38, 361.
28. H.J. Zimmermann, Journal of Macromolecular Science and Physics, 1993, B32,
141.
29. S. Laihonen, U.W. Gedde, P.E. Werner, M. Westdahl, P. Jääskeläinen and

J. Martinez-Salazar, Polymer, 1997, 38, 361.
30. L. Dujourdy, J.P. Bazile and J.P. Cohen Addad, Polymer International, 1999, 48,
561.
31. M. Avrami, Journal of Chemical Physics, 1941, 9, 177.
32. J.P. Cohen Addad, A. Guillermo and C. Lartigue, Physics Review Letters, 1995,
74, 3820.
1996, 29, 6222.
34. A. Labouriau and J.P. Cohen Addad, Journal of Chemical Physics, 1991, 94,
3242.
35. A. Labouriau and J.P. Cohen Addad, Journal of Chemical Physics, 1991, 94,
3242.
36. P. Rouse, Journal of Chemical Physics, 1953, 21, 1272.
37. M.G. Brereton and M.E. Ries, Macromolecules, 1996, 29, 2644.
319
38. J.P. Cohen Addad and A. Guillermo, Journal of Chemical Physics, 1999, 111,
7131.
39. A. Guillermo, J.P. Cohen Addad and D. Bytchenk, Journal of Chemical Physics,
2000, 113, 5098.
40. J.P. Cohen Addad and A. Guillermo, Physics Review Letters, 2000, 85, 3432.
41. J.P. Cohen Addad and L Dujourdy, Polymer Bulletin, 1998, 41, 253.
42. M. Aubouy, Physics Reviews B, 1997, 56, 3370.
43. J.P. Cohen Addad and N. Morel, Journal Physique, 1996, 6, 267.
44. J.P. Cohen Addad and O. Girard, Macromolecules, 1992, 25, 593.
45. J.P. Cohen Addad and P. Frebourg, Polymer, 1996, 37, 4235.
46. V.J. McBrierty and K.J. Packer, Nuclear Magnetic Resonance in Solid Polymers,
Cambridge University Press, Cambridge, UK, 1993.
47. V.M. Litvinov and P.A.M. Steeman, Macromolecules, 1999, 32, 8476.
320
Chemical Characterisation of Vulcanisates by High-Resolution Solid State NMR
9
Chemical Characterisation of Vulcanisates by
High-Resolution Solid State NMR
Dallas D. Parker and Jack L. Koenig
9.1 Introduction
It is in the vulcanisation process of elastomeric materials that the elastic properties are
generated. The formation of crosslinks in rubbers prevents permanent deformation under
load and ensures elastic recovery on removal of the load. Because the type of crosslinking
process (sulfur, peroxide, radiation, etc.) and the number of crosslinks can radically
change the physical properties of rubber, an understanding of the mechanism and structure
of vulcanisation is extremely important. However, from a molecular characterisation
viewpoint, the introduction of crosslinking limits the ability to characterise those systems
because of the insolubility of the resulting network. In addition, the network structures
formed during vulcanisation are complex and diverse in nature. Additionally, the extremely
low concentration of the chemically modified structures induced by the vulcanisation
approach the detection limits of many traditional analytical techniques. Many approaches
including physical tests and chemical analysis have been attempted in an effort to relate
the chemical microstructure to the physical properties of both raw and cured elastomers.
Nuclear magnetic resonance (NMR) spectroscopy has become one of the most important
analytical techniques used in the characterisation of materials. In the field of
macromolecules, its use extends from monomer characterisation, through polymerisation
kinetics and mechanisms, to direct observation of the chemical structures of polymeric
materials [1]. Using NMR, information obtainable on polymer structures includes main
chain microstructures (conformation, geometric isomerisation, spatial distances, etc.)
comonomer composition and sequence, end and side group analysis, branching and
crosslinking, abnormal structures (cyclic and isomerised structures), bonding, region
enchainment, and tacticity.
One advantage of NMR spectroscopy is that it can be used on almost all polymeric
materials through either solid-state or solution NMR methods. However, the techniques
and resolution of the two methods are radically different. For example, the proton NMR
spectrum of the water is sharp and narrow with a bandwidth of 1 Hz, while the proton
NMR spectrum of ice is extremely broad with a bandwidth of 20 KHz. The differences
321
in the NMR spectra of solids and liquids are due to a motional averaging of interactions
[2, 3]. In liquids and solutions, local interacting fields are averaged to zero by the rapid
isotropic motions of the nuclei (termed incoherent averaging) resulting in narrow
linewidths. Anisotropic interactions, such as dipolar and quadrupolar interactions and
chemical shift anisotropy, are averaged to zero by the molecular motions to effectively
remove them from broadening the resonances in the observed spectra. In solution NMR,
because of the lower frequency motions of polymers, incomplete averaging of anisotropic
effects is observed to a certain extent resulting in some peak broadening. In solids, however,
this effect is highly magnified. Because there is not sufficient motion to average the
anisotropic interactions extremely broad lines are present often encompassing much of
the entire spectrum. Because the incoherent averaging (molecular motion) does not narrow
the NMR lines, coherent-averaging techniques such as dipolar decoupling (DD) and
magic-angle spinning (MAS) must be used to produce narrow line widths [4].
In solids, the major contribution to 13C NMR line broadening is due to heteronuclear
dipole coupling with attached protons. By using broad band irradiation in the proton
field, the effect of C-H dipole decoupling is drastically reduced. In addition, rapid sample
spinning at the ‘magic angle’ of 54°44´, causes the term (3cos2 θ –1) to become zero in
the dipolar Hamiltonian, thus further decreasing heteronuclear interaction. In 1H solid
state NMR, CRAMPS (Combined Rotation And Multiple-Pulse Spectroscopy) techniques
combine MAS and multiple pulses (WAHUHA, MREV-8, BR-24, etc.) to reduce
homonuclear dipolar interactions. The low sensitivity of the 13C nucleus is improved by
transferring polarisation from the magnetisation-rich protons to the 13C nuclei by using
cross-polarisation. By combining the techniques of high-power decoupling (DD), MAS,
and cross-polarisation (CP) experiments into one grand experiment [5], narrow lines
and enhanced sensitivity can be obtained for polymers in the solid state. As a consequence,
high-resolution NMR spectroscopy has become an important tool in the structural
investigation of polymers in the solid state [6, 7, 8, 9].
Recently, solid-state 13C NMR has been widely applied for the characterisation of
vulcanised rubber systems [10] with several reviews published covering the solid-state
NMR analysis of crosslinked elastomers [11, 12].
9.2 Sulfur Vulcanisation Mechanism
Although sulfur vulcanisation was discovered over one hundred and fifty years ago, the
exact mechanism of vulcanisation is still being examined. This arises not only from the
complexity of the reactions and products formed but also to the fact that the mechanism
of accelerated sulfur vulcanisation changes is dependent on the class of accelerators/
activators used. Typically, benzothiazole or sulfenamide are used as accelerators, zinc
322
oxide as the activator, and a fatty acid such as stearic acid used as the coactivator. A
generally accepted sequence of reactions is as follows [13, 14]:
• An interaction of the curatives occurs to form the active sulfurating agent, Ac-Sx- Ac
by a reaction of accelerator (Ac) and activator with sulfur:
Ac + S8 → Ac - Sx – Ac [active sulfurating agent]
• The polymer chains interact with the sulfurating agent to form polysulfidic pendent
groups terminated by accelerator groups:
Ac - Sx - Ac + RH → R - Sx - Ac + AcH [pendent sulfurating agent]
where RH is the rubber chain.
• Polysulfidic crosslinks are formed;
R - Sx - Ac + RH → R - Sx – R + AcH [crosslinks]
• Network maturing and competing side reactions also occur which do not lead to
effective crosslinks. Thermal decomposition leads to the following reactions:
R -Sx- Ac → cyclic sulfides + dienes + ZnS [degradation]
R -Sx- Ac → R - S - Ac + Sx-1 [desulfuration]
R - Sy - R → R - S - R + Sy-1 [monosulfidic crosslinks]
R - Sx+y - R + Ac - Sz - Ac → R - Sx - R + Ac - Sy+z - Ac [sulfur exchange]
During the curing and network maturing periods, there are at least three competing
reactions: crosslinking, desulfuration, and degradation reactions. Desulfuration is a
reduction in sulfur rank of the crosslinked chain while decomposition or degradation
refers to a complete loss of the crosslink with formation of conjugated dienes, cyclic
structures, thiols, etc. The network structures formed depend not only on temperature
but accelerator types and concentration. The ratio of poly-, di-, and monsulfidic crosslinks
strongly depends on the ratio of sulfur to accelerator in the formulation.
Thus, the reactions of diene rubber with sulfur produce a variety of sulfurised structures.
A schematic representation of different types of sulfurised structures in natural rubber
(NR) is shown in Figure 9.1. The type of sulfurisation has been designated as A1, A2,
B1, B2 and C1-types depending on the positions of sulfide attachment as seen in Figure 9.1
[15]. The c and t indicate cis and trans isomers of the structures. The main chain, saturated
323
Figure 9.1 Designation of nomenclature for structures occurring upon vulcanisation of

NR and cis-polyisoprene
and unsaturated structures are also indicated. Crosslink structures may also be classified
as efficient (intermolecular) or inefficient (intramolecular, cyclic, pendent) in nature.
9.3 The NMR Methods for Assigning Resonances to Chemical Structure
One of the earliest studies of vulcanisate structure by solid-state NMR was the sulfur
vulcanisation of NR with samples compounded under different formulations of accelerator
and sulfur [15]. The 13C NMR spectra of NR samples cured with 10% sulfur at different
cure times acquired with MAS and scalar decoupling is shown in Figure 9.2. Line
broadening and decreased resolution are observed with longer cure times. A dominant
new resonance was observed at 58 ppm that is assigned to the polysulfide or vicinal
crosslink structure. The spectral resolution of the solid-state spectra was found to improve
on application of high-power dipolar decoupling along with MAS [16]. With the increased
resolution, new resonances were observed in this experiment upon vulcanisation at 14.0,
16.1, 18.0, 30.0, 40.1, 44.6, 50.2 and 57.5 ppm. The peaks at 40.1 and 16.1 ppm are
generated from the C-1 and C-5 carbons of the trans-isoprene units. The increase in
these two resonances with cure suggests the occurrence of cis-to-trans isomerisation during
the curing process. Intensities of the peaks at 14.0, 44.6 and 57.5 ppm also increased
with cure time.
To aid in the NMR peak assignments of those 13C NMR spectra, the NR model
compounds, 2-methyl-2-pentene, 2-methyl-1-pentene, and 4-methyl-2-pentene, were
324
Figure 9.2 Stack plots of 10% sulphur-cured NR for different times of cure. B:
butadiene-like species, P; polysulfidic crosslinks Carbon numbers in 1,4-isoprene unit
in NR have been designated as follows;-C1-C2(-C5)=C3-C4- or -Cγ-Cα(-Ce)=Cβ-Cδ-
sulfur-vulcanised and studied by solution 13C NMR spectroscopy [16]. In addition to the
13
C NMR observations of the model compounds, the chemical shifts of the carbons for
these structures were calculated. The aliphatic additivity constants for the sulfur groups
were obtained from ‘The Sadtler Guide to Carbon-13 NMR Spectra’ [17] and added to
the carbons of polyisoprene and ethylene-propylene copolymer. Figure 9.3 shows the
calculated 13C chemical shifts of the model structures. These data indicate that the
monosulfidic and disulfidic crosslinks should be distinguishable in the 13C NMR spectra
of sulfur-vulcanised NR. However, it may not be possible to distinguish disulfidic crosslinks
from crosslinks of higher sulfur rank (polysulfide).
325
Figure 9.3 The calculated 13C chemical shifts of the model vulcanised structures
The chemical shifts due to the monosulfidic crosslinks are influenced not only by the
position on the monomeric unit to which it belongs, but also by the position of the
carbon atom of the monomeric unit on the other side of the bridge. The shielding
parameters of monosulfide substitution on the individual carbons of the isoprene unit
have also been determined. It is shown that resolvable polysulfidic crosslink resonances
exist in all positions of the backbone carbons while monosulfidic crosslinks appear only
between C-1 and C-4 carbons with detectable intensity [18].
The complementary application of the NMR inversion recovery measurements (180o-τ-

90o-tR) and the computer fitting of the overlapping regions was found to be a useful
method for structural analysis of vulcanised NR in the solid state [19]. Since the line
widths in 13C NMR spectra of solids are relatively broad compared with the differences
between chemical shifts, some weak signals are completely obscured in the resulting
spectra. If the resonances have sufficiently different relaxation times, such as with methyl
and methylene carbons, it is possible to detect neighbouring overlapped signals by using
the inversion recovery delay τ value at which the interfering strong resonance has null
intensity.
326
9.4 Unaccelerated Sulfur-vulcanisation of NR
There is wide variety of vulcanisation agents and methods available for crosslinking
rubber materials including peroxide, radiation, urethane, amine-boranes, and sulfur
compounds [20]. Because of its superior mechanical and elastic properties, ease in use,
and low cost, sulfur vulcanisation is the most widely used. Although vulcanisation with
sulfur alone is not practical compared to the accelerated sulfur vulcanisation in terms of
the slower cure rate and inferior physical properties of the end products, many
fundamental aspects can be learned from such a simply formulated vulcanisation system.
The use of sulfur alone to cure NR is typically inefficient, i.e., requiring 45~55 sulfur
atoms per crosslink [21], and tends to produce a large portion of intramolecular (cyclic)
crosslinks. However, such ineffective crosslink structures are of interest in the
understanding of complex nature of vulcanisation reactions.
The solid-state 13C NMR investigation of unaccelerated, 10% sulfur vulcanisation at

different cure times was performed [16]. Using analysis at 80 °C and swelling of the
vulcanisate samples to increase chain mobility, major peaks were observed at 57.5, 56.5,
44.6, 43.0, 40.1, 38.6, 36.1, 30.0, 18.0, 16.1, and 14.0 ppm. The peaks at 57.5, 44.6,
40.1, 16.1, and 14 ppm all increase in intensity at approximately the same rate with
increasing cure time while the peak at 50.2 ppm appears at early cure times and disappears
after 30 minutes of cure. The remaining peaks show greater increase with increasing
cure time. All NMR peaks show increased intensities with increasing sulfur contents of
10%, 20%, and 30% by weight. The peak at 50 ppm was assigned to the carbons involved
in the polysulfidic crosslinks of the A1-type structure. The resonance peak, which appears
at roughly 45 ppm in the spectra, was assigned to the polysulfidic crosslinks of B1-type
structure although the monosulfidic A1-type crosslinks may also overlap with this
resonance peak. The peak at 58 ppm is assigned to the crosslink point found in B1-type
polysulfide structure. The splitting of this peak observed in the high temperature
measurement was postulated to arise from differences in carbon configurations or
conformations in the same type of structure. The small peaks found at 30.0, 36.1 and
38.6 and at the shoulder of 43 ppm, which were resolved in the high temperature
measurement, were assigned to cyclic sulfides.
9.5 Accelerated Sulfur-vulcanisation of NR and IR
Accelerated sulfur formulations are the most common vulcanisation systems used in
commercial and industrial applications. Therefore, research on both the fundamental
and applied aspects of accelerated sulfur vulcanisation is ongoing. Several reviews of the
chemistry and/or physics of accelerated sulfur-vulcanisation of elastomers have been
published [13, 14, 22, 23].
327
The structure of NR vulcanised with sulfur contents of 0, 2, and 5 pph and the accelerators
of tetramethyl thiuram disulfide (TMTD), N-cyclohexyl-2 benzothiazole sulfenamide (CBS),
and N-oxydiethylene-2-benzothiazole sulfenamide (MOR) was studied by solid-state 13C
NMR [23]. MOR is categorised as a benzothiazolesulfenamide and is more popularly
called 2-(4-morpholinothio)benzothiazole (MBS). On examination of the spectrum
generated by the sulfur donor TMTD and 0 pph added sulfur, no crosslink peaks were
observed implying that the concentration of monosulfidic crosslinks was lower than the
NMR detection limits. A significant amount of cis-trans isomerisation was noted by the
increase in the trans isomer peaks at 16.1 and 40.1 ppm. On comparison of the TMTD
vulcanised spectrum with unaccelerated sulfur vulcanisation from a previous study [16],
many similarities were noted especially the formation of prominent peaks at 44 and 58 ppm.
However, the accelerator generated spectra shows fewer peaks implying a simpler network
structure with fewer side reactions. With accelerated systems, new resonances were observed
upon vulcanisation at 15, 31, 34, 45, 58 and 130 ppm and at the shoulder of 30 and
41 ppm along with the trans isomers appearing at 17 and 41 ppm.
Model compounds based on 2-methyl-2-pentene were studied to supplement the 13C

chemical shift assignments of the products from accelerated sulfur vulcanisation of NR.
It is observed in the model compound data that it may not be possible to distinguish
between a 13C NMR resonance which is due to disulfidic crosslinks and a peak due to
pendent accelerator groups, while a large chemical shift difference (~3 ppm) is observed
for the monosulfidic bonds. The MBS-accelerated sample shows similar new resonances
as seen in the TMTD accelerated systems. In this comparison however, the quantitative
aspects of the data might be obscured due to the differences in the state of cure among
the different formulations.
The 13C chemical shifts were assigned in more detail for monosulfidic and polysulfidic
crosslinks occurring in the accelerated sulfur vulcanisation of NR [18]. The NR was
cured with a pure thiuram formulation (TMTD alone) in order to predominantly prepare
monosulfidic bridges in the network. The distortionless enhancement by polarisation
transfer (DEPT) experiments, in which the carbons with different level of protonation
can be distinguished [22-24], were performed for the NR cured with extended levels of
sulfur. Based on the DEPT results and previously reported model compound results [20],
the chemical shifts of the resonances occurring in the spectra were assigned.
In the DEPT experiments, both peaks around 50 and 58 ppm are divided into three
components and the levels of the protonation for these six individual resonances are
evaluated. The peaks at 57.4, 58.0 and 58.6 ppm are assigned to the polysulfidic crosslinks
in the A1, B1 and B2-type structures, respectively. The peaks at 37.2 and 50.7 ppm are
due to A1-type polysulfidic crosslinks. There was no apparent structural match for the
quaternary peak at 50.2 ppm.
328
Another approach in the 13C NMR technique has been utilised to assign the peaks for
the sulfurised structures occurring in the accelerated sulfur-vulcanisation of NR [25].
The NR formulations with a variation of accelerator (CBS) to sulfur ratio were cured at
150 °C to t90 (90% of the rheometer torque increase) plus 10 minutes. The Attached
Proton Test (APT) experiment [13, 14] was performed for those vulcanisates, in which
the carbon with even number of hydrogen (methylene or quaternary carbon) and the
carbon with odd number of hydrogens (methine or methyl carbon) can be distinguished.
In addition, the 13C spin-lattice relaxation times (T1) for the individual carbons in both
backbone and sulfurised structures were measured. The carbons with odd number of
hydrogens can be easily distinguished between methine and methyl carbons by observing
the region where the resonance is occurring. The carbons that have even numbers of
hydrogens with long T1 (1s~) are assignable to the quaternary carbons while those with
short T1s are due to the methylene carbons.
The resonances at 57.5 and 50.4 ppm are assigned to the polysulfide and monosulfide of
B2-type structures, respectively. The peaks at 50.8 and 55.6 ppm are due to the polysulfidic
bonds of A1 and B1-type structures, respectively, however the monosulfidic counterparts
for these structures, which fit the APT results, are missing. None of the structures can
explain the splitting of the peaks near 64 ppm. The T1 value of 1.1 seconds of these
coupled peaks suggests that the resonances are due to oxidation products. The trans
structures increase with increasing number of sulfide attachments to the polymer
backbone. The peaks appearing at 12.9 and 14.3 ppm are possibly due to the end groups
of hydrocarbon fragments formed during the mixing and crosslinking steps of sample
preparation.
The crosslink densities of the rubber networks were determined by the quantitative
treatment of the 13C NMR spectra [26]. In this study, the peaks at 37.3, 44.7, 48.2, 49.6,
50.6, 52.5, 54.8 and 57.5 ppm are involved in crosslinks according to the peak
assignments. One-half the sum of these areas should give the crosslink density of the
sample as determined by 13C MAS NMR. The results were compared to the network
chain densities obtained by n-heptane swelling methods to examine the efficiency of the
intermolecular crosslinking reactions for the individual formulation. For the efficient
cure system (EV); higher accelerator concentration relative to sulfur), nearly 100% of
the sulfurisation is monosulfide and is involved in crosslinks. However for conventional
cure systems or systems containing predominantly polysulfidic bonds, about 50% of the
sulfide units are not involved in crosslinking.
Solid-state 13C MAS NMR has been applied for quantitative determination of crosslink
density in accelerated sulfur-vulcanised NR [27]. The concepts used to calculate the
crosslink density by 13C NMR are the same as the one mentioned above, but different
resonances were used for the quantitative treatment based on the different assignments
329
of the 13C resonances [28]. The crosslink density of the network is evaluated by the
following formulas:
I(SX ) ρ
μC = ⋅
Io Mo
1
I(SX ) = [I(58.6) + I(58.0) + I(57.4) + I(44.4) + I(44.5) + I(40.7)]
2
where Io is the reference intensity of the monomer unit, ρ is the density of the vulcanisates
and Mo is the molar mass of the monomer unit. The peaks at 40.7 and 44.5 ppm are due
to monosulfides and the others are assigned to polysulfides. If the overlapping peak
around 45 ppm can be precisely decomposed to polysulfidic (44.5 ppm) and monosulfidic
(44.4 ppm) contributions, the densities of poly- and mono-sulfidic crosslinks are
individually determined by the 13C NMR.
The 13C NMR crosslink density results were compared with the crosslink density obtained
by the mechanical measurements. In the determination of the crosslink density by
mechanical methods, the contributions of the topological constraints on the results were
neglected and the density was expressed as G/2RT. The 13C and mechanical-crosslink
densities were obtained for both sulfur and dicumyl peroxide (DCP)-cured samples to
ensure the effect of wasted crosslinks (pendent or intramolecular type sulfurisations),
which are expected in the typical sulfur-vulcanisation of NR. In the major range of
crosslink densities, the crosslink densities for those two systems are described by the
same linear function with a slope of 1.0. Based on these observations, it is shown that
the crosslink density of the sulfur-vulcanised NR as determined by 13C is identical with
the true crosslink density, and the influence of the wasted or ineffective crosslinks (pendent
and cyclic crosslinks) and chain ends is negligible. However, this conclusion seems to be
only valid if the effect of topological constraints or entrapped entanglements on the
mechanical modulus is negligible which is rarely the case in real systems.
The chemical microstructures of cis-polyisoprene (IIR) vulcanised with sulfur and N-t-
butyl-2-benzothiazole sulfenamide (TBBS) accelerator were studied as a function of extent
of cure and accelerator to sulfur ratio in the formulations by solid-state 13C NMR
spectroscopy at 75.5 MHz [29]. Conventional (TBBS/Sulfur=0.75/2.38), semi-efficient
(SEV=1.50/1.50) and efficient (EV=3.00/1.08) vulcanisation formulations were prepared,
which were cured to different cure states according to the magnitude of increase in rheometer
torque. The order and types of the sulfurisation products formed are constant in all the
formulation systems with different accelerator to sulfur ratios. However, the amount of
sulfurisation has been found to vary directly with the concentration of elemental sulfur.
330
Similar vulcanisation chemistry is observed with the N-t-butyl-2-benzothiazole sulfenimide

(TBSI) accelerated sulfur-vulcanisation of IIR [26] compared to the TBBS accelerated systems
[29]. At low cure times, only A1c and A2c polysulfidic structures (50 ppm) are observed. At
longer cure times, A1c and A2c polysulfidic structures reduce in sulfur rank to monosulfide
(45 ppm), and B1c (58 ppm). B1t (64 ppm) and C1c (45 ppm) polysulfidic structures are
observed. A small amount of cis-to-trans isomerisation was observed, which increased with
sulfur content. The reversion reactions of TBSI-accelerated systems result in a lower degree
of sulfurisation as opposed to TBBS-accelerated samples. Based on the equilibrium swelling
measurements, TBSI is found to be a less efficient accelerator than TBBS.
The 13C NMR chemical shifts occurring in the 40 to 70 ppm region of the TBBS-cured
IR system were further examined by the DEPT experiments [30]. Based on the DEPT
data along with the APT [31] and chemical measurements, the peak assignments for the
resonances at 45.0, 50.2, 50.7 and 58.0 ppm in the previous study were confirmed. The
peaks at 50.2 and 50.7 ppm are due to polysulfidic A1c and A2c structures, respectively.
The monosulfidic counterparts of these structures are expected at approximately 45 ppm.
Thus, the 45 ppm resonance contains contributions from A1c and A2c monosulfides as
well as C1c polysulfides. According to observations in the chemical network analysis, it
is postulated that the peak at 45 ppm is most likely due initially to the A1c and A2c
structures, while at longer cure times, it reflects the formation of polysulfidic C1c
structures. The peak at 58 ppm has been assigned to a B1c polysulfide. The resonances
at 64 ppm, which has been assigned to a B1t polysulfide, could not be detected in the
DEPT experiments probably due to the low concentration.
A mechanism of the network formation was proposed based on the data of 13C NMR
[30] and chemical network analysis. Figure 9.4 illustrates the concentration profile of
(A) accelerator intermediates and (B) accelerator, sulfur and by-products during the
vulcanisation of IIR, obtained by high pressure liquid chromatography (HPLC) analysis
[32]. Figure 9.5 gives the structures and abbreviations used for the accelerators and cure
intermediates. Initially, BtSxBt, which is the product of TBBS/sulfur reaction, reacts with
rubber to form A1c and A2c-polysulfidic intermediates along with BtSH. For the later
extent of cure, the BtSxZnSxBt structure is produced by the consumption of BtSH with
Zn/sulfur, which reacts with the rubber to form B1c and B1t-polysulfidic intermediates.
In the reversion period, reduction in sulfur rank and the formation of C1c structure
occurs via the following schemes:
B1c-Sx-B1c + [BtS]2Zn ←→ B1c-S(x-y)-SBt + B1c-SyZnSBt
B1c-S(x-y)-SBt + Rubber → B1c-S(x-y)-C1c + BtSH
331
Figure 9.4 The concentration profile of (A) accelerator intermediates and (B)
accelerator, sulphur and by-products during the vulcanisation of IR, obtained by the
HPLC analysis
332
Figure 9.5 Structures and abbreviations used for the accelerators and cure
intermediates
The network chain density (νe) and the density of the chains between sulfurisations
(2[S]chem) for the NR [27] and high-cis-IIR vulcanised with TBBS were determined by
equilibrium swelling and solid-state 13C NMR measurements, respectively. The proportion
of the intermolecular crosslinks over a whole range of sulfurisation reactions (Ec) was
estimated by the comparison of the chemical network chain density νchem) calculated
from the νe using Mullins’s approach with the 2[S]chem obtained by NMR. The Ec in the
NR system plotted as a function of cure is shown in Figure 9.6. The Ec is zero until the
scorch cure (t10; 10% of rheometer torque increase), increases with cure during the
curing periods and hits the maximum, 59.8%, at the optimum cure state (t90). The Ec
then decreases with time in the overcuring and drops to 23.4% at 120 minute cure.
9.6 Sulfur-vulcanisation of BR
Solid-state 13C NMR spectroscopy was used to study accelerated [33] and unaccelerated
[34] sulfur-vulcanisation and sulfur-donor (TMTD) [35] vulcanisation of cis-polybutadiene
(BR). Olefinic and methylene carbons of the cis-BR repeating unit typically resonate at
129.5 and 27.5 ppm, respectively. The dominant products occurring in the vulcanisation
333
Figure 9.6 The Ec in the NR system plotted as a function of cure
with sulfur alone are the new resonances at 33, 38 and 50 ppm. Similar new resonances
are observed in the TMTD accelerated sulfur-cure systems. In the system where zinc
oxide was formulated in addition to the elemental sulfur and TMTD, simpler spectra are
obtained in which the peaks at 38 and 50 ppm are missing.
The peak at 33 ppm is assigned to the trans structure of 1,4-BR. An increasing intensity
at 33 ppm peak with cure in both sulfur-cured and accelerated sulfur-cured BR postulates
the occurrence of cis-to-trans chain isomerisation in these systems. The resonances at 38
and 50 ppm are assigned to cyclic monosulfide and polysulfidic crosslink structures. The
expected monosulfidic junctions are not detected in this study possibly due to the low
concentration of these species [33].
Doing the experiment at 75.5 MHz, improved resolution attained in the gated high
power decoupling (GHPD) spectra of sulfur-vulcanised BR. Figure 9.7 is the magnified
spectrum of the high-cis BR cured with 10 phr sulfur at 150 °C for 30 minutes. Twenty-
five new peaks are obtained with cure that are labelled in the Figure 9.7. According to
the DEPT experiments, the peaks of no. 1 through 7 are due to methine carbons and all
other resonances are due to methylene carbons [34].
Similar vulcanisation products with reduced structural modifications are obtained in the
sulfur-donor (TMTD) vulcanised BR [35]. By comparing the GHPD and DEPT results of
sulfur-vulcanised BR with the results of sulfur-donor vulcanisation, detailed chemical
shift assignments are possible where the peaks from inter- and intra-molecular
sulfurisations and those from accelerator fragments can be distinguished.
334
Figure 9.7 The magnified spectrum of the high-cis BR cured with 10 phr sulfur at 150 °C
for 30 min. Twenty-five new peaks are obtained with cure which are labelled in the figure
Side reactions including cis-to-trans isomerisation and sulfidic cyclisation are observed
along with the formation of crosslinks in the BR cured with sulfur alone. In the sulfur-
donor vulcanisation of BR, cis-to-trans isomerisation is the predominant feature of the
vulcanisation reaction sequence and seems to obey first-order kinetics with respect to the
concentration of accelerator.
The chemical crosslink densities determined from the NMR were obtained for both sulfur-
cured and sulfur-donor cured BR by computing the crosslink densities as one-half the
sum of all sulfurised structures, which were then compared with the corresponding
crosslink densities obtained by the swelling experiments (modified Flory-Rehner equation).
The NMR crosslink densities of the vulcanisates are routinely an order of magnitude
greater than the swelling crosslink densities in both systems. It was concluded that sulfur-
or TMTD-vulcanised BR networks must possess a significant proportion of mechanically
ineffective vicinal and/or cyclic crosslink structures. Either a free-radical chain reaction
mechanism of crosslink induction or the phase-solubility isolation of crosslinking within
335
discrete localised domains in the rubber would account for the formation of a high
percentage of elastically inefficient crosslink structures.
The chemistry of the TBBS-accelerated sulfur-vulcanisation of high-vinyl BR was studied

by solid-state 13C NMR at 75.5 MHz [36]. The 13C NMR study of the high-vinyl BR (or
high-styrene styrene-butadiene rubber (SBR)) is challenging since the regions between
30 and 45 ppm are typically overlapped by the backbone resonances arising from the
complex tacticity distribution in the vinyl-vinyl (or vinyl-styrene, styrene-styrene)
sequences. Furthermore, the inherent rigid nature of the vinyl-BR (or styrene sequence)
structure broadens the resonance lines, which decreases the resolution in the particular
region where the sulfurisation products are expected.
The effect of the accelerator to sulfur ratio on the vulcanisation chemistry was also
investigated by comparing the vulcanisation products from conventional, SEV and EV
formulation systems. With increase in accelerator to sulfur ratio (from conventional to
EV) there is a lowering in the sulfur rank. Also, the cis-to-trans isomerisation increases
with the amount of accelerator.
High resolution MAS techniques of 13C, DEPT, correlated spectroscopy (COSY), total
correlation spectroscopy (TOCSY), heteronuclear chemical shift correlation (HETCOR)
were used to examine conventional CBS and efficient TMTD vulcanisation of
polybutadiene [37]. In conventional CBS vulcanisation, the major vulcanisate 13C NMR
peak occurred at 44.9 ppm and was assigned to a trans allylic structure (–C=C-C-Sx
with X=3 or 4). The efficient TMTD vulcanisation yielded as main product a 13C NMR
peak at 54.0 ppm and was assigned to a cis allylic vulcanisate (-C=C-C-Sx; x=1). While
cyclic sulfur by-products were observed in both vulcanisation systems, the CBS
formulations gave rise to a higher percentage postulated to be formed via a episulfide
intermediate.
The vulcanisate structure of polyoil 110 after curing with CBS was examined using 1D
and 2D NMR experiments of 1H, 13C, COSY, and HETCOR [38]. Polyoil 110 is a
butadiene oligomer containing 72% cis, 27% trans, and 1% vinyl structures. Assignments
were made in the 1H spectrum using model data and COSY NMR results and then
correlated to the 13C spectrum. Assignment in the 13C spectrum showed monosulfidic
crosslink peaks at approximately 41 ppm, disulfides between 50-52 ppm, and polysulfides
between 52-55 ppm. The author concluded that 13C shift increments obtained from
previous model studies were not useful in examining 13C network structure.
The vulcanisation of high cis-BR was studied in unfilled and carbon black filled samples
vulcanised with TBBS at 148 °C at variable curing times of 0, 20, 30, and 60 minutes
[39]. On examining the individual peaks in the 13C spectrum (Figure 9.8), with the onset
336
of vulcanisation, two large peaks at 31.4 and 40.9 ppm appear simultaneously, increase
and than decrease with time. The two peaks closely integrate to the same values and
probably represent a CH2-CHS- pair. From chemical shift addition values reported in
the literature [36] the closest assignment would be a monosulfide-linked carbon. This
appears an unlikely assignment as monosulfides are not generally formed early in the
cure process. Later in the vulcanisation, grouped methine peaks between 48.9-51.8 and
51.8-55.8 ppm appear although with their corresponding b-CH2 groups between 35.8-
39.2 ppm. The 40.9 ppm peak appears at greater concentration earlier and decreases
faster in the carbon black CB filled samples. After 60 minutes cure, the concentration in
both unfilled and (CB) filled is the same. In the vulcanisate carbon concentrations between
35.8-39.2, 48.9-51.8, and 51.8-55.8 ppm, this same trend is observed with unfilled and
filled samples. Cis-trans isomerisation increased by approximately 400% during
vulcanisation from the starting (0 minutes) and ending (60 minutes) cure times as
determined from the trans peak areas at 33.0 ppm.
Vulcanisate structures of BR crosslinked with cyclic disulfides was studied by NMR.36

Using high resolution MAS techniques of DEPT, COSY, TOCSY, and HETCOR, the
resulting spectra showed that crosslinking gave an addition product to the double bond
and not the allylic structure found in typical sulfur vulcanisations.
Figure 9.8 The 13C NMR spectra of TBBS vulcanised BR at cure times of 0, 20, 30 and
60 minutes
337
NMR studies of blends of NR and BR have also been reported [37, 40]. It was found that
peaks characteristic of BR crosslinking appear at early stages in vulcanisation and that NR
crosslink peaks appear much later [37]. Cis-trans isomerisation was found to occur in both
rubbers but to a greater extent in BR. Interfacial reactions are observed [38].
9.7 Sulfur-vulcanisation of SBR

The NMR characterisation of SBR has also been studied although to a lesser extent than
NR and BR.
One study was carried out on the model compounds of cyclooctene, cis-3-heptene, trans-
4-octene, cis-1,7-diphenyl-3-heptene, and cis-1-phenyl-3-heptene [38]. The model
compounds were vulcanised with 4.5 phr sulfur, 4.5 phr CBS, and 13.6 phr zinc oxide at
150 °C. The BR models of cyclooctene, cis-3-heptene and trans-4-octene gave vulcanisation
products via allylic hydrogen substitution yielding both cis and trans structures. The
SBR models of cis-1,7-diphenyl-3-heptene and cis-1-phenyl-3-heptene gave no
vulcanisation products under these conditions. Only when the more reactive accelerator,
TMTD, and a longer vulcanisation time of 1 hour were used did cis-1-phenyl-3-heptene
form vulcanisation products consisting exclusively of allylic trans configuration. Thus
vulcanisation near the phenyl group is severely restricted.
The 13C NMR examination of CBS cured SBR was performed on unfilled, silica filled,
and carbon black filled samples [41]. Several different SBR samples with respect to styrene
content and cis, trans, and vinyl BR content were used. The unfilled SBR samples gave 3
major peaks that appeared at 51.0, 50.2, and 49.3 in a spectrum similar to BR
vulcanisation. Unfortunately, peaks below 45 ppm are obscured by the different main
chain structural peaks of SBR. A difference was seen in the rate of formation of these
peaks in filled samples with silica inhibit the vulcanisation rate compared to carbon
black filled samples.
9.8 Peroxide, Radiation, and High Pressure Vulcanisation

Dicumyl peroxide vulcanisation of high-cis BR and NR samples at different peroxide
levels were investigated by solid-state 13C NMR [42]. As the peroxide level increased,
decreased signal intensity and peak broadening was observed in the main chain peaks in
both BR and NR due to decreased segmental motion. In addition, cis-trans isomerisation
was observed during the vulcanisation process in both rubbers. New peaks at 35 and
44 ppm were observed in the BR spectra while peaks at 14.9, 21.4, 30.6, 37.5, and 45.0
338
were observed in the NR spectra. While no attempt was made to assign the peaks to
crosslink structures, the number and position of the peaks indicate that the vulcanisation
mechanism is more complicated than a simple combination of allylic radicals.
Network structure and reaction mechanisms in high pressure vulcanisation (HPV) and
peroxide vulcanisation of BR was studied by 13C solid-state NMR [43]. Different samples
of polybutadiene (51% trans, 38% cis, and 11% vinyl) were peroxide cured with dicumyl
peroxide on a silica carrier and by the HPV conditions of 250 °C and 293 MPa. The 13C
NMR spectra from peroxide and HPV cures were compared to a control samples heated
to 250 °C for 6 minutes under atmospheric pressure. Although no new isolated strong
peaks were detected in either the peroxide or HPV vulcanisations, small increases in
both spectra were observed at 29.5, 36.0, 46.5, and 48.0 ppm. These peaks compare
favourably with calculated shifts from structures that arise from main chain radical
addition to the pendent vinyl groups. These assignments are further reinforced by the
observation that the vinyl carbon concentration is substantially reduced during
vulcanisation in both peroxide and HPV curing. Two peaks at 39.5 and 42.5 ppm appear
only in the peroxide spectrum. Cis-trans isomerisation was absent in both cures.
The networks of dicumyl peroxide cured cis-polyisoprene in the presence and absence of
the coagent, triallyl cyanurate, were investigated by solid-state NMR techniques [44]. In
peroxide only curing, peak broadening and cis-trans isomerisation increased with
increasing peroxide level in the cure formulation. In addition, an increase in the fractions
of addition to the double bond is observed with increasing peroxide levels as observed
by the rate of decrease of the olefin peaks in the NMR spectra. When the coagent was
added, new peaks at 22.8, 36.0, and 47.6 ppm were observed in the 13C spectra and
assigned to different structures formed between coagent and polymer backbone.
Solid-state 13C NMR was used to characterise the structure produced in gamma irradiated
NR [45]. Cis-trans isomerisation increases from 5.1% at 40 Mrad to 10.3% with
161 Mrad. Peaks in the NMR spectra were observed during irradiation due to the
formation of a quaternary carbon (44 ppm) and its attached methyl group (21 ppm), a
methine carbon (38 ppm), and a g-carbon of a cis isomeric unit adjacent to a trans
isomeric unit (30.1 ppm).
The mechanism of radiation vulcanisation of NR with 2-ethyl hexyl acrylate (EHA) was
examined by 13C NMR [46]. All peaks in spectrum due to the acrylate decrease in
intensities and broaden with increasing radiation dose. Using the solid-state CP technique,
all peaks corresponding to EHA and NR were observed. The olefin peak at 129 ppm in
the acrylate was absent, however, indicating that all the EHA had polymerised.
339
9.9 Vulcanisation of Other Elastomer Systems
Microstructural changes of an accelerated sulfur vulcanisation of IIR with TMTD/ZnO/

sulfur has been studied by solid-state 13C NMR spectroscopy [47]. The IIR containing
2% isoprene and 98% isobutylene were formulated using EV and cured at 160 oC for
several cure times. The resonances at 20.3 and 24.4 ppm, which are due to trans isoprene
units in the IIR, decrease with cure, while the resonances at 26.9 and 25.2 ppm which
arise from cis isoprene units increase with cure times. The cis:trans ratio increases up to
a maximum ratio of approximately 4:1 at a cure time of 60 minutes. New resonances are
observed at 15, 21, 23.6 and 49 ppm. The peak at 49 ppm is assigned to the mixture of
the isoprene units in cis-IIR, polysulfidic A1t and polysulfidic A1c structures. The
resonance peaks at 15, 21 and 23.6 ppm are assigned to the isoprene units in mono- and
polysulfidic B1t, mono- and polysulfidic B1c and polysulfidic A1t structures, respectively.
No reaction occurs in the isobutylene units. No migration of the double bond saturation,
internal cyclisation or sulfurisation resulting in C1t and C1c structures is observed.
The curing of brominated poly(isobutylene-co-4-methylstyrene) by pure ZnO was studied

by solid-state 1H and 13C MAS NMR [48]. From NMR results, crosslinking appears to
occur by aromatic ring addition of the benzylic carbocation generated from an intermediate
PhCH2-Br-Zn complex.
Solid-state 13C NMR has been used to identify elastomers in binary blends of chloroprene
(CR) and NR, CR and CSM, NR and CSM, and SBR and acrylonitrile-butadiene rubber
(NBR). The type of NBR can be determined by identifying the sequences of acrylonitrile
and butadiene. The tertiary blend of NR/SBR/BR was also studied [49]. High-temperature
13
C solid-state NMR identified ethylene-propylene diene terpolymer (EPDM) and fluoro
and nitrile rubbers [50].
Elastomeric components and compositions in BR/SBR and NR/BR/SBR blends have been
studied by 13C solid-state NMR. The MAS spectra are of sufficient quality for polymer
identification of the carbon black filled vulcanisates in most cases [51].
Vulcanised SBR/EPDM diblends were quantitatively characterised by solid-state 13C NMR

spectroscopy. The SBR/EPDM blend ratio can be determined as well as the cis-1,4, trans-
1,4 and vinyl-1,2 butadienes and styrene ratios in the SBR and the ethylene and propylene
contents in the EPDM. No evidence for homo- and co-vulcanisation has been obtained
in these systems. No evidences are found for the change in cis:trans ratio in the SBR
upon the vulcanisation [52].
340
9.10 Effect of Carbon Black on Vulcanisation of Elastomers

The incorporation of carbon black into elastomeric systems is a process of significant
commercial importance. However, the additional stiffness of the sample imparted by the
reinforcement effect of fillers is not favourable in terms of the experimental conditions
for high-resolution NMR spectroscopy. Electric conductivity of the carbon black may
also interfere to some extent. Under these circumstances, filled formulations are not
widely used for the study of elastomer vulcanisations where high resolution and signal-
to-noise ratios are required to detect small amounts of vulcanisation products.
Solid-state 13C NMR spectra of carbon black filled, uncured and sulfur-vulcanised IIR
were recorded at 22.6 MHz. The line broadening of the filled polymer relative to the
unfilled polymer is attributed to incomplete motional narrowing of the NMR lines [53,
54] Incorporation of filler also results in a decrease in the signal-to-noise ratios in the
spectra, but fundamentally it does not obscure the qualitative and quantitative nature of
the spectra for the moderately cured elastomer systems.
Carbon black filled NR vulcanisates have been studied by high-resolution solid-state 13C
NMR at 75.5 MHz. It is shown that the detection of poly- and mono-sulfidic crosslinks
with 13C MAS spectroscopy is also possible in the case of carbon black filled NR
vulcanisates. Quantitative results suggest that the addition of carbon black leads to a
decrease in crosslink density with increasing filler content [55].
The effect of carbon black on the vulcanisation chemistry has been studied for sulfur-
cured NR systems by 13C solid-state GHPD experiments at 75.5 MHz [27]. Figure 9.9
shows the 13C NMR spectra of carbon black (N347) filled NR cured to t50 (50% of
rheometer torque increase)(A) and for 120 minutes (B). Similar new resonances are
observed as compared to the results for unfilled cis-polyisoprene systems. The increase
in the 16 ppm peak in the overcuring period is attributed to the cis-to-trans isomerisation.
Incorporation of carbon black does not affect the level of this isomerisation reaction.
The peaks at 51 and 45 ppm are due mainly to polysulfidic A1c and monosulfidic A1c
structures. Figure 9.10 illustrates the change in peak intensities at 51 ppm (A) and 45 ppm
(B) for four formulations with different carbon black loading as a function of cure. The
amount of A1c polysulfide decreases with increase in carbon black (Figure 9.10 (A)),
while more monosulfidic A1c structures occur after t90 with higher amounts of filler
(Figure 9.10 (B)). The occurrence of the B1-type sulfides increases with the level of carbon
black loading. These changes in the quantitative nature of the vulcanisation are partially
explained by the preferential adsorption of the rubber molecules on the surface of the
carbon black aggregates [27].
341
Figure 9.9 The 13C NMR spectra of carbon black (N347) filled NR cured to t50 (50%
of rheometer torque increase) (A) and for 120 min (B)
342
Figure 9.10 The change in peak intensities at 51 ppm (A) and 45 ppm (B) for four
formulations with different carbon black loading as a function of cure
Similar trends have been observed in the carbon black (N347) filled, TBBS accelerated
sulfur-vulcanisation of high-cis-IIR. In contrast to the NR/CB system, the reversion
reactions, i.e., the cis-to-trans isomerisation and the chain scission at 3,4-isoprene units,
increase with black content during the overcuring.
The amount of the sulfurisation occurring in the carbon black (N347) filled, TBBS
accelerated sulfur-vulcanisation of NR and IR have been determined by quantitative
343
measurements in the solid-state 13C NMR [27]. The density of the chains between
sulfurisations (2[S]chem) obtained from the 13C NMR was then compared with the effective
network chain density (ne) determined by the equilibrium swelling measurements (modified
Flory-Rehner equation). The 2[S]chem constantly increases with cure even in the overcuring,
while the ne hits the ceiling after t90. The 2[S]chem are almost constant for all formulations
with different content of carbon black, suggesting that the incorporation of carbon black
does not affect the total amount of chemical sulfurisation reactions.
The physical network chain density (nent) was estimated from the comparison of the
NMR and swelling results. A linear relationship of the nent in the NR system as a function
of carbon black content exists. This result suggests that the physical adsorption plays a
major role in the polymer-filler interactions.
Similar NR formulations, but using N110, N220, N326, N330, N550 and N765 carbon
blacks, have been studied by solid-state 13C NMR and equilibrium swelling measurements
[27]. Changes in the structural parameters (surface area and structure) of carbon black do
not affect the amount of sulfurisation reactions. The nent at t90 are obtained for each
formulation and plotted as a function of carbon black content, see Figure 9.11. The comparison
of the slopes of the plots and the 300% modulus data (ASTM D1765 [56], D3192 [57])
suggests the use of the plot as a method of evaluating filler reinforcement effects.
Figure 9.11 Physical network chain densities of NR cure to t90 with a variety of carbon
black as a function of carbon-black content
344
The new resonances, which appear with vulcanisation in the presence of carbon black,
are qualitatively the same as observed for the unfilled formulation systems, i.e., new
resonances at 50.7, 50.2, and 37 ppm were detected at lower cure times. At longer cure
times these three resonances disappeared and new resonances were detected at 58, 64,
and 45 ppm along with new peaks at 17 and 12 ppm and the growing of the intensity of
the peak at 14 ppm. As the amount of carbon black content increased, the latter features
were observed at an earlier stage of cure.
At low cure times, doublet peaks at 50.7 and 50.2 ppm were detected along with the
peak at 37 ppm, and these peaks disappeared at longer cure times. The intensities of
these two peaks were comparable for every detected pair. In the previous studies, the
peaks at 50.7, 50.2, and 37 ppm were assigned to a quaternary carbon of A1c polysulfide,
a methine carbon of A1c polysulfide, and a β methylene carbon of A1c polysulfide,
respectively.
While at longer cure times, resonances at 58, 64, 45, 17, and 12 ppm were detected
along with increasing intensity of the 14 ppm peak. The peaks at 58, 64, 17 and 12, and
14 ppm are assigned to B1t monsulfide (α methine), B1t polysulfide (α methine), B1c
polysulfide (β methylene), and B1t polysulfide (β methylene), respectively.
Accordingly, two phenomena are observed involving the effect of carbon black. One is
the reduction in sulfur rank for the A1c sulfide structure, and the other, is the formation
of the A1 type product at an earlier cure stage and an increase in the B1 products at a
later stage.
The reduction in sulfur rank was interpreted as due to an entropy change. The disulfide
crosslinks have less mobility than the monosulfide crosslinks due to the rigid nature of
-S-S- linkage. Similarly, the mobilities of the polysulfidic crosslinks are probably lower
than that of the monosulfidic crosslinks. Therefore the polysulfidic crosslink would occupy
more space in the network structure than the monosulfidic crosslink. Considering this
fact, the change in sulfur rank from polysulfide to monosulfide would occur along with
the increase in entropy when the curing process is extended. Since the system becomes
more rigid by the incorporation of the carbon black, the reduction in sulfur rank to
monosulfide would eventually be accelerated.
The second phenomenon due to carbon black on a proposed scheme based on the
vulcanisation process shown in Figure 9.12. During the vulcanisation process, the reactions
(1) and (2), and (3) and (4) form polysulfidic A1c and B1t, respectively. Considering the
fact that A1c polysulfide reduces its sulfur rank to monosulfide and reaction (3) and (4)
is the predominant process in the later stage of cure, the system may favour reaction (3)
and (4) rather than (1) and (2) in the whole reaction scheme. BtSH is formed from the
345
reaction producing polysulfidic A1c at the earlier stage of cure, which is the reactant for
reaction (4) producing the B1 structure. As previously mentioned, the sulfur and
accelerator to rubber ratio may be higher for the system with high carbon black content.
As a result, reactions (1) and (2) would occur more rapidly and, consequently, produce
enough BtSH for the initiation of reactions (3) and (4) , which means that the B1t
polysulfides initiate at the earlier stage of cure for the sample with higher filler content.
Since reactions (3) and (4) would occur until the elemental sulfur is consumed, the total
amount of B1 structure would be larger for the sample with higher carbon black content.
The network chain density measured by equilibrium swelling includes both the chemical
and physical crosslinks while the 13C NMR measurements determine only the chemical
crosslinks. The observed differences in network chain density between the results of the
two measurements (νSWELL - νNMR) reflect differences in the physical crosslinks induced
by entanglements involving polymer-polymer interactions and polymer-carbon black
interactions. The magnitude of the differences between the two methods of measuring
the network chain density increases with the volume fraction of the carbon black for the
same stage of cure. The differences in network density values reach a maximum near the
90% cure stage for the filled formulations. These results coincide to similar observations
from tensile tests where the maximum filler reinforcement effect occurs at an optimum
Figure 9.12 Mechanism of vulcanisation
346
(90%) stage of cure. Consequently, the comparisons of the network densities determined
by the swelling and 13C NMR measurements give a measure of the effectiveness of the
filler reinforcement at the molecular level.
9.11 Effect of Silica Filler on Vulcanisation Chemistry
Solid state 13C NMR spectroscopy was used to identify and characterise the effects that
silica has on the network structures that form during the vulcanisation of zinc activated,
sulfur-cured cis-1,4 IIR [58]. It was determined that cis-trans isomerism and chain scission
were enhanced when silica was added as a filler. The monosulfidic linkages, especially
the A1cis and the B1trans type monosulfide linkages, increased in concentration in the
silica-filled systems. Also, the polysulfidic linkages, namely the A1cis and B1trans type
linkages, decreased in concentration. Finally, it was also determined that the intermolecular
crosslinks decreased in the presence of silica, while the intramolecular structures, including
pendent side groups and cyclic sulfides, increased in the presence of silica, resulting in a
lower overall crosslink density when silica is used as a filler.
When silica is used as a filler, the cure ingredients are adsorbed, causing the initial overall
lower cure rate. Several ways exist to counteract this effect including inhibiting the silica
absorption by adding a silane coupling agent, or polyethylene glycol (PEG). Both additives
have been shown in the literature to improve the vulcanisate properties. Therefore, solid
state 13C NMR spectroscopy was used to identify and characterise the effects on the
network formation when additives, such as a coupling agent and PEG, were added along
with silica filler [59]. It was determined that cis-trans isomerisation was inhibited by the
inclusion of both the coupling agent and the PEG. However, there was only a slight
influence on the chain scission of the main chain carbons. The presence of coupling
agent or PEG both caused an increase in the concentration of the polysulfidic structures,
especially the A1cis and B1trans type linkages, as well as a decrease in the monosulfidic
structures, namely the A1cis and B1trans type structures. Therefore, the intermolecular
crosslinks, mainly the polysulfidic structures, increased with the use of the additives,
while the intramolecular structures, which are mainly monosulfidic in nature, decreased.
Therefore, the network chain density of the silica-filled rubber system increased when
additives were included in the formulation.
9.12 Thermal-Oxidation of Network Structures
Elastomeric materials undergo both thermal and oxidation degradation over time. Main
chain scission and loss of sulfur crosslinks can occur with either factor or by both factors
by a thermo-oxidative mechanism.
347
The chemical changes that occur in thermal and oxidative degradation of sulfur cured
NR were investigated using 13C NMR analysis [58]. The TBBS vulcanised samples of
NR were heated at 100 °C and 150 °C under both air (oxidative) and nitrogen (inert)
environments to assess both temperature and oxidative effects. Changes were observed
by measuring peak areas under the crosslinked carbon resonance peaks. A and B-type
monosulfide linkages were found to decompose much faster by oxidative factors at both
100 °C and 150 °C than under nitrogen at either temperature. The polysulfide linkages,
however, showed the opposite result of thermal degradation predominating over oxidative
factors. Cis-trans isomerisation of the main chain carbons was observed at the higher
temperature while absent at 100 °C.
Another study used 13C NMR to examine the thermal ageing of NR with both peroxide
and conventional and EV sulfur vulcanised samples [59]. The samples were heat aged at
70 °C in air with 13C NMR recorded after 3 and 21 days. The vulcanisate structures
vulcanised by peroxide and the EV sulfur system showed low sensitivity to ageing effects
while peaks in the conventional sulfur system almost disappeared after 21 days of ageing.
In addition, the main chain isoprene carbon peaks showed considerable broadening due
to increasing stiffness of the network.
9.13 Summary
From this chapter, it is clear that high resolution 13C NMR spectroscopy has made a
large number of contributions to our knowledge of the structure of vulcanised crosslinked
elastomers as well as the mechanism by which the vulcanisation process occurs. It is
anticipated that further NMR measurements of these systems will continue to generate
new structural and mechanistic information.
Acknowledgements
The authors wish to acknowledge the support of the National Science Foundation as
well as the support of the Yokohoma Rubber Company and the Pirelli Rubber Company.
References
1. F.A. Bovey, Chain Structure and Conformation of Macromolecules, Academic
Press, New York, 1982.
2. M. Mehring, High Resolution NMR Spectroscopy in Solids, Springer-Verlag,

New York, 1983.
348
3. C.A. Fyfe, Solid State NMR for Chemists, C.F.C. Press, Guelph, 1984.
4. H.W. Spiess, Annu. Rev Mater. Science, 1991, 21, 131.
5. J. Schaefer and E.O. Stejskal, Journal of the American Chemistry Society, 1976,
98, 1031.
6. D.E. Axelson, Solid State Nuclear Magnetic Resonance of Fossil Fuels: An

Experimental Approach, Multiscience, 1983.
7. J.R. Havens and J.L. Koenig, Applied Spectroscopy, 1983, 37, 226.
8. R.A. Komoroski, High Resolution NMR of Synthetic Polymers in Bulk, Ed.,

VCH Publishers, Deerfield Beach, 1986.
9. L. Mathias, Solid State NMR of Polymers, Ed., Plenum Publishers, New York,
1989.
10. M. Mori and J.L. Koenig, Annual Reports in NMR Spectroscopy, Academic Press
Ltd., Vol. 34 (1997)
11. M. Andreis and J.L. Koenig, Advances in Polymer Science, 1989, 89, 71.
12. R.A. Kinsey, Rubber Chemistry Technology, 1990, 63, 407.
13. M.R. Krejsa and J.L. Koenig, Rubber Chemistry and Technology, 1993, 66, 376.
14. M.R. Krejsa and J.L. Koenig, Elastomer Technology Handbook, CRC Press,
1993, Chapter 11.
15. J.L. Koenig and D.J. Patterson, Elastomers Rubber Technology, 1987, 32, 31.
16. A.M. Zaper and J.L. Koenig, Rubber Chemistry and Technology, 1987, 60, 252.
17. The Sadtler Guide to Carbon-13 Spectra, Ed., W.W. Simons, Sadtler Research
Laboratories, 1983.
18. W. Gronski, H. Hasenhindl, B. Freund and S. Wolff, 1991, Kautschuk und

Gummi Kunststoffe, 44, 119.
19. M. Andreis, J. Liu and J.L. Koenig, Journal of Polymer Science, Phys. Ed., 1989,
27, 1389.
20. Natural Rubber Science and Technology, Ed., A.D. Roberts, Oxford University
Press, New York, 1988, Chapters 11-12.
349
21. C.G. Moore, L. Mullins and P.M. Swift, 1961, Journal of Applied Polymer
Science, 5, 293.
22. A.Y. Coran, Rubber Chemistry Technology, 1995, 68, 351.
23. A.Y. Coran, Science and Technology of Rubber, Academic Press, 1994, Chapter 7.
24. A.M. Zaper and J.L. Koenig, Rubber Chemistry and Technology, 1987, 60, 278.
25. D. Parker and J.L. Koenig, Journal of Applied Polymer Science, 1998, 70, 1371.
27. M. Mori and J.L. Koenig, Journal of Applied Polymer Science, 1998, 70, 1391.
28. W. Gronski, U. Hoffmann, G. Simon, A. Wutzler and E. Straube, Rubber

31. R.C. Hirst, ACS Rubber Division Meeting, October 1991, Paper No. 69.
32. A.B. Sullivan, C.J. Hahn and G. H. Huhls, Rubber Chemistry and Technology,
1992, 65, 488.
33. A.M. Zaper and J.L. Koenig, Macromolecular Chemistry, 1988, 189, 1239.
34. R.S. Clough and J.L. Koenig, Rubber Chemistry Technology, 1989, 62, 908.
35. S.R. Smith and J.L. Koenig, Rubber Chemistry and Technology, 1992, 65, 176.
36. M.A. Rana and J.L. Koenig, Rubber Chemistry and Technology, 1993, 66, 242.
37. R. Hulst, R.M. Seyger, J.P.M. van Duynhoven, L. van der Does, J.W.M.
Noordermeer, and A. Bantjes, Macromolecules, 1999, 32, 22, 7521.
38. J. Hahn, M. Runk, M. Schollmeyer, U. Theimer and E. Walter, Kautschuk und

39. Dallas D. Parker, Makio Mori and J.L. Koenig, submitted to Rubber Chemistry
Technology,
350
40. B. Klei and J.L. Koenig, Rubber Chemistry and Technology, 1997, 70, 231.
41. L. Pellicioli, S.K. Mowdood, F. Negroni, D.D. Parker and J.L. Koenig, submitted
to Rubber Chemistry and Technology,
42. D.J. Patterson, J.L. Koenig and J.R. Shelton, Rubber Chemistry and Technology,
1983, 56, 5, 971.
43. M. Bellander, B. Stenberg and S. Persson, Journal of Applied Polymer Science,

1999, 73, 14, 2799.
44. S.J. Oh and J.L. Koenig, Journal of Polymer Science, Part B: Polymer Physics,
2000, 38, 11, 1417.
45. D.J. Patterson and J.L. Koenig, Applied Spectroscopy, 1987, 41, 3, 441.
46. D.J.T. Hill, J.H. O’Donnell, M.C.S. Perera, P.J. Pomery and P. Smetsers, Journal
of Applied Polymer Science, 1995, 57, 10, 1155.
48. R.R. Eckman, I.J. Gardner and H.-C. Wang, Rubber Chemistry and Technology,
1993, 66, 1, 109.
49. D. Gross and J. Kelm, Kautschuk and Gummi Kunststoffe, 1987, 40, 13.
50. D. Gross and J. Kelm, Kautschuk und Gummi Kunststoffe, 1985, 38, 1089.
51. R.A. Komoroski, Rubber Chemistry and Technology, 1983, 56, 959.
52. G.P.M. Van Del Velden and J. Lelm, Rubber Chemistry and Technology, 1990,
63, 215.
53. Schaefer, Macromolecules, 1972, 5, 427.
54. J. Schaefer, S.H. Chin and S.I. Weissman, Macromolecules, 1972, 5, 798.
55. M. Mori and J.L. Koenig, Rubber Chemistry and Technology, 1997, 70, 671.
56. ASTM D1765-01, Standard Classification Systems for Carbon Blacks used in
Rubber Products, 2001.
57. ASTM D3192-00, Standard Test Methods for Carbon Black Evaluation in NR
(Natural Rubber, 2000.
351
58. C. Hill and J.L Koenig, Polymer Bulletin, 1998, 40, 275.
59. M.L. Kralevich and J.L. Koenig, Composite Interfaces, 1997, 5, 125.
60. J.Y. Buzare, G. Silly, J. Emery, G. Boccaccio and E. Rouault, European Polymer
Journal, 2001, 37, 1, 85.
352
Characterisation of Chemical and Physical Networks in Rubbery Materials Using …
10
Characterisation of Chemical and Physical
Networks in Rubbery Materials Using
Proton NMR Magnetisation Relaxation
Victor M. Litvinov
Abstract
The use of solid-state nuclear magnetic resonance (NMR) transverse magnetisation

relaxation experiments for characterising various types of viscoelastic materials and
rubbery phases in blends is reviewed in this chapter. The methods applied for quantitative
analysis of the density of chemical crosslinks, temporary and trapped chain entanglements
and physical network junctions, which are formed in filled rubbers, semi-crystalline and
ionic containing elastomers, are discussed. Several examples of the determination of the
molecular-scale heterogeneity of polymer networks are given. Use of NMR magnetisation
relaxation methods in real-time NMR experiments and quality control are discussed.
10.1 Introduction
The density of chemical crosslinks and physical network junctions largely affects the
mechanical properties of rubbery materials and polymer blends containing rubbery
components. Despite the apparent simplicity, the network structure has a complex
topology, which can significantly affect functional properties. The type of curing may
often cause significant differences in network topology and properties of cured materials.
Different types of network heterogeneity will be found in cured materials if no precautions
are taken to control the curing chemistry and conditions. The following types of molecular-
scale heterogeneity may occur: heterogeneity in the distribution of network junctions,
polymer chains unattached to the network, and/or dangling chains and loops. In addition,
the type of network junction, i.e., the functionality and bulkiness, which determine the
network junction’s ability to fluctuate (affine versus phantom networks), may also affect
the mechanical properties. A difference in curing conditions in the sample volume, for
example a difference in temperature or in the concentration of vulcanisation agents, will
result in spatial heterogeneity of the network structure. Besides chemical crosslinks the
following types of physical network junctions may occur in rubbery materials:
353
1) Temporary and trapped chain entanglements,
2) Junctions that are formed due to chain adsorption at the surface of active fillers,
3) Junctions that are formed by crystallites,
4) Junctions deriving from strong hydrogen and ionic bonds, and
5) Junctions that are formed at the interface of polymer blends and those in materials
that reveal nano-scale phase separation.
These are variables in the network structure that can be utilised to modify the properties
of cured materials. However, they also cause difficulties in the analysis of network
structures and complicate efforts to determine structure-property relationships.
To be able to relate mechanical properties to the composition of rubbery compounds

and curing chemistry it is essential to understand the network topology of the resultant
cured materials.
The methods that are used to analyse network structures can be generally subdivided
into three categories on the basis of their methodology:
1) Characterisation of physical properties of cured materials in relation to volume-average

network density. The most common methods are equilibrium swelling and mechanical
measurements [1-6]. Besides information on the network structure, the mechanical
methods provide information that is useful for practical applications, such as
information on the modulus of elasticity, ultimate tensile properties and glass transition
temperature (Tg). Rubber elasticity theory and phenomenological theories are used to
relate a measured quantity to the density of chemical and physical crosslinks. Several
molecular models have been developed for ‘ideal’ defect-free networks [1, 7-9]. The
validity and applicability of these models is however a heavily debated topic and
several discrepancies between theory and experiment still remain, e.g., the role of
chain entanglements, network defects and network heterogeneity. It is generally
acknowledged that traditional methods are not capable of providing fully reliable
information on network topology [9, 10, 11].
2) Analysis of chemical conversion and cure chemistry is another way of studying network
structures. Several techniques are used for this purpose, e.g., optical spectroscopy [12],
high-resolution NMR spectroscopy and titration of non-reacted functional groups. The
spectroscopic methods can be used for quantitative analysis of crosslinks [13-15]. Chemical
conversion is usually closely related to the network density. However, no exact quantitative
information on the network structure can be obtained because reacted groups can form
354
not only chemical crosslinks but also viscoelastically ineffective chains, such as chain
branches and chain loops. Furthermore, side-reactions, which may readily lead to the
formation of additional crosslinks, complicate data interpretation due to the overlapping
of signals from different types of chemical groups in complex mixtures. Moreover, physical
network junctions are virtually not detected by spectroscopic methods.
3) Analysis of molecular mobility of polymer chains is used to study network density

and its heterogeneity. Dynamic mechanical analysis (DMA) and dielectric spectroscopy
provide information on the mobility of polymer chains, which is linked to the network
density [16]. One of the most informative and sensitive methods of network analysis
is solid-state NMR [17-23]. Furthermore NMR imaging or microscopy is used to
determine the spatial heterogeneity of rubbery materials on a scale of 15-50
micrometres [19,24-26]. Different types of NMR magnetisation relaxation experiments
are used to analyse local and long-range spatial mobility of polymer chains. Since
chain motion is closely coupled to the length of network chains, chemical information
on network structure and network defects can be obtained in this way. The mobility
of polymer chain units of different chemical origins can be determined by means of
selective NMR magnetisation relaxation experiments using high-resolution solid-state
NMR techniques. These techniques can also be used for the analysis of the network
structure in rubbery blends. Relationships between NMR relaxation parameters,
dielectric and mechanical properties have been established [27-29].
Evidently, the most comprehensive information on network structure in relation to

properties can be obtained by using these three complementary methodologies.
The use of solid-state NMR magnetisation relaxation experiments to characterise network

structures in various rubbery materials is reviewed in this chapter. Comprehensive reviews
of high-resolution NMR techniques can be found elsewhere [21-23, 30-35].
10.2 Network Structure Analysis by Means of NMR Transverse

Magnetisation Relaxation
Solid-state NMR magnetisation relaxation experiments provide a good method for the
analysis of network structures. In the past two decades considerable progress has been
made in the field of elastomer characterisation using transverse or spin-spin (T2) relaxation
data [36-42]. The principle of the use of such relaxation experiments is based on the
high sensitivity of the relaxation process to chain dynamics involving large spatial-scale
chain motion in elastomers at temperatures well above the Tg and in swollen networks.
Since chain motion is closely coupled to elastomer structure, chemical information can
also be obtained in this way.
355
The temperature dependence of the 1H T2 relaxation time of well-defined end-linked

poly(propylene oxide) (PPO) networks [43] is shown in Figure 10.1 [44] 1. The
distinguishing feature of T2 relaxation in the case of viscoelastic networks is the high
temperature plateau that is observed at temperatures well above the Tg [45-49].
At these temperatures, the temperature independence of T2 is attributable to constraints

limiting the number of possible conformations of a network chain to those of a free
chain. The slight increase in T2 with an increasing temperature, which is observed for
some networks in the plateau temperature range, may be attributable to the presence of
network defects, such as dangling chains, long chain loops and residual non-vulcanised
material [49]. A value of T2 at the plateau, T2pl, is determined by the asymmetry of
random rotations of monomer units, and does not depend on the mechanism and
frequency of motions of network chains. The theory of the transverse relaxation in
elastomeric networks relates T2pl to the number of statistical segments, Z, between
chemical and physical network junctions [45, 46]:
Z = (T2pl)/[a(T2rl)]
where a is the theoretical coefficient, which depends on the angle between the segment
axis and the internuclear vector for the nearest nuclear spins at the main chains. In the
case of polymers containing aliphatic protons in the main chain the coefficient a is close
to 6.2 ± 0.7 [46]. T2rl is the relaxation time measured at temperatures below the Tg of the
swollen polymer. A solvent free of hydrogen atoms (fully deuterated, chlorinated or
fluorinated solvent) is used in this experiment. The weight average molecular mass of
network chains between chemical and physical junctions, <Mc+p>w, can be calculated
using the number of backbone bonds in the statistical segment, C∞:
<Mc+p>w = ZC∞Mu/n
where Mu is the molecular mass per monomer unit and n is the number of rotatable
backbone bonds in the monomer units. For copolymers, the average molecular mass per
rotatable backbone bond is used in the Mu calculation.
The maximum relative error of this NMR network density determination was estimated to
be about 15%–25% [49, 50]. Comparative analysis of the network density measured in
the same samples with the aid of the NMR method and other techniques, such as stress-
1
Tg for a linear, high-molecular-mass PPO, which was prepared from a polypropylene glycol precursor
(with a molecular mass of 4000 g/mol) using a chain extender with a chemical structure similar to that of
the crosslinker, is –62.5 °C. The Tg was determined from DMA measurements performed at 0.215 Hz [43].
356
Figure 10.1 Temperature dependence of the 1H T2 relaxation time of well-defined end-

linked (PPO) networks with narrow molecular mass distributions between chemical
crosslinks [44]. The molecular mass of network chains (in g/mol) is shown in this
figure. The temperature dependence of a linear, high-molecular-mass poly(propylene
oxide) prepared from a polypropylene glycol precursor (with a molecular mass of
4000 g/mol) using a chain extender with a chemical structure similar to that of the
crosslinker is shown for comparison. The synthesis of the model networks has been
described elsewhere [43]
357
strain measurements, equilibrium swelling and 13C NMR spectroscopy, proves that T2
relaxation data provide quantitative data on the network density [15, 49-52]. As an example,
the results obtained for UV-cured acrylates with a broad crosslink density range were
compared using the NMR method and DMA (see Figures 10.2 and 10.3 [52]).
A fairly good correlation between the storage modulus and the relaxation rate (1/T2)
was observed. Both methods gave similar values for the mean molecular mass of network
Figure 10.2 The relaxation rate (1/T2s)max measured for a cured mixture of a
poly(ethylene glycol) diacrylate (Mn = 700 g/mol) and 2-ethylhexyl acrylate as a
function of the storage modulus at 273 K (-0.1 °C) [52]. The rubbery plateau was
observed for all samples at 273 K (-0.1 °C). (1/T2s)max corresponds to the relaxation
component with short decay time that was measured at 323 K (50 °C) for partially
swollen in 1,1,2,2-C2D2Cl4 samples. This relaxation component corresponds to the
relaxation of network chains. The line represents the result of a linear regression
analysis: intercept = 1.1 ± 0.3 ms-1; slope = 0.34 ± 0.02 ms-1(MPa)-1. The correlation
coefficient equals 0.992
Reprinted with permission from V.M. Litvinov and A.A. Dias, Macromolecules, 2001, 34,
12, 4051, Figure 10. Copyright 2001 American Chemical Society
358
chains at small values of Mc+e. A substantial difference in the Mc+e value at low crosslink
density in caused by the effect of network defects, which decrease volume average network
density as determined by DMA. The effect of network imperfections was excluded from
the NMR analysis, as will be discussed below (see Section 10.3).
Figure 10.3 Mean molecular mass between chemical crosslinks and trapped chain
entanglements Mc+e in a cured mixture of a poly(ethylene glycol) diacrylate (PEGDA)
and 2-ethylhexyl acrylate (EHA) as a function of the EHA content [52]. Mc+e values
were determined from (1/T2s)max and the plateau modulus (see Figure 10.2). A
substantial difference in Mc+e value, as determined by these two methods at low
crosslink density, is caused by the effect of network defects which decrease volume
average network density determined by DMA (see Section 10.3). The molecular mass
of PEGDA (Mn = 700 g/mol) is indicated by an arrow. The molecular mass of network
chains in cured PEGDA is about three times smaller than that of the initial monomer.
The molecular origin of this difference is discussed in Section 10.3
359
The effect of chain entanglements on T2 was studied using series of ethylene-propylene-

diene (EPDM) vulcanisates with a broad range of chemical crosslink densities [49]. The
network density in the vulcanisates was determined with the aid of the following methods:
1
H NMR transverse magnetisation relaxation, Mooney-Rivlin analysis of stress-strain
curves [2-4] and equilibrium swelling experiments. It was shown that in the case of the
original vulcanisates, the NMR method measures the total network density, which
comprises chemical crosslinks (CC) and temporary entanglements (EN/TE) and trapped
chain entanglements (EN/TR). For partially swollen samples, the NMR method determines
the network density comprising CC and EN/TR. A difference in the T2 values obtained
for the original vulcanisates and the swollen ones was analysed with respect to
entanglement density. The estimated molecular mass of EPDM chains between apparent
chain entanglements (EN) was 1900 ± 200 g/mol. The value obtained is in good agreement
with the values obtained using other techniques [49]. The CC density was determined
from the total density of network junctions and the measured entanglement density on
the assumption that CC and chain entanglements are decoupled and additive. The
contribution of CC, EN/TE and EN/TR to the total network density was estimated for
EPDM vulcanisates with a broad crosslink density range.
The results of the T2 relaxation studies prove that this method is a very useful technique
for the quantitative characterisation of network structures, while the more
sophisticated NMR techniques, which also determine the residual dipole-dipole
interactions [31, 53-60], provide specific information for the chemical structure and
molecular mobility, which may be useful in determining mechanisms of molecular
motions and refining interpretations of the non-selective T 2 relaxation method,
especially for composite materials.
10.3 Characterisation of Network Heterogeneity and Network Defects
Mechanical properties of crosslinked elastomers are influenced not only by the volume-
average crosslink density but also by network heterogeneity. The influence of structural
defects (such as residual sol, dangling chains, chain loops and the heterogeneity of the
junction distribution) on the viscoelastic properties and the equilibrium swelling data is
still under discussion. Local methods which probe molecular properties are very suitable
for the determination of the degree of network heterogeneity [11].
Several types of heterogeneity may occur in rubbery materials:
1) molecular-scale heterogeneity, which is caused by the chemical heterogeneity of uncured

elastomers, network defects and heterogeneous distribution of network junctions on
a molecular level;
360
2) morphological heterogeneity of rubbery compounds due to a spatially heterogeneous

distribution of components and filler in the compound;
3) spatial heterogeneity due to differences in curing conditions such as temperature and

concentration of vulcanisation agents throughout the sample volume.
A significant difference between the large spatial-scale mobility of network chains and
that of network defects allows us to determine the degree of network heterogeneity. The
most reliable data are obtained for swollen samples, because an increasing solvent content
results in the disentanglement of network defects from network chains [52, 61, 62]. The
molecular mobility of network chains is consequently decoupled from that of network
defects, resulting in a major distinction in the relaxation behaviour.
The effect of a solvent on the T2 relaxation decay of cured acrylates is shown in

Figure 10.4 [52].
The monoexponential T2 decay of a cured acrylate is split into two components upon
swelling. This behaviour is most clearly observable at high solvent concentrations [52,
62]. One component has short decay time (T2s), which is comparable with that of the
original sample.
This component apparently derives from network chains. Starting at a low volume solvent
content Vs, (see Figure 10.5), T2s shows an increase, which may be attributable to the
following phenomena:
1) the disentanglement of network chains,
2) an increase in the frequency of the large spatial-scale chain motion, and
3) a slight decrease in the strength of the inter-chain proton dipole-dipole interactions.
At Vs ≈ 40 - 50 vol.%, T2s reaches a maximum value.
A value of T2s at the maximum, (T2s)max, is related to the molecular mass of the network
chains between chemical crosslinks and trapped chain entanglements [49]. At higher
values of Vs, T2s decreases until the state of equilibrium swelling is reached. This decrease
in T2s is thought to reflect the increase in the inter-chain proton dipole-dipole interactions
as a result of the network chain elongation following a progressive increase in the solvent
fraction in a swollen gel [20].
The long decay time (T2l) of the other component is typical of semi-diluted polymer
solutions. The T2l value continuously increases with an increasing solvent content. This
361
Figure 10.4a
Figure 10.4b
362
Figure 10.4 1H T2 decay measured for a cured mixture of a poly(ethylene glycol)

diacrylate and 2-ethylhexyl acrylate (10:90 wt%) without a solvent (a) and with 46
vol.% (b) and 90 vol.% (c) 1,1,2,2-C2D2Cl4 [52]. The solid line represents the result of
a least-squares adjustment of the decay with a liner combination of two exponential
functions. The dotted lines represent separate components
component apparently originated from the relaxation of network defects, which are
disentangled from network chains in a swollen state. At the equilibrium swelling degree,
the relative fraction of the T2l relaxation component could be used as a measure of the
fraction of highly mobile network defects. The described behaviour of a T2 relaxation
decay, following a progressive increase in the solvent fraction, is typical of networks
containing a significant fraction of network defects [52, 62]. However, precautions should
be taken in the event of such analyses of T2 relaxation with respect to network
heterogeneity, as will be discussed later in this section.
Distinct T2 relaxation components with widely differing mean decay times suggest
molecular or macroscopic heterogeneity of the material. In such cases the submolecule
concept can be used to describe the relaxation behaviour [20]. In a simplified
interpretation, the overall T2 relaxation decay of a heterogeneous elastomer is the weighted
sum of the decays originating from the submolecules, which are defined as the network
363
Figure 10.5 The T2 relaxation time for a polymer network with defects against the
volume fraction of a good solvent
chains that are formed by the chemical and physical junctions, and network defects, i.e.,
chains that are not attached to the network, chain loops and dangling chain ends. The
large spatial-scale mobility of these submolecules differs substantially, and so does their
relaxation behaviour. The relative contribution of the submolecules to the overall decay
is proportional to the number of protons attached to these chain fragments.
A quantitative analysis of the shape of the decay curve is not always straightforward due
to the complex origin of the relaxation function itself [20, 36, 63-66] and the structural
heterogeneity of the long chain molecules. Nevertheless, several examples of the detection
of structural heterogeneity by T2 experiments have been published, for example the analysis
of the gel/sol content in cured [65, 67] and filled elastomers [61, 62], the estimation of
the fraction of chain-end blocks in linear and network elastomers [66, 68, 69], and the
determination of a distribution function for the molecular mass of network chains in
crosslinked elastomers [70, 71].
The applicability of T2 experiments for determining the molecular mass distribution of

network chains is illustrated in Figure 10.6.
364
Figure 10.6 The distribution function of the chain length for two EPDM sulfur
vulcanisates whose crosslink densities differ by a factor of about two and that for the
physical mixture (50:50 mass%) of these vulcanisates [72]
Figure 10.6 shows the distribution function of the chain length for two EPDM sulfur
vulcanisates whose crosslink densities differ by a factor of about two and the distribution
factor for the physical mixture (50:50 mass %) of these vulcanisates [72]. The vulcanisates
show a rather broad distribution in molecular mass of network chains, which is apparently
caused by the statistical distribution of the third monomer in the original EPDM and
possibly by incomplete conversion of the third monomer’s double bonds during
vulcanisation. The physical mixture of the vulcanisates shows a bimodal distribution
function with maxima at about the same chain length as that for the two vulcanisates in
the mixture. This shows that the method is suitable for analysing heterogeneous networks.
The analysis of computer-simulated T2 decay curves of bimodal networks shows that a
difference in the molecular masses of network chains greater than a factor of two can be
determined with the aid of this method.
Bulky crosslinks or side-groups in the network chains, e.g., dendritic wedges [73], may
also influence molecular mobility and viscoelastic properties of polymer networks. For
example, UV curing of difunctional acrylates results in the formation of zip-like network
junctions, which may be regarded as extreme cases of bimodal networks [52]. Results
obtained with the NMR T2 relaxation method agree well with those of mechanical tests
365
for molecular mass between crosslinks, Mc+e, as was discussed in Section 10.2 (see
Figure 10.3) [52]. However, the values of Mc+e obtained are significantly lower than the
molecular mass of the initial diacrylate. So, network junctions of this type evidently
strongly affect elastic properties and T2 relaxation. As anticipated, classical rubber
elasticity theories cannot be used to characterise heterogeneous networks.
10.4 Network Structure in Oil-Extended Rubbers - Effect of Chain

Entanglements
Knowledge of the physical state of oil in oil-extended rubbers and their vulcanisates is of
great importance, because the processing of oil-extended rubbers and the mechanical properties
of such rubbers are strongly affected by the rubbers’ oil content and the type of oil concerned.
1
H T2 relaxation of oil-extended EPDM revealed two distinct relaxation components whose
characteristic decay times are comparable with those of initial rubber and paraffinic oil
(Figure 10.7) [74]. This suggests that the components with a short and long decay time
mainly originate from the relaxation of rubbery chains and oil molecules, respectively.
Up to 400 K (127 °C), the fraction of the component with a long decay time was smaller than
the concentration of oil hydrogen in the oil-extended rubber. Apparently, a small fraction of
oil molecules shows a molecular mobility comparable with that of EPDM chains. The fraction
of these physically trapped oil molecules decreased with an increasing temperature, and
Figure 10.7 The T2 relaxation decay of oil-extended EPDM rubber containing 12.2
(o), 33.7 (❑) and 50.0 mass% (Δ) paraffinic oil [74]. The line corresponds to the best
fit of experimental data points with a linear combination of two exponential functions
366
above 400 K (127 °C) the molecular mobilities of the EPDM chains and oil molecules were
largely decoupled. It was shown that 1H T2 relaxation experiments above this temperature
can be used to quantitatively determine both the oil content of EPDM rubbers and the
molecular mass of network chains in oil-extended vulcanisates. The results obtained suggest
that the molecular diffusion of oil molecules depends largely on:
1) the mesh size of network holes with respect to the size and shape of oil molecules,
2) the molecular mobility of EPDM chains, and
3) intermolecular forces of cohesion.
The network densities of the oil-extended EPDM rubbers were determined by means of
1
H T2 relaxation experiments [74]. The EPDM samples containing varying amounts of
paraffinic oil (from 5 to 100 phr) were cured under the same conditions, using the same
vulcanisation recipe. The network density was nevertheless found to decrease substantially
with an increasing oil concentration in the vulcanisates (Figure 10.8). Analysis of these
data using scaling laws for polymer melts and solutions [75-77] suggests that the decrease
in the overall network density, comprising chemical crosslinks and chain entanglements,
is due mainly to a decrease in the entanglement density with an increasing oil content.
Figure 10.8 The density of network junctions as a function of the volume fraction
of paraffinic oil in EPDM/oil vulcanisates [74]. The solid line represents the result
of a linear regression analysis of the dependence: (intercept = 453 ± 5 mmol/kg;
slope = -6.2 ± 0.0.3 mmol/kg; the correlation coefficient = 0.996). Maximum
torque in the rheometer curve for the vulcanisates is shown on the right ordinate
367
10.5 Network Structure in Filled Rubbers - Rubber-Filler Interface and

the Structure of the Physical Network
10.5.1 NMR Relaxation of Filled Rubbers
The incorporation of fillers in rubbers is of significant commercial importance, since

use of fillers not only enhances the end product’s mechanical properties but can also
decreases its costs. Although much research has been carried out on filled rubbers
using different techniques, the molecular origin of the filler’s reinforcing effect is still
unclear. The reinforcing effect at a moderate strain (<7%) is primarily the result of
disruption of the continuous network of filler particles, which interpenetrates the rubber
matrix [78]. It is generally assumed that the nature of the elastomer-filler interactions
is of major importance for improving the mechanical properties of filled rubbers at
high elongation values. The elastomer-filler interactions are evidently of such great
importance due to the large total elastomer-filler contact area. However, there is a
distinct lack of techniques that are capable of providing information on the physical
properties and the fraction of the interface between the rubber matrix and filler particles
in filled rubbers.
A low-resolution proton NMR method is one of the few techniques that have so far
proved to be suitable for studying elastomer-filler interactions in carbon-black-filled
conventional rubbers and silica-filled silicon rubbers [20, 62, 79]. It was pointed out
by McBrierty and Kenny that ‘Many of the basic characteristics of filled elastomers are
revealed by low resolution spectra while the more sophisticated techniques and site
specific information refine interpretations and clarify motional dynamics’ [79].
Experimental studies of filled rubbers are complicated by several things, such as the
effect of the magnetic susceptibility of the filler, the effect of free radicals present at the
surface of carbon black, the complex shape of the decay of the transverse magnetisation
relaxation of elastomeric materials due to the complex origin of the relaxation function
itself [20, 36, 63-66], and the structural heterogeneity of rubbery materials.
One of the most important phenomena affecting the NMR line width, line position and
T2 relaxation decay in filled rubbers is the large heterogeneity of the static magnetic field
Bo across the sample due to magnetic susceptibility differences between rubbers and
fillers. Several authors have considered this effect [79-83]. Since polymer chains do not
diffuse throughout a large volume in the sample on the time scale of the NMR experiments,
line broadening is observed near the filler surface, and chains close to filler particles
experience a different resonance frequency. Spinning of samples at the magic angle (MAS)
to a large extent averages out the effect of Bo heterogeneity [84]. On the other hand, the
368
spinning causes an increase in the T2 relaxation of rubbery materials due to averaging of

the residual dipole-dipole interactions [20]. This effect could hamper quantitative analysis
of NMR line width with respect to the network density.
According to estimates of magnetic field gradients, a T2 relaxation time in the order of

milliseconds is predicted from magnetic field inhomogeneities introduced by carbon
black particles [85]. A T2 relaxation component with a much shorter decay time of
about 0.02-0.03 ms was observed in the case of various carbon-black-filled rubbers
[20, 62, 79]. This relaxation component was attributed to immobilised chain units
adjacent to the filler surface. It should be added that the results of these studies, which
were performed using low-field NMR equipment, were affected by the magnetic field
gradients to a lesser extent than results of high-field NMR experiments. It was concluded
that magnetic susceptibility inhomogeneities do not contribute significantly to the short
T2 values observed in the case of filled rubbers, which supports the above assumption
of chain immobilisation at the filler surface.
The presence of free radicals deriving from carbon black could also complicate the
interpretation of NMR data in the case of filled rubbers, because radicals may cause a
substantial decrease in T2. Two types of radicals have been detected in carbon-black-
filled rubbers: localised spins attributable to the carbon black and mobile spins deriving
from rubbery chains [86]. Mobile spins are formed because of the mechanical breakdown
of polymer chains when a rubber is mixed with carbon black. The concentration of
mobile spins increases linearly with carbon black loading [79, 87].
The amount of radicals in carbon black filled rubbers decreases significantly upon
extraction of free rubber with the aid of a solvent containing a free radical scavenger.
The extraction nevertheless causes a substantial increase in the fraction of the T2
relaxation component with the decay time of about 0.02-0.03 ms [62]. This increase is
apparently caused by an increase in the total rubber-carbon black interfacial area per
volume unit of the rubber due to the removal of free rubber. The T2 relaxation
component with a short decay time is also observed in poly(dimethyl siloxane) (PDMS)
filled with fumed silicas [88], whose particles contain a minor amount of paramagnetic
impurities. Apparently, free radicals hardly influence the interpretation of NMR data
obtained for carbon-black rubbers in any drastic way [62, 79].
10.5.2 Carbon-Black-Filled Rubbers
Several NMR relaxation studies using carbon-black-filled natural rubber (NR), EPDM
and butadiene (BR) rubbers have shown that a layer of immobilised, tightly bound
rubber is formed on the carbon black surface [20, 62, 79, 87, 89] (Figure 10.9).
369
Figure 10.9 A simplified graphic representation of EPDM chains at the carbon black
surface [62]. Monomer units with low mobility in the interface and mobile chain units
outside of interface are represented by solid and open points, respectively. The
rotational and translational mobilities of a few chain units next to the adsorption layer
along the chain (dashed points) are hindered somewhat more than those of the chain
units in the matrix. The chain fragments with low mobility in the interface provide
adsorption network junctions for the rubber matrix. At the bottom of the figure, the
spatial profile of the correlation time τc of the chain motion is schematically
represented as a function of the distance, r, from the carbon black surface. The τc is the
‘average’ time of a single reorientation of a chain unit
Reprinted with permission from V.M. Litvinov and P.A.M. Steeman, Macromolecules,
1999, 32, 25, 8476, Figure 5. Copyright 1999, American Chemical Society
370
At temperatures well above Tg, the mobility of chain units in the rubber-filler interface is
greatly hindered and is comparable with that in unfilled rubber at temperatures slightly
above the Tg [62, 79]. The degree of immobilisation is not greatly affected by the presence
of a good solvent [62, 90, 91], whereas chain mobility outside the interfacial layer increases
rapidly with an increasing solvent concentration [62, 91]. The mobility of short chain
loops at the carbon black surface in filled BR rubber is hindered to a greater extent than
that of chain bridges between carbon black particles [92]. The estimated thickness of the
interfacial layer is in the range of one to two diameters of the monomer unit of the
rubber (≈1 nm) [62, 79, 90].
It is suggested that the sites of interaction between carbon black and rubbery chains
cause physical (adsorption) network junctions in the rubber matrix [20, 62] (Figure 10.10).
The average molecular mass of network chains between adjacent adsorption junctions,
<Mad>, has been estimated for carbon-black-filled EPDM. The <Mad> value obtained for
Figure 10.10 Schematic representation of the physical network structure in a carbon-

black-filled elastomer [62]. The symbol … indicates elastomer - carbon black
adsorption junctions. The length scales in this figure and the EPDM/carbon black
volume ratio are fictional. For simplicity, none of the contacting carbon black
aggregates, which form agglomerates, have been included
371
the bound EPDM rubber was about 1800-2500 g/mol and was found to depend on
carbon black aggregation and on the content and type of filler [62]. The mean end-to-
end distance between adsorption junctions was comparable to the average distance
between adjacent carbon black aggregates [62]. This suggests that carbon black aggregates
in the bound rubber are interconnected by EPDM tie chains, and a continuous EPDM/
carbon black physical network is formed in the bound rubber. It was shown that the
physical EPDM-carbon black network is largely responsible for determining the modulus
at high strains [62].
The results obtained for unvulcanised EPDM and NR filled with carbon black provide
convincing evidence that the physical network has a ‘bimodal’ structure [62, 79]. Two
types of EPDM chains and/or chain fragments with widely differing densities of EPDM-
carbon black adsorption junctions are present in the rubbery matrix outside the EPDM-
carbon black interface (tightly bound rubber) (Figure 10.11) [62].
The first is an EPDM fraction which is loosely bound to the carbon black due to adsorption
interactions. This loosely bound rubber has numerous adsorption network junctions,
similar to those in bound rubber. The second EPDM fraction, consisting of extractable
rubber, contains a relatively small number of adsorption network junctions and can
apparently be extracted from the compounds. The fraction of loosely bound EPDM
chains determined with the aid of NMR increases with an increase in the maximum
possible EPDM-carbon black contact area per unit volume of the elastomer, regardless
the type of carbon black used, and is relatively close to the content of bound rubber [62].
Several authors have studied the effect of various structural characteristics of carbon
black on the rubber-filler interactions. They found that the interfacial fraction is larger
at rough carbon black surfaces with numerous edges and ledges [93]. This suggests that
edges and ledges are the active sites whose surface energy is quite a lot higher than that
of a smooth quasi graphitic scales of carbon black [94]. The strength of rubber-filler
interactions can be varied by modifying the carbon black surface by means of oxidation
or chemical treatment. Such treatments result in the formation of hydroxyl and acid
groups at the carbon black surface. The mobility of the NR chains at the modified carbon
black surface is greater than that at untreated surfaces, which suggests that such
modification results in a decrease in the number of active sites and the strength of the
adsorption, resulting in the observed increase in mobility [95-97]. Oxidation moreover
causes a change in the structure of the physical network in the rubbery matrix.
Selective 1H T2 relaxation experiments were used to determine specific sites of styrene-

butadiene (SBR) rubbery chains, which interact with the carbon black surface [84, 98,
99]. The T2 of all the different kinds of protons in SBR chains was found to consist of at
least two components attributable to a highly immobilised, tightly bound rubber around
the filler particles and an outer region of loosely bound and free rubber. The relative
372
Figure 10.11 The structure of an EPDM/N550 (phr=100) vulcanisate according to the

results of NMR and extraction studies - (A) and mechanical data for the case of pure
hydrodynamic reinforcement - (B) [62]. The volume fraction of microphases/
components is given in vol.%. According to the NMR data, the total network density
in the rubber phase, 1/2Mc+e+ad, equals 425 mmol/kg, where subscripts c, e and ad
stand for chemical crosslinks, chain entanglements and adsorption rubber-filler
junctions. The density of the adsorption junctions in the loosely bound rubber, 1/
2Mad, equals 17 mmol/kg
373
immobilisation of different types of chain units at the carbon black surface suggests that
double bonds of butadiene chain units are preferentially adsorbed by carbon black, unlike
the aromatic and methylene chain fragments. Advanced 2-dimensional NMR experiments
will have to be carried out to determine the conformational state of polymer chains at
the carbon black surface and the exact nature of the rubber-filler interactions.
Several T2 relaxation studies have shown that the fraction of rubber-filler interface and
the structure of the physical network is affected by dispersion of carbon black in the
rubbery matrix, which is improved with increasing the length of the mixing time [62,
100-104]. In the first stage of mixing, the fraction of the rubber-filler interface and the
density of rubber-filler physical junctions increase considerably with the length of the
mixing time. The mobility of rubber chains at the interface decreases rapidly upon mixing
[103,104]. In the second stage, the amount of interface increases at a slower rate, but the
chain mobility becomes independent of the mixing time. In the third stage, no appreciable
change is observed in the amounts of each component. The tightly bound rubber appears
to be formed very quickly when the rubber segments come into contact with a new active
carbon black surface, resulting in rapid growth of the loosely bound rubber component
around the rubber-filler interface.
10.5.3 Silica-Filled Silicon Rubbers
Physical interactions between PDMS and silica surfaces have been studied in close detail
using model compounds containing a large fraction of silica with a large specific surface
[88, 105-113]. Using 1H, 2H, 13C T1 and T2 relaxation experiments, unambiguous
information on the short-range (local) chain dynamics near the surface of hydrophilic
and silylated silicas was obtained in these studies. 1H and 2H spin-lattices or longitudinal
(T1) magnetisation relaxation experiments show two types of PDMS chain units with
widely differing local chain mobilities [88, 105, 109]. 1H T1 data obtained for PDMS
containing different amounts of a fumed silica are shown in Figure 10.12.
The frequency of the chain motions of one of the two types of chain units, inferred from
the minimum T1 around 195 K (–78 °C), is similar to that observed in unfilled PDMS.
The mobility of the chain units of the other type is greatly hindered, resulting in a T1
minimum at higher temperature (at about 7 °C). This minimum becomes more pronounced
in proportion to the total surface of filler particles in the mixture. It was suggested that
the observed chain immobilisation is caused mainly by physical adsorption of PDMS
chain units adjacent to the silica surface and to a lesser extent by entropy constraints at
the silica surface [113].
The estimated thickness of the PDMS-silica interface is about 1 nm [105, 109, 110],
which is comparable with that of carbon-black-filled conventional rubbers. The adsorption
374
Figure 10.12 Temperature dependence of the 1H T1 relaxation time of unfilled and

filled PDMS [113]. Experimental values are indicated by points and the contributions
to T1 deriving from the C3- rotation of CH3 groups [T1 minimum at ~ 100 K (-173
°C)], from chain motion in the PDMS matrix and at the PDMS-silica interface (T1
minima at ~ 195 K (-78 °C) and ~ 280 K (7 °C), respectively) are indicated by solid
lines. Also indicated is the volume fraction of hydrophilic silica (the specific surface
area is 286 m2 g-1) in mixtures. The experiments were performed on PDMS-silica
mixtures, which were prepared by adding silica to a dilute PDMS solution in CCl4
followed by solvent evaporation after prolonged stirring of the suspension. The
measurements were performed using amorphisised samples [113]
Reproduced with permission from V.M. Litvinov in Organosilicon Chemistry II. From
Molecules to Materials, Eds., N. Auner and J. Weis, VCH, Weinheim, Germany, 1996,
779, Figure 6. Copyright 1996, VCH Publishers
375
energy per single SiO(CH3)2 bond at the surface of hydrophilic silica is about 8-12 kJ/
mol [108]. It was further suggested that adsorption of polymer chains results in extended
chain conformation at the silica surface [113, 114]. The total energy per binding site,
which involves 10-15 hydrogen bonds between PDMS chain units and the surface of
hydrophilic silica, was estimated to be about 120 kJ/mol [115, 116].
13
C NMR T1 and T2 relaxation experiments using adsorbed PDMS at submolecular
coverage have shown that chain dynamics within the interface is dependent on the extent
of surface coverage [110, 111]. It was shown that the chain mobility increases as the
chain surface density increases. It was suggested that these changes reflect the
conformational state of the adsorbed chains. At a very small degree of surface coverage
the chains apparently adopt an extremely flat profile, with few loops or tails longer than
approximately 0.5-1.5 nm. As the surface density (degree of surface coverage) increases
above two monolayers, the polymer chain segments are driven from the surface into
loop and tail configurations. The degree of immobilisation at a fixed distance from the
surface is determined by the size of the loops and tails. The mobility of loops and tails
increases with increasing their length, because of a decrease in the number of constraints.
It was suggested that the backbone architecture is the key factor in restricting local chain
mobility at a surface. Polymer chains of greater stiffness would probably show both a
larger boundary region and a greater decrease in chain mobility at the filler surface.
The relative number of immobilised adsorbed chain units in a single type of silica is
proportional to Mn-0.5 [117]. The fraction of immobilised chain units is larger in hydroxyl-
terminated PDMS [117]. This suggests that the majority of hydroxyl chain ends are
linked to the silica surface due to the formation of double hydrogen bonds.
Contrary to carbon-black-filled conventional rubbers, which form a semi-rigid interface

at the carbon black surface, PDMS chain units at the silica surface are not rigidly linked
to the silica surface. Two types of dynamic processes are thought to occur at the interface:
relatively fast anisotropic reorientation of chain units in the interfacial layer and slow
adsorption-desorption of chain units (Figure 10.13) [108, 113].
When the temperature increases, both the fraction of less mobile, adsorbed chain units
and the lifetime of the chain units in the adsorbed state decrease. The lifetime of PDMS
chain units in the adsorbed state approaches zero at approximately –73 °C and 227 °C
at the surface of hydrophobic and silylated silicas, respectively [108, 113].
The frequency of PDMS chain mobility in the interfacial layer is largely dependent on
the type of silica surface. Water adsorption by the hydrophilic silicas causes the degree of
chain immobilisation in the interface to decrease [107]. Sililation of the silica surface
results in a significant decrease in the strength of adsorption interactions at the silica
376
Figure 10.13 Schematic representation of an adsorbed fragment of a PDMS chain

[113]. The solid arrows denote the anisotropic motion of chain units during the
lifetime in the adsorbed state. The open arrows denote the adsorption-desorption of
chain units. The length of the adsorption-desorption jumps is possibly in the order of
0.1-0.5 nm
Reproduced with permission from V.M. Litvinov in Organosilicon Chemistry II. From
Molecules to Materials, Eds., N. Auner and J. Weis, VCH, Weinheim, Germany, 1996,
779, Figure 8. Copyright 1996, VCH Publishers
surface, which causes a decrease in the fraction of the interface and a significant increase
in chain dynamics in the interfacial layer [107, 109, 113].
A good understanding of the structure of the network in filled rubbers is of great

importance, because the rubber’s elastic properties are determined primarily by the density
of chemical and physical network junctions and their ability to fluctuate. The following
types of network junctions occur in filled rubbers:
1) chemical crosslinks between rubbery chains and between rubbery chains and the filler
surface,
2) adsorption junctions at the filler surface,
3) temporary and trapped chain entanglements, and
4) topological hindrances of elastomeric chains due to the confinement of chains in a

restricted geometry or elastomer-filler entanglements [113, 118].
The relative contribution of each of these types of network junctions to the overall crosslink
density in silica-filled PDMS was estimated by means of 1H T2 relaxation experiments
[113, 118-121]. To determine the relative contributions of the different types of network
377
junctions to the overall network density, the relaxation experiments were performed using
original samples, swollen samples and samples swollen in the presence of ammonia. In the
case of the original samples all the different types of network junctions were found to
contribute to the overall network density. The effect of temporary chain entanglements
and topological constraints diminishes largely in the swollen samples. Since ammonia is
preferentially adsorbed onto the silica surface [113] its molecules cause desorption of
physically adsorbed PDMS chains. In this case the network density was determined mainly
by chemical crosslinks and trapped chain entanglements. Combined analysis of the T2 data
obtained for the same sample under three different conditions showed that adsorption
junctions and topological constraints are largely responsible for the overall crosslink density
in the case of silica-bound PDMS. The apparent number of elementary chain units between
the topological hindrances in bound rubber is estimated to be approximately 40-80.
As in carbon-black-filled EPDM and NR rubbers, the physical network in silica-filled

PDMS has a ‘bimodal’ structure [61]. A loosely bound PDMS fraction has a high density
of adsorption junctions and topological constraints. Extractable or free rubber does
virtually not interact with the silica particles. It was found that the density of adsorption
junctions and the strength of the adsorption interaction, which depends largely on the
temperature and the type of silica surface, largely determine the modulus of elasticity
and ultimate stress-strain properties of filled silicon rubbers [113].
10.5.4 Silica-Filled Conventional Rubbers
The use of silicas for reinforcing conventional rubbers is receiving ever more attention
due to the specific properties of the composites, in particular for their use in tyres [122,
123]. Hydrogen-bond interactions between surface silanol groups in silica agglomerates
are very strong, in particular in comparison with the interactions that take place between
the silanol groups of the silica and the commonly used non-polar olefinic hydrocarbon
rubbers. This makes it very difficult to mix silica with rubber. Chemical modification of
the silica surface using functional organosilanes enhances the compatibility of hydrocarbon
rubbers and silica, leading to marked improvements in the mechanical properties of
silica-reinforced rubber vulcanisates.
1
H NMR transverse magnetisation relaxation experiments have been used to characterise
the interactions between NR, isoprene rubber, BR, EPDM and polyethylacrylate rubbers
with hydrophilic silica and silicas modified with coupling agents [124-129]. These studies
showed that the physical interactions and the structures of the physical networks in
rubbers filled with carbon black and rubbers filled with silicas are very similar. In both
cases the principal mechanism behind the formation of the bound rubber is physical
adsorption of rubber molecules onto the filler surface.
378
The mobility in both tightly and loosely bound BR and isoprene rubbers increases, and
the fraction of bound rubber decreases with a decreasing concentration of silanol groups
on the silica surface [124]. This led to the suggestion that the silanol groups on the silica
surface are active sites for the chain adsorption. The grafting of aliphatic chains to the
silica surface leads to a decrease in BR-silica interactions [125]. The effect is less
pronounced in BR filled with carbon black containing aliphatic chains at the surface.
The number of immobilised rubbery chains on the silica surface seems to be dependent
on the coupling agent used and the grafting density. A significant number of immobilised
NR chains were found on the surface of hydrophilic silica and silica grafted with 3-
mercaptopropyl triethoxysilane (MPTES) [128, 129]. Strong NR-filler interactions were
observed at 100 °C even in the presence of a good solvent, suggesting that a quasi-
permanent rubber-silica network formed in the samples due to physical adsorption of
chain units at the silica surface. Mixtures of NR filled with silicas that were treated with
bis(triethoxysilyl propyl)tetrasulfide (TESPT) and triethoxysilyl benzothiazole (TESBT)
contain only a small amount of immobilised NR [128, 129]. Since the fraction of
immobilised chains is determined by the strength of the adsorption interactions and the
degree of dispersion of silica particles in the rubber matrix, NMR experiments using
well-defined model compounds will have to be carried out to obtain a good understanding
of the effect of silica treatment on rubber-filler interactions.
10.6 Chains Grafted onto a Filler Surface
Chain grafting is widely used to modify the surface of active fillers and to improve filler
dispersion in the matrix material. Fumed silicas are used in the reinforcement of rubbers
and the thickening of polymeric liquids. By varying the grafting density, its heterogeneity
and the average length of the grafted chains, the silica-silica interactions and interactions
between silica particles and the matrix material can be influenced to a large extent.
Therefore, knowledge of the molecular structure of the grafted layer is necessary for a
proper understanding of the molecular mechanisms responsible for the reinforcing and
thickening effects.
The molecular mobility of high-molecular-mass PDMS chains grafted to the silica surface
at one end has been studied with the aid of various solid-state NMR techniques [130-132].
The use of 2H NMR methods for this purpose is reviewed in the present book [133].
A PDMS layer chemically attached at both chain ends to the surface of hydrophilic silica
has been studied by means of 1H T2 relaxation experiments [112]. The average length of
the grafted chains in samples studied was varied from about 4 to 8 Si-O bonds. These
experiments showed that the T2 relaxation method is very sensitive to the heterogeneous
379
mobility of PDMS chains on a silica surface. Since molecular mobility is closely coupled
to chain length, grafting density and chain length distribution, T2 experiments can provide
valuable information on the structure of grafted layers.
The grafted PDMS layer was found to consist of immobilised chain segments at the
PDMS-silica interface and mobile chain portions outside the interface similar to that of
adsorbed PDMS. The interface fraction increases proportionally with a decreasing chain
length (Figure 10.14).
About four dimethylsiloxane pendent chain units next to the grafting site were immobilised
due to chain anchoring to the silica surface.
Figure 10.14 The fraction of the PDMS-silica interface, as determined by the fraction of
the T2 relaxation component (%T2in) with a decay time of about 80 μs, against the length
of the grafted chains, Nt, in silica samples swollen in C2Cl4 in the presence of NH3 [112].
The Nt value corresponds to the total length of any chains grafted at one end and the half-
length of chain loops and chains attached to neighbouring silica particles. The line
represents the result of a linear regression analysis of the data using the following
equation: %T2in = 1 – (Nt – Nad)/Nt, where Nad is the number of immobilised -SIO(CH3)2-
chain units at the PDMS-silica interface. The correlation coefficient equals 0.89
Reproduced with permission from V.M. Litvinov, H. Barthel and J. Weis, Macromolecules,
2002, 35. Copyright 2002, American Chemical Society
380
The fraction of immobilised PDMS chain units decreased slightly when the swollen samples
were treated with ammonia, which molecules cleft adsorption bonds between PDMS
chain units and silanol groups at the silica surface [113]. This suggests that a small
fraction of -SiO(CH3)2- chain units was immobilised as a result of physical adsorption at
the silica surface. The fraction of physically adsorbed chain units appeared to be
proportional to the number of residual silanol groups on the silica surface. The mobility
of the chain portions outside the interface was found to differ significantly in the different
samples studied, and to increase with an increasing average length of the grafted chains.
The NMR method allowed us to distinguish a grafted layer with a dense ‘brush-like’
structure containing chains of a fairly uniform length and a layer containing a significant
fraction of long chain loops.
10.7 Semi-Crystalline Elastomers
Rubbery chains in various rubbery materials, such as elastomeric polyolefins and EPDM
rubbers with a high ethylene content, form a crystalline micro-phase. Chain anchoring
to crystals leads to the formation of a physical network, with the crystallites acting as
multifunctional crosslink junctions [134-136]. Above the Tg, the mean chain length
between anchoring points to the crystalline phase largely determines the T2 relaxation
time of the amorphous phase [136-139]. The effect of chain confinement of amorphous
segments by crystallites moreover induces additional constraints of segmental motions
in the amorphous phase [139]. At a low degree of crystallinity, T2 relaxation experiments
can provide information on the mean molecular mass of polymer chains interconnecting
adjacent crystals [136, 138].
Wide-line 1H and 2H NMR spectra and T2 relaxation experiments have been used to
determine the composition of the phases in semi-crystalline polymers [133, 136, 138-
144]. The experiments were also used to obtain real-time information on the kinetics of
crystallisation and melting [143-148]. The use of high-resolution NMR methods to
characterise semi-crystalline polymers is reviewed elsewhere [17, 18, 30, 34, 149].
The T2 data obtained for semi-crystalline polymers are often analysed with the aid of a
three-phase model comprising a crystalline phase, a semi-rigid interphase2 and a soft
2
The degree of molecular mobility in the interphase lies between that in the rigid crystalline phase and
that in soft amorphous phase. The fraction of the intermediate phase was found to depends on the
employed technique, the temperature and the method of data evaluation. In many respects, the
intermediate phase has a kinetic origin and may not be regarded as a true thermodynamical phase. It
would apparently be more correct to define the third phase as an interface or a semi-rigid fraction of
the amorphous phase.
381
amorphous phase. Good correlation was observed between crystallinity measured by

means of NMR and with the aid of other methods [140-144]. The fraction of highly
disordered nano-size crystals can also be determined with the aid of the T2-relaxation
method [138]. In the case of elastomeric materials with a low degree of crystallinity a
simple two-phase model can yield a good description of the T2 relaxation behaviour, as
can be seen in Figure 10.15 [136, 138]. This figure shows that the 1H T2 decay for
ethylene/1-octene copolymer can be well described with two components, whose
characteristic decay time is typical for crystalline and amorphous phases.
It appears in this case that the two-phase model may yield reliable results, because
1) the mobility of chains adjacent to the surface of small crystals is greater than that at
the surface of large crystals due to the larger disorder at the surface of small crystals
and crystal disorder that promotes chain mobility, and
Figure 10.15 The decay of the transverse magnetisation (points) for ethylene-octene
copolymer at different temperatures [136]. The decay was measured using the solid-
echo pulse sequence. The solid lines represent the result of a least-squares adjustment
of the decay using a linear combination of Weibull and exponential functions. The
dotted lines represent the relaxation component with a long decay time. In the
experiments the sample was heated from room temperature to 343 K (70 °C)
Reproduced with permission from V.M. Litvinov, Macromolecules, 2001, 34, 24, 8468.
Copyright 2001, American Chemical Society
382
2) the broad distribution of the correlation times of chain motions in the amorphous
phase can cause ‘apparent’ single exponential relaxation.
10.8 Ionic Viscoelastic Materials
Polymers with small concentrations of ion-containing groups (up to about 15 mol%) are
known as ionomers, as opposed to polymer electrolytes, which contain larger amounts
of ionic groups. Elastomers containing ionic groups can be used as thermoplastic rubbers.
The presence of ionic groups in polymer chains greatly affects their physical properties
[150-152]. Ionic aggregation causes a marked increase in viscosity and the strength of
such polymer materials. The main mechanism responsible for the increase in strength is
dependent on strong electrostatic interactions between ionic groups, resulting in the
formation of immobilised ionic multiplets and their aggregates and clusters. Polymer
chains interconnect multiplets, resulting in the formation of a physical network
(Figure 10.16) [153].
Figure 10.16 A schematic representation of the morphology of a telechelic ionomer

where (1) represents the ion-containing core, (2) an immobilised interfacial layer and
(3) a phase containing network chains, dangling chains and free chains [153]
Reprinted with permission from V.M. Litvinov, A.W.M. Braam and A.F.M.J. van der Ploeg,
Macromolecules, 2001, 34, 3, 489, Figure 17. Copyright 2001 American Chemical Society
383
Multiplets are supposed to contain a relatively small number of ionic pairs and consist of
ionic material only, whereas clusters are large structures that are rich in ion pairs and
contain a considerable quantity of an organic material.
Recent proton NMR relaxation studies have shown that the method can be used to study
the formation of ionic clusters and the structure of the physical network in ionomers [153].
10.9 Rubbery Phases in Blends and Emulsions
The most comprehensive information on the phase structure and molecular mobility in
polymer materials can be gained with high-resolution NMR techniques [17, 18, 30, 31,
34]. Low-resolution NMR techniques may offer some advantages in the following cases:
1) characterisation of phase structure and network density in polymers with a complex

chemical composition that result in substantial overlapping of resonances from different
chemical groups, hampering the interpretation of high-resolution NMR data,
2) fast quantitative analysis of phase composition, phase transition and network structure,
and
3) when accurate control of sample temperature is required, which can be easily realised
with the aid of low-field NMR equipment, allowing analysis of phase transitions.
Proton NMR magnetisation relaxation experiments are often used to determine microphase
structure and molecular motions in polymer materials containing rubbery phases [17].
The results of proton NMR transverse magnetisation relaxation experiments for
heterogeneous materials are easier to interpret than those of T1 and T1ρ (spin-lattice
relaxation in the rotating frame) experiments, because 1H T2 data are not complicated by
the effect of spin-diffusion [17, 18]. Moreover, the T2 relaxation process is very sensitive to
even small changes in molecular mobility, as can be inferred from the T2 range from about
10 μs observed for glassy and crystalline phases/components to a few seconds measured
for low-molecular-mass liquids. This great difference in T2 shows that this experiment is
highly sensitive to changes in the chemical structure, composition and morphology of
polymer materials. However, care should be taken in analysing T2 relaxation decay curves
with respect to the different phases/components contained in the polymer (see Section
10.3).
Analysis of 1H T2 relaxation as a function of temperature yields information on molecular

motions. Side-groups and local chain motions (secondary relaxations) cause a change in
T2 below the Tg. The value of T2 increases with the amplitude and the frequency of molecular
motions. The glass transition that occurs in the time scale in the NMR T2 relaxation
384
experiments, i.e., 10-100 μs, causes a sharp increase in T2. The Tg measured in such NMR
experiments is comparable with the dynamic Tg measured by means of DMA at frequencies
of about 10-50 kHz and is usually observed at a temperature 30-50 °C above Tg measured
with differential scanning calorimetry.
T2 experiments are used to determine the concentration of (inter)phases/components in

polymer materials if polymer chains in these (inter)phases/components reveal a significant
difference in molecular mobility [17, 34]. In such cases, the T2 relaxation function is the
weighted sum of the T2 decays of different components/phases. The relative fraction of
these components is proportional to the concentration of hydrogen in these (inter)phases/
components (see Sections 10.4 and 10.7). The characteristic decay time, T2, is related to
molecular mobility in different phases.
Several examples of such studies are given in this section. The 1H T2 relaxation method has
been used to determine the crosslink density in the rubbery phases in (styrene)-b-(butadiene)-
b-(styrene) tri-block copolymers (SBS) [154, 155], BR rubber in polystyrene (PS) used to
improve impact strength [156] and acrylic rubber in impact-modified a styrene acrylonitryl
copolymer (SAN) [157]. It was found that a high degree of toughness can be obtained only
if the rubber phase has a low crosslink density, resulting in low cavitation resistance of the
rubber particles under strain [157].
1
H T2 experiments have also been used to see whether they are suitable for analysing
interfaces between different phases in polymer blends, such as SBS block copolymers [158]
and (methylmethacrylate)-b-(n-butylacrylate) block copolymers [159].
The distribution of crosslinks in blends of different rubbers has been determined by means
of line-width analysis of 1H high-resolution NMR spectra obtained using swollen samples
[33,160-166]. It should be noted that the results of these experiments suffer from systematic
errors, since the line width of the NMR spectra is affected not only by the crosslink density,
but also by differences in the magnetic susceptibility of the different phases [79-83] and
even by the effects of slow sample rotation [20].
1
H T2 relaxation experiments have moreover been used to investigate the phase
composition, interfaces and phase diagrams of polymer solutions, emulsions and core-
shell lattices [167-173].
10.10 Real-Time NMR Experiments

Two practical aspects are of importance in vulcanising rubber compounds:
1) the crosslinking kinetics, which determine the production cycle time or throughput,
and
385
2) the crosslink density, which partly determines the mechanical and elastic properties
of the vulcanisate.
The crosslinking kinetics and the final state of cure are commonly studied with the aid of
rheometers. NMR relaxation experiments can offer several advantages for the
characterisation of the crosslinking kinetics in complex materials because of high method
selectivity with respect to the rubbery chains/phases in polymer blends, filled and oil
extended rubbers.
Network structures have been quantitatively determined by means of real-time 1H NMR

T2 relaxation experiments for several polymers [174-178]. The effect of the curing
conditions on sol and gel fractions and the spatial heterogeneity of the network structure
has been studied for polyethylene [174], polyacrylamide [175], PDMS [176], BR [177],
epoxy resins [178] and EPDM [179].
A change in 1H T2 relaxation during accelerated sulfur vulcanisation of EPDM rubber is

shown as an example in Figure 10.17 [179].
Figure 10.17 Dependency of T2 relaxation time as a function of the actual

vulcanisation time during accelerated sulfur vulcanisation of EPDM [179]. The line
has been included to guide the eye. The vulcanisation temperature was 413 K (140 °C)
Reproduced with permission from V. Litvinov and M. van Duin, Kautschuk und Gummi
Kunststoffe, 2002, 55. Copyright 2002, Huthig Verlag
386
The time it takes for the temperature of a sample to stabilise in an NMR probe can also
be determined in real-time 1H NMR T1 relaxation experiments, because at vulcanisation
temperatures T1 is only slightly affected by a moderate crosslink density [22, 180, 181].
The T2 data obtained in the aforementioned experiments were used for determining an
increase in the density of chemical crosslinks upon vulcanisation time (Figure 10.18)
[179]. The method can also be used in kinetic studies of the vulcanisation of filled and
oil-extended rubbers.
Figure 10.18 Dependency of the chemical crosslink density (1/2Mc) on the actual
vulcanisation time during accelerated sulfur vulcanisation of EPDM [179]. The line
has been included to guide the eye. The vulcanisation temperature was 413 K (140 °C)
Reproduced with permission from V. Litvinov and M. van Duin, Kautschuk und Gummi
Kunststoffe, 2002, 55. Copyright 2002, Huthig Verlag
10.11 Low-Field NMR Magnetisation Relaxation Experiments for

Quality Control Purposes
The high selectivity of solid-state NMR with respect to different phases/components in

polymer materials can offer several advantages for quality tests over other analytical
techniques. Although access to information on the use of NMR for quality-control
purposes is rather limited, several applications for fast analysis of rubbery materials
have been published, for example for determining the concentration of oil in rubbers
387
[182, 183], the distribution of carbon black in rubber matrices [100,101], the solids
content of rubber lattices [184], the viscosity of oil [185], the crystallinity of polyolefins
[186], molecular masses and melt flow indices [186] and the component contents of
copolymers and polymer blends [186]. The method is also used for in-process analysis in
polymer industry [186, 187].
The NMR method was found to offer the following advantages for quality tests:
• It yields quantitative data on network density and phase composition
• It shows a high selectivity with respect to phases/components in blends
• Mineral fillers do not interfere with the results
• Samples do not have to meet any special requirements; swollen gels, solid materials,
emulsions, suspensions, granules and powders can all be analysed
• The method measures volume-average properties
• It yields results quickly
• It is environmental friendly because small amounts of chemicals are used.
Quality tests are usually performed using bench-top, low filed NMR spectrometers.
Volume-average properties are determined with this equipment. Surface layer of samples
can be analysed using recently developed NMR-MOUSE (mobile universal surface
explorer) [26, 188]. The NMR-MOUSE is a relatively small NMR device suited for the
investigation of surface-near volume elements. Lateral surface heterogeneity of elastomeric
materials can be scanned with this device. Possible applications of the NMR-MOUSE
for the characterisation of rubbery materials were demonstrated [26,189-191].
10.12 Conclusions
We may safely conclude that proton NMR magnetisation relaxation experiments are a valuable
tool for characterising various types of viscoelastic materials and rubbery phases in blends,
emulsions and core-shell lattices. The NMR method can be used to quantitatively analyse
the structure of chemical and physical networks as well as network defects. The great advantage
of the NMR method over traditional macroscopic methods is due to the high selectivity of
the NMR experiments with respect to different (inter)phases/components in materials intended
for practical use. This information is essential for obtaining a better understanding of structure-
property relationships. NMR relaxation experiments may also be more advantageous in
routine quality control than traditional methods.
388
References
1. P.J. Flory, Principles of Polymer Chemistry, Cornell University Press, Ithaca, NY,
USA, 1953.
2. L.R.G. Treloar, The Physics of Rubber Elasticity, 3rd Edition, Clarendon Press,
Oxford, UK, 1975.
3. J.D. Ferry, Viscoelastic Properties of Polymers, 3rd Edition, Wiley, New York,
NY, USA, 1980.
4. J.E. Mark and B. Erman, Rubberlike Elasticity. A Molecular Primer, Wiley, New
5. J.F. Rabek in Radiation Curing in Polymer Science and Technology:

Fundamentals and Methods, Eds., J.P. Fouassier and J.F. Rabek, Elsevier Applied
Science, London, UK, 1993, 1, 329.
6. D.J.P. Harrison, W.R. Yates and J.F. Johnson, Journal of Macromolecular Science
C, 1985, 25, 4, 481.
7. G. Heinrich, E. Straube and G. Helmis, Advances in Polymer Science, 1988, 85,

33.
8. T.A. Vilgis and G. Heinrich, Die Angewandte Makromolekulare Chemie, 1992,

202/203, 243.
9. T.A. Vilgis and G. Heinrich, Kautschuk und Gummi Kunststoffe, 1992, 45, 12,
1006.
10. K-H. Schimmel and G. Heinrich, Colloid and Polymer Science, 1991, 269, 1003.
11. T.A. Vilgis and G. Heinrich, Makromolekulare Theory and Simulations, 1994, 3,
2, 271.
12. A.A. Dias, H. Hartwig and J.F.G.A. Jansen, Surface Coatings International, 2000,
83, 382.
13. D.R. Bauer, Progress in Organic Coatings, 1986, 14, 45.
14. J.L. Koenig, Journal of Applied Polymer Science, 1998, 70, 14, 1359.
15. W. Gronski, U. Hoffman, G. Simon, A. Wutzler and E. Straube, Rubber

389
16. N.G. McCrum, B.E. Read and G. Williams, Anelastic and Dielectric Effects in
Polymeric Solids, Wiley, New York, NY, USA, 1967.
17. V.J. McBrierty and K. Packer, Nuclear Magnetic Resonance in Solid Polymers,
Cambridge University Press, Cambridge, UK, 1993.

Polymers, Academic Press, London, UK, 1994.
19. P.T. Callaghan, Principles of Nuclear Magnetic Resonance Microscopy,

Clarendon Press, Oxford, UK, 1991.
20. J.P. Cohen Addad, Progress in NMR Spectroscopy, 1993, 25, 1.
21. Annual Reports on NMR Spectroscopy, Volume 34, Eds., G.A. Webb and I.
Ando, Academic Press, San Diego, CA, USA, 1997.
23. R.A. Kinsey, Rubber Chemistry and Technology, 1990, 63, 3, 407.
24. P. Blümler and B. Blümich in NMR Basic Principles and Progress, Eds., P. Diehl,
E. Fluck, H. Günther, R. Kosfeld and J. Seelig, Springer-Verlag, Berlin, Germany,
1994, Volume 30.
25. P. Adriaensens, A. Pollaris, D. Vanderzande, J. Gelan, J. White, A.J. Dias and M.

Kelchtermans, Macromolecules, 1999, 32, 14, 4692.
26. B. Blümich and D.E. Demco in Spectroscopy of Rubbery Materials, Eds., V.M.
Litvinov and P.P. De, Rapra Technology, Shawbury, UK, 2002, Chapter 7.
27. D.W. McCall, Accounts of Chemical Research, 1971, 4, 6, 223.
28. J. Schaefer, E.O. Stejskal and R. Buchdahl, Macromolecules, 1977, 10, 2, 384.
29. A.A. Parker, J.J. Marcinko, P. Rinaldi, D.P. Hedrick and W.M. Ritchey, Journal of
Applied Polymer Science, 1993, 48, 4, 677.
30. High Resolution NMR Spectroscopy of Synthetic Polymers in Bulk, Ed., R.A.
Komoroski, VCH, Deerfield Beach, FL, USA, 1986.
31. D.E. Demco, S. Hafner and H.W. Spiess in Spectroscopy of Rubbery Materials, Eds.,
V.M. Litvinov and P.P. De, Rapra Technology, Shawbury, UK, 2002, Chapter 14.
390
32. D.D. Parker and J. Koenig in Spectroscopy of Rubbery Materials, Eds., V.M.
33. A.K. Whittaker in Spectroscopy of Rubbery Materials, Eds., V.M. Litvinov and
P.P. De, Rapra Technology, Shawbury, UK, 2002, Chapter 13.
34. NMR Spectroscopy of Polymers, Ed., R.N. Ibbett, Blackie Academic &
Professional, London, UK, 1993.
35. N.J. Clayden in NMR Basic Principles and Progress, Volume 29, Eds., P. Diehl,
E. Fluck, H. Günther, R. Kosfeld and J. Seelig, Springer-Verlag, Berlin, Germany,
1993,
36. J.P. Cohen Addad in Spectroscopy of Rubbery Materials, Eds., V.M. Litvinov and
P.P. De, Rapra Technology, Shawbury, UK, 2002, Chapter 8.
37. J.P. Cohen Addad, Journal de Physique (Paris), 1982, 43, 10, 1509.
38. J.P. Cohen Addad and C. Schmit, Polymer, 1988, 29, 5, 883.
39. R. Kimmich, G. Schnur and M. Köpf, Progress in NMR Spectroscopy, 1988, 20,
385.
40. M.G. Brereton, I.M. Ward, N. Boden and P. Wright, Macromolecules, 1991, 24,
8, 2068.
41. W.G. Hiller, H. Schneider and V.D. Fedotov, Journal of Polymer Science: Polymer
Physics, 1992, 30, 9, 931.
42. R. Kimmich, G. Schuz and M. Köpf, Progress in NMR Spectroscopy, 1988, 20,
385.
43. J.P. Cohen Addad, L. Pellicioli, and J.J.H. Nusselder, Polymers Gels and
Networks, 1997, 5, 201.
44. V.M. Litvinov and J.J.H. Nusselder, unpublished results.
45. Yu.Ya. Gotlib, M.I. Lifshits, V.A. Shevelev, I.A. Lishanskii and I.V. Balanina,
Polymer Science USSR, 1976, 18, 10, 2630.
46. C.G. Fry and A.C. Lind, Macromolecules, 1988, 21, 5, 1292.
47. Ye.R. Gasilova and V.A. Shevelev, Polymer Science USSR, 1986, 31, 7, 1683.
391
48. D. Geschke and K. Pöschel, Colloid and Polymer Science, 1986, 264, 482.
49. V.M. Litvinov, W. Barendswaard and M. van Duin, Rubber Chemistry and
Technology, 1998, 71, 1, 105.
50. G. Simon, K. Baumann and W. Gronski, Macromolecules, 1992, 25, 14, 3624.
51. U. Hoffmann, W. Gronski, G. Simon and A. Wutzler, Die Angewandte

Makromolekulare Chemie, 1992, 202/203, 283.
52. V.M. Litvinov and A.A. Dias, Macromolecules, 2001, 34, 12, 4051.
1996, 29, 19, 6222.
54. R. Graf, D.E. Demco, S. Hafner and H.W. Spiess, Solid State Nuclear Magnetic
Resonance, 1998, 12, 2-3, 139.
55. E. Fischer, F. Grinberg and R. Kimmich, Journal of Chemical Physics, 1998, 109,
2, 846.
56. P.T. Callaghan and E.T. Samulski, Macromolecules, 1997, 30, 1, 113.
57. R.C. Ball, P.T. Callaghan and E.T. Samulski, Journal of Chemical Physics, 1997,
106, 17, 7352.
60. M. Garbarczyk, F. Grinberg, N. Nestle and W. Kuhn, Journal of Polymer Science:

Part B Polymer Physics, 2001, 39, 18, 2207.
61. V.M. Litvinov, International Polymer Science and Technology, 1988, 15, T/28.
62. V.M. Litvinov and P.A.M. Steeman, Macromolecules, 1999, 32, 25, 8476.
63. M.G. Brereton, Macromolecules, 1990, 23, 4, 1119.
64. M.G. Brereton, Macromolecules, 1991, 24, 8, 2068.
65. J.P. Cohen Addad and O. Girard, Macromolecules, 1992, 25, 2, 593.
66. T.P. Kulagina, V.M. Litvinov and K.T. Summanen, Journal of Polymer Science:
Part B Polymer Physics, 1993, 31, 3, 241.
392
67. A. Charlbesy in Radiation Effects on Polymers, Eds., R.L. Clough and S.W.
Shalaby, ACS Symposium Series, Washington, USA, 1991, 475, 193.
68. H.W. Weber and R. Kimmich, Macromolecules, 1993, 26, 10, 2597.
69. R. Kimmich, M. Köpf and P.T. Callaghan, Journal of Polymer Science: Part B
70. G.I. Sandakov, V.P. Tarasov, N.N. Volkova, Y.A. Ol’khov, L.P. Smirnov, L.N. Erofeev
and A.K. Khitrin, Vysokomolekulyarnye Soedineniya, Seria B, 1989, 31, 821.
71. N.N. Volkova, G.I. Sandakov, A.I. Sosikov, Y.A. Ol’khov, L.P. Smiznov and K.T.
Summanen, Polymer Science USSR, 1992, 34, 2, 127.
72. V.M. Litvinov and K.T. Summanen, to be published.
73. S. Jahromi, V.M. Litinov and B. Coussens, Macromolecules, 2001, 34, 4, 1013.
74. V.M. Litvinov, to be published.
75. D.S. Pearson, Rubber Chemistry and Technology, 1987, 60, 3, 439.
76. T.A. Kavassalis and J. Noolandi, Macromolecules, 1989, 22, 6, 2709.
77. S.P. Obukhov, M. Rubinstein and R.H. Colbe, Macromolecules, 1994, 27, 12,
3191.
78. G. Kraus, Advances in Polymer Science, 1971, 8, 155.
79. V.J. McBrierty and J.C. Kenny, Kautschuk und Gummi Kunststoffe, 1994, 47, 5,
342.
80. J. Jeener, B.H. Meier, P. Bachmann and R.R. Ernst, Journal of Chemical Physics,
1979, 71, 11, 4546.
81. A.P.M. Kentgens, W.S. Veeman and J. van Bree, Macromolecules, 1987, 20, 6,
1234.
82. J. Roland and D. Michel, Magnetic Resonance in Chemistry, 2000, 38, 587.
83. P. Ekanayake, H. Menge, H. Schneider, M.E. Ries, M.G. Brereton and P.G. Klein,
84. N.K. Dutta, N.R. Choudhury, B. Haidar, A. Vidal, J.B. Donnet, L. Delmotte and
J.M. Chezeau, Polymer, 1994, 35, 20, 4293.
393
85. E.M. Cashell, D.C. Douglass and V.J. McBrierty, Polymer Journal, 1978, 10, 5,
557.
86. E.M. Cashell and V.J. McBrierty, Journal of Materials Science, 1977, 12, 2011.
87. J. O’Brien, E. Cashell, G.E. Wardell and V.J. McBrierty, Macromolecules, 1976,
9, 4, 653.
88. V.M. Litvinov and A.A. Zhdanov, Polymer Science USSR, 1987, 29, 5, 1133.
89. S. Kaufman, W.P. Slichter and D.D. Davis, Journal of Polymer Science: Part A2,
1971, 9, 829.
90. T. Nishi, Rubber Chemistry and Technology, 1978, 51, 5, 1075.
91. T. Nishi, Journal of Polymer Science: Polymer Physics Edition, 1974, 12, 4, 685.
92. J.P. Cohen Addad and P. Frébourg, Polymer, 1996, 37, 19, 4235.
93. W. Niedermeier and B. Freund, Kautschuk und Gummi Kunststoffe, 1999, 52,
10, 670.
94. J-B. Donnet, Kautschuk und Gummi Kunststoffe, 1994, 47, 9, 628.
95. H. Serizawa, T. Nakamura, M. Ito, K. Tanaka and A. Nomura, Polymer Journal,

1983, 15, 3, 201.
96. H. Serizawa, T. Nakamura, M. Ito, K. Tanaka and A. Nomura, Polymer Journal,

1983, 15, 7, 543.
97. S. Asai, H. Kaneki, M. Sumita and K. Miyasaka, Journal of Applied Polymer

Science, 1991, 43, 7, 1253.
98. R. Mansencal, B. Haidar, A. Vidal, L. Delmotte and J-M. Chezeau, Polymer

International, 2001, 50, 387.
99. N. K. Dutta, N.R. Choudhury, B. Haidar, A. Vidal, J-B. Donnet, L. Delmotte and
J.M. Chezeau, Rubber Chemistry and Technology, 2001, 74, 2, 260.
100. G.E. Wardell and V.J. McBrierty, inventors; The Provost, Fellows and Scholars of
the College of the Holy and Undivided Trinity of Queen Elizabeth, Dublin,
Ireland, assignee; US 4, 301,411, 1981.
394
101. G.E. Wardell and V.J. McBrierty, inventors; The Provost, Fellows and Scholars of
the College of the Holy and Undivided Trinity of Queen Elizabeth, Dublin,
Ireland, assignee; GB2 043 262 A, 1982.
102. G.E. Wardell and V.J. McBrierty, Rubber Chemistry and Technology, 1982, 55,
4,1095.
103. M. Furuta, T. Hikasa and E. Kato, Journal of Applied Polymer Science, 1986, 31,
7, 2325.
104. H. Serizawa, M. Ito, T. Kanamoto, K. Tanaka and A. Nomura, Polymer Journal,

1982, 14, 2, 149.
105. V.M. Litvinov and A.A. Zhdanov, Doklady Physical Chemistry, 1985, 283, 4-6,
811.
916.
107. V.M. Litvinov, M. Wobst, D. Reichert, H. Schneider and A.A. Zhdanov, Acta
Polymerica, 1988, 39, 5, 244.
108. V.M. Litvinov, Polymer Science USSR, 1988, 30, 10, 2250.
109. V.M. Litvinov and H.W. Spiess, Makromolekulare Chemie, 1991, 192, 12, 3005.
110. J. van Alsten, Macromolecules, 1991, 24, 19, 5320.
111. D.G. Rethwisch, J. van Alsten and C.R. Dybowski, Macromolecular Symposia,
1994, 86, 171.
112. V.M. Litvinov, H. Barthel and J. Weiss, Macromolecules, 2002, 35, 11, 4356.
113. V.M. Litvinov in Organosilicon Chemistry II. From Molecules to Materials, Eds.,
N. Auner and J. Weiss, VCH, Weinheim, Germany, 1996, 779.
114. V.L. Fernandez, J.A. Reimer and M.M. Denn, Journal of the American Chemical
Society, 1992, 114, 24, 9634.
115. J.P. Cohen Addad, C. Roby and M. Sauviat, Polymer, 1985, 26, 8, 1231.
116. J.P. Cohen Addad, P. Huchot, P. Jost and A. Pouchelon, Polymer, 1989, 30, 1, 143.
395
117. J.P. Cohen Addad and R. Ebengou, Polymer, 1992, 33, 2, 379.
118. V.M. Litvinov and V.G. Vasilev, Polymer Science USSR, 1990, 32, 11, 2231.
759.
120. V.M. Litvinov and A.A. Zhdanov, Polymer Science USSR, 1988, 30, 5, 1000.
121. V.M. Litvinov and H.W. Spiess, Makromolekulare Chemie, 1992, 193, 5, 1181.
122. M.P. Wagner, Rubber Chemistry and Technology, 1976, 49, 3, 703.
123. S. Wolff, Tire Science and Technology, 1987, 15, 4, 276.
124. M. Ito, T. Nakamura and K. Tanaka, Journal of Applied Polymer Science, 1985,
30, 8, 3493.
125. A.P. Legrand, N. Lecomte, A. Vidal, B. Haidar and E. Papier, Journal of Applied
Polymer Science, 1992, 46, 12, 2223.
126. S. Ono, Y. Kiuchi, T. Sawanobozi and M. Ito, Polymer International, 1999, 48,
1035.
127. S. Ono, M. Ito, H. Tokumitsu and K. Seki, Journal of Applied Polymer Science,
1999, 74, 10, 2529.
128. J.W. ten Brinke, V.M. Litvinov, J.E.G.J. van Wijnhoven and J.W.M. Noordermeer,
Proceedings of the International Rubber Conference, Birmingham, UK, 2001, 426.
129. J.W. ten Brinke, V.M. Litvinov, J.E.G.J. van Wijnhoven and J.W.M. Noordermeer,
Macromolecules, submitted.
130. M. Zeghal, P. Auroy and B. Deloche, Physical Review Letters, 1995, 75, 2140.
131. M. Zeghal, B. Deloche and P. Auroy, Macromolecules, 1999, 32, 4947.
132. M. Zeghal, B. Deloche, P-A. Albouy and P. Auroy, Physical Review E, 1997, 56,
603.
133. B. Deloche and P. Sotta in Spectroscopy of Rubbery Materials, Eds., V.M.

396
134. S. Bensason, E.V. Stepanov, S. Chum, A. Hiltner and E. Baer, Macromolecules

1997, 30, 8, 2436.
135. Yu.K. Godovsky and N.P. Bessonova, Polymer Science A, 2001, 43, 8, 880.
136. V.M. Litvinov, Macromolecules, 2001, 34, 24, 8468.
137. V.J. McBrierty and D.C. Douglass, Journal of Polymer Science: Macromolecular
Reviews, 1981, 16, 295.
138. W. Barendswaard, V.M. Litvinov, F. Souren, R.L. Scherrenberg, C. Gondard and

C. Colemonts, Macromolecules, 1999, 32, 1, 167.
139. L. Dujourdy, J.P. Bazile and J.P. Cohen Addad, Polymer International, 1999, 48,
558.
140. R. Kitamaru, F. Horii and S-H. Hyon, Journal of Polymer Science: Polymer
Physics Edition, 1977, 15, 5, 821.
141. V.D. Fedotov and N.A. Abdrashitiva, Polymer Science USSR, 1985, A27, 2, 287.
142. R.R. Eckman, P.M. Henrichs and A.J. Peacock, Macromolecules, 1997, 30, 8,
2474.
143. E.W.Hansen, P.E.Kristiansen and B.Pedersen, Journal of Physical Chemistry B,

1998, 102, 28, 5444.
144. E.W.Hansen, P.E.Kristiansen and B.Pedersen, Journal of Physical Chemistry B,

1999, 103, 18, 3552.
145. V.M. Litvinov, B.D. Lavrukhin, A.A. Zhdanov and K.A. Andrianov, Polymer
Science USSR, 1976, A18, 11, 2515.
Science USSR, 1976, A20, 11, 2758.
147. P.E. Kristiansen, E.W. Hansen and B. Pedersen, Polymer, 2000, 41, 1, 311.
148. P.E. Kristiansen, E.W. Hansen and B. Pedersen, Polymer, 2001, 42, 5, 1969.
149. A.E. Tonelli, M.A. Gomez, H. Tanaka and M.H. Cozine in Solid State NMR of
Polymers, Ed., L.J. Mathias, Plenum Press, New York, NY, USA, 1991, 81.
397
150. Ionomers. Characterization, Theory, and Applications, Ed., S. Schlick, CRC

Press, Boca Raton, FL, USA, 1995.
151. W.J. MacKnight and R.D. Lundberg in Thermoplastic Elastomers, 2nd Edition,
Eds., G. Holden, N.R. Legge, R.P. Quirk and H.E. Schroeder, Hanser Publishers,
Munich, Germany, 1996, 271.
152. A. Eisenberg and J-S. Kim, Introduction to Ionomers, John Wiley and Sons, New
153. V.M. Litvinov, A.W.M. Braam and A.F.M.J. van der Ploeg, Macromolecules,
2001, 34, 3, 489.
154. K. Bergmann and K. Gerberding, Colloid and Polymer Science, 1981, 259, 10,
990.
155. C. Fülber, D.E. Demco, O. Weintraub and B. Blümich, Macromolecular

Chemistry and Physics, 1996, 197, 2, 581.
156. E.R. Gasilova, V.A. Shevelev, L.I. Ginzburg and E.M. Tarkova, Polymer Science
A, 1993, 35, 3, 324.
157. A.C. Steenbrink, V.M. Litvinov and R.J. Gaymans, Polymer, 1998, 39, 20, 4817.
158. K. Fukumori, T. Kurauchi and O. Kamigaito, Journal of Applied Polymer

Science, 1989, 38, 7, 1313.
159. R. Soltani, F. Lauprêtre, L. Monnerie and P. Teyssie, Polymer, 1998, 39, 15,
3297.
160. M.J.R. Loadman and A.J. Tinker, Rubber Chemistry and Technology, 1989, 62,
2, 234.
161. P.S. Brown, M.J.R. Loadman and A.J. Tinker, Rubber Chemistry and Technology,
1992, 65, 4, 744.
162. A.J. Tinker, Rubber Chemistry and Technology, 1995, 68, 3, 461.
163. V.A. Shershnev, I.K. Shundrina, V.D. Yulovskaya and I.A. Vasilenko, Polymer
Science A, 1993, 35, 10, 1428.
164. P.S. Brown and A.J. Tinker, Kautschuk und Gummi Kunststoffe, 1995, 48, 9, 606.
165. S. Cook, Kautschuk und Gummi Kunststoffe, 1999, 52, 5, 350.
398
166. S. Groves, Rubber Chemistry and Technology, 1999, 71, 5, 958.
167. A. Viallat, M.M. Margulies, J.P. Cohen Addad and M. Thomas, Macromolecular
Chemistry and Physics, 1997, 198, 7, 2035.
168. J. Kriz, B. Marar and D. Doskosilová, Macromolecules, 1997, 30, 15, 4391.
169. F. Zeng, Z. Tong and H. Feng, Polymer, 1997, 38, 22, 5539.
170. J.P. Cohen Addad, J.P. Marchand and A. Viallat, Polymer, 1994, 35, 8, 1629.
171. Y. Uemura and P.M. MacDonald, Macromolecules, 1996, 29, 1, 63.
172. J.P. Cohen Addad, F. Desbat and J. Richard, Macromolecules, 1994, 27, 8, 2111.
173. J. Kriz, B. Marar, H. Pospisil, J. Plestil, Z. Tuzar and M.A. Kiselev,

174. J.P. Cohen Addad and C. Schmidt, Journal of Polymer Science: Part C, Polymer
Letters, 1987, 25, 487.
175. H. Tanaka, K. Fukumori and T. Nishi, Journal of Chemical Physics, 1988, 89, 5,
3363.
176. O. Girard, A. Guilermo and J.P. Cohen Addad, Die Makromolekulare Chemie:
Macromolecular Symposia, 1989, 30, 69.
177. J.P. Cohen Addad, E. Soyez, A. Viallat and J.P. Queslel, Macromolecules, 1992,
25, 1259.
178. T. Nishi and T. Hayashi, Progress in Pacific Polymer Science, 2, Proceedings of

the 2nd Pacific Polymer Conference, 1992, 141.
179. V.M. Litvinov and M. van Duin, Kautschuk und Gummi Kunststoffe, 2002, 55, 6.
Science USSR, 1977, A19, 9, 2331.
181. A.A. Askadskii, V.M. Litvinov, A.A. Zhdanov and G.L. Slonimskii, Polymer
Science USSR, 1985, A27, 11, 2705.
182. R.W. Tomlinson and D.F. Sheridan, Jr, Rubber Chemistry and Technology, 1989,
62, 4, 643.
399
183. G.N. Ghebremeskel, N. Westendorf and G. Hendrix, Rubber World, 1999, 220,
52.
184. S.A. Jones and H.J. Stronks, Minispec Application Notes, No. 20, Bruker
Analytik GmbH, Karlsruhe, Germany.
185. S.A. Jones, Minispec Application Notes, No. 21, Bruker Analytik GmbH,
Karlsruhe, Germany.
186. C.I. Tanzer and A.K. Roy, Proceedings of the SPE ANTEC ’95, Boston, MA,
USA, 1995, Volume 2, 2700.
187. P.J. Prado, B. Blümich and B.J. Balcom in Spectroscopy in Process Analysis, Ed.,
J.M. Chalmers, Sheffield Academic Press, Sheffield, UK, 2000, 4, 234.
188. G. Eidmann, R. Savelsberg, P. Blümler and B. Blümich, Journal of Magnetic

Resonance, 1996, A122, 104.
189. A. Guthausen, G. Zimmer, P. Blümler and B. Blümich, Journal of Magnetic

Resonance, 1998, 130, 1, 1.
190. G. Guthausen, A. Guthausen, F. Balibanu, R. Eymael, K. Hailu, U. Schmitz and

B. Blümich, Macromolecular Materials Engineering, 2000, 276/277, 25.
191. A. Wiesmath, C. Filip, D.E. Demco and B. Blümich, Journal of Magnetic

Resonance, 2001, 149, 258.
400
11
High Resolution NMR of Elastomers
Yasuyuki Tanaka, Eng Aik-Hwee and Jitladda Sakdapipanich
11.1 Introduction
The structure of elastomers gives an important information to polymer scientist and

technologist because it is closely related to the properties and hence performance of the
polymer. The information on the structure of elastomers also provides important clues
on the mechanism of polymer synthesis, or biosynthesis in the case of naturally occurring
elastomer. This knowledge would certainly contribute to a better control in producing
elastomers with certain desirable properties. Therefore, structural characterisation of
elastomers, whether natural or synthetic, is a subject of paramount important in many
research works. In this aspect, high-resolution nuclear magnetic resonance (NMR) is
one of the most powerful tool for the structural characterisation, because of its high
sensitivity, applicability for quantitative measurements without calibration curve, and
versatility applications to elastomers in solid, solution, latex and swollen states.
The structure of elastomers can be very complex but it must satisfy two basic requirements,
i.e., low glass transition temperature and vulcanisability. In the case of synthetic elastomers,
the former can be fulfilled by copolymerisation of common monomers to give higher
molecular mobility. There are many diene homopolymers that can satisfy these
requirements. Diene copolymers contain double bonds that serve as a crosslinking point
for vulcanisation with sulfur or peroxides. The sequence distribution of comonomer
units in these copolymers has to be controlled to provide desirable properties as elastomers.
With increasing demand in the versatility of physical properties, the structure of
copolymers has now been made more and more complicated. For example, styrene-
butadiene copolymer (SBR) synthesised by emulsion polymerisation has been recognised
to be a random copolymer having good properties for general purpose. However, recent
progress in anionic polymerisation makes it possible to produce SBR of a controlled
sequence distribution having the random, partial block, or S-B-S triblock structures and
the corresponding partial hydrogenated products. These copolymers exhibit excellent
properties satisfying individual requirements. More recently, it has been substantiated
that certain functional terminal groups of SBR could improve the physical properties of
the copolymer dramatically.
401
In the case of naturally occurring elastomers, such as natural rubber, it is necessary to get
a thorough understanding of the biosynthesis process in the plants. The process could be
eventually controlled through genetic engineering so that a highly desirable elastomer could
be obtained. Knowledge on structure of these elastomers that have certain superior properties
also enables scientists to produce the synthetic analogue with the similar properties.
High-resolution NMR has been applied to the characterisation of these elastomers in

many areas, such as chemical composition analysis, extent of chemical modification,
isomeric structure of diene monomer units, sequence distribution, small amounts of
modified structure, end-group, etc. However, this technique is less useful in the analysis
of cured rubber samples, long block sequences, inhomogeneity of chains such as chemical
composition distribution, and very small amounts of structural irregularity. Recent
progress in solid-state high-resolution NMR has partly solved the first problem. High-
resolution NMR at 500-750 MHz has been shown to improve the sensitivity dramatically
and has enabled determination of very small signal derived from terminal group or
irregular structures.
Valuable reviews and books of structural analysis of elastomers have been published by
several authors [1-6]. Some of these reviews provide excellent explanation on the basic
theory of sequence distribution of copolymer and NMR techniques applicable to
elastomers. Typical high-resolution 1H- and 13C-NMR spectra of various vulcanisates
and raw rubbers are depicted in a book written by Kelm [6]. The assignments and
references shown for each rubber are very useful for structural studies of elastomers. In
view of recent progress in the hardware and software of NMR, this chapter describes
some of the more recent applications of high-resolution NMR to the structural
characterisation of elastomers, after a brief description on the fundamental structural
features of elastomers.
11.2 Structural Feature of Elastomers
The analysis of chemical composition of copolymers can be done without difficulty using
1
H-NMR and 13C-NMR techniques, provided that the signals of each component are
assigned correctly. However, most of the signals in copolymers reflect the sequence
distribution of comonomer units from diad to tetrad or longer sequences. NMR instrument
with magnetic field above 500 MHz for 1H-NMR and 125 MHz for 13C-NMR has
facilitated the detailed assignment of the signals. In this case the chemical composition is
obtained by using the sequence equations such as diad and triad.
Diene monomers such as butadiene, isoprene, 2-chlorobutadiene give rise to the following
isomeric structures in the polymers shown in Figure 11.1. It should be noted that the
402
Figure 11.1 Microstructures of polybutadiene and polyisoprene
configurational sequences of 1,2 and 3,4 units reflect the arrangement of R and S
configuration such as RR or SS (meso) and RS or SR (racemic or racemo) configurations.
The ratio of these isomeric structures, which is sometimes referred to as microstructure,

can be determined by 1H-NMR and 13C-NMR. For example, the microstructure of
polybutadiene can be measured by 1H-NMR with magnetic field higher than 300 MHz
or 13C-NMR with magnetic field higher than 25 MHz. The microstructure of polyisoprene
can be analysed using similar techniques. However, most of synthetic diene polymers are
a kind of copolymers composed of these isomeric units. Signal-split reflects the sequence
distribution of three to four isomeric units, configurational sequence of 1,2 and/or
3,4 units, and head-to-tail arrangements.
Rubber from Hevea brasiliensis is a typical naturally occurring cis-1,4 polyisoprene termed
as natural rubber. Many higher plants and fungi produce cis-1,4 polyisoprene. Natural
polyisoprenes appear to be a simple polymeric system, considering that they are produced
by enzymatically controlled polymerisation process. However, the reason why only natural
rubber exhibits outstanding properties has yet to be clarified. The analysis of detailed
structure of natural rubber, in comparison with other natural cis-1,4 polyisoprene, could
provide direct information on the origin of the characteristic properties of natural rubber.
Gutta percha and Balata from Paraquim gutta and Mamusops balata, respectively, are
typical naturally occurring trans-1,4 polyisoprene. Chicle resin from Achras zapota
contains both cis-1,4 and trans-1,4 polyisoprenes. Biochemical studies since 1950s have
not elucidated the biosynthesis mechanism on the formation of cis-1,4 and trans-1,4
polyisoprenes. The detailed structural analysis would disclose the longstanding puzzles
on the biosynthesis mechanism of these natural polyisoprenes.
403
The analysis of end-groups and irregular structures, e.g., branch-points and grafting
points, is a difficult task because of the high molecular weight of elastomers. Elastomers
usually have higher molecular weight than ordinary plastics except for some thermoplastic
elastomers, such as S-B-S triblock copolymers of styrene and butadiene. For such analysis,
a high magnetic field NMR is required in order to obtain a high signal-to-noise ratio.
In addition to the sequence distribution, copolymers usually have the distribution of

chemical composition, i.e., the difference of chemical composition among polymer chains.
Unfortunately, conventional NMR techniques are not sensitive to the chemical
composition distribution, because NMR signals reflect an average structure of each
polymer chain. Therefore, solvent cross-fractionation or HPLC separation technique
followed by composition analysis of each fraction is usually adopted for this purpose.
However, the development of on-line high performance liquid chromatography (HPLC),
including size exclusion chromatography (SEC), makes it possible to analyse the difference
of structural features among polymer chains.
In the following section, certain topics are selected to introduce the recent applications
of high-resolution NMR to the structural characterisation of elastomers, together with
basic information on signal assignments.
11.3 Analysis of Chemical Composition and Sequence Distribution
11.3.1 Accuracy of NMR Measurements of Chemical Composition
A series of works has been made on the reliability of NMR data; chemical shifts, intensity
and resolution [7], spin-lattice relaxation time and nuclear Overhauser enhancement
factor (NOE) [8,9], and composition and comonomer sequence distribution [10], by
round robin method using poly(methyl methacrylate) (PMMA), solanesol, and copolymer
of MMA and acrylonitrile (AN). The standard deviations (σ) of chemical shift
measurements for 1H- and 13C-NMR signals of PMMA were 0.0036-0.0071 and 0.054-
0.307 ppm, respectively. Here, the measurements were made with 90 to 500 MHz
instruments for 1H-NMR. The relative intensities of 3(CH2 + α-CH3)/OCH3 were 4.94
(σ = 0.156), showing a good agreement with the theoretical value of 5. In the case of
13
C-NMR, the relative signal intensities corrected by the NOE value agreed with the
theoretical value within ± 10% relative errors, with an appropriate pulse repetition time.
The composition of MMA-AN copolymer was determined with the σ value of 3.7-9.5%
404
for both 1H- and 13C-NMR measurements. The monomer reactivity ratio for a penultimate
model was determined with the σ value 5-14%.
The quantitative analysis of end-groups provides information on the number average molecular
weight, number of branch-points per chain, and mechanisms of initiation, termination and
chain transfer reactions. These are possible only if the structure of end-groups is identified.
Direct measurements of the relative intensity for small signals depend on the signal-to-noise
ratio and dynamic range of digitiser (ADC). In a model experiment, the determination of
CH3NO2 in the solution of PMMA is possible even at a level as low as 1/250,000, with
accuracy of 2.5-10.2%, using 1H-NMR at 500 MHz. However, it is necessary to do
accumulation of 18,000 scans [11,12]. This will be the limit of quantitative measurement
from relative intensity between large and small signals, by considering the dynamic range of
ADC. The limit of quantitative measurement can be improved to 1/106 by the use of a coaxial
tube, containing a standard sample in the internal tube [12]. Various types of end-groups
were analysed quantitatively for PMMA by using 1H-NMR at 750 MHz [13].
One of the advantages of using NMR technique is its ready application to quantitative
analysis. Usually, it does not require authentic sample nor calibration curve. However, it
is necessary to check the accuracy and precision of the measurement, as mentioned above.
It is also useful to check these by comparison with another direct method such as titration
method that is applied frequently for the determination of functionality of telechelic
diene polymers, as mentioned in the following section. The unsaturation in butyl rubber
has been analysed by the epoxy titration method using 3-chloroepoxybenzoic acid or
simply iodine index test. The disappearance of all the olefinic signals in butyl rubber was
clearly indicated after the epoxy titration, as shown in Figure 11.2. The epoxy titration
method showed little deviation such as σ =0.0128 for butyl rubber containing 1.79%
unsaturation. However, it is remarkable that a side reaction may occur after reaching the
end-point of titration. The unsaturation values with 1H-NMR at 500 MHz showed a
good agreement with those obtained by the epoxy titration method as well as those
estimated by copolymerisation kinetics [14]. 1H-NMR measurements showed a similar
good agreement with analysis by epoxy titration method, by mass balance of monomers,
and by predicted method from copolymerisation equations [15].
The molecular composition of ethylene-propylene copolymer (EPM) was analysed using

the round robin analysis on a set of 10 EPM samples by 13C-NMR measurements mainly at
75 MHz [16]. According to the signal assignments given in Table 11.1, it is possible to set
up several methods to determine the content of E and P. Here, the following four methods
were applied [16]. The consistent and reproducible molecular composition results were
obtained independent of the measuring laboratory and the following calculation method.
405
Figure 11.2 1H-NMR spectra of butyl rubber, (a) olefinic region, (b) full spectrum
after epoxidation, observed at 500MHz [14]
Reproduced with permssion from J.E. Puskas and C. Wilds, Rubber Chemistry and
Technology, 1994, 67, 329, Figure 1. Copyright 1994, Rubber Division, ACS
Method 1:
S = all secondary carbons

T = all tertiary carbons
P = all primary carbons
Mol % ethylene = [3-(S + T - 2 x P)/(S + T)] x100
Mol % propylene = 100 - mol % ethylene
Method 2:
P′ = Sαα+ 0.5 x (Sαβ + Sαδ )

E′ = 0.5 x [3-Sββ + Sβγ + Sβδ + Sγγ + Sγδ + Sδδ + 0.5 x (Sαβ + Sαγ + Sαδ )]
Mol % ethylene = [3-P′/(E′ + P′)] x 100
Mol % propylene = [3-E′/(E′ + P′)] x 100
Method 3:
406
Table 11.1 13C-NMR signal assignment for EPM in tretrachloroethane-d2

[16]
Peak Carbon Integration
Chemical Shift Sequence
No. Type Limit
1 Sαα 48.1 ÷ 45.3 Sαα 48.50-44.50
2 Sαγ 38.8 r-Sαγ
3 Sαδ 38.4 r-Sαδ
4 Sαγ 37.96 m-Sαγ + m-Sαγ 40.00-36.50
5 S αδ 37.58 m-Sαδ + m-Sαδ
6 Sαβ 35.7 r-Sαβ
7 Sαβ 34.9 Sαβ + Sαβ 36.20-34.30
8 Tγγ 33.9 Tγγ
9 Tγδ 33.6 Tγδ 34.29-32.80
10 T δδ 33.3 Tδδ
11 Tβγ 31.2 Tβγ (m)+ Tβγ(r)
12 T βδ 30.9 Tβδ (m)
13 Sγγ 30.8 Sγγ +Tβδ (r) 31.91-30.61
14 Sγδ 30.4 Sγδ 30.61-30.23
15 S δδ 30.0 Sδδ 30.23-29.32
16 Tββ 28.8 ÷ 28.5 Tββ (mm)+Tββ (mr+rr) 29.15-28.22
17 Sβγ 27.85 Sβγ 28.22-27.63
18 S βδ 27.45 ÷ 26.3 Sβδ 27.63-26.63
19 Sββ 24.9, (24.7, 24.6) Sββ 25.60-23.95
20 Pββ 22.0 ÷ 21.3 Pββ (mm)
21 Pβγ 21.3 ÷ 20.6 Pββ (mr)+Pβγ(m)+ Pβδ(m) 22.50-19.00
22 Pγγ 20.6 ÷ 19.5 Pββ(rr)+Pβγ(r)+Pβδ(r)+ Pγγ
Reproduced with permission from S.D. Martino and M. Kelchtermans Journal of Applied
Science, 1995, 56, 1781, Table 2. Copyright 1995, John Wiley & Sons
407
N = total no. methylenes = (Sαα+ Sαβ + 3 x Sββ + 2 x Sβγ + 5 x Sγγ

+ 3 x Sγδ + Sδδ )
N1 = total no. methyls = (Pββ + Pβγ + Pγγ )
Mol % ethylene = [3-(N0 - N1)/(N0 + N1)] x 100
Mol % propylene = [3-2 x N1/(N0 + N1)] x 100
Method 4:
n0 = number-average sequence length of uninterrupted methylenes

(Sαα+ Sαβ + 3 x Sββ + 2 x Sβγ + 5 x Sγγ+ 3 x Sγδ + Sδδ )
= ⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯⎯
(Sαα+ 0.5 x Sαβ + Sββ + 0.5 x Sβγ + Sγγ+ 0.5 x Sγδ)
Mol % ethylene = [3-2/(1 + n0)] x 100

Mol % propylene = 100 - mol % ethylene
The composition analysis of SBR, i.e., styrene, cis-1,4, trans-1,4 and 1,2 (vinyl) units,
has been carried out by 13C-NMR based on signal assignments indicated in Table 11.2.
The integrated intensities of regions A to I are used to calculate percent composition by
using following equations [17].
%S = 100 x (A + D)/(A + B + C + D + E)
%V = 100 x (B + E)(A + B + C + D + E) (11.1)
%T = (100 - %S - %V) x (F + G)/(F + G + H + I)
%C = (100 - %S - %V) x (H + I)/(F + G + H + I)
Determination of the composition of many samples is time-consuming, since integration

technique often requires careful check on each spectrum, due to the variation of signal
overlap reflecting the composition. Here, principal component analysis (PCA) was applied
to the analysis of four components using 20 integral regions [18] indicated in Figure 11.3.
Automated analysis was carried out by reading 27 sets of data from SBR and BR samples
with 20 integral values, into a spreadsheet program to construct a 20 x 27 independent
matrix. Then, a 4 x 27 dependent matrix was created using the composition data obtained
according to the equations (11.1). PCA using the partial-least squares regression algorithm
was performed on the data set. The result of composition analysis by PCA was almost
equivalent within experimental error to that obtained by ordinary 13C-NMR method.
Automated sample classification is also achieved by PCA analysis.
408
Table 11.2 Integral regions for PCA analysis of SBR and their structural
meaning [18]
13
Integral C-NMR Chemical Resonance Assignment Integral in
Shift Region (ppm) Figure 11.3
1 146.0-145.0 Styrene C1 A
2 143.2-142.2 Butadiene vinyl CH B
3 130.5-130.05 Butadiene trans CH C
4 130.05-129.8 Butadiene trans CH C
5 129.8-129.5 Butadiene cis CH C
6 129.5-129.0 Butadiene cis CH C
7 126.2-125.6 Styrene C4 D
8 114.7-113.7 Butadiene vinyl CH2 E
9 46.2-45.2 Styrene CH
10 44.0-43.0 Butadiene vinyl CH
11 40.6-39.6 Butadiene trans CH2 styrene CH
12 38.6-37.5 Butadiene trans CH2 vinyl CH
13 36.2-35.2 Styrene CH2
14 34.6-33.6 Butadiene vinyl CH2
15 33.3-32.1 Butadiene trans CH2 F
16 31.0-30.25 Butadiene trans CH2 styrene CH2 G
17 30.25-29.5 Butadiene trans CH2 vinyl CH2 G
18 27.9-26.9 Butadiene cis CH2 H
19 25.7-25.2 Butadiene cis CH2 styrene CH2 I
20 25.2-24.7 Butadiene cis CH2 vinyl CH2 I
Reproduced with permission from N.G. Walsh, J.K. Hardy and P.L. Rinaldi, Applied
Spectroscopy, 1997, 51, 889, Table 1. Copyright 1997, Society for Applied Spectroscopy
409
Figure 11.3 13C-NMR spectrum of SBR including poly(ethylene glycol) as internal

standard, observed at 75 MHz. The signals in the region A-I are used to calculate the
integral intensity according to the equations (11.1) [18]
Reproduced with permission from N.G. Walsh, J.K. Hardy and P.L. Rinaldi, Applied
Spectroscopy, 1997, 51, 889, Figure 1. Copyright 1997, Society of Applied Spectroscopy
11.3.2 Analysis of Chemical Composition Distribution using SEC-NMR
Copolymers usually have three types of distribution, i.e., the molecular weight distribution,
sequence distribution and chemical composition distribution. The chemical composition
distribution means the molecular weight dependence of chemical composition.
Fractionation techniques, such as solvent fractionation and HPLC, including SEC, are
commonly used for the analysis of molecular weight distribution. So-called cross
fractionation, using two sets of solvents or special HPLC columns, is applied to the
fractionation followed by chemical composition analysis for each fraction. The direct
coupling of HPLC (SEC) and NMR spectroscopy will be a powerful tool for the analysis
of chemical composition distribution. This is due to the development of high-field NMR
instruments and commercially available HPLC-NMR proves, which ensure high sensitivity
and resolution for a flowing sample [19, 20]. The applicability of on-line SEC-NMR was
demonstrated for the analysis of molecular weight dependence of the chemical composition
[21] and tacticity [22]. The quantitative measurement of end-group of a polymer chain
has enabled the direct measurement of molecular weight distribution [23].
On-line SEC-NMR has successfully applied to the determination of chemical composition

distribution in ethylene-propylene-diene terpolymers (EPDM) by the use of 1H-NMR at
750 MHz [24]. EPDM containing 2-ethylidene-5-norbornene (ENB), as a diene
410
component, shows the 1H-NMR signals characteristic of each component. The assignment
of signals were made by means of PFG-heteronuclear quantum coherence (HMQC)
measurements, using the assignments of 13C-NMR signals of the terpolymer thus far
reported [25, 26]. As shown in Figure 11.4, the continuous-flow SEC-NMR gives high-
resolution signals almost comparable to those recorded under ordinary NMR
measurements. It also gives enough signal-to-noise ratio (S/N) for the signals assigned to
β+ -CH2 and P-CH3 at 1.25 and 0.83 ppm, respectively. However, the signals due to a
small amount of ENB are difficult to use for quantitative measurements in the continuous-
Figure 11.4 1H-NMR spectra of EPDM in deuterochloroform, observed at 750MHz

[24]. (Sample contains E 78.9, P 19.1, and ENB 1.1 mol%): (a) Continuous-flow SEC-
NMR measurement, (b) normal measurement using 5 mm o.d. sample tube and
(c) stop-and-flow SEC-NMR measurement
Reproduced with permission from K. Ute and K. Hatada, International Journal of Polymer
Analysis and Characterisation, 1999, 5, 47, Figure 3. Copyright 1999, Academic Press
411
flow experiments. In the stop-and-flow experiments, NMR scans are accumulated in

each point until an enough S/N is obtained for the signals due to ENB component. Here,
it is remarkable that the SEC-NMR spectra at 750 MHz instrument improves the S/N
ratio by 9.2 times compared with those at 500 MHz one [23].
Figure 11.5 shows a stacked trace plot of the continuous-flow SEC-NMR data for a
mixture of EPDM and EP, containing 31.5 mol% P and 0.6 mol% ENB on an average.
Cross sections of the stacked plot at 1.25 ppm and 0.83 ppm give the 1H-NMR detected
SEC curves for E and P components, respectively, as a function of elution volume, using
the following equations:
Ip = I(P-CH3)/3
Ie = [3-I(α-CH2) + I(β+ -CH2) - 2Ip]/4
[3-E] = 100Ie /(Ie + Ip)
The relationship between the elution volume and chemical composition can be obtained
by using a calibration curve of standard samples, such as standard polystyrene. It was
clearly shown that the ENB and ethylene contents in this sample decreased with decreasing
molecular weight. Here, it should be noted that the elution volume reflects the molecular
size including the molecular weight and composition. This measurement was made using
deuterochloroform, but it was suggested to use o-dichlorobenzene as eluent at 135 °C to
separate mainly by molecular size of EPDM. Temperature gradient fractionation followed
by NMR analysis also gives chemical composition distribution for EPDM containing
Figure 11.5 A stacked trace plot of 750MHz continuous-flow SEC-NMR spectra [24]
Reproduced with permission from K. Ute and K. Hatada, International Journal of Polymer
Analysis and Characterisation, 1999, 5, 47, Figure 6. Copyright 1999, Academic Press
412
higher ENB units [27]. This measurement has shown that ENB content is independent of
molecular weight, while ethylene content decreases in high molecular weight fraction.
11.3.3 Analysis of Sequence Distribution

Sequence distribution studies on several types of rubber by 13C-NMR technique have
been reported. Some of the more recent reports include silicone rubbers [28-30], SBR
[31], acrylonitrile-butadiene rubber (NBR) [32,33], polyurethane [34,35],
polyepichlorohydrin [36], ethylene-norbonene [37] and ethylene-propylene rubber [4,
16, 25, 38-44]. The NMR studies on EPDM have been carried out extensively, because it
is one of the important parameters, which control the physical properties of the elastomer.
For example, ethylene sequence can influence the crystallisation kinetic and melting
behaviour of the rubber [38].
The early definitive work on comonomer sequence distribution was based on mathematical
model of ethylene-propylene polymerisation, where propylene added by either primary
or secondary insertion [43]. The improvement [39] was made using higher field 13C-NMR
to correct several mistakes, which were made in the earlier signal assignments [40-42]. A
typical 13C-NMR spectrum of EPM is given in Figure 11.6. The corresponding signal
assignments are indicated in Table 11.1.
Figure 11.6 13C-NMR spectrum of EPM in tetrachloroethane-d2 [16]
Reproduced with permission from S.D. Martino and M. Kelchtermans, Journal of Applied
Polymer Science, 1995, 56, 1781, Figure 1. Copyright 1995, John Wiley & Sons
413
Later, it was indicated that 13C-NMR signals of the third monomer such as 1,4-hexadiene,
ENB and 5,8-dicyclopentadiene might overlap with the main chain EPM carbons,
particularly in the Sαα region [43]. The difference in ethylene content obtained by different
methods of calculation could be as high as 6 mol% with increasing the ENB content.
Using the correction of the third component, the standard deviation dropped from 3%
to 0.5%. The technique on sequencing experiments can also be improved by introducing
paramagnetic relaxation agent such as Cr(acac)3 to shorten the 13C spin lattice relaxation
times. The sensitivity of measurement is further enhanced via NOE by allowing the 1H
decoupling during the recycle time because the NOE is expected to be constant for all
carbons in the polymer chain. With these improvements, the experimental time was
reduced from more than 12 hours to 3-6 hours [43].
An improved method was proposed for the determination of triad distribution, by

considering the inherent uncertainties in the quantitative analysis of EP copolymer in
EPM from 13C-NMR integral intensities [44]. The 13C-NMR signals due to triad sequence
of E and P units were grouped into 8 regions and 8 simultaneous equations were derived
to describe six triad sequence distribution numbers. Using a statistical Z-test, the triad
concentrations were obtained by solving directly the over-determined set of equations.
These results showed smaller standard deviation than those derived from the equations
that take into the consideration of systematic errors in the derived triad due to spectral
overlap. This was further confirmed by the uncertainty analysis using simulation, assuming
a Gaussian error distribution. The results also showed that the uncertainty in a triad
concentration is affected by the relative distribution of the other triad concentrations.
This effect is larger for the sequence with higher ethylene content.
The content of 1,2 (vinyl), cis and trans isomeric units in polybutadiene (PB) can be
obtained by 1H-NMR [1]. A better spectral resolution of 13C-NMR permits the sequence
analysis to be used to quantify the microstructure of PB. Thus, the assignments for aliphatic
[45] and olefinic [46] carbons in triad sequence have been reported. As shown in
Figure 11.7, the olefinic carbon resonances were assigned to the triad sequences of 1,2
(vinyl), cis and trans in the 13C-NMR of PB [47]. A good agreement was observed between
the experimental peak intensity and calculated one assuming the Bernoullian model. The
assignment of peaks in the aliphatic and olefinic regions was further tested using 1,2′,4,4′-
tetradeuteriopolybutadiene (PB-d4h2) and perdeuterated polybutadiene (PB-d6) in
comparison with the nondeuterated PB (PB-h6) [48, 49]. The trans content of deuterated
and partially deuterated PB determined from the aliphatic region was higher than that of
olefinic region of partially deuterated PB due to signal overlap in the region [49]. Since,
only aliphatic region is amenable for quantitative analysis of PB-d6, {2H}13C-INEPT NMR
with simultaneous 1H and 2H decoupling during data acquisition was used to characterise
the sequence distribution of deuterated PB [50]. By using this technique, spectral
simplification is achieved, which is useful for performing quantitative analysis and
developing signal assignment.
414
Figure 11.7 Expanded 13C-NMR spectra of the olefinic resonances of polybutadiene,

due to cis-1,4 and trans-1,4 centred triads, observed at 50 MHz [47]. The content of
1,2 units is 1% (A), 12% (B), 13% (C), 22% (D), and 49% (E)
Reproduced with permission from G. van der Velden, C. Didden, T. Veermans and J. Beulen,
415
Synthetic cis-1,4 polyisoprene has the structural feature similar to PB. As to the NMR
analysis of sequence distribution, however, little work has been done after 1978. Detailed
assignment may be possible by using modern NMR techniques and magnetic field higher
than 100 MHz for 13C-NMR.
The microstructure of polychloroprene, polymerised at +12 to +70 ºC, was analysed

using 1H- and 13C-NMR [32]. Signal assignments were made for head-and-tail
arrangements of trans-1,4 units, which was the major component of polychloroprene,
and for other isomeric units. Polymerisation at high temperature resulted in a slight
increase in head-to-head and tail-to-tail linkage of trans-1,4 units as well as the increase
in cis-1,4 units.
The assignment of 13C-NMR signals in carboxyl-terminated acrylonitrile-butadiene

copolymers was done by using one and two-dimensional (1D and 2D) NMR techniques
[32]. The fundamental assignment of 1H-NMR signals of NBR done in 1970s was
confirmed based on 2D COSY-45 and 2D J-resolved spectra. Detailed assignment of
13
C-NMR signals were successfully made with spectral editing of carbon peaks using
DEPT technique as well as C-H COSY and long-range two-dimensional heteronuclear
chemical shift correlation (COLOC) experiments, as tabulated in Table 11.3. Block
polybutadiene units were found in this polymer, but no block acrylonitrile sequence was
detected. The absence of diad sequence of acrylonitrile units in NBR containing
acrylonitrile less than 30-33% was also confirmed by 1H-NMR [52] and 13C-NMR [33].
Segmented polyurethane-urea elastomer is a block copolymer composed of soft and hard

segments, see Figure 11.8. The composition and length of both segments as well as
sequence length distribution are the primary factors to govern the mechanical properties.
The mean sequence lengths of polyurethane block and polyurea block, were analysed by
13
C-NMR, based on the assignment of aromatic carbon signals by using model
compounds, as depicted in Figure 11.9 [34]. Quantitative analysis indicated that the
polymer is almost a random copolymer. A similar analysis of segmented polyurethane-
urea elastomer found that two-step polymerisation gave the mean sequence length longer
than one-step polymerisation [35].
The combination of lanthanide shift reagents and 2D NMR techniques (COSY and
NOESY) was applied to the 1H- and 13C-NMR assignments of ethylene-vinyl acetate
Figure 11.8 Structure of segmented polyurethane-urea
416
Table 11.3 Assignments of 13C-NMR peaks of carboxyl-terminated NBR [32]

Chemical Carbon Chemical Carbon
Sequence Sequence
Shift (δ) Type Shift (δ) Type
176.40 I COOH 143.46 VV -CH=
142.45 VC -CH= 142.28 VT -CH=
140.56 VA -CH= 140.36 VA′ -CH=
140.24 TV′I -CH= 134.31 TTA -CH=
134.07 TTAT -CH= 133.06 TV′I -CH=
133.01 TV′A -CH= 132.41 ATA -CH=
131.73 TTTA -CH= 131.27 ATTA -CH=
131.17 A′TTA -CH= 130.15 VTT -CH=
129.93 CTT -CH= 129.80 TTT -CH=
129.59 TTV -CH= 129.40 TVT -CH=
129.22 TCT -CH= 129.05 TCV -CH=
128.20 TTTA -CH= 127.85 ATTA -CH=
126.81 CyCA -CH= 126.53 ATA -CH=
126.40 ATAT -CH= 124.85 TTA -CH=
124.75 TTAT -CH= 124.66 TTA -CH=
124.25 ATV′ -CH= 124.02 ATV′ -CH=
123.11 CyCA, I -CH=,CN 122.20 CyCA CN
122.05 AC CN 121.64 TAT CN
121.51 AT CN 120.26 AA CN
117.15 AV =CH2 116.44 TV′I =CH2
115.71 AV′ =CH2 114.31 VT =CH2
114.09 VC =CH2 43.31 TVT CH
42.44 TV′I CH, e or ta 41.28 TV′I CH, e or ta
42.03 TV′I =CH2 41.88 VA CH, e or ta
41.03 VA CH, e or ta 40.43 TVI CH
417
Table 11.3 Continued...

Chemical Carbon Chemical Carbon
Sequence Sequence
Shift (δ) Type Shift (δ) Type
38.40 AVT CH 2 37.96 VT CH2
37.20 TV′I CH2 36.34 TA′I CH2
36.09 I C 35.45 TVV CH 2
34.86 TA′ CH2 34.70 AAT CH2
33.86 TV CH2 33.21 I CH2
33.11 I CH2 32.49 TTT, VC CH2
32.36 TTA CH2 32.14 TTTA CH2
32.05 TTAA CH2 31.34 AT CH2
31.21 TAT CH2 31.05 TAT + AT CH
30.85 A′VT CH 2 30.77 AV′T CH
30.11 I CH 2 29.95 TV CH2
29.78 TA CH2 29.54 AC CH2
29.41 CI CH2 29.10 VAT CH, e or ta
28.10 A′I CH, e or ta 27.94 CyCA CH2
27.28 CC CH2 27.16 CT CH2
26.94 CCA CH2 26.78 CC CH2
25.10 CyCA CH2 24.95 CI CH2
24.85 CVA CH2 24.66 CV CH2
24.60 CA CH2 24.41 CyCA CH
23.28 TV′I CH 3 22.78 CyCA CH2
22.15 TVI CH3
A: acrylonitrile unit, B: butadiene unit, V: 1,2-butadiene unit, C: cis-1,4 butadiene unit,
primed units: reversed arrangement, T: trans-1,4 butadiene unit, CyCA: 2-cyclohexene-
1-carbonitrile and I: 4,cyanopentanoic acid end group
a
Erythro or threo defined by a Newman projection of this carbon and the other nearest
asymmetric centre
Reproduced with permission from T. Fang, Macromolecules, 1990, 23, 2145, Table 2.
418
Figure 11.9 13C-NMR spectrum of segmented polyurethane-urea, observed at

75.5 MHz [34]
Reproduced with permission from A. Kajii and M. Murano, Polymer Journal (Japan),
1990, 22, 1065, Figure 3. Copyright 1990, The Society of Polymer Science, Japan
copolymer [53]. The addition of shift reagents causes a number of resonances to emerge
and shift downfield from the large resonances centered at 1.26 ppm, as summarised in
Table 11.4. The complete 13C-NMR interpretation was obtained for the sequences.
11.3.4 Analysis of Chemically Modified Structure and Graft Polymers
Chemical modification of rubber such as hydrogenation, chlorination, epoxidation,

cyclisation, addition of maleic anhydride, is a popular technique to give a desirable
419
Table 11.4 13C- and 1H-NMR assignments in EVAc copolymers [53]

1 13
H C
No. Shift Shift Sequence Sequence Shift
[ρ=0] [Yb(fod)3]a
1 1.26 ca. 1.26 Sγγ+Sγδ+Sδδ Acetate 21.0
2 1.33 1.60 S βδ Sββ 21.5
3 1.40 ca. 1.70 Sββ Sβδ 25.7
4 1.50 1.93 Sαγ+Sαδ Sγγ+Sγδ 29.8
5 1.73 2.08 S αα Sδδ 30.0
6 1.92 2.58 Acetate Sαδ 34.7
2.74 Sαγ 35.2
2.90 Sαα 39.6
7 5.15 5.35 Tββ (VVV)
Tββ ca. 68
8 5.03 5.75 Tβδ (VVE)
9 4.90 5.90 Tδδ (VEVEV) Tβδ 71
6.00 Tδδ (VEVEE)
Tδδ 73
6.30 Tδδ (EEVEE)
a Shift observed with Yb(fod)3 added, reagent/substrate molar ratio ρ = 0.00235
Reproduced with permission from H.N. Cheng and G.H. Lee, Polymer Bulletin, 1988,
19, 89, Table 1. Copyright 1988, Springer-Verlag GmbH & Co KG
property. Structural characterisation of the chemically modified rubbers is done to decide

the extent of reaction, sequence distribution of modified units and structural change by
side reactions.
Acid-catalysed cyclisation of polyisoprene and polybutadiene forms resin-like polymer containing

cyclised segments consisting of one or more fused rings as shown in Figure 11.10. The average
number of rings per segment, termed cyclicity, degree of cyclisation and segment end-groups
have been analysed by the use of 1H-NMR and 13C-NMR. The cyclicity was presumed to be
from three to five in extensively cyclised polyisoprene with TiCl4/CCl3CO2H catalyst [54].
420
Figure 11.10 Cyclised polyisoprene with tetracyclic segment
Hydrogenation of NBR has been developed to improve thermal stability. Perfect

hydrogenation of the double bonds in butadiene unit, without any hydrogenation of
the acrylonitrile units, using RhClH2(P(C6H5)3)2 catalyst was confirmed with 1H- and
13
C-NMR [55]. In the 13C-NMR spectrum of hydrogenated NBR, CN carbon atom
showed splitting reflecting the pentad sequences of acrylonitrile and ethylene units
[56]. The 1H-NMR analysis showed clearly the progress of ring-opening reaction of
the epoxide group to form diols in partially epoxidised natural rubber [57]. Formation
of fluoroesters of cis-polyisoprene (Guayule rubber) by ring opening with fluoroacids was
confirmed with 1H- and 13C-NMR [58]. The 13C-NMR signals of epoxidised low molecular
weight polybutadiene have been assigned to diad sequences of epoxide units based on
the calculation of chemical shifts by empirical rule. The cis-1,4 butadiene units
epoxidised preferentially and no reaction was observed for 1,2 units [59].
The functionalisation of polyethylene, polypropylene and EPM with maleic anhydride

(MA) has been done to give compatibilisation and enhanced adhesion to polar materials.
The free-radical initiated grafting of 13C-enriched [2,3-13C2]MA on to polyethylene (PE),
polypropylene (PP) and EPR was studied using noise-decoupled and 1D INADEQUATE
(incredible natural abundance double-quantum transfer experiment) 13C-NMR [60,61].
The result indicated the formation of following graft structures shwon in Figure 11.11,
where MA grafts were in the form of single succinic anhydride part in the case of EPR
and PP. In EPM with high propylene content, a chain scission reaction could occur,
yielding a ring attached to the chain-ends via a double bond.
The structure of curing points has been studied using model compounds of rubber.
Figure 11.11 Presumed structure of polyethylene and polypropylene modified with

maleic anhydride
421
Vulcanisation of EPDM with sulfur systems was studied by 1H-NMR using ethylidene
norbornane (ENBH) as a model of ENB [62]. The use of ENBH was also effective to
elucidate the curing reaction of EPDM with phenol-formaldehyde resin [63]. Similarly,
halogenation reaction of IIR was studied by 1H-NMR using 2,2,4,8,8-pentamethyl-4-
nonene as a model [64].
11.4 Analysis of End-groups and Branching
11.4.1 Assignment of Small Signals
Structural characterisation of chain end-groups provides information on the mechanism

of polymer formation and degree of polymerisation. The importance of functional end-
groups on the performance of elastomers has been recognised recently. For example,
chemical modification of end-groups has led to the improvement of dynamic properties
of anionically polymerised SBR [65,66]. Recent progress of NMR instruments enabled
the quantitative analyses of end-groups in high molecular weight polymers with less
difficulty. This is because the limit of detection is sufficiently high to analyse the number
average molecular weight of a polymer based on the end-group analysis, if the polymer
molecule contains a known amount of end-groups per chain. For instance, the end-
group analysis makes it possible to determine the M n value higher than 50 x 104 for
PMMA, by the use of 1H-NMR at 500 MHz [11]. Chemical modification of chain end-
groups is an ordinary technique to obtain a special functional group. The extent of
reaction can be evaluated by the functionality of polymer chain, i.e., the average number
of functional group per chain. Long-chain branching in polymer chain could also give
rise to small NMR signals. The signal assignment for branch-points is sometimes more
difficult than that of end-groups, because of difficulty in the preparation of low molecular
weight model compounds.
One of the major challenges would be assignment of small signals. However, the use of
modern NMR instruments at high magnetic field has enabled one to apply various NMR
techniques such as 2D NMR and DEPT, for the assignment of very small signals. Other
difficult tasks may include the purification of polymer samples. Sometimes it is important
to distinguish between the signals due to end-groups or branch-points in a polymer chain
and signals from impurity, especially in the case of naturally occurring rubbers. It is preferable
to remove impurities, which could give rise to signals near the resonance in question. This
is to ensure that the signal intensity of impurities is much smaller than that of chain-ends,
abnormal groups or branch-points, in order to get unequivocal assignment.
422
11.4.2 Functionality of Telechelic Diene Polymers
The chain-ends of anionically polymerised so-called “living” polybutadiene and

polyisoprene can be modified to form hydroxyl or carboxyl group at the final stage of
polymerisation. The functionality of telechelic diene polymers is a very important factor
to establish a uniform network structure by a coupling reaction. A simple chemical titration
or NMR analysis can be applied to the determination of functionality, because telechelic
rubbers are normally of low molecular weight. The functionality of hydroxy-telechelic
polybutadiene prepared by the reaction of ethylene oxide was reported to be 2.03 ± 0.1,
by measurement at 100 MHz in the early stage [67]. Here, the functionality was determined
according to the following equation, using the relative intensity of signals in Figure 11.12.
Here, x and y stand for the weight fraction of 1,2 and 1,4 units, respectively. On the
other hand, commercial hydroxy-telechelic polybutadiene obtained by radical
Figure 11.12 1H-NMR spectrum of hydroxy-telechelic polybutadiene, observed at

100 MHz [67]
Reproduced with permission from Y. Camberlin, J.P. Pasucault and Q.T. Pham, Die
Makromolekulare Chemie, 1979, 180, 397, Figure 1. Copyright 1979, Wiley-VCH
423
polymerisation had a distribution of functionality. The Fn value increased from 1.9 to

3.5 (average Fn =2.1) with increasing the molecular weight, as it was shown by the analysis
of fractions with SEC fractionation [68].
The functionality of hydroxy-telechelic polybutadiene and polyisoprene was determined

by 1H-NMR after modification with phenyl isocyanate or naphthyl isocyanate. Liquid
telechelic polybutadiene and polyisoprene, prepared by choosing appropriate reaction
conditions, showed Fn value of about 1.9, which agreed with the values determined by
titration method [69].
Amino-terminated telechelic polybutadiene was prepared by LiAlH4 reduction of amidino

end-group in polybutadiene, which was polymerised by a water-soluble initiator, 2,2′-
azobis(amidinopropane)dihydrochloride. The structure was analysed by 1H- and 13C-NMR,
but functionality of 2.0 was obtained by a titration method [70]. Synthesis of ω-epoxy-
functionalised polyisoprene was carried out by the reaction of 2-bromoethyloxirane with
living polymer that was initiated with sec-butyllithium. The functionality of the resulting
polyisoprene was 1.04 by 1H-NMR and 1.00 by thin layer chromatography detected
with flame ionisation detection [71].
The terminal hydroxyl groups in hydroxy-telechelic polybutadiene are assumed to be

predominantly primary alcohols linked to cis-1,4, trans-1,4, and 1,2 butadiene units
[72]. Detailed structure of the butadiene or isoprene units at both terminals was studied
by the use of 2H-NMR [73], as shown in Section 6.1
11.4.3 Structure of Terminal Groups
The 13C-NMR of low molecular weight ethylene-propylene rubber (EP) exhibited 4 olefinic
and 33 aliphatic signals. These signals were assigned to the diad and triad sequences of E
and P units including initiating and terminating units. Empirical additive rules have been
adopted to estimate the chemical shifts for the structures expected from the proposed
polymerisation mechanism [42]. A complete triad sequence distribution, including all of
the possible end-groups, was obtained for EP with metallocene catalyst as shown in
Figure 11.13 [74]. A statistical approach was used, based on a first-order Markovian
distribution, which indicated that initiating units were almost exclusively propyl groups.
Mixing and extrusion techniques are usually applied to elastomers for processing and
improvement of physical properties. These mechanical processes normally lead to the
occurrence of radical reactions such as chain scission, chain coupling, and crosslinking.
The 13C-NMR analysis of EP showed that the shearing of polymer resulted in hydrogen
abstraction. This was followed by disproportionation reaction to form olefins [75].
424
(a) (b)
Figure 11.13 Structures and 13C-NMR signal assignments for: (a) olefinic end-groups,
and (b) aliphatic end-groups. P denotes the polymer chain and polymerisation occurs
from left to right [74]
Reproduced with permission from J.C. Randall and S.P. Rucker, Macromolecules, 1994,
27, 2120, Figures 3 and 4. Copyright 1994, American Chemical Society
Detailed assignment of signals was carried out for various types of end- group of EP with
molecular weigh ca. 1.7 x 103, obtained after high temperature extrusion [76]. In this
case, 2D NMR measurements at 300 Hz were able to be applied due to the low molecular
weight of the samples.
A mechanism involving hydrogen abstraction followed by disproportionation is presumed

to form different types of olefins. The 1H-NMR signals in the olefinic region were assigned
based on information from 1H-13C 2D correlation measurement. The molecular weight
was estimated based on the relative intensity of olefinic proton signals from the end-
groups. A reasonable agreement was observed between the molecular weight by
1
H-NMR and by SEC for EP samples decomposed by shearing at various temperatures
[76]. The concentration of end-groups increased in the order of vinylidene>vinyl>vinylene
for all samples prepared at extruder shearing temperature from 260° to 450 °C.
425
The mechanism of γ-ray irradiation-induced scission of polyisobutylene was studied,

based on the structural characterisation of end-groups by 13C-NMR as well as GC, GC/
MS, and SEC [77]. The assignments of signals were made by comparison with those
from model compounds and predictions based on empirical rules. Quantitative 13C-NMR
measurements of chain-ends allowed the determination of radiation yield of products
and of chain scission.
Carboxyl-terminated butadiene-acrylonitrile copolymer, mentioned above, was found

to have 4-cyanopentanoic acid end-group originating from 4,4,7-azobis(4-cyanopentanoic
acid) initiator [32].
11.4.4 Structure of Branch-points and Coupling Points
The incorporation of a special functional group at the chain-end of rubber molecules

was recognised as a way to improve dynamic properties of vulcanised rubber with carbon
black, such as resilience, heat build-up, and tan δ [65]. Some coupling agents, used to
combine chain-ends [66] or chains to carbon black [78], are also known to play a similar
role. Living styrene-butadiene copolymer having butadienyl-Li end-groups was prepared
with n-butyllithium/THF initiator by adding a small amount of butadiene at the final
stage of polymerisation. The resulting living copolymer was coupled with tin tetrachloride
[66]. The presence of terminal butadienyl carbon linked to tin was presumed based on
Sn-C-C=C- signals around 11 and 14 ppm in the 13C-NMR spectrum, which were also
observed in model compounds. The resulting rubber showed a bimodal molecular weight
distribution, suggesting the presence of unreacted chains, although quantitative
measurements were not presented. It is remarkable that the coupling of butadienyl-tin is
more effective than that of styryl-tin for the improvement of dynamic properties. A similar
modification at the chain-end of neodymium catalysed living cis-1,4 polybutadiene was
also attempted [79,80].
Butyl rubber is a copolymer of isobutylene with 1-3 mol% isoprene. The isoprene units
are presumed to be predominantly in the trans-1,4 addition and in head-to-tail
arrangement [81,82]. However, a very small doublet signal appears at 4.93 ppm in 1H-
NMR, in addition to a triplet signal at 5.05 ppm, which was assigned to the olefinic
proton of isoprene units in 1,4 addition, as shown in Figure 11.14. The doublet signal
was tentatively assigned to the isoprene unit in 1,2 or 4,3 addition [14,83]. Detailed
assignment of this signal was made by using 2D methods at 500 MHz, such as 1H-1H
COSY and hypercomplex phase-sensitive inversely detected heteronuclear multiple-
quantum coherence (HMQC) [84]. These experiments indicated that the signal at
4.93 ppm corresponds to =CH coupled with a signal at 2.44 ppm having equal intensity
and 1.72 ppm having 3 times in the intensity. As a model, isobutylene-d 8 was
426
Figure 11.14 1H-NMR spectrum of (a) commercial butyl rubber and

(b) polyisobutylene, observed at 500 MHz [84]
Reproduced with permission from J.L. White, T.D. Shaffer, C.J. Ruff and J.P. Cross,
copolymerised with isoprene. The aliphatic methine signal at 2.44 ppm appeared as a
sharp doublet with the same splitting as that of 4.93 ppm signal. These findings indicate
that the signals are derived from the isoprene unit in Figure 11.15, consisting of a branch-
point. The presence of long-chain branching was supported by SEC and rheological
measurements.
Figure 11.15 Presumed structure of the branch point of butyl rubber
427
The 1H-, 13C- and 2D NMR techniques were applied to the analysis of short chain
branching of ethylene copolymers synthesised at high pressure by free-radical initiator
[85]. The level of short chain branching (branches of length C5 or shorter) increases
from about 6 to 20 branches per 1000 CH2 for ethylene homopolymer with increasing
the polymerisation temperature from 165 °C to 270 °C. By comparison with the spectrum
of ethylene homopolymer, the assignment of 13C-NMR signals due to branches was made
for ethylene copolymers, such as n-butyl acrylate, methyl acrylate, vinyl acetate, n-butyl
methacrylate, acrylic acid, and methacrylic acid. Confirmation of the assignments has
been obtained in a 2D HSQC-TOCSY (heteronuclear single quantum correlation-total
correlation spectroscopy) experiments. Ethylene-vinyl acetate copolymer (EVA) showed
two additional signals in the ethyl region of the 13C-NMR spectrum, as shown in
Figure 11.16. In contrast to the acrylate copolymers, branching structure of EVA is more
similar to ethylene homopolymer as summarised in Table 11.5.
Figure 11.16 13C-NMR signals due to short chain branches in ethylene-vinyl acetate
copolymers, observed at 100 MHz: (a) ethylene homopolymer (195 °C), (b)14 wt%
VAc (200 °C), (c) 22 wt % VAc (200 °C) and (d) 38 wt % VAc (200 °C) [85]. (The
notation of carbons is given in Table 11.4)
Reproduced with permission from E.F. McCord, W.H. Shaw Jr. and R.A. Hutchinson,
428
Table 11.5 Short chain branch (SCB) structures in ethylene-vinyl acetate

copolymers from 13C-NMR analysis a [85]
Trct (°C) wpVAc VActot 1B2 2B4 1A3B4+ 2B5+ SCBtot
165 0.000 0 0.9 3.3 0 3.2 7.4
165 0.141 26 1.1 3.7 0.7 3.3 9.0
165 0.235 48 0.7 3.4 0.7 3.3 8.3
165 0.386 93 1.4 2.8 1.6 4.0 10.0
195 0.000 0 1.7 4.3 0 3.7 9.7

200 0.140 26 1.9 5.7 0.7 4.9 13.8
200 0.219 44 1.4 4.8 1.2 4.9 12.8
200 0.230 46 1.5 5.3 1.2 5.1 13.4
200 0.381 91 0.9 3.7 2.2 4.9 12.0
a
All numbers (columns 3-8) per 1000 CH2 units
The second carbon from the end of a butyl branch is 2B4. xBy+ refers to branches of
length y and longer and includes end-of-chain (EOC) contributions. xAQBy indicates that
the comonomer unit located off the backbone carbon at the branch point, making it
quaternary
Reproduced with permission from E.F. McCord, W.H. Shaw and R.A. Hutchinson,
Macromolecules, 1997, 30, 246, Table 3. Copyright 1997, American Chemical Society
11.5 Structural of Naturally Occurring Polyisoprenes
11.5.1 Structure of Natural cis- and trans-Polyisoprenes
Naturally occurring polyisoprenes are composed of isoprene-units in the cis or trans

configuration. Natural rubber from H. brasiliensis (NR) and Gutta percha from P. gutta
are typical cis-1,4 and trans-1,4 polyisoprene, respectively. On the other hand, Chicle
from A. zapota contains both cis-1,4 and trans-1,4 polyisoprenes. NR has been presumed
to be a pure cis-1,4-polyisoprene, containing no detectable amount of trans-1,4 and
3,4 units as indicated by 1H-NMR at 100 MHz [86]. Natural isoprene oligomers consisting
of 9-23 isoprene-units, referred to as polyprenol, can be a good model for structural
429
analysis of naturally occurring polyisoprenes. Polyprenol is classified into three categories,

i.e., all-trans (Solanesol-type), two-trans and poly-cis (Betulaprenol-type), and three-trans
and poly-cis (Ficaprenol-type). However, a polyprenol of all-cis configuration has not
been reported [87].
The sequence distribution of both trans and cis isoprene-units in polyprenol has been
analysed by 13C-NMR, using acyclic terpenes of all the arrangements of both trans and
cis isoprene-units, corresponding to isoprene dimer to tetramer. The chemical shifts of
signals reflecting the sequence distribution are independent of the degree of polymerisation
for polyprenols longer than 10-mer [88-90], as shown in Figure 11.17. Here, the
dimethylallyl group and hydroxylated isoprene-unit are referred to as ω and α terminals,
respectively, and the carbon atoms in isoprene-unit are numbered as indicated in
Figure 11.17. The absence of cis-trans linkage in all the polyprenols and the presence of
only one trans-cis linkage per chain indicate that the trans-cis polyprenols are di-block
copolymer, aligned in order of ω-(trans)2 or 3-(cis)n-OH. Solanesol is a homologue of all-
trans acyclic terpenes.
Figure 11.17 Relationship between length of isoprene-units and chemical shifts of 13C-
NMR signals reflecting the arrangement of isoprene-units, observed at 50 MHz [88]
Reproduced with permission from Y. Tanaka in NMR and Macromolecules, ACS

Symposium Series No. 247, Ed., J.C. Randall, 1984, ACS, Washington, DC, p241,
Figure 3. Copyright 1984, American Chemical Society
430
Polyisoprene containing both cis and trans isoprene-units can be prepared by UV

irradiation in the presence of thiobenzoic acid [91]. The 13C-NMR study has indicated
that cis-trans isomerised polyisoprene is a random copolymer of both isomeric units
[92]. Trans-1,4 polyisoprene from Chicle, separated by crystallisation from hexane
solution, exhibits the same 13C-NMR signals as those of solanesol, including both terminal
groups [93]. This indicates that trans polyisoprene in Chicle is a high molecular weight
homologue of solanesol, consisting of about 100 trans-1,4 isoprene-units. Cis polyisoprene
in Chicle showed no signal due to the cis-trans linkage, suggesting that Chicle polyisoprene
is a mixture of both pure trans-1,4 and cis-1,4 polyisoprenes.
Gutta percha and Balata contain high molecular weight ( M n = 1.4-1.7 x 105) trans-1,4
polyisoprene [94,95]. Both shows 13C-NMR signals due to the dimethylallyl group and
hydroxylated terminal group. It is remarkable that the diphosphate terminal group was
observed in the low molecular weight trans-1,4 polyisoprene from Eucommia ulmoides
[96]. The C-4 methylene protons of terminal trans isoprene-unit showed a splitting by
coupling from both =CH and 31P in the 1H-NMR spectrum. The splitting pattern is the
same as that observed for solanesyl diphosphate.
The presence of dimethylallyl group (ω-terminal) and trans isoprene-units in cis-1,4

polyisoprene was clearly observed in the 13C-NMR spectrum of rubber from Lactarius
volemus, a rubber-producing mushroom [97,98]. The structure of mushroom rubber
was determined to be a high molecular weight homologue of two-trans polyprenol,
based on the signal assignments mentioned above, as depicted in Figure 11.18 (1) [89,
93, 95]. The presence of trans isoprene-units and both terminal groups was also confirmed
in the low molecular weight rubbers from the leaves of Goldenrod (Solidago altissima)
and Sunflower (Helianthus annuus) [94,99].
The arrangement of trans isoprene-units is observed more in detail from the 13C-NMR
signal splitting of C-1 methylene carbon [94, 97], as shown in Figure 11.19. The
mushroom rubber shows two peaks due to the trans isoprene-unit in dimethylallyl-
trans and trans-trans linkages similar to those of the two-trans polyprenol. On the
other hand, the rubbers from Goldenrod and Sunflower exhibit three peaks similar to
the overlap of peaks of two-trans and three-trans polyprenols. This suggests the presence
Figure 11.18 The structure of rubber from Lactarius volemus
431
Figure 11.19 Splittings of the C-1 methylene carbon signal of: (a) rubber from L.
volemus, (b) three-trans Ficaprenol-11, (c) two-trans Betulaprenol-18, (d) rubber from
Sunflower, (e) rubber from Goldenrod, (f) three-trans Ficaprenol-12 and (g) two-trans
Betulapreno-16. (13C-NMR measurements at 100 MHz for (a) to (c) at room
temperature for (d) to (f) at 50 °C)
of two types of rubber chains containing two- or three-trans isoprene-units as a mixture,

which could be derived from different initiating species of rubber formation such as
trans,trans-farnesyl diphosphate (FDP) and trans,trans,trans-geranylgeranyl
diphosphate (GGDP) [100].
432
11.5.2 Structure of Natural Rubber
11.5.2.1 Trans Isoprene-units in Natural Rubber
Natural rubber from H. brasiliensis showed clear 13C-NMR signals due to trans isoprene-
units in the trans-trans linkage, the relative intensity of which decreases with increasing
the molecular weight of fraction, as shown in Figure 11.20. These trans isoprene-units
were presumed to be derived from an initiating species of rubber formation and not
isomerisation of cis isoprene-units, because the trans isoprene-units were in the trans-
trans linkage and not in the cis-trans linkage [101]. The 1H-NMR spectrum of polyprenols
at 500 MHz gives isolated signals for the methyl protons of trans isoprene-units in trans-
trans and trans-trans-trans sequences [102]. The methyl proton signals of NR shows
striking resemblance to two-trans polyprenol, although the signal from dimethylallyl
group itself is not detected as shown in Figure 11.21. Accordingly, the structure of NR
was postulated to be a kind of two-trans and poly-cis rubber as depicted in Figure 11.22,
where ω- and α-ends are unidentified terminal group [103].
Wild rubbers, occurring as latex from higher plants such as Sorvinha (Coma utilis),
Indian Laurel (Ficus retusa) and Jelutong (Dyera costulata), show no signal due to trans
isoprene-units as well as both terminal groups in the 13C-NMR spectrum [94, 104]. The
absence of these groups was presumed to be due to a certain chemical or biochemical
oxidative degradation of the rubber molecule during storage in the tree. A similar oxidative
degradation of rubber in the tree was observed for NR from firstly tapped Hevea tree,
i.e., virgin tree. The rubber from a virgin tree showed no signal due to trans isoprene-
units [105,106].
NR contains about 2% proteins and 1% lipids. Deproteinisation of NR with a proteolytic

enzyme in latex state reduced the gel content, indicating the decomposition of branch-
points consisting of proteins [107,108]. Transesterification of deproteinised NR (DPNR)
with sodium methoxide decomposes fatty acid ester and phosphoric ester linkages in NR
to form linear rubber chains [107-109]. The amount of trans isoprene-units per rubber
chain was determined from the ratio of cis and trans isoprene-units by 13C-NMR, by
considering the number-average molecular weight obtained by osmometry. The observed
number of trans isoprene-units in transesterified DPNR was about two independent of
the molecular weight [103,109]. The presence of exactly two trans isoprene-units was
also confirmed by comparing the 1H-NMR spectrum with those of two-trans and three-
trans polyprenols as mentioned above (cf. Figure 11.21).
433
Figure 11.20 13C-NMR spectra of fractionated deproteinized NR from fresh field latex
with M n of (a) 2.7 x 105, (b) 0.93 x 105 and (c) 0.30 x 105, observed at 100 MHz [103]
Reproduced with permission from Y. Tanaka, A.H. Eng, N. Ohya, N. Nishiyama, J.

Tangpakdee, S. Kawahara and R. Wititsuwannakul, Phytochemistry, 1996, 41, 1501.
Copyright 1996, Elsevier Science
434
Figure 11.21 1H-NMR spectra of (a) low molecular weight fraction of deproteinized
NR (b) two-trans Betulapreno-16 and (c) three-trans Ficaprenol-12, observed at
400 MHz [103]
Figure 11.22 Presumed structure of natural rubber
435
11.5.2.2 Structure of Terminal Groups
The absence of dimethylallyl-group in NR indicates that the initiating species for rubber
formation in Hevea tree is not FDP, but FDP modified at the dimethylallyl-group, which
is abbreviated here as ω′ [103,109,110]. This was confirmed by 13C-NMR analysis of in
vitro polymerised rubber by incubation of the bottom fraction of fresh latex and
isopentenyl diphosphate (IDP) [111]. The newly synthesised in vitro rubber formed in
the presence of FDP and IDP showed the dimethylallyl group derived from FDP. On the
other hand, no dimethylallyl group was detected in the in vivo rubber prepared without
the addition of FDP [112].
It was disclosed that NR contained about two long-chain fatty acids per rubber chain
linked to rubber chain, which were able to remove after transesterification or saponification
of NR [107,113]. Most of NR molecules were presumed to be terminated with a
phospholipid, taking into the account of the fact that most of rubber particles in fresh latex
have no activity for chain elongation reaction [113]. Based on these findings NR was
presumed to contain peptide group and phospholipid or phosphate group at the initiating
and terminating chain-ends, respectively. Both functional terminal groups can form branch-
points, which are able to decompose by deproteinisation and transesterification (or
saponification) [109]. Outstanding properties of NR was presumed to be derived from the
long-chain fatty acids included in phospholipid at the α-terminal [109].
11.6 Application of High-Resolution NMR
11.6.1 Multinuclear High-resolution NMR
11.6.1.1 2H-NMR
Multinuclear high-resolution NMR other than 1H, 19F and 13C has a versatile application
for the structural analysis of polymer, because some nuclei give information about the
structure of polymer chain liked to terminal groups, coupling points, and mechanisms of
initiation and termination reactions. The structure of chain-end group of living
polybutadiene and polyisoprene, quenched with deuteromethanol, was analysed by 2H-
NMR. The following terminal groups of butadiene and isoprene oligomers, synthesised
with ethyllithium in hexane, were confirmed by 2H-NMR [73] (see Figure 11.23).
A catalyst system composed of neodymium chloride/trialkyl aluminium is presumed to

give living polymer of butadiene containing cis-1,4 units higher than 98%. This living
436
Figure 11.23 Presumed structure of chain-end groups of polybutadiene and

polyisoprene initiated with ethyllithium
polymerisation of isoprene is reported to accompany side reactions such as chain transfer

to monomer. The polyisoprenes quenched with deuteromethanol showed signals due to
3,4 (36-45%) in addition to cis-1,4 terminal units [73], as shown in Figure 11.24. In
polybutadiene, trans-1,4 terminal units were predominant, but the formation of 1,2
terminal group was also observed. The initial unit of polyisoprene polymerised with
neodymium catalyst in the presence of DAl(i-Bu)2, which is known to act as chain transfer
agent, was predominantly trans-1,4 units. It was explained that the principal site of the
chain transfer reaction was the Al-D bond.
The 2H-NMR was applied to the analysis of cross-linking and branching structure in
copolymers composed of butyl acrylate, methyl methacrylate, and small amounts of
methacrylic acid and allyl-d5 methacrylate (ALMA-d5) to generate crosslinks. The solution
state 2H-NMR spectrum of the copolymer showed clear signals due to the ALMA-d5
units in branch-points. The allyl group in ALMA-d5 was presumed to react with a growing
radical chain predominantly by the reaction of the vinyl group at the side chain as shown
in Figure 11.25 [114].
11.6.1.2 7Li-NMR
7
Li-NMR has been widely applied to the structural characterisation of lithium initiators
for anionic polymerisation. Anionic polymerisation of diene monomer with lithium
437
Figure 11.24 2H-NMR spectra of polyisoprenes obtained with Nd/Al catalyst and
quenched by CH3OD, observed at 76.77 MHz [73]
Reproduced with permission from K.D. Skuratov, M.I. Lobach, A.N. Shibaeva, L.A.
Chunyaeva, T.V. Erokhina, L.V. Osetrova and V.A. Kormer, Polymer, 1992, 33, 5197,
Figure 1. Copyright 1992, Elsevier Science
438
Figure 11.25 Formation of branch point in polybutylacrylate
initiators gives living polymer. Bifunctional lithium compounds are very important
initiators for the preparation of telechelic diene polymers and triblock copolymers. Diene
polymers composed of 1,4 units are obtained through organolithium compounds in non-
polar solvents. However, the aggregation or association of both initiators and living
chains sometime causes the formation of a broad molecular weight distribution. Initiator
systems can be modified by the addition of a special amine or ether and so-called seeding
reactions to form polydienyllithium by the addition of small amounts of a diene monomer.
7
Li-NMR gives information on the association, i.e., the most associated species show a
signal at the lowest chemical shift.
A dilithium initiator, 1,3-phenylene-bis(3-methyl-1-[methylphenylene]-pentylidene)

dilithium produced triblock copolymers of styrene-butadiene-styrene (SBS) and styrene-
isoprene-styrene (SIS) in hydrocarbon solvent. However, resulting copolymers showed
broad and bimodal molecular weight distribution [115]. 7Li-NMR indicated that this
compound exists as a complex mixture of aggregated species, showing wide chemical
shift ranges, while a monolithium compound ((1,1-diphenyl-3-methylpentylidene)lithium)
and sec-butyllithium gave a sharp signal. Here, the monolithium compound was presumed
to be in dimeric aggregation. The addition of equivalent amounts of N,N,N′,N′′,N′′-
pentamethyldiethylenetriamine (PMDETA) to the dilithium compound changed these
signals to a sharp signal shifted downfield, showing the deaggregation.
The addition of tert-butyllithium to 1,2-bis(isopropenyl-4-phenyl)ethane in hexane leads

to the formation of a dilithium compound. The bifunctional initiator in hydrocarbon solvent
provided diene polymers with unimodal molecular weight distribution. The molecular
weight of resulting polymer was the same as that of theoretical one [116]. 7Li-NMR of the
resulting α, ω-diithio-polyisoprenes showed a major signal at 0.03 ppm, which was almost
the same as the signal of monolithio-polyisoprene, and small signals at -1.64, -1.85 and
-1.92 ppm, as given in Figure 11.26. The major signal was assigned to dimeric aggregation
and small signals tentatively to cyclic aggregates involving cyclic and linear species involving
four lithium atoms. The aggregate species were presumed to be dependent on the chain
length. It was also observed that α,ω-diithio-polyisoprenes of high degree of polymerisation
exhibited 7Li-NMR similar to those of monofunctional ones [117].
439
Figure 11.26 7Li-NMR spectrum of α,ω-dilithio-polyisoprenes (DPn = 3), 4 x 10-4 M in

hexane/C6D6 at 7 °C, observed at 194.3 MHz [116]
Reproduced with permission from J.M. Boutillier, J.C. Favier and P. Hemery, Polymer,
1996, 37, 5197, Figure 4. Copyright 1996, Elsevier Science
11.6.1.3 29Si-NMR
29
Si-NMR has been used to determine chain and block length of oligo- and polysiloxanes.
The 29Si chemical shift ranges for some structural units in polysiloxanes have been studied
[118]. Poly(dimethylsiloxane) and poly(dimethylsiloxane-co-methyl(trifloropropyl)siloxane)
were analysed using 13C- and 29Si-NMR [119]. A relatively long spin-lattice relaxation
time (T1) and negative NOE were overcome by the addition of a shiftless relation reagent
440
such as Fe(acac)3 and by using inverse gated decoupling technique, respectively, to make
quantitative 29Si-NMR measurement. The sequence length in poly(dimethylsiloxane)/
poly(tetramethyl-p-silphenylene-siloxane) (PDMS-PTMPSS) copolymers have been analysed
by 29Si-NMR [29]. The 29Si-NMR spectrum of copolymer exhibits five signals reflecting
the sequence distribution, as shown in Figure 11.27. The assignments were made based on
known 29Si chemical shifts. It was shown that redistribution of siloxane bonds proceeded
during polymerisation by the presence of water.
Sequencing of siloxane units was also investigated for anionic ring-opening polymerisation
of 1,1-diphenyl-2,2,3,3-tetramethylcyclotetrasiloxane by 29Si-NMR [120].
Figure 11.27 29Si-NMR spectrum of a silphenylene-siloxane block copolymer

(67 mol % DMS), observed at 59.6 MHz [29]
Reproduced with permission from E.A. Williams, J.H. Wengrovius, V.M. van Valkenburg
and J.F. Smith, Macromolecules, 1991, 24, 1445, Figure 1. Copyright 1991, American
Chemical Society
441
11.6.1.4 31P-NMR
Chemical degradation of microporous polyurethane elastomer was investigated by using

1
H-, 13C- and 31P-NMR [121]. The degradation products polyester-polyurethane by
phosphoric acid esters, such as (CH3CH2O)3PO and (ClCH2CH2O)3PO, gave a major
31
P-NMR peak at 0.24 ppm and -0.98 ppm as a septet, respectively. These peaks were
assigned to phosphorous atom in the oligomer product with terminal phosphate groups,
formed by exchange reaction as shown in Figure 11.28:
Figure 11.28 Structure of terminal phosphate groups of polyester-polyurethane by

degradation with phosphoric acid esters
11.6.1.5 119Sn-NMR
Polybutadiene and SBR can be anionically synthesised with tributyltin lithium (TBTLi)
initiator in hexane/THF [122]. The 119Sn-NMR chemical shift of trialkyltin compounds
is sensitive to the structure of substituent. For example, Bu3Sn-CH2CH=CHCH3 shows
a signal at -17.79 and -13.69 ppm from Me4Sn for trans and cis isomers, respectively.
Low molecular weight polybutadiene initiated with TBTLi and terminated with tributyltin
chloride (TBTCl) exhibited signals reflecting the structure of butadiene units at both
terminals as shown in Figure 11.29. It is clear that the signals a and e are observed only
after termination with TBTCl and that the signal d is predominantly formed by initiation.
Each signal was assigned to the sequences of 1,4 and 1,2 units, based on information
from model compounds, as tabulated in Table 11.6. This indicates that the initiation
starts from trans-1,4 and both 2,1 and 1,2 additions of butadiene units to tin-lithium
bond and subsequent butadiene units can be added to form either cis-1,4, 2,1, 1,2 or
trans-1,4 units. It is remarkable that no styrene unit was observed at the initiating terminal
in the case of copolymerisation with butadiene and styrene.
442
Figure 11.29 119Sn-NMR of polybutadiene initiated with tributyltin Li (a) and

terminated with tributyltin chloride (b), both 15% solutions in CDCl3, observed at
111.9 MHz [122]
Reproduced with permission from W.L. Hergenrother, J.M. Doshak, D.R. Brumbaugh,
T.W. Bethea and J. Oziomek, Journal of Polymer Science, Polymer Chemistry Edition,
1995, 33, 143, Figure 3. Copyright 1995, John Wiley & Sons
11.6.2 NMR of Swollen State and Latex
11.6.2.1 Swollen State NMR
High-resolution NMR in solution requires the sample to be soluble in a solvent such that
the various nuclear spin interactions can be averaged or removed by molecular micro-
Brownian motions. Unfortunately, elastomers used in various applications are normally
crosslinked materials and therefore not soluble in any solvent. Thus, solid state NMR with
magic angle-spinning technique has been used with great success in the study of cured
elastomers. However, this technique demands extended instrument facilities and expertise.
It is well known that if a cured rubber sample were to be analysed by a conventional

solution NMR spectrometer, very broad signals would be obtained. The resolution of
the spectrum can be improved through swelling the rubber sample in a solvent. This
443
Table 11.6 Assignment of 119Sn-NMR signals of tributyltin-end-capped

polybutadiene [122]
Compound ppm from Me4Sn Peak Region
P—CH2CH=CHCH2 —(trans) -18.17 to -18.34 f
CH=CH2
P—CH2CHCH2CHCH2CH -17.39 to -17.47 e
CH=CH2 CH=CH2
CH=CH2
-15.97 to -16.33 d
P—CHCH2—
CH=CH2
P—CH2CHCH2 CH— -15.46 to -15.51 c
CH=CH2
P—CH2CH—
-15.17 to -15.26 b
CH=CH2
P—CH2CH=CHCH2 —(cis) -14.60 to -14.76 a

Reproduced with permission from W.L. Hergenrother, J.M. Doshak, D.R. Brumbaugh,
T.W. Bethea and J. Oziomek, Journal of Polymer Science: Polymer Chemistry Edition,
1995, 33, 143, Table 5. Copyright 1995, John Wiley & Sons
technique has been applied to the identification, analysis of molecular structure [123],
and analysis of crosslink density of blends of vulcanised rubber [124]. Figure 11.30
shows the 13C-NMR spectra of polychloroprene rubber, obtained in solution (uncured),
solid (cured), and swollen (cured) states. The rubber in swollen state has enhanced
molecular mobility and therefore it showed improved spectral resolution. The enhanced
spectral resolution makes it possible to calculate the ‘head-to-tail’ and ‘tail-to-head’ ratios
for the triads composing of 1,4-trans unit in the polychloroprene rubber [123].
In the case of crosslinked EPDM, the spectral resolution was sufficiently good to carry
out the monomer sequence analysis [123]. A similar technique was also applied to NBR.
The triad sequence distribution of monomers was determined, by simplifying the
copolymer to be consisting of acrylonitrile and butadiene, without discriminating between
cis, trans, and vinyl isomeric units [123]. The results indicated the anticipated distribution
in the rubber, reflecting the reliability of this technique in the analysis. Based on the
444
Figure 11.30 13C-NMR spectra of polychloroprene: (a) solution state, (b) solid state,
(c) swollen state, observed at 100 MHz [123]
Reproduced with permission from I. Fonao, L. Gonzalez, M.L. Jimeno and A. Marcos,
Kautshuk und Gummi Kunststoffe, 1993, 46, 431, Figure 3. Copyright 1993,
Hüthig GmbH
445
intensity ratios, the acrylonitrile content for several NBR rubbers was calculated and
found to be close to the data supplied by the manufacturer. Analysis on NBR/PVC blends
revealed that the weight fraction of each polymer was comparable to the known value
[123]. For cured blends, however, the use of this technique to the composition analysis
was less satisfactory due to the uneven distribution of crosslink density in the two rubber
components [125]. Thus, the rubber with higher crosslink density gave smaller 13C-NMR
signal area, resulting in less detection of this rubber in the blend.
The signal broadening in 1H-NMR and 13C-NMR spectra, which caused by a reduction
in molecular mobility in a rubber upon crosslinking, has been applied to the estimation
of crosslink density in rubber blends [123]. The signal broadening effect could be correlated
with crosslink density of the individual rubber, estimated by other methods such as
equilibrium swelling or rheometer torque values [125]. For rubber blends prepared from
unsaturated rubber, 1H-NMR can be used because the signal of the olefinic protons
could be clearly observed without interference. In the case of blends containing saturated
rubber, 13C-NMR will be the better choice because there are too many 1H-NMR signals
in the alkyl proton regions [126,127]. The factors affecting the accuracy of this technique
were also investigated [128].
11.6.2.2 Latex State NMR
Many commercial elastomers are being produced in the form of latex. Characterisation
of these samples in solution state might be difficult because these elastomers normally
contain both sol and gel fractions. Therefore, attempts have been made to characterise
directly these latices using high-resolution NMR spectroscopy. Early work on latex NMR
involved the studies of acrylic latex [129,130]. The 13C-NMR signal is related to the
polymer Tg and the temperature of measurement. The structure of polymer in the latex
particles was also investigated in terms of temperature dependence of 13C-NMR line-
widths [131]. The plot of line width against temperature yields a curve with two linear
regressions. The intersection of the two linear plots is defined as breakpoint temperature.
This temperature varies systematically with the interpenetrating network structure. The
changes in 13C-NMR line-width are sensitive to the variation of crosslink density in an
interpenetrating polymer network [131].
High mobility hydrophilic interfacial polymer chain will expand in the aqueous phase and
therefore exhibit narrow lines in the NMR spectrum. The N-methylol acrylamide (NMA) in
the functionalised butyl acrylate latex showed narrower carbonyl line than that of butyl
acrylate [132]. The concentration of these hydrophilic groups is affected by the monomer
concentration of NMA and also the timing of the addition of the monomer during the reaction.
446
Recently, a novel method for determining the microstructure of crosslinked polybutadiene
in latex using solution 13C-NMR technique was reported [133]. The surfactant and
polymer concentrations in the latex were adjusted to give a good signal resolution of the
latex sample, as indicated by half-width of the resonance peak at 32.7 ppm. Under these
conditions, the S/N ratio was almost identical to that of sample in solution, as shown in
Figure 11.31. The microstructure of sol and gel fractions in a radical initiated
polybutadiene, determined by this technique, was similar to that of solution measurements.
Figure 11.31 13C-NMR spectra of polybutadiene containing 80% gel: (a) Solution
measurement (10 w/v% CDCl3 solution), (b) latex measurement (10 w/v % dry rubber
content), and (c) solid state measurement [133]
Reproduced with permission from S. Kawahara, S. Bushimata, T. Sugiyama, C.

Hashimoto and Y. Yanaka, Rubber Chemistry and Technology, 1999, 72, 2, 848,
Figure 2. Copyright 1999, American Chemical Society
447
The effect of particle size and spinning of the NMR tube were studied for the latex state
13
C-NMR of natural rubber latex fractionated by particle size [134]. High-resolution
spectrum was obtained by measurement without sample spinning. The diffusion constant
of Brownian motion was found to be a dominant factor governing the intensity and half-
width of the signals. As the particle size decreased and temperature of measurement was
raised, the intensity of signals increased and was comparable to the theoretical value,
which was observed by the addition of triethylene glycol as an internal standard.
11.7 Conclusion
High-resolution NMR has expanded its applicability to high molecular weight elastomers,
with increasing the magnetic field from 60-100 MHz in the early stage to 500-750 MHz
in modern 1H-NMR spectrometer. Improvement of the signal-to-noise ratio as well as
the resolution of NMR spectrum makes it possible to detect very small amounts of sample
and small signals derived from irregular structures of elastomers. This induced new
application fields such as on-line HPLC-NMR measurement and analysis of end groups
and branch points in addition to the conventional analysis of chemical composition,
sequence distribution, microstructure of diene polymers. Applicability of many NMR
pulse techniques, which have been widely applied in the case of low molecular weight
compounds, to the assignment of small signals in elastomers increased the number of
works in the field of structural characterization of elastomers. Now it is possible to
analyse the structure of terminal groups, branch points, and crosslink points of elastomers
with molecular weight higher than 105 by the use of modern NMR pulse techniques.
These analyses provide direct information between the structure and physical properties
as well as polymerisation mechanism including biosynthesis mechanism of natural rubber.
It should be noted, however, the reliability of NMR analysis depends on the accuracy
and precision of measurement in addition to the validity of signal assignment. Various
traditional techniques, such as use of model compounds, partially deuterated polymers,
13
C-enriched polymers, and shift reagents, are very useful for signal assignment by
combination with NMR pulse sequence techniques. Usage of nuclei other than 1H, 13C,
and 19F will develop new application field of NMR of elastomers. Swollen state and
latex state NMR will be just the method specialized to elastomers, because of no
disturbance effect by the presence of gel fraction for these methods.
References
1. H.J. Harwood, Rubber Chemistry and Technology, 1982, 55, 769.
2. D.A. W. Wendisch, Applied Spectroscopy Reviews, 1993, 28, 165.
448
3. J.L. Koenig, Spectroscopy of Polymers, American Chemical Society, Washington,
DC, USA, 1992.
4. J.C. Randall, Polymer Sequence Determination. Carbon-13 NMR Method,

Academic Press, New York, USA, 1977.
5. F.A. Bovey and P.A. Mirau, NMR of Polymers, Academic Press, New York, NY,
USA, 1996.
6. J. Kelm, Forschungsbericht 213 - Carbon and Proton NMR Spectra Catalogue,

Wirtschaftsverlag, Germany, 1995.
7. R. Chujo, K. Hatada, R. Kitamaru, T. Kitayama, H. Sato and Y. Tanaka, Polymer

Journal (Japan), 1987, 19, 413.
8. R. Chujo, K. Hatada, R. Kitamaru, T. Kitayama, H. Sato, Y. Tanaka, F. Horii and

Y. Terawaki, Polymer Journal (Japan), 1988, 20, 627.
9. K. Horii, M. Nakagawa, R. Kitamaru, R. Chujo, K. Hatada and Y. Tanaka,

Polymer Journal (Japan), 1992, 24, 1155.
10. K. Hatada, T. Kitayama, Y. Terawaki, H. Sato, R. Chujo, Y. Tanaka, R. Kitamaru,

I. Ando, K. Hikichi and F. Horii, Polymer Journal (Japan), 1995, 27, 1104.
11. K. Hatada, T. Kitayama, K. Ute and Y. Terawaki, Polymer Preprints Japan, 1987,
36, 3112.
12. K. Hatada, Y. Terawaki and T. Kitayama, Kobunshi Ronbunshu, 1992, 49, 335.
13. K. Hatada, T. Kitayama, K. Ute, Y. Terawaki and T. Yanagida, Macromolecules,

1997, 30, 6754.
14. J.E. Puskas and C. Wilds, Rubber Chemistry and Technology, 1994, 67, 329.
15. J. Puskas, J. Schmidt, P. Collart and M. Verhelst, Kautschuk und Gummi

16. S.D. Martino and M. Kelchtermans, Journal of Applied Polymer Science, 1995,
56, 1781.
17. D.D. Werstler, Rubber Chemistry and Technology, 1980, 53, 1191.
18. N.G. Walsh, J.K. Hardy and P.L. Rinaldi, Applied Spectroscopy, 1997, 51, 889.
19. K. Albert, Journal of Chromatography A, 1995, 703, 123.
449
20. J.C. Lindon, J.K. Nicholson and I.D. Wilson in Advances in Chromatography,
Volume 36, Eds., P.R. Brown and E. Grushka, 1996, 315.
21. K. Hatada, K. Ute, T. Kitayama, M. Yamamoto, T. Nishizawa and M.

Kashiyama, Polymer Bulletin, 1989, 21, 489.
22. K. Hatada, K. Ute, T. Kitayama, T. Nishimura, M. Kashiyama and N. Fujimoto,

Polymer Bulletin, 1989, 22, 549.
23. K. Ute, R. Niimi, S. Hongo and K. Hatada, Polymer Journal (Japan), 1998, 30, 439.
24. K. Ute and K. Hatada, International Journal of Polymer Analysis and

Characterisation, 1999, 5, 47.
25. C.J. Carman, R.A. Harrington and C.E. Wilkes, Macromolecules, 1977, 10, 536.
26. G. van der Velden, Macromolecules, 1983, 16, 85.
27. O. Chiantore, P. Cinquina and M. Guaita, European Polymer Journal, 1994, 9,

1043.
28. J. Liu, S.G. Park and S.W. Ko, Polymer, 1998, 39, 1051.
29. E.A. Williams, J.H. Wengrovious, V.M. Van Valkenburgh and J.F. Smith,
30. G.N. Babu, and R.A. Newmark, Macromolecules, 1991, 24, 4503.
31. X. Chen, L. Hu, B. Yan and S. Jiao, Chinese Journal of Polymer Science, 1990, 8,
269.
32. T. Fang, Macromolecules, 1990, 23, 2145.
33. L. Li, C.M. Chan and L.T. Weng, Macromolecules, 1997, 30, 3698.
34. A. Kaji and M. Murano, Polymer Journal (Japan), 1990, 22, 1065.
35. N. Luo, D.N. Wang and S.K. Ying, Journal of Polymer Science, Polymer
36. K.R. Lindfors, P. Sheng and P. Dreyfuss, Macromolecules, 1993, 26, 2919.
37. I. Tritto, L. Boggioni, M.C. Sacchi, P. Locatelli, D.R. Ferro and A. Provasoli,
Macromolecular Rapid Communications, 1999, 20, 279.
450
38. Y. Feng and J. N. Hay, Polymer, 1998, 39, 6589.
39. H. N. Cheng, Macromolecules, 1984, 17, 1950.
40. W.V. Smith, Journal of Polymer Science, Polymer Physics Edition, 1980, 18,
1573.
41. W.V. Smith, Journal of Polymer Science, Polymer Physics Edition, 1980, 18,
1587.
42. J.C. Randall, Macromolecules, 1978, 11, 33.
43. A.C. Kolbert and J.G. Didier, Journal of Applied Polymer Science, 1999, 71, 523.
44. E.W. Hansen, K. Redford and H. Oysaed, Polymer, 1996, 37, 19.
45. S. Bywater, Polymer Communications, 1983, 24, 203.
46. K.F. Elgert, G. Quack and B. Stutzel, Polymer, 1975, 16, 154.
47. G. van der Velden, C. Didden, T. Veermans and J. Beulen, Macromolecules, 1987,
20, 1252.
48. H. Sato, K. Takebayashi and Y. Tanaka, Macromolecules, 1987, 20, 2418.
49. G. van der Velden and L.J. Fetters, Macromolecules, 1990, 23, 2470.
50. L. Li, D.R. Hensley, H.J. Harwood, L.J. Fetters and P.L. Rinaldi,
51. R. Petiaud and Q. T. Pham, Journal of Polymer Science, Polymer Chemistry

Edition, 1985, 23, 1343.
52. D.J.T. Hill, J.H. O’Donnell, M.C.S. Perera and P.J. Pomery. Journal of Polymer
Science, Polymer Chemistry Edition, 1996, 34, 2439.
53. H.N. Cheng and G.H. Lee, Polymer Bulletin, 1988, 19, 89.
54. D.B. Patterson, D.H. Beebe and J. Lal, Polymer Science and Technology, 1983,
21, 383.
55. N.A. Mohammadi and G.L. Rempel, Macromolecules, 1987, 20, 2362.
56. A. Kondo, H. Ohtani, Y. Kosugi, S. Tsuge, Y. Kubo, N. Asada, H. Inaki and A.

Yoshioka, Macromolecules, 1988, 21, 2918.
451
57. S.N. Gan and Z.A. Hamid, Polymer, 1997, 38, 1953.
58. S.F. Thames and S. Gupta Journal of Applied Polymer Science, 1997, 63, 1077.
59. Maenz, H. Schütz and D. Stadermann, European Polymer Journal, 1993, 29, 855.
60. W. Heinen, C.H. Rosenmöller, C.B. Wenzel, H.J.M. de Groot, J. Lugtenburg and
M. van Duin, Macromolecules, 1996, 29, 1151.
61. W. Heinen, M. van Duin, C.H. Rosenmöller, C.B. Wenzel, H.J.M. de Groot and
J. Lugtenburg, Macromolecular Symposia, 1998, 129, 119.
62. J.H.M. van den Berg, J.W. Beulin, E.F.J. Duynstee and H.L. Nelissen, Rubber
63. M. van Duin and A. Souphanthong, Rubber Chemistry and Technology, 1995,
68, 717.
64. R. Vukov, Rubber Chemistry and Technology, 1984, 57, 275.
65. N. Nagata, T. Kobatake, H. Watanabe, A. Ueda and A. Yoshioka, Rubber

66. F. Tsutsumi, M. Sakakibara and N. Oshima, Rubber Chemistry and Technology,

1990, 63, 8.
67. Y. Camberlin, J.P. Pasucault and Q.T. Pham, Makromolekulare Chemie, 1979,
180, 397.
68. I. Descheres, O. Paisse, J.N. Colonna-Ceccaldi and Q.T. Pham, Makromolekulare

Chemie, 1987, 188, 583.
69. R. Santos Mauler, G.B. Galland, D. Samios and S. Tokumoto, European Polymer
Journal, 1995, 31, 51.
70. A. Xu, V.L. Dimonie, E.D. Sudol and M.S.J. El-Aasser, Journal of Polymer
Science Polymer Chemistry Edition, 1995, 33, 1353.
71. K. Takenaka, A. Hirao and S. Nakahama, Polymer International, 1995, 37, 291.
72. W.D. Vilar, S.M.C. Menezes and L. Akcelrud, Polymer Bulletin, 1994, 33, 557.
73. K.D. Skuratov, M.I. Lobach, A.N. Shibaeva, L.A. Churyaeva, T.V. Erokhina, L.V.
Osetrova and V.A. Kormer, Polymer, 1992, 33, 5197.
452
74. J.C. Randall and S.P. Rucker, Macromolecules, 1994, 27, 2120.
75. A.C. Kolbert, J.G. Didier and L. Xu, Macromolecules, 1996, 29, 8591.
76. A.C. Kolbert and J.G. Didier, Journal of Polymer Science, Polymer Physics, 1997,
35, 1955.
77. T. Bremner, D.J.T. Hill, J.H. O’Donnell, M.C.S. Perera and P.J. Pomery, Journal
of Polymer Science, Polymer Chemistry, 1996, 34, 971.
78. T. Yamaguchi, I. Kurimoto, K. Ohashi and T. Okita, Kautschuk und Gummi

79. I. Hattori, F. Tsutumi, M. Sakakibara and K. Makino, Journal of Elastmers and

Plastics, 1991, 23, 135.
80. I. Hattori, M. Sakakibara, K. Makino and Y. Hongu, Proceedings of the 139th

ACS Rubber Division Meeting, 1991, Toronto, Canada, Paper No. 93.
81. C.Y. Chu and R. Vukov, Macromolecules, 1985, 18, 1423.
82. I. Kuntz and K.D. Rose, Journal of Polymer Science, Part A, 1989, 27, 107.
83. D. M. Cheng, I.J. Gardner, H.C. Wang, C.B. Fredrick and A.H. Dekmezian,
84. J.L. White, T.D. Shaffer, C.J. Ruff and J.P. Cross, Macromolecules, 1995, 28,
3290.
85. E.F. McCord, W.H. Shaw, Jr. and R.A. Hutchinson, Macromolecules, 1997, 30,
246.
86. H.Y. Chen, Journal of Polymer Science, 1966, B4, 891.
87. F.W. Hemming in Natural Substances Formed Biologically from Mevalonic Acid,
Ed., T.W. Goodwin, Academic Press. London, UK, 1970, p.105.
88. Y. Tanaka in NMR and Macromolecules, ACS Symposium Series No. 247, Ed.,
J.C. Randall, American Chemical Society, Washington, DC, USA, 1984, p.233.
89. Y. Tanaka, H. Sato and A. Kageyu, Polymer, 1982, 23, 1087.
90. Y. Tanaka in Methods in Plant Biochemistry, Volume 7, Terpenoids, Eds., P.M.

Dey and J.B. Harborne, Academic Press, London, UK, 1991, p.519.
453
91. J.L. Cunneen, G.M.C. Higgins and W.F. Watson, Journal of Polymer Science,
1959, 40, 1.
92. Y. Tanaka and H. Sato, Polymer, 1982, 17, 113.
93. Y. Tanaka, K. Nunogaki, A. Kageyu, M. Mori and Y. Sato, Journal of Natural

Rubber Research 1988, 3, 177.
94. Y. Tanaka, Journal of Applied Polymer Science, Applied Polymer Symposia,

1989, 44, 1.
95. Y. Tanaka, M. Mori, A. Takei, P. Boochathum and Y. Sato, Journal of Natural

Rubber Research 1990, 5, 241.
96. J. Tangpakdee, Y. Tanaka, K. Shiba, S. Kawahara, K. Sakurai and Y. Suzuki,

Phytochemistry, 1997, 45, 75.
97. Y. Tanaka, M. Mori, K. Ute and K. Hatada, Rubber Chemistry and Technology,
1990, 63, 1.
98. Y. Tanaka, M. Mori and A. Takei, Journal of Applied Polymer Science, Applied
Polymer Symposium, 1992, 50, 43.
99. Y. Tanaka, H. Sato and A. Kageyu, Rubber Chemistry and Technology, 1983, 56,
299.
100. Y. Tanaka, S. Kawahara, A.H. Eng, K. Shiba and N. Ohya, Phytochemistry,

1995, 39, 779.
101. A.H. Eng, S. Kawahara and Y. Tanaka, Rubber Chemistry and Technology, 1994,
67, 159.
102. Y. Tanaka and H. Hirasawa, Chemistry and Physics of Lipids, 1989, 51,183.
103. Y. Tanaka, A.H. Eng, N. Ohya, N. Nishiyama, J. Tangpakdee, S. Kawahara and

R. Wititsuwannakul, Phytochemistry, 1996, 41, 1501.
104. Y. Tanaka and A.H. Eng, Trends in Polymer Science, 1993, 3, 493.
105. J. Tangpakdee and Y. Tanaka, Journal of Rubber Research, 1988, 1, 77.
106. J.T. Sakdapipanich, Y. Tanaka, J.L. Jacob and J. d’Auzac, Rubber Chemistry and
Technology, 1999, 72, 299.
107. J. Tangpakdee and Y. Tanaka, Journal of Natural Rubber Research, 1997, 12, 112.
454
108. J. Tangpakdee and Y. Tanaka, Journal of Rubber Research, 1998, 1, 14.
109. Y. Tanaka, S. Kawahara and J. Tangpakdee, Kautschuck und Gummi

110. J. Tangpakdee and Y. Tanaka, Phytochemistry, 1998, 48, 447.
111. J. Tangpakdee, Y. Tanaka, K. Ogura, T. Koyama, R. Wititsuwannakul and N.

Chareonthiphakorn, Phytochemistry, 1997, 45, 269.
112. J. Tangpakdee, Y. Tanaka, N. Ohya, T. Koyama, R. Wititsuwannakul and N.

Chareonthiphakorn, Phytochemistry, 1997, 45, 275.
113. A.H. Eng, S. Ejiri, S. Kawahara and Y. Tanaka, Journal of Applied Polymer
Science, Applied Polymer Symposium, 1994, 53, 5.
114. S.K. Wolk and E. Eisenhart, Macromolecules, 1993, 26, 1086.
115. A.L. Gatzke and D.P. Green, Macromolecules, 1994, 27, 2249.
116. J.M. Boutillier, J.C. Favier, P. Hemery and P. Sigwalt, Polymer, 1996, 37, 5197.
117. A.E. Madani, J. Belleney, J-C. Favier, P. Hemery and P. Sigwalt, Polymer
International, 1993, 31, 169.
118. E.A. Williams in The Chemistry of Silicon Compounds, Eds., S. Patai and
Rappoport, John Wiley Sons, New York, USA, p.511, 1989.
119. L.R. Herbert and A.D.H. Clague, Macromolecules, 1989, 22, 3267.
120. E.A. Williams, J.H. Wengrovius, V.M. Van Valkenburg and J.F. Smith,
121. K. Troev, V.L. Atanassov and R. Tzevi, Journal of Applied Polymer Science,
2000, 76, 886.
122. W.L. Hergenrother, J.M. Doshak, D.R. Brumbaugh, T.W. Bethea and J. Oziomek,
Journal of Polymer Science, Polymer Chemistry Edition, 1995, 33, 143.
123. I. Fontao, L. Gonzalez, M.L. Jimeno and A. Marcos, Kautschuck Gummi und
234.
455
125. C.D. Hull, C.D.O. Jackson and M.J.R. Loadman, Journal of Natural Rubber
Research, 1994, 9, 23.
1992, 65, 744.
127. P.S. Brown and A.J. Tinker, Journal of Natural Rubber Research, 1996, 11, 227.
129. C.J. McDonald, Journal of Dispersion Science and Technology, 1984, 5, 365.
130. P.J. Tarcha, R.M. Fitch, J.J. Dumis and L.W. Jelinski, Journal of Polymer Science,
131. C.J. McDonald, P. Smith, J.A. Roper, D.I. Lee and J.G. Galloway, Colloid and
132. B. Bonardi, P. Christou, M.F. Llauro-Darricades, J. Guillot, A. Guyot and C.

Pichot, New Polymer Materials, 1991, 2, 4, 295.
133. S. Kawahara, S. Bushimata, T. Sugiyama, C. Hashimoto and Y. Tanaka, Rubber

134. S. Kawahara, K. Washino, T. Morita, Y. Isono and Y. Tanaka, Rubber Chemistry

456
129
Xe NMR of Elastomers in Blends and Composites
12
129
Xe NMR of Elastomers in Blends and
Composites
Wiebren S. Veeman
12.1 Introduction to 129Xe NMR Spectroscopy of Materials
12.1.1 129Xe NMR Spectroscopy
Most non-crystalline materials, and especially elastomers, readily absorb xenon (Xe)
atoms. The Xe atoms can then be used to probe the local surrounding of the Xe atom in
the material via nuclear magnetic resonance (NMR) spectroscopy. The direct information
from standard NMR experiments usually comes from the chemical shift of the nucleus
involved. The chemical shift of a nucleus is determined by the local surrounding of the
atom (radius < approximately 1 nm). The additional information Xe NMR experiments
can provide is because of the fact that, especially in elastomers, the Xe atom is very
mobile. Depending on the type of NMR experiment, the characteristic time scale of a
NMR experiment may be between 1 millisecond and 1 second and during this time a Xe
atom may diffuse over several μm. An NMR investigation of Xe atoms in a material
therefore provides information about the structure of the material averaged over a length
scale of this magnitude. Compared to other branches of NMR, Xe NMR is unique in
this sense.
Several reviews on xenon NMR spectroscopy have appeared [1, 2], therefore in this
introduction only the most important aspects will be discussed. The xenon atom has two
isotopes which are suited for NMR studies, the 129Xe and 131Xe isotopes. For most studies
129
Xe is more convenient than 131Xe, while the former nucleus with spin I=1/2 does not
have an electrical quadrupole moment. In some cases, however, the quadrupolar
interaction can provide additional (spectral and relaxation) information. Here we will
only consider the 129Xe isotope.
For the probing of void spaces in and surfaces of materials Xe gas is ideal while Xe is an
inert atom with a large (0.44 nm) and highly polarisable electron cloud. This is reflected
by a very wide 129Xe chemical shift range (over 7500 ppm), although the larger part of
this range results from Xe covalently bonded in compounds. The effective chemical shift
457
range for Xe atoms sorbed in materials is about 300 ppm. This is of course still very
large in comparison to, for instance, the 1H chemical shift range.
In the experiments described here, the material investigated is contained in a glass tube
in a Xe atmosphere at ca. 1,000,000 Pa. As a result each Xe spectrum usually contains at
least two lines, one from Xe in the gas surrounding the material but inside the NMR coil,
and one for Xe, either absorbed in the material or adsorbed at pore and outer surfaces of
the material (Figure 12.1). Since the Xe NMR resonance of the free gas is used as the
spectral reference, the NMR of Xe gas will be discussed shortly.
12.1.2 NMR of 129Xe in the Gas Phase
Because of the collisions between the Xe atoms in the gas phase, the chemical shift of
129
Xe in the gas phase is temperature and pressure dependent. Quantitatively, a relationship
between the 129Xe chemical shift in ppm and the density of the gas has been found to
exist [3]:
δ ( T , p ) = δ 0 + δ1( T )ρ + δ 2 ( T )ρ 2 + ........... (12.1)
where δ0 is a reference shift set to 0 ppm; δ1, δ2,..... are coefficients and ρ is the density.
When the density is expressed in the unit Amagat (1 Amagat is the density of the gas at
1 atmosphere at 0 °C and corresponds to approximately 2.7 x 1019 atoms/cm3, assuming
an ideal gas) the coefficients at 25 °C are δ1 = -0.548 ppm/amagat and δ2 = -0.169 x 10-3
ppm/amagat2. For the pressures used in the experiments described here higher terms in
Equation 12.1 need not be considered.
Figure 12.1 A typical 129Xe NMR spectrum of a polymer at a temperature above the
glass transition temperature (Tg) (here ethylene-propylene diene terpolymer (EPDM))
in a ca. 1,000,000 Pa Xe atmosphere. The signal of the free gas is used as an internal
chemical shift resonance
458
129
For our experiments the sample is located in a sealed glass tube in a Xe atmosphere and
the signal of the free Xe gas is taken as the 0 ppm reference. The initial pressure is
carefully adjusted to 1,000,000 Pa, but slight deviations (50,000 Pa) may occur during
the sealing process. Using Equation 12.1 and assuming that the Xe gas behaves like an
ideal gas, a pressure variation of 50,000 Pa leads to chemical shift variations of ±0.2 ppm.
In view of the width of the lines, this variation of the chemical shift reference with
pressure is neglected here.
Materials at temperatures above the glass transition temperature (Tg) absorb Xe readily,
therefore the fact that after sealing of the sample tube the density of the Xe gas could
decrease with time when more Xe is absorbed by the material is neglected. For samples
which absorb Xe slowly, like some polymers below their Tg, this effect may not be
negligible. Also for temperature dependent measurements the variation of the coefficients
in Equation 12.1 with temperature should be taken into account [4, 5, 6], but in view of
the widths of the resonance lines investigated here as a function of temperature, the
temperature dependence of the reference has also been neglected.
12.1.3 129Xe NMR of Polymers
Before describing 129Xe NMR experiments on polymers and polymer composites, it is

worthwhile to show with a few selected examples, some of the general aspects of 129Xe
NMR of polymers. The first question that needs to be addressed is where sorbed Xe
atoms are located in a polymer material, or rather where Xe atoms are not located. With
a diameter of 0.44 nm the Xe atom is clearly larger than the interchain distance for most
crystalline polymers. In general it means, that if Xe atoms are found in crystalline domains,
they must reside in areas which contain defects. In the examples studied, no evidence of
Xe in crystalline domains of polymers was ever found.
A clear example for the absence of Xe in crystalline polymers is shown in Figure 12.2 for
highly stretched polyethylene (PE) fibres.
From X-ray experiments it is known that the crystallinity of the PE fibres increases with
stretching. In Figure 12.2 the NMR signal of 129Xe absorbed in the PE disappears with
increasing stretching, which shows that the Xe is practically absorbed only in the
amorphous domains, and possibly in the interface between crystalline and amorphous
domains, of PE.
The absence of Xe in crystalline polymers is also due to the fact that the polymer chains
in crystalline polymers are quite rigid. The energy needed to deform the chains enough
so that a Xe atom can be incorporated between the chains, is too high.
459
Figure 12.2 129Xe NMR spectra of PE fibres as a function of stretching. The Young’s
modulus increased from 38 GPa for PE1 to 131 GPa for PE5
The mobility of the chains plays an important role for the line width of the Xe resonance
in amorphous polymers. Figure 12.3 shows the line width of the 129Xe resonance of
129
Xe absorbed in polymethylmethacrylate (PMMA) as a function of temperature [7].
The line width drastically reduces when the Tg is approached. The mobility of the polymer
chains in amorphous domains affects the 129Xe line width while it makes it possible for
the Xe atom to move fast from one location to another, thereby averaging out local
differences in Xe chemical shift and dipolar interactions with proton spins [8]. The chain
mobility also influences the rate of Xe absorption in an amorphous polymer, because it
was noted that for PMMA at room temperature it takes several weeks before enough Xe
is absorbed in the material so that a NMR signal can be detected [7].
From Figure 12.3, it can be concluded that room temperature 129Xe NMR spectroscopy
is especially well suited for the investigation of amorphous domains in polymer materials
460
129
Figure 12.3 The line width of the 129Xe NMR resonance of PMMA as a function of
temperature. The sharp decrease of the line width for temperatures below Tg shows
that the mobility of the Xe atoms already increases at temperature for below the Tg
Reproduced with permission from K. Sperling-Ischinsky and W.S. Veeman, Journal of the
Brazilian Chemical Society, 1999, 10, 293, Figure 11. Copyright 1999, The Brazilian
Chemical Society
for which the Tg is clearly below room temperature. In such materials the Xe atoms are
very mobile, therefore the Xe diffusion coefficient can also be considered as a characteristic
parameter of the material. The determination of the Xe self-diffusion coefficients allows
the calculation of the root mean square displacement of diffusing xenon atoms with time
in the material.
461
12.1.4 129Xe Pulsed Field Gradient Echo (PFGE) Spectroscopy
The PFGE experiment is a NMR imaging experiment [9, 10], in which pulsed magnetic
field gradients are used to make the Larmor precession frequency of the 129Xe spins
during the time of the gradient pulses dependent on the spatial coordinates parallel to
the direction of the field gradient, The scheme of the radiofrequency (rf) and magnetic
field pulses is shown in Figure 12.4. In addition to the rf pulses (representing a stimulated
spin echo experiment) two pulsed gradients, identical in amplitude g, width δ and direction,
and separated by a time Δ >> δ are applied.
The first gradient pulse causes a phase jump of the magnetisation transverse to the static
external magnetic field. This phase jump linearly depends on the coordinate of the spin
position in the direction, e.g., x, of the field gradient. For spins which do not diffuse
during the time change (Δ) between the two gradient pulses, or which at the time of the
second gradient pulse have returned to the same x-coordinate value they had at the time
of the first pulse, this phase jump is exactly compensated by the second (identical) gradient
pulse thanks to the spin echo sequence. They do not suffer an overall phase loss and
contribute fully to the total intensity of the spin echo signal.
For spins which did diffuse from their original position (or more accurately who changed
their x-coordinate value) during the time change, the phase jump their magnetisation
vector acquired during the first pulse is not compensated by the phase jump during the
second pulse and their contribution to the echo intensity is decreased.
Figure 12.4 The pulse scheme of the three-pulse echo sequence to determine Xe
diffusion coefficients in polymers and other porous systems. The shaded areas are
magnetic field gradient pulses with amplitude g and length δ. The time between the
two gradient pulses Δ determines the time during which the diffusion path length is
determined (the diffusion time)
Reproduced with permission from F. Junker and W.S. Veeman, Macromolecules, 1998,
20, 7010, Figure 1. Copyright 1998, American Chemical Society
462
129
In the limit that δ<<Δ the echo intensity ratio E(g)/E(g=0) is given by [11]:
E( g )
= exp[ −γ 2δ 2 g2 DΔ ] (12.2)
E( 0 )
where E(g) = echo intensity with gradient amplitude g
γ = gyromagnetic ratio of the nucleus involved
δ = width of the gradient pulse (s)
g = gradient amplitude (T/m)
D = self-diffusion coefficient (m2/s)
Δ = time between the gradient pulses (s)
At fixed values of δ and Δ the slope of ln[E(g)/E(0)] versus g2 equals ΔD and the diffusion
coefficient D can be determined (Figure 12.5).
To determine the absolute values of the gradient strength g diffusion experiments on

pure deuterium oxide or water with known diffusion coefficients can be used.
For normal (free) diffusion, D should be independent of the diffusion time, Δ. For cases
where the free diffusion of the Xe atoms in a material is restricted by, for example,
crystalline barriers or pore walls, the diffusion coefficient can be diffusion time dependent.
This will be discussed again in Section 13.5.
12.2 Experimental
The ethylene-propylene (EP), present as a component with 20 wt% in a copolymer or

blend with isotactic polypropylene (iPP), are random copolymers with varying
composition.
Two types of EPDM were investigated. The random terpolymer EPDM (46 wt% ethylene,
49 wt% propylene and 5 wt% norbornene), present as the elastomer component in the
carbon black composites investigated, was not vulcanised and has been mixed with three
types of carbon black N110, N330 and N550 with different particle size (Table 12.1) in
a laboratory rolling mill at 100 °C. The content of carbon black for all three samples
was 85.5 phr (for other parameters of the carbon black samples, see Table 12.1). For the
preparation of the bound rubber samples EPDM/N110 and EPDM/N330 the material
has been boiled in o-xylene for 8 hours and then dried at 120 °C and 500 Pa.
463
Figure 12.5 A graphic representation of Equation 12.2. Usually the diffusion

coefficient is determined from the slope of the curve ln[E(g)/E(O)] versus g2 for a
constant value of the diffusion time Δ
The EPDM, investigated by PFGE NMR, is a commercial product BUNA AP 437; it

contains 70% ethylene and has as diene, 5-ethylidene–2-norbornene. The crystallinity of
this EPDM is given as 30%.
Finally, a thermoplastic elastomer polybutyleneterephthalate/polytetramethyleneoxide

(PBT/PTMO) block copolymer was also investigated, because it is a partly crystalline
elastomer with large spherulites. Details of the composition can be found in [12].
The samples are put into a NMR tube and evacuated for 2 hours. After filling the tube
with Xe gas, the tube was sealed off. In all samples the Xe pressure at room temperature
was 1,000,000 ± 50,000 Pa.
464
129
Table 12.1 Properties of carbon black and composition of the

carbon black filled EPDM samples
Carbon Particle Specific Weight Carbon Bound Carbon black in
black size in nm area in % black rubber bound rubber
m2/g volume weight fraction volume
% % %
N110 18 143 85.5 0.284 40.6 0.495
N330 30 83 85.5 0.284 29.6 0.573
N550 56 30 85.5 0.284 - -
12.3 129Xe NMR of Ethylene/Propylene Copolymers in Blends and

Block-copolymers with Polypropylene
Many elastomers are used as a component in blends, copolymers or composites. Next,

129
Xe experiments on EP copolymers with varying composition, which are present as the
elastomer component in a blend or block copolymer with iPP are discussed [13, 14].
Figure 12.6 shows the spectra of a statistical EP copolymer with 50/50 composition
(weight%), as a single component material (Figure 12.6a), in a block copolymer with
iPP (80 wt% iPP, 20 %wt EP) (Figure 12.6b) and in a blend with iPP (80 wt% iPP,
20 wt% EP) (Figure 12.6c). Figure 12.7 shows the peaks of PE, iPP and one of the iPP/EP
copolymers.
From the spectra in Figures 12.6 and 12.7 immediately three conclusions can be drawn:
(a) for the sorbed Xe atoms the EP copolymer is a homogeneous material, while only one
Xe resonance is found;
(b) the Xe chemical shift depends on the composition of the EP copolymer and is in
between that of iPP and PE;
(c) both the iPP/EP copolymer and blend are heterogeneous for the Xe experiment, i.e.,
during the characteristic time of the NMR experiment (roughly equal to the Xe spin-
spin relaxation time T2, which for these materials is between 1-10 milliseconds) some
Xe atoms have the characteristic chemical shift of Xe sorbed in iPP and others the
shift of Xe in an EP surround.
465
Figure 12.6 The 129Xe spectra of (a) a 50/50 ethylene/propylene (EP) copolymer, (b)
the copolymer consisting of 80% iPP and 20% EP and (c) the blend consisting of 80%
iPP and 20% EP
Electron microscopy has shown that in the samples studied here, iPP forms a matrix in
which small EP domains are embedded, even for the block-copolymers. The Xe NMR
results show that the Xe mobility is not so large that during the NMR characteristic
time, Xe atoms move many times between the EP domains and the iPP matrix.
The NMR characteristic time depends, however, on the type of experiment. Figure 12.8
shows the result of a two-dimensional exchange experiment [13], for which the
characteristic time can be varied from a value of the order of T2 to a value of the order of
the spin-lattice relaxation time T1, for Xe usually several seconds (depending for example
on the amount of oxygen in the Xe gas used). The presence of cross-peaks in the 2-
dimensional spectrum indicates that Xe atoms move during the mixing time (here
5 seconds) from the rubbery EP phase into the iPP matrix and vice versa. The cross
peaks first appear already after a mixing time of 50 milliseconds.
466
129
Figure 12.7 The 129Xe NMR spectra of (a) high density PE, (b) iPP and (c) of a blend
consisting of 80% iPP and 20% EP copolymer with the composition 33% propylene
and 67% ethylene. The line positions show that the Xe absorbed in the iPP matrix is
identical to Xe in iPP (Figure 12.7b), but that the Xe in the EP domains have a
chemical shift in between that of Xe in PE (Figure 12.7a) and of Xe in iPP
(Figure 12.7b)
Reproduced from M. Mansfeld, A. Flohr and W.S. Veeman, Applied Magnetic
Resonance, 1995, 8, 573, Figure 2. Copyright 1995, Springer Verlag
467
Figure 12.8 Two-dimensional 129Xe NMR exchange spectrum of the blend of

Figure 12.7 with an exchange time of 5 seconds. The clear cross-peaks between the
peaks due to absorbed Xe shows that within the exchange time the Xe atoms change
location very often. During 5 seconds no exchange with the gas surrounding the
material is detected
With the Einstein relationship <r2>=6DΔ, where <r2> is the average squared distance
particles with a self diffusion coefficient D cover in a time Δ, the maximum average size
of the domains can be estimated as ≈5 μm, when we use Δ=50 milliseconds as the smallest
mixing time for which cross-peaks appear, and 10-10 m2/s for D, the value found for
468
129
elastomers like EP (vide infra). Since two separate lines are detected in the normal NMR
experiment of Figures 12.6b, 12.6c and 12.7 with a Δ of about 1 millisecond the minimum
average EP domain size is ≈1 μm.
Using the results of 12 iPP/EP blends, 4 iPP/EP block copolymers and 1 pure EP copolymer
[13, 14], the Xe chemical shift of Xe absorbed in EP can be correlated to the chemical
composition of EP, i.e., the ethylene fraction. Figure 12.9 shows an approximately linear
relationship between the Xe chemical shift and the weight fraction ethylene in the EP
copolymer. The values for the pure components PE (ethylene content = 1) and iPP (ethylene
content = 0) are included. Of course, this also implies that a linear relationship exists
between the Xe chemical shift and the density of the EP copolymer.
In the model of Miller and co-workers [15], the Xe chemical shift is proportional to the
difference in interaction energy between units of neighboring polymer chains with and
without an enclosed xenon atom. The higher chemical shift for Xe in iPP or in EP with
low PE content then shows that to incorporate a xenon atom between iPP chains costs
more energy than between PE units. Approximately the same energy difference is needed
Figure 12.9 The 129Xe chemical shift as a function of the composition of the EP
copolymer. The two values for pure iPP result from two different iPP samples
469
to transform a rigid amorphous polymer, i.e., a polymer at a temperature below the Tg,
into a polymer in which the polymer chains are more or less free to move (the polymer at
a temperature above Tg). This qualitative picture is confirmed by the higher Tg of
amorphous iPP (approximately –0 °C) relative to that of amorphous PE (T g=
approximately –43 °C). It is worthwhile to investigate whether 129Xe NMR can be used
to accurately determine the Tg for amorphous polymers.
The difference in Tg between iPP and PE also is an important factor in the explanation of
the temperature dependence of the Xe spectra. Figure 12.10 shows the Xe spectra between
20 °C and –40 °C of a iPP/EP blend. By spectral simulation of the doublet by two lines
the temperature dependence of the line width (Figure 12.11) and of the chemical shift
(Figure 12.12) for each line can be determined.
The temperature dependence of the line widths in Figure 12.11 can be explained by the
difference in Tg of the iPP matrix and of the elastomer EP: the lowest temperature in
Figure 12.11 is above the Tg of EP, but below the Tg of iPP. Neglecting the relative small
variations in line width of Xe in EP and by referring to Figure 12.3 for PMMA, one can
state that for EP for all temperatures the line width has reached the low plateau value for
polymers at temperatures above Tg , while for the iPP matrix the line width shows exactly
the same behaviour as for PMMA in Figure 12.3.
Surprisingly at room temperature the iPP line is narrower than the EP line.
The temperature dependence of the chemical shift for both lines, as shown in Figure 12.12,
is quite similar: until at approximately –0 °C both chemical shifts are constant (neglecting
the ‘wiggle’ in the iPP chemical shift curve) and above –0 °C the chemical shifts decrease
linearly with temperature with practically the same slope. At first sight it seems that the
Xe chemical shifts for both components are mainly determined by the free volume, which
increases linearly with temperature above Tg [16], while the line widths of the Xe
resonances are mainly determined by the mobility of the polymer chains. This would
imply that the increase of the free volume with temperature of (amorphous) iPP and EP
must be approximately the same in the temperature range investigated here.
The last experiment discussed here, shows that the mobility of Xe in the iPP matrix is
much lower than in the elastomer domains.
Figure 12.13 shows the Xe spectra of a iPP/EP blend obtained with single pulse excitation
and with 1H ⇒ 129Xe cross-polarisation at –33 °C [13]. In the cross-polarisation spectrum
only the Xe line from Xe in iPP is observed, because the dipolar interaction between the
polymer 1H spins and the 129Xe spins due to the Xe mobility is too weak in EP. Also the
cross-polarisation signal of iPP disappears at temperatures higher than the Tg of iPP.
470
129
Figure 12.10 The temperature dependent 129Xe NMR spectra of the blend of Figure 12.7
471
Figure 12.11 The temperature dependent spectra of Figure 12.10 are fitted with two
lines with variable width and chemical shift. Here the line width of both lines are
shown as a function of temperature
472
129
Figure 12.12 The temperature dependence of the chemical shifts which result from
fitting of the spectra in Figure 12.10 by two lines
473
Figure 12.13 129Xe spectra of the blend of Figure 12.7 obtained with single pulse
excitation (top spectrum) and with 1H-129Xe cross-polarisation at –30 °C. The cross-
polarisation time was 3 milliseconds and recycle time of 3 seconds. Cross-polarisation
is effective only for Xe in the iPP matrix, the Xe atoms in the EP domains are too
mobile. At higher temperatures also the CP resonance of Xe in iPP disappears
12.4 129Xe NMR of an Ethylene/Propylene/Diene Terpolymer in Carbon

Black Composites
Next 129Xe experiments on an EPDM terpolymer, which is present as the elastomer

component in a composite material with carbon black will be discussed. The question
investigated for these materials is whether the existence of any polymer-filler interaction
can be detected by 129Xe NMR. This interaction influences the mobility of the elastomer
chains in a relatively large shell around the filler particles. This fraction is called the
bound rubber fraction. It is generally believed that the bound rubber fraction influences
the mechanical and frictional properties of the filled elastomer [17, 18].
474
129
Elastomer-filler interactions were the subject of many intensive investigations. Kaufmann

and co-workers [17] investigated carbon-black-filled EPDM by nuclear spin relaxation
time measurements and found three distinct regions in the material. These regions are
characterised by different mobility of the elastomer chains: a mobile region in which the
polymer chains have no interaction with the filler particles, loosely bound rubber in an
outer shell around the carbon black particles and an inner shell of tightly bound elastomer
chain with limited mobility.
In proton T2 experiments conducted by Litvinov [18] a strong dynamic heterogeneity in

the fraction of EPDM bound to carbon black could also be detected: an immobilised
EPDM layer covering the carbon black surfaces and a mobile EPDM part outside of this
interfacial layer. In dynamic-mechanical experiments by Haidar [19], it was found that
the polymer layer on the filler surface had glass-like mechanical properties.
Figure 12.14 shows the room temperature 129Xe NMR spectra of the unfilled (and
uncured) and carbon-black-filled EPDM samples.
Figure 12.14 129Xe NMR spectra at room temperature of EPDM (a) and of carbon
black filled EPDM (85.5 phr). Three different carbon blacks were used: (b) carbon
black N110, (c) N330 and (d) N550
475
In spite of the high concentration of carbon black particles the only observable effect of
carbon black is a slight broadening of the resonance of Xe absorbed in EPDM at 205 ppm.
The temperature dependence of this shift and the line width of the 205 ppm line are
slightly dependent on the presence and the type of carbon black. The temperature
dependence is shown in Figure 12.15: at lower temperatures (–33 °C), the sample with
the largest carbon black particles (EPDM/N550) has the smallest chemical shift.
Bound rubber samples were prepared by boiling the original material in o-xylene. In this
way the fraction of EPDM which is not bonded to the carbon black particles is removed.
The Xe spectra of the bound rubber fraction, show a large contrast to those of the
corresponding original samples, compare Figure 12.14 and Figure 12.16. The relatively
Figure 12.15 The temperature dependent chemical shift of absorbed Xe in the four
materials of Figure 12.14
Chemical Society
476
129
Figure 12.16 The temperature dependent 129Xe spectra of the bound rubber fraction of
the EPDM/N110 sample. A weak broad resonance is observed for Xe in this bound
fraction
Chemical Society
narrow line at 205 ppm has been replaced by a broad resonance between 100 and
200 ppm, depending on the temperature. The line width of this line decreases somewhat
at lower temperatures.
Now it is of interest to compare these spectra to the 129Xe spectra of Xe adsorbed on the
carbon black surfaces of pure carbon black samples. Since the effect of three carbon
blacks as filler in EPDM was studied here, it limits the study here to the three carbon
blacks N110, N330 and N550. For the investigation of other carbon blacks with 129Xe
NMR, see [20].
The room temperature 129Xe spectra of the three carbon blacks are shown in Figure 12.17
and, surprisingly, the line width of these Xe resonances is much smaller than that of Xe in the
composites. Although it is known that magnetic impurities and susceptibility effects excessively
broaden the 13C line width of carbon blacks, these effects clearly are not as serious for Xe
adsorbed at the outer and inner surfaces of the carbon black aggregates [20].
This must be due to a high mobility of Xe on the carbon black surface. For the three
carbon blacks shown here the 129Xe chemical shift is proportional to the average particle
diameter [14].
477
Figure 12.17 Room temperature 129Xe spectra of Xe adsorbed at the surface of the
three carbon blacks investigated here
Reproduced with permission from J Sperling and W.S. Veeman, Kautschuk und Gummi
Kunststoffe, 1997, 50, 804, Figure 1. Copyright 1997, Huthig Verlag GmbH
Figure 12.18 shows the temperature dependence of the 129Xe chemical shifts of Xe in
EPDM/carbon black N110, in the bound rubber fraction of EPDM/N110 and in the
carbon black N110 itself. It shows that the temperature dependence of the chemical shift
of the broad resonance in the bound rubber sample shows a similar temperature
dependence as the chemical shift of the pure EPDM and that of Xe adsorbed on the
carbon black.
The chemical shift of Xe in the bound rubber fraction is always between that of the pure
carbon black and of pure EPDM. Figure 12.19 shows that the Xe chemical shift in the
bound rubber composites depends on the amount and type of the carbon black.
These results prove that Xe is in contact with the surface of the carbon black particles
and that consequently the carbon black particles are not completely covered by the EPDM
chains, there are still adsorption sites for the Xe. The fact that the Xe chemical shift of
478
129
Figure 12.18 The temperature dependent chemical shift of 129Xe in EPDM (solid
squares), in the bound rubber fraction of EPDM/N110 and in the carbon black N110
Chemical Society
Xe in the bound rubber is between the value for pure carbon black and for pure EPDM
for the temperature range investigated here, suggests that the chemical shift results from
an exchange process, in which the Xe atoms exchange quickly between adsorption sites
at the carbon black surface and absorption between the EPDM layers. The large line
width also suggests that the EPDM chains in the bound rubber fraction are strongly
immobilised by the carbon black particles, as expected.
479
Figure 12.19 The chemical shifts of 129Xe in carbon black filled EPDM as a function of
the type of carbon black and the volume fraction of the carbon black in the material
This exchange model is supported by experiments on samples with more carbon black
(Figure 12.19). In samples with more carbon black the Xe chemical shift shifts in the
direction of pure carbon black, just as expected for simple exchange effects. More work
is needed to clarify, for example, the decrease of the line width with decreasing temperature.
The results show, however, that 129Xe NMR has the potential to characterise the interface
between the polymer and filler particles in carbon black filled materials, for which it is
known that they are difficult to characterise by other spectroscopic techniques.
12.5 The Self-diffusion Coefficient of Xe in Elastomers
12.5.1 General
As discussed in Section 12.4 the PFGE experiment allows the measurement of the Xe
self-diffusion coefficient in arbitrary materials.
As schematically shown in Figure 12.5 the self-diffusion coefficient D can be determined

from the PFGE experiment for different values of Δ. For normal diffusion, for which
Fick’s law holds*, D is independent of Δ and the average square distance the diffusing
particles cover during the diffusion time Δ is directly proportional to Δ (Einstein
relationship):
480
129
<r2>=6DΔ
For many materials, for example semi-crystalline polymers, it is expected that the Xe
atoms are absorbed and diffuse into the amorphous domains and that the crystalline
domains act as diffusion barriers. In general this should make the effective diffusion
coefficient dependent on the diffusion time Δ [10]. For (very) short values of Δ only a
small fraction of the diffusing particles experience the restrictions by the diffusion barriers
and the diffusion coefficient is approximately D0, the diffusion coefficient in the absence
of diffusion barriers. With increasing Δ the diffusion of an increasing fraction of the
particles is hindered by the barriers and <r2> is smaller than results from the Einstein
relationship. For such cases, by setting <r2>eff = 6Deff Δ, a Δ-dependent effective diffusion
coefficient Deff can be defined. The time dependence of Deff provides information on the
length scale of the objects that restrict the diffusion. For diffusion in pores the short-time
behaviour of Deff depends on the surface/volume ratio of the pores and the long-time
diffusion is determined by the tortuosity factor [21, 22] (see Figure 12.20). For diffusion
in enclosed pores the effective diffusion coefficient approaches zero when Δ→∞.
Figure 12.20 A schematic representation of the effective diffusion coefficients as a

function of the diffusion time in the case of hindered diffusion. For short diffusion
times Deff equals the diffusion coefficient of the particle in the material in the absence
of diffusion barriers. For longer diffusion times the diffusion is slowed down by the
barriers. In the long time limit the diffusion coefficients approaches Ttortuosity
481
It is important to know for which Δ the transition from the time dependent Deff to the
constant Dtort occurs. Using the relationship <r2> = 6D0Δ with D0≈10-11 m2/s, which is a
typical value for polymers, and Δ between 10-2 and 1 second, the range accessible for
most PFGE experiments, one finds (<r2>)1/2 ≈ 0.8-8 μm. Consequently, the PFGE
experiment is able to detect in a material diffusion barriers if the space between these
barriers has dimensions in this range.
12.5.2 Xe Diffusion in a iPP/EPDM Blend
The iPP/EPDM system investigated here is very similar to the iPP/EP blends described
before. A scanning electron microscope (SEM) image is shown in Figure 12.21 and makes
clear that the blend consists of a iPP matrix and EPDM domains (light spots). The 129Xe
NMR spectra of iPP, EPDM and the iPP/EPDM blend are shown in Figure 12.22 [23].
Because the spectrum of the blend shows two lines, for absorbed Xe in the iPP matrix
and for Xe in the EPDM domains, this enables separate determination of the Xe diffusion
coefficient for Xe in the matrix and for Xe in the EPDM domains. These can then be
compared to the diffusion coefficients for Xe in the pure materials. Figure 12.23 and
Table 12.2 show the results [23].
Figure 12.21 An electron microscope image of the iPP/EP blend investigated by Xe

diffusion NMR
Reproduced with permission from F. Junker and W.S. Veeman, Macromolecules, 1998,
20, 7010, Figure 2. Copyright 1998, American Chemical Society
482
129
Table 12.2 Xe diffusion coefficients in iPP, EPDM and the iPP/EPDM

blend
Xe diffusion coefficient in m2/s Xe in iPP Xe in EPDM
iPP 3.8 x 10-12
EPDM 78 x 10-12
iPP/EPDM blend 8.6 x 10-12 12 x 10-12
Figure 12.22 The room temperature 129Xe NMR spectrum of the iPP/EP blend
investigated by Xe diffusion NMR
*In general the diffusion of a solute like Xe is not driven by the concentration gradients
but by the chemical potential. In polymer blends with incompatible components, as we
investigate here, in the equilibrium situation it may occur that different domains have
different concentrations of Xe. Then the usual Fick laws do not hold.
In spite of the not so great signal-to-noise ratio of the results in Figure 12.23 it can
clearly be seen, that for the blend, the Xe diffusion coefficients for Xe in the iPP matrix
and in the EPDM domains are quite similar and in between the values for the pure
components.
483
Figure 12.23 The echo intensity E(g)/ε(0)] as a function of the square of the gradient
amplitude for the copolymer EP (stars), for iPP (squares ), for the Xe line of Xe in the
iPP matrix of the iPP/EP blend (circles) and for Xe in the EP domains of the iPP/EP
blend (triangles)
The Xe atoms during the diffusion time Δ apparently exchange many times between the
matrix and the EPDM domains and the difference between the two diffusion coefficients
averages out. In spite of this exchange between matrix and EPDM domains the absorbed
Xe shows two chemical shifts. During the characteristic time for the NMR spectrum,
here approximately 1 millisecond, the exchange is not fast enough to average the two
resonance frequencies, but it does average the diffusion coefficient during the much longer
diffusion time Δ = 1.2 seconds.
In Section 12.3 it was described how, with the help of a two-dimensional exchange
spectrum, domain sizes in the iPP/EP blend are estimated. There an estimated value for
the Xe diffusion coefficient was used. Now with experimental data from Table 12.2 the
average EPDM domain size for the iPP/EPDM blend can be calculated. It was assumed
that the structure of the EPDM in the blend is the same as in the pure material, (i.e., the
Xe diffusion coefficient in the EPDM domains in the blend is equal to the measured D
for pure EPDM and likewise for iPP), then <r2> for Xe in the EPDM domain during Δ =
1.2 seconds is 20 μm. In the same time <r2> for Xe in the iPP matrix is approximately
5 μm. These distances for a diffusion time of 1.3 milliseconds, the inverse of the frequency
difference 770 Hz of the two lines in the Xe NMR spectrum of the blend, are 0.6 and
0.2 μm, respectively. The average size of the EPDM domains in the iPP/EPDM blend is
484
129
therefore clearly larger than 0.6 μm and clearly smaller than 20 μm. By shortening the
diffusion time Δ, the upper limit of the average domain size can be further decreased,
thereby making the method applicable for the determination of average domain sizes.
The average domain size range determined here agrees with SEM measurements from
which domain sizes of 1-3 μm can be estimated.
The above treatment also suggests that most Xe in the iPP matrix is absorbed in a shell
around the EPDM domain with thickness d, where d must be larger than 0.2 μm and
significantly smaller than 5 μm.
12.5.3 Xe Diffusion as a Function of the Diffusion Time
Here the PFGE results for two elastomers: EPDM and polyisoprene [24] are compared.
The EPDM investigated here is, for an elastomer, highly crystalline (30%). While
crystalline domains are expected to behave as diffusion barriers for Xe, this was thought
to be an interesting case for the determination of the diffusion coefficient as a function of
the diffusion time Δ. As a comparison the completely amorphous polyisoprene was used.
The results of the time dependent PFGE experiments are shown in Figure 12.24. This
figure shows that the Xe diffusion in polyisoprene is faster than in the EPDM sample
Figure 12.24 The effective diffusion coefficient as a function of the diffusion time Δ
for polyisoprene (top curve) and EPDM (lower curve)
485
and that for both samples the effective diffusion coefficient shows no significant time
dependence. Since the shortest value for Δ in our case is 10 milliseconds, if any time
dependence would exist for Δ< 10 milliseconds, for example for the EPDM sample with
the high crystallinity, then the corresponding characteristic dimensions of the space
between the crystallites must be smaller than approximately 1 μm.
Another system which was investigated to find time dependent diffusion coefficients is a
PBT/PTMO block copolymer. This system, which was investigated before with 13C NMR
[12], contains large crystalline spherulites with dimensions in a range which could give
restrictions for the Xe diffusion in the time range of Δ>10 ms. Figure 12.25 shows the effective
diffusion coefficient as a function of the diffusion time Δ for untreated PBT/PTMO with
large spherulites and of PBT/PTMO which was melted at 200 °C and then quickly cooled
down with liquid nitrogen. This sample contained fewer and smaller spherulites.
The results of Figure 12.25 show very little difference between the two samples for Xe
diffusion. The presence of large spherulites in the untreated sample does not seem to
affect the diffusion of Xe. The reason maybe that for the copolymer investigated here the
spherulites are only partly crystalline because the block chain lengths [12] dictate that
the spherulites also contain the elastomer component.
Figure 12.25 The 129Xe diffusion coefficient for Xe absorbed in the copolymer
material PBT/PTMO as a function of the diffusion time. One sample was slowly
cooled down from the melt and contained many large spherulites. The other was
quickly cooled down and contained only small spherulites
486
129
For both samples a small decrease of the effective diffusion coefficient for short Δ is
found. This could indicate some diffusion restriction. However, the samples consisted of
small pieces of material pressed together in the NMR tube. Unfortunately, at the moment
the possibility that the initial decrease of Deff is caused by a motion (approximately 1 μm
would be sufficient!) of some of the particles caused by the pulsed field gradients cannot
be excluded. For short diffusion times such a motion has a large effect, for longer Δ such
motions can be neglected.
12.6 Conclusions
129
Xe can be used to probe the structure of polymer blends and polymer composites. A
material with an heterogeneous structure with different amorphous phases in general
absorbs Xe atoms in all phases. It depends on the size of the domains, the Xe diffusion
coefficient and the characteristic time scale of the NMR experiment, whether in the
NMR spectrum one or more resonances from absorbed Xe are found. A particular good
example of this principle is found in the iPP/EP blend or copolymer, where the EP domains
are large enough so that two Xe resonancs are found, one for Xe in iPP and one for Xe
in EP. The single resonance for Xe in EP makes clear that the EP copolymer for Xe is
homogeneous. From the diffusion coefficients, which can also be determined by NMR,
the average size of domains can be estimated and this method represents an alternative
to electron microscopy techniques.
As is shown in the first part of the paper it is believed that Xe absorbs only in non-
crystalline regions of polymers. Therefore it can be expected that crystalline domains
form a diffusion barrier for Xe. For the right sizes of crystallites this would imply that
the Xe diffusion coefficients are dependent on the diffusion time. Such effects have been
found in some catalysts [25]. Attempts to detect similar phenomena in semi-crystalline
polymers so far failed, possibly because the systems chosen here do not have the right
internal crystalline structure (in the PBT/PTMO case) or the crystallites are too small (in
the case of EPDM). Results on semi-crystalline non-elastomer polymers will be published
elsewhere.
References
1. D. Raftery and B.F. Chmelka in Solid-State NMR 1, NMR Basic Principles and
Progress, Eds., B. Bluemich, P. Bluemler, B.F. Chmelka, G. Fleischer, F. Fujara, A-
R. Grummer, F. Laupretre and D. Raftery, Springer Verlag, Berlin, 1994, 30, 111.
487
2. C. Ratcliffe in Annual Reports on NMR Spectroscopy, Ed., G.A. Webb, Academic

Press, 1998, 36, 123.
3. A.K. Jameson, C.J. Jameson and H.S. Gutowsky, Journal of Chemical Physics,
1970, 53, 2310.
4. C.J. Jameson, A.K. Jameson and S.M. Cohen, Journal of Chemical Physics, 1973,
59, 4540.
5. C.J. Jameson, A.K. Jameson and S.M. Cohen, Journal of Chemical Physics, 1975,
62, 4224.
6. C.J. Jameson, Journal of Chemical Physics, 1975, 63, 5296.
7. A. Flohr, 129Xe-NMR-spektroskopie und PVT-Untersuchungen am Blends von

PVDF und PMMA, Duisburg 1995. [PhD Thesis]
8. A.P.M. Kentgens, H.A. van Boxtel, R.J. Verweel and W.S. Veeman,
9. J. Kärger and D.M. Ruthven, Diffusion in Zeolites and Other Microporous

Solids, John Wiley & Sons, New York, 1992.
10. P.T. Callaghan, Principles of Nuclear Magnetic Resonance Microscopy, Oxford

University Press, Oxford, 1991.
11. E.O Stejskal and J.E. Tanner, Journal of Chemical Physics, 1965, 42, 288.
12. A. Schmidt, W.S. Veeman, V.M. Litvinov and W. Gabriëlse, Macromolecules

1998, 31, 1652.
13. M. Mansfeld, A. Flohr and W.S. Veeman, Applied Magnetic Resonance, 1995, 8,
573.
14. K. Sperling-Ischinsky, 129Xe-NMR-spektroskopische Untersuchungen an

Ungefullten und Gefullten Polymeren, Duisburg 1999. [PhD Thesis]
15. J.B. Miller, J.H. Walton and C.M. Roland, Macromolecules, 1993, 26, 5602.
16. G.R. Strobl, The Physics of Polymers, Concepts for Understanding their
Structures and Behaviour, Springer Verlag, Berlin, 1996.
17. S. Kaufman, W.P. Slichter and D.P. Davis, Journal of Polymer Science, A, 1971, 2,
9, 829.
488
129
19. B. Haidar, Proceedings of the ACS Rubber Division Meeting, Las Vegas, NV,
USA, 1990, Paper No.64.
20. K. Sperling-Ischinsky and W.S. Veeman, Journal of the Brazilian Chemical

Society, 1999, 10, 293.
21. P.P. Mitra, P.N. Sen, L.M. Schwartz and P. le Doussal, Physical Review Letters,
1992, 68, 3555.
22. L.L. Latour, P.P. Mitra, R.L. Kleinberg and C.H. Sotak, Journal of Magnetic
Resonance A, 1993, 101, 342.
23. F. Junker and W.S. Veeman, Macromolecules, 1998, 20, 7010.
24. F. Junker, Materialforschung an Porosen Festkorpern Mittles Xenon –

Diffusionsmessungen: eine 129Xe-PFGE-NMR-studie, Duisburg, 2000. [PhD
Thesis]
25. F. Junker and W.S. Veeman, Chemical Physics Letters, 1999, 305, 39.
489
490
13
Swollen Rubbery Materials: Chemistry
and Physical Properties Studied by NMR
Techniques
Andrew K. Whittaker
13.1 Introduction - The Swelling of Crosslinked Rubbers
One of the most characteristic properties of crosslinked rubbers is the ability to swell in
appropriate solvents to a constant volume. Not only is this property exploited for
estimation of parameters such as crosslink densities and polymer-solvent interaction
parameters, but the resultant change in nuclear magnetic resonance (NMR) parameters
allows a large number of new and interesting NMR experiments. It is the aim of this
chapter to introduce some simple concepts of polymer swelling and to examine the
information obtainable for the range of NMR experiments possible on swollen gels.
13.1.1 The Theory of Rubber Swelling
The seminal work of Flory and Rehner [1-3] set the framework for analysis of swollen
polymer gels. In their approach it is assumed that in a crosslinked rubber at equilibrium
the free energy of mixing and the change in elastic component are separable and additive.
The chemical potential change on dissolution at equilibrium is written as:
μ1 − μ10 = (μ1 − μ10 )mix + (μ1 − μ10 )el = 0
The mixing term can be expressed using the liquid-lattice theory of Flory and Huggins
[4, 5] as:
(μ1 − μ10 )mix = RT[ln(1 − υ2 ) + υ2 + χυ22 ]
where ν2 is the volume fraction of polymer in the swollen state, and χ is the polymer-
solvent interaction parameter. The change in the elastic part of the chemical potential
can be written as [6]:
V1 ∂ΔA
(μ1 − μ10 )el =
λ 2 ∂λ
491
where R is the gas constant, T is temperature, V1 is the molar volume of the solvent, λ is
the stretching ratio and is equal to the inverse cubed root of the rubber volume fraction,
and ΔA is the elastic free energy per unit volume of dry rubber. For a discussion of the
elastic potential the reader is referred to the works of Flory [7] and Mark [8].
There has been much recent discussion on the validity of application of the Flory-Huggins
equation to crosslinked polymers, in particular whether the interaction parameter for a
network is identical to that for a linear polymer [9 and references therein]. For example
the work of McKenna and co-workers [10, 11] on crosslinked natural rubber (NR)
swollen with a variety of solvents, showed that the polymer-solvent interaction parameter
was indeed dependent on crosslink density as (χc-χ0)/ χ0 = αν. The proportionality
constant, a, relating the crosslink density with the interaction parameter in the crosslinked
network was nearly independent of solvent type. McKenna and co-workers used
measurements of the torque and normal force as described by Kearsley and Zapas [12].
More recently, Hrnjak-Murgic and co-workers [13] reported a similar relationship for
ethylene-propylene-diene rubber (EPDM) swollen with a number of solvents using
crosslink densities derived from the use of the Mooney-Rivlin equation. On the other
hand, Malone and co-workers [14] used measurements of the shear modulus to show
that χ was independent of crosslink density for polydimethyl siloxane (PDMS) swollen
with di- and tri-methylpentane. In summary it appears that for polymers in poor solvents,
the interaction parameter is often dependent on volume fraction, while for many good
solvents the original Flory-Huggins relationship holds.
13.1.2 Relationship with NMR Parameters
It is largely accepted that the dominant mechanism of nuclear spin relaxation in condensed
polymers is due to dipolar interactions between the spins. The truncated homonuclear
dipolar Hamiltonian has the form [15]:
1
HD ∝ ∑ ∑ 2(3 cos 2
θ − 1) / r 3 (3I z I z − I.I) (13.1)
The sum must be made over all spin pairs in the proton-rich solid. In the absence of
large-amplitude molecular motion this Hamiltonian describes a line shape of width of
up to 100 kHz. In the presence of molecular motion the angular part of Equation 13.1
becomes time-dependent, and the partial averaging of this term results in reduced
linewidths. In polymers the geometry of main-chain motion is limited by the structure of
the polymer chain, and is inherently anisotropic. As a general rule, as the measurement
temperature is increased the motion tends to become more isotropic in nature as the free
volume increases, and the extent of averaging of the dipolar Hamiltonian increases. This
492
Swollen Rubbery Materials: Chemistry and Physical Properties Studied by NMR Techniques
effect is most pronounced on passing through the glass transition temperature (Tg) into
the rubbery phase, where the linewidths may decrease by around two orders of magnitude.
It follows that the dipolar couplings are averaged even more effectively by swelling the
polymer with a low molecular weight solvent. It must be stated, however, that at high
swelling ratios the elongation of the network chain segments can result in local ordering,
and hence increased anisotropy of motion.
Crosslinking of polymer chains has a profound effect on many important physical

properties, as reviewed for example by Nielsen [16]. In the following sections the effect
of crosslinking and swelling on the NMR properties are discussed. As a starting point
studies of the chain motion in uncrosslinked rubbers in the bulk and solution-states are
reviewed (Section 13.2). A feature of these studies is the calculation of correlation times
of co-operative chain motion in close agreement to those predicted by the Williams Landel
Ferry (WLF) equation. Motion of the low molecular weight solvents is next examined
(Section 13.3). Measurements of self-diffusion of the solvent molecules has demonstrated
the close relationship with the relaxation of the polymer chain, and is again consistent
with the free volume theories of chain motion. Diffusion under the presence of a
concentration gradient as studied by NMR Imaging is reviewed in Section 13.4.1.
The NMR studies reviewed in this chapter do not only provide insights into the
mechanisms of polymer swelling and the effect of swelling on NMR parameters, but also
allow the opportunity of routine characterisation of network properties. This is especially
crucial for filled or blended materials which are especially challenging materials for most
analytical techniques. The use of NMR imaging to provide high-resolution images of
network densities in swollen materials is briefly reviewed (Section 13.4.2). Measurements
of 1H and 13C spectra and 1H relaxation times which can provide a rapid estimate of
crosslink densities in complex materials are reviewed in Section 13.5. Finally the use of
high-resolution magic-angle spinning (MAS) NMR to measure crosslink structures is
suggested as a future area of much interest.
13.1.3 Reviews of NMR of Crosslinked Rubbers and Polymers
A number of reviews of the use of NMR to study crosslinked polymers have appeared
over the past 15 years, and provide a range of perspectives on this interesting field.
Andreis and Koenig [17] provided an extensive review of the literature in 1989, including
a comprehensive summary of background theory. Harrison and co-workers [18] have
summarised the range of possible techniques for analysis of crosslinked polymers, including
NMR spectroscopy. Bauer [19] and more recently Kinsey [20] have also reviewed the use
of NMR to characterise crosslinked polymers. The most recent reviews have been provided
by Whittaker [21] and Mori and Koenig [22] in 1997.
493
13.2 Motion of Polymer Chains in Polymer Solutions and Swollen

Networks
13.2.1 Background
It has long been recognised that the classical theory of NMR relaxation [23], derived for
simple isotropic motion of small molecules, is insufficient to describe the complexity of
relaxation of nuclear spins in polymers. This complexity is of course a result of the
coupled anisotropic motion of the chain segments. The model of Dejean, Laupretre and
Monnerie [24] provides the most successful description of relaxation of 13C spins in
polymers via the dipolar mechanism. The model derives from the autocorrelation function
of Hall and Helfand for conformational jumps [25]. Dejean and co-workers [24] found
the Hall-Helfand model unable to describe the temperature dependence of 13C relaxation
times of polyvinyl methyl ether in the bulk and solution, and recognised that an additional
rapid motion, namely librations (small angle oscillatory motions) of the C-H vectors
about their equilibrium positions, needs to be accounted for. Their model yielded values
of the correlation time of correlated jump motion with a temperature dependence
consistent with that predicted by the WLF equation. In an accompanying paper, the
authors examined motion in linear and crosslinked polyethers [26]. For polyethylene
oxide (PEO) crosslinked through a tri-functional aromatic urethane linkage the average
correlation times of segmental motion increased with decreasing PEO segment length.
The spectra were collected using high-resolution solution-state techniques, that is scalar
proton decoupling, and so the signal from segments removed from the crosslink points
were only detected. For these units faster molecular motion results in partial averaging
of the dipolar couplings. The use of cross-polarisation to measure spectra of 13C nuclei
experiencing near-static dipolar couplings with near-neighbour protons was demonstrated.
It was estimated that up to four methylene units adjacent to the urethane crosslinks are
detected in the cross-polarisation experiment.
13.2.2 Solutions and Bulk Rubbers
Laupretre and co-workers [27-29] have studied the local motions of polyisobutylene
(PIB) and polyisoprene in the bulk and in solution using both measurements of 13C T1
relaxation times, and simulations using the co-operative kinematics (CK) theory. Ediger
and co-workers have also contributed to this field [30, 31]. Figure 13.1 shows the spin-
lattice relaxation times of bulk polyisoprene at two field strengths across a range of
temperatures, as well as the results of fitting of the data to the theory of Dejean de la
Batie and co-workers [24]. The correlation times of the co-operative segmental motions
494
Figure 13.1 13C spin-lattice relaxation times (nT1) for methylene carbons in bulk
polyisoprene at two different field strengths (62.5 (O) and 25.15 (●) MHz) as a
function of inverse temperature. The lines are the best fits of the data to the theory of
Dejean de la Batie and co-workers [24] (long dashes) and for the modified relaxation
function described in reference [27] (dotted lines)
Reprinted with permission from Macromolecules, 1989, 22, 122, Figure 5. Copyright
1989 American Chemical Society
were by an order of magnitude shorter for polybutadiene (BR), however, the activation
energies for both polymers were very similar. Very recently this group reported a study
of motion using the CK theory [29], and indeed obtained a result in close agreement
with the NMR data. The longer correlation times for polyisoprene were ascribed to a
number of effects including differences in excluded volume and short-range
conformational potentials between the two polymers.
Another study was conducted by this group of the effect of polymer microstructure on
the segmental mobility. In the first of two papers [32] they compared the rates of 13C T1
relaxation for five samples of BR having varying proportions of 1,2-vinyl, cis-1,4- and
trans-1,4 double bonds. The high-resolution solution-state spectra can be resolved into
495
peaks due to sequences up the triad level. The T1 relaxation times of methylene units are
relatively high when there is a high proportion of adjacent 1,4-units, and remained constant
with decreasing levels of substitution of 1,4-units. Below a mole fraction of 1,4-unit
content of about 0.6, T1 decreased, indicating reduced segmental mobility. This reflects
the presence of 1,2-vinyl units which slow down segmental motion of the 1,4-unit. In a
later publication [33] the authors again use the CK theory to examine the mobility of cis-
and trans-1,4 units in BR. Shorter correlation times for segmental motion were observed
in BR with high cis-1,4 content. Comparison with CK results indicated that the NMR
measurements were consistent with larger amplitude motions involving cis-1,4 units
resulting in more rapid loss of the original orientation.
13.3 Motion of Small Molecules in Swollen Rubbers
13.3.1 Self-Diffusion of Small Molecules in Rubbers
The diffusion of small molecules in rubbers is of both theoretical and practical importance.
The theories of diffusion based on consideration of free volume can be tested by
measurement of self-diffusion using methods such as pulsed field gradient NMR. Self-
diffusion of small molecules must be understood for applications of rubbers as seals in
contact with solvents, and for example for diffusion of plasticisers and other small
molecules.
The pulsed field gradient NMR (PFG NMR) experiment was introduced in 1965 by
Stejskal and Tanner [34]. In brief the experiment consists of two matched gradient pulses
which have the effect of applying to the nuclear spins a well-defined change in phase of
the magnetisation which depends on the position of the spins in real space. The two
gradient pulses are applied during a spin-echo sequence in which a 180° pulse results in
a change in sign of the phase of all spins within the sample. In the stimulated echo
sequence the two 90o pulses on either side of the storage time have the same effect. As a
consequence, if the spins have changed position during the interval separating the gradient
pulses, i.e., if they have undergone diffusion or flow to a new position, the change in
phase due to the second pulse does not match that of the first. When averaged over the
whole sample there will be a net loss of phase coherence. Stejskal and Tanner [34] showed
that the decay in echo amplitude was described by the following expression, provided
that the self-diffusion was due to unrestricted Brownian motion:
δ
I(2τ) = I(0).exp(− γ 2 Dδ 2 g 2 (Δ − ))
3
496
where I is the echo amplitude, D is the self-diffusion coefficient, δ is the width of the
gradient pulse, Δ is the separation of the two gradient pulses, and g is the strength of the
gradient pulses. For a detailed discussion of this experiment the interested reader is referred
to the standard text of Callaghan [35].
The self-diffusion of benzene in PIB [36], cyclohexane in BR [37] and toluene in PIB [38-
40] has been investigated by PFG NMR. In addition more recently Schlick and co-workers
[41] have measured the self-diffusion of benzene and cyclohexane mixtures in
polyisoprene. In the first reported study of this kind, Boss and co-workers [36] measured
the self-diffusion coefficients of benzene in polyisoprene at 70.4 °C. The increase in Dself
with increasing solvent volume fraction could be described by the Fujita-Doolittle theory
which states that the rate of self-diffusion scales with the free volume which in turn
increases linearly with temperature. At higher solvent volume fractions the rate of self-
diffusion deviates from the Fujita-Doolittle theory, as the entanglement density decreased
below the critical value.
A later study of Guillermo and co-workers [37] examined the diffusion of cyclohexane
in BR as a function of temperature and solvent concentration. Figure 13.2 shows the
increase in self-diffusion coefficient with increasing temperature at several solvent
volume fractions, and the fits of the data to the WLF equation (solid lines). The change
in Dself with concentration at constant temperature was equally well described by the
Vrentas-Duda equation, which is also a free volume model. This latter model provided
an estimate of the size of the polymer unit undergoing motion during the critical diffusion
step being approximately 1.5 times the size of the monomer unit. More recently again,
Bandis and co-workers [36, 37] conducted a similar analysis of the self-diffusion of
toluene in PIB, and found both the WLF and Vrentas-Duda theories described the
behaviour with temperature and solvent volume fraction well. These authors went on
to measure the dependence of the 13C spin-lattice relaxation times of both the PIB and
toluene under identical conditions, and at different field strengths [38]. A spectral
density function based on the Hall-Helfand function was used to describe motion of
the methylene units of the PIB. The motion was described in this model by two
correlation times, for single conformational transitions and correlated transitions,
respectively. The authors considered both an Arrhenius and WLF temperature
dependence of these correlation times; use of the WLF equation resulted in estimates
of the fractional free volumes close to those determined from the measurement of
translational diffusion of toluene in their previous paper. The correlation times of
diffusion of toluene and for single conformational transitions were comparable at low
toluene volume fractions. At higher toluene concentrations the diffusion of the small
molecules is less strongly coupled to polymer chain motion, and so the correlation
times differed by up to two orders of magnitude.
497
Figure 13.2 Self-diffusion coefficients of cyclohexane in BR solutions as a function of

temperature and solvent volume fraction (ν1 = 0.05 (O), 0.19 (●), 0.29 (❑), 0.375 (■)
and 0.45 (Δ) [37]. The solid and dashed lines are best fits to the WLF equation [37]
1993 American Chemical Society
13.3.2 Self-Diffusion of Rubbers
While PFG NMR has been applied extensively to the study of polymers in solution and
in the molten state, there has been just one report of the diffusion of chain segments in a
swollen polymer network. This is a result of the requirements of very high gradient field
strengths to measure the very small displacement which occur in crosslinked gels. In
addition, as described elsewhere in this chapter, the spin-spin relaxation times of the
protons on the polymer chain segments are reduced on crosslinking, hence limiting the
possible gradient and diffusion times. The groups of Skirda [for example 42-45] and
Fleischer [46] has equipment capable of gradient strengths up to 50 T/m; this compares
with gradients of 1 T/m typically available from commercial equipment. Since the range
of accessible values of Dself is proportional to the square of the maximum gradient strength,
Fleischer and co-workers [46] are able to measure values of Dself as low as to 2x10-16 m2/s.
498
The mean-square displacement of the chain segments of BR swollen with deuterated

benzene was observed to be independent of diffusion time, indicating restricted diffusion
around an attractive centre. The mean-squared displacement decreased with increasing
crosslinking density and was approximately equal to the mean-squared collective
fluctuations calculated for these polymers.
13.4 NMR Imaging of Swollen Rubbers
Chapter 7 of this text details the use of NMR imaging methods to determine structure
and function of crosslinked rubbers. In this section we review the NMR imaging of
rubbers swollen with small molecules. Two sets of experiments are described, namely
the study of diffusion of small molecules into a previously unswollen rubber (macroscopic
diffusion), and the measurement of images of the solvent and/or polymer after equilibrium
has been achieved.
13.4.1 Macroscopic Diffusion of Small Molecules in Swollen Rubbers
NMR imaging is becoming increasingly acknowledged as an important tool for the study
of the structure of polymeric materials [47-54], including the kinetics of diffusion of
small molecules. In many cases details of the kinetics of diffusion are only discernible
from a knowledge of the concentration profile of the solvent in the swelling polymer.
An example of the power of the NMR imaging technique has been given by Halse and
co-workers [55] who studied the diffusion of decane into peroxide-crosslinked NR. They
use a gradient echo sequence and exploited the differences between the T2 relaxation
times of the solvent and the polymer to obtain an image of the solvent only within the
rubber. In a second experiment they collect the echo after 1.8 milliseconds, and by repeating
the experiment every 100 milliseconds they acquired a signal largely from the rubber,
which has a short T1 relaxation time (approximately 25 milliseconds) compared with the
solvent (approximately 1 second). Concentration profiles from images of the solvent
taken at two diffusion times were consistent with simple Fickian kinetics, however, the
calculated value of diffusion coefficient was two times smaller than that obtained from
measurement of increase of mass of solvent. The profiles were then fitted by assuming a
concentration-dependent diffusion coefficient and results consistent with the mass uptake
experiments were obtained. The macroscopic diffusion coefficient was assumed to depend
on concentration in the same manner as the self-diffusion coefficient. In the second part
of the paper [55], images of the rubber phase were obtained during swelling. The T2
relaxation time of the rubber was found to vary across the width of the swelling sheet, as
expected from the discussion above.
499
The diffusion of acetone into crosslinked NR was measured by Webb and Hall [56] and
shown to follow Fickian diffusion kinetics, despite the shape of the profiles of
concentration of acetone in the rubber giving evidence for the diffusion coefficient being
dependent on concentration. This concentration dependence of the diffusion coefficient
is ascribed to the plasticisation of the polymer chains by the small molecules. In the
mixed solvent system [57], benzene and acetone, diffusing into crosslinked natural rubber,
images of the individual solvents were obtained by selective excitation of the 1H resonances
of each solvent by replacing the 90o excitation pulse with a binomial excitation pulse
sequence. The rates of ingress were identical for the two solvents, but differed from the
values measured for the pure solvents; the rate of diffusion was 0.8 times that for pure
benzene, and 2.5 times that for pure acetone. The authors also demonstrated that in the
case of imaging of liquids having overlapping peaks in the 1H NMR spectrum, the
chemical-shift selection method is not appropriate, and that selectivity via the creation
of multiple quantum coherence is possible [57]. A more complete analysis involving a
larger number of diffusants is reported in a third paper by these authors [58].
13.4.2 NMR Imaging of Swollen BR and Polyisoprene Rubbers
Rubbery materials are particularly suitable for imaging experiments since the T2 relaxation
times of rubbers are relatively long, and so simple spin-echo imaging sequences result in
images having high signal-to-noise ratios. This is especially true when the rubbers are
swollen with small molecules.
The properties of swollen rubbers have been investigated in several papers by Koenig
and co-workers [59-61]. Smith and Koenig [59] have obtained well-resolved images of
sulfur-vulcanised BR swollen with deuterated cyclohexane. The residual soluble fraction
of the materials had previously been extracted to eliminate contributions from free chains
with longer T2 relaxation times. On increasing the extent of cure, an overall reduction in
signal-to-noise ratio was observed, consistent with a decrease in the T2 relaxation time.
The authors also collected T2-maps, i.e., images in which the intensity of the image is
proportional to the T2 at each pixel, and from these constructed histograms of T2 values
as a function of cure time (Figure 13.3). These images are highly heterogeneous,
presumably reflecting the heterogeneous nature of the crosslinking reaction. The values
of T2 were related to the equilibrium dimensions of the crosslinked network, and
histograms of local swelling ratios produced. As expected the swelling ratios decreased
with longer curing time. In a more recent paper, Mori and Koenig [61] measured spin-
echo images of cured carbon black-filled polyisoprene, and related the image intensity to
the T2 relaxation time of the chains within each image volume element. In this case T2
relaxation times, and the width of the distribution of relaxation times, decreased with
increasing filler content.
500
Figure 13.3 Histograms of values of T2 relaxation times derived T2-resolved NMR

images of BR swollen with deuterated cyclohexane [59]. The material was cured with
2 phr tetramethylthiuram disulfide at 150 °C for the times shown on the figure [59]
1991, American Chemical Society
Another approach to the imaging of heterogeneous structures in crosslinked rubbers is

to swell the material with a fully-protonated small molecule, and measure the distribution
of the small molecule. Examples of this approach have again been given by Koenig and
co-workers. Clough and Koenig [62] studied the heterogeneous distribution of dioxane
in BR. Rana and Koenig [60] have measured images of cyclohexane in sulphur-cured BR
having 70% 1,2-vinyl groups. Areas of low image intensity indicated low cyclohexane
content and hence high crosslink density. There was some evidence of reversion in samples
cured at 150 oC for two hours. More recently Oh and Koenig [63] have measured T2
images of the cyclohexane in peroxide-crosslinked polyisoprene. When zinc diacrylate
was used as a crosslinking co-agent, the spin-spin relaxation time of the cyclohexane
was found to decrease with increasing crosslinking level, as confirmed by high-resolution
501
NMR spectroscopy. On the other hand it was found that T2 of the solute in the rubber
cured in the presence of triallyl cyanurate (TAC) increased on curing, despite the
equilibrium swelling ratio decreasing. This surprising result was explained by a change
in the polymer-solvent interactions on copolymerisation with TAC.
13.5 Studies of Network Density in Swollen Rubbers and Blends
13.5.1 Measurements of Transverse Relaxation Times
A major focus of this chapter is the effect of network formation and network density on
the relaxation times of the 1H and 13C nuclei in the NMR experiment. The changes in
relaxation times can be exploited to obtain information about crosslink density and
chain motion, and must be taken into account in the design of experiments to determine
changes in chemical structure. In this section we examine how crosslinking changes the
transverse (T2) relaxation times of 1H nuclei, and how this information can be of use.
Two different approaches have been taken in the literature, namely changes in T2 can be
used to estimate crosslink density, or used to develop and verify models of chain motion.
13.5.1.1 Measurement of Crosslink Density from Transverse Relaxation Decays
It has long been known that the decay of transverse magnetisation in entangled polymers
is non-exponential. In the most general sense the rate of T2 relaxation can be expressed
as the sum of the individual relaxation decays for different parts of the molecules. In
early work the details of the motion of the polymer chains were neglected, and a empirical
model of the behaviour of the spin system as a function of crosslink density developed.
Later workers, in particular Gotlib, Cohen-Addad and Brereton and their colleagues
have made much progress in developing descriptions of the relaxation based on models
of motion of the entangled and crosslinked chains.
Despite the complexity of the decay functions generated from consideration of the
hierarchical motion of the chain segments, a more phenomenological approach has met
with some success. In concentrated solution (C > C*), or in the molten state when the
molar mass is greater than the critical mass for entanglements, there is often observed
two separate decays with largely-different time constants of decay:
M x, y (t)
= f1R1(t) +f2 R 2 (t)
M x, y (0)
502
The fractions of protons decaying according to relaxation functions R1 and R2 are given
by f1 and f2. In molten polymers this relationship has long been exploited to provide a
measure of the crosslink density in many polymer systems [64-83]. The form of the
decay functions has been the subject of much discussion, however, it is often observed
that R1 and R2 can be approximated by simple exponential decay functions. It is generally
accepted that the protons with short relaxation times are those directly attached to or
adjacent to crosslink points. As an example Figure 13.4 shows the decays of transverse
Figure 13.4 1H spin-spin relaxation decays recorded using the Hahn-echo sequence
for polyisoprene at 150 °C, crosslinked with radiation doses of 0 (●), 0.3 (O),
0.58 (▲), 1.5 (■), 2.25 (❑) and 5.2 kGy [65]
Reprinted with permission from R. Folland and A. Charlesby, Polymer, 1979, 20, 211,
Figure 1, Copyright 1979, Elsevier Science
503
magnetisation in polyisoprene crosslinked with increasing dose of gamma radiation [65].

Clearly the rates of decay increase with increasing radiation dose. In addition the
proportion of protons decaying more slowly decreases with dose. When the irradiation
dose is corrected for a ‘virtual dose’ to account for the presence of physical entanglements
present prior to irradiation, the fraction of slowly decaying protons agrees very closely
with the soluble fraction measured by Soxhlet extraction. Although this and similar
studies were performed on molten materials, the same information could be obtained
from swollen samples. Indeed the range of crosslink densities that could be examined
using this method would be extended by swelling in a deuterated solvent, since the
difference between the rates of fast and slow relaxation would be increased [84], and
hence resolution of the two decays would become more facile.
The effect of swelling on the rates of transverse relaxation has been examined by Fukumori
and co-workers [84]. T2 relaxation times were measured for carbon black-filled nitrile
rubber (NBR) during swelling with carbon tetrachloride. The relaxation decay was
resolved into a two exponential process, the slower of which became more predominant
during swelling. The fraction of protons relaxing slowly, i.e., the more mobile chains,
was found to increase with increasing swelling ratio, and was therefore directly
proportional to the squared root of swelling time, as predicted by Fickian models of
diffusion. In addition there is evidence of more effective rubber-filler interactions, and
restricted pathways to diffusion with increased filler content persistent at long diffusion
times.
While it is generally accepted that the slow relaxation decay in Equation 13.2 can be
represented by an exponential decay, it is well known that the rapid decay of magnetisation
conforms more closely to a gaussian decay function. Folland and Charlesby [65] recognised
this but were more concerned with changes in the relative proportions of the rapidly and
slowly decaying functions. In 1976 Gotlib and co-workers [85] developed an expression
for the second moment of the NMR line shape by considering the anisotropic motion of
constrained polymer chains. They were the first to recognise that the average moment
will not depend on the details of the chain motion, but rather the average conformations
of the polymer chains. In their case the conformational mean was calculated using the
results of Kuhn and Grün for freely jointed chains [86]. They were able to show that the
plateau value of T2 of crosslinked polystyrene swollen in toluene was inversely
proportional to the crosslink density. Fry and Lind [76] later showed that these calculations
could be applied to heavily crosslinked epoxy resins. More recently, Litvinov and co-
workers have used these calculations to determine the density of chemical crosslinks and
chain entanglements in cured EPDM [87], the interactions between carbon black and
polymer chains in filled EPDM [88], and the average length of amorphous chains spanning
inter-lamellar spaces in semi-crystalline polyvinyl chloride [89].
504
Simon and co-workers [71-75] later developed a variation of this model of T2 relaxation
in which the faster decay process arises from an ‘inter-crosslink’ component of the network,
in which motion is rapid and anisotropic and does not eliminate completely the dipolar
interactions. The rapid motion, described by a correlation time τf, occurs simultaneously
with a slower isotropic motion of the complete inter-crosslink segment described by the
correlation time τs. The longer relaxation decay arises from dangling chain ends and is
exponential in nature. The total decay of transverse magnetisation if then given by:
M x, y (t) −t −t t −t
= A.exp[ − qM 2 τ s2 (exp( ) + − 1)] + B.exp[ ]
M x, y (0) T2 τs τs T2
where M2 is the second moment of the rigid-lattice line shape, q is the fraction of the
rigid-lattice second moment unaveraged by the anisotropic motion, and 1/T2 = M2τf. An
additional long relaxation decay having approximately 5% of the total signal amplitude
was ascribed to uncrosslinked or sol chains. It was argued that the parameter q, the
degree of averaging of the rigid-lattice second moment, could be related to the molecular
weight between crosslinks Mc as:
3cMru
Mc =
5n q'
where q′ is parameter q for a crosslinked network minus the contribution from physical
entanglements (qo) measured for an uncrosslinked material.
This method was recently applied by Menge and co-workers [90] to the study of swollen
PDMS networks. While the analysis of the transverse relation behaviour of the unswollen
network conformed well with the theory of Simon and co-workers, on swelling more
complex behaviour was observed. For weakly swollen materials the parameter q, the
extent of averaging of the second moment decreased, however, for strongly swollen
networks relaxation became more efficient at short times. This was indicative of increased
anisotropy of motion of inter-crosslink segments due to deformation of the network.
The relationship between Mc and q thus breaks down in this situation. Application of
the empirical relationship proposed by Schmidt and Cohen-Addad [91] demonstrates
the departure at higher swelling ratios from Gaussian behaviour.
In more recent years extensive work by Cohen-Addad [92-95] and Brereton [96-98] has
lead to a more comprehensive understanding of the effect of chain entanglements on the
decay of transverse relaxation in the NMR experiment. In their scale-invariant model,
the polymer chain is considered to consist of a series of sub-units, the smallest of which
505
consists of a small number of monomer units. The dipolar interactions experienced by

the protons are scaled by the number of units in these elemental units. It was further
shown [96] that in the limit of fast local motion, the dipolar interactions can be further
scaled over larger segments which consist of a number of the smaller units, and in
crosslinked polymers these units are connected by crosslinks and chain entanglements. It
was found that the decay of transverse magnetisation is insensitive to the correlation
time of motion of these larger units, but depends on the number of Rouse units between
the crosslink points, and hence the molar mass between crosslinks. These methods have
been applied with success to molten linear polymers, however, the effects of free chain
ends in crosslinked polydisperse materials limits its applicability [99]. This topic is
discussed in detail in Chapter 8.
13.5.2 Estimation of Crosslink Density from NMR Linewidths
13.5.2.1 1H NMR Linewidths
In 1989 Loadman and Tinker [100] published a study of the variation of linewidth in the
1
H NMR spectra of NR and NBR with crosslink density. In this and subsequent papers
their aim was to measure the crosslink density in blends of two rubbers. For this reason
they developed an empirical method for determining crosslink density from the peak
widths. The width was estimated from the ratio of the height of the peak at +/- 0.205 ppm
from the centre of the peak to the maximum height of the peak. This parameter, which
they called H%, was measured on the low field side of the olefinic peak of NBR, and on
the high field side of the olefinic peak of NR, to minimise interference from the overlapping
peaks in the blends. None-the-less, there is significant overlap in spectra of blends which
must be corrected for by reference to the spectra of the pure crosslinked swollen rubbers.
An iterative procedure is described in this initial paper [100] which results in a measure
of the equilibrium swelling of each component. The relationship between H% and the
volume fraction of rubber in swollen NR is shown in Figure 13.5. Although it is not
possible, given the information provided, to construct a comprehensive theory relating
the mechanism of line broadening with the extent of network formation, the value of
this work lies in the ability to provide a measure of crosslink densities in blended and
filled systems.
In a second paper Brown and Tinker [101] examined the effects of a number of parameters,
such as the accelerator used in the sulfur curing, peroxide versus sulfur curing, and
swelling ratio at constant crosslink density. The results for cis-polyisoprene show that
the value of H% is independent of accelerator. However, lower values of H% were seen
for peroxide-cured materials. The results for BR were independent of curant; the reasons
506
Figure 13.5 The relationship between the linewidth parameter H% and swelling ratio for
swollen NR [100]. The measurements were made in deuterochloroform at 23 °C [100]
Adapted with permission from M.J.R. Loadman and A.J. Tinker, Rubber Chemistry and
Technology, 1989, 62, 234, Figure 4. Copyright 1989 Rubber Division, American
Chemical Society
for this were not clearly explained. In another series of experiments, measurements were
made on partially swollen BR rubbers; a weaker dependence of H% on the swelling
ratio compared to the equilibrium swelling was observed. It was concluded that the
method is more sensitive to the crosslink density rather than the degree of swelling and is
hence a robust method for determining crosslink density.
The presence of carbon black filler increases the linewidths in the spectra from both
phases of a blend, and as a result comparison could only be made of the relative crosslink
density [102]. Later the authors showed that H% in the presence of filler in blends of
NR and EPDM was linearly related to H% in the unfilled blend [103]. In blends of NR
and BR a small excess of the sulfur reacted in the NR phase compared with the BR.
Measurement in unfilled blends allowed crosslink densities to be determined, and
confirmed the greater yield of crosslinking in the NR, is shown in a plot of H% against
crosslink density [104].
507
Separate work on a Fourier transform (FT) NMR spectrometer revealed the presence of
two peaks from the solvents from both within and outside the swollen gel [104]. See
below for a discussion of the origin of the NMR linewidths. The chemical shift of the
olefinic peaks was found to shift down field with increasing crosslink density, and hence
a modified method for determining H% was introduced. In this paper they also introduced
the first 13C NMR measurements of swollen rubber blends, and again found a systematic
increase in linewidth with increasing crosslink density. The higher resolution in the 13C
spectrum compared with 1H NMR allows the potential of more detailed information on
rubber mixtures.
The effect of partitioning of curatives on the crosslinking reactions in NR/BR blends was
explored in more detail in a later paper [105]. The ultimate extent of curing observed in
the individual phases of the blends was identical to that obtained for the pure components,
however, for the blend faster curing was initially observed in the BR phase. This was
related to the greater affinity of sulfur and accelerator for BR compared with NR. Other
systems examined include blends of epoxidised NR and cis-BR [106], NR blended with
cis-BR [107] and NR blended with EPDM [108]. The work of Tinker and co-workers
has been discussed at length by Cook [109].
13.5.2.2 13C NMR Linewidth
Before leaving this subject it is of value to discuss the possible use of swollen-state 13C
NMR to determine crosslink densities in swollen rubbers. As briefly mentioned previously,
Brown and co-workers [104] showed that measurements of the crosslinking density could
be obtained using 13C NMR. They examined in much greater detail the curing of blends
of NR and EPDM, including chemically-modified EPDM, in a later paper [108]. The
much greater resolution afforded by 13C spectroscopy has the potential to provide more
detail of the mechanism of reactions occurring during curing.
The work of Whittaker and O’Donnell [110] on radiation-crosslinked ethylene-propylene

rubber (EPR) showed that the overall 13C NMR signal intensity decreased with increasing
crosslink density (Figure 13.6). This was also observed earlier by Ford and co-workers
for crosslinked polystyrene (PS) [111]. The scalar decoupling used in most solution-state
NMR experiments is insufficient to remove strong dipolar couplings to 1H nuclei present
at 13C nuclei involved directly in crosslinks or adjacent to crosslinks. As a consequence
the NMR signal is observable only from nuclei well removed from the crosslink points.
Measurement of the chemical structure of crosslinks by solution-state NMR is therefore
restricted to materials crosslinked to levels below the gel point [110, 112-114]. It also
follows that for higher crosslink densities the 13C NMR signal is not representative of
the whole sample, but that rather the motion of the chains well removed from the crosslinks
reflects the crosslink density of the entire sample.
508
Figure 13.6 Decrease in 13C NMR signal intensity (●) and soluble polymer fraction
(■) for crosslinked EPR containing 36 mol% propylene, as a function of radiation
dose [110]. The NMR experiments were conducted in deuterated chloroform at 25 °C.
The solid lines are drawn as a guide to the eye
Adapted with permission from J.H. O’Donnell and A.K. Whittaker, Polymer, 1992, 33,
62, Figure 3, Copyright 1992 with permission from Elsevier Science
13.5.2.3 Origin of Changes in Linewidth
The origin of the linewidths in the 1H and 13C spectra of crosslinked polymer has received
some attention. Mohanraj and Ford [111] found an approximately linear increase in 13C
linewidth with magnetic field strength (2.3 T and 7 T) for swollen crosslinked PS, and
therefore suggested the linewidth was determined by chemical shift dispersion, although
at higher crosslink densities a smaller increase in linewidth was recorded. It was concluded
that other mechanisms, for example dipole-dipole couplings could become important
for highly-immobilised gels. This is certainly reasonable when using standard solution-
state NMR equipment.
Bain and co-workers [115] studied the field dependence of 13C linewidths for polyacrylates
crosslinked to low levels. The linewidth was directly proportional to field strength,
consistent with a significant contribution from a dispersion of chemical shifts arising
from new chain structures. In addition the authors used the Delays Alternating with
509
Nutations for Tailored Excitations (DANTE) pulse sequence [116] to selectively irradiate
a small band of frequencies in the centre of the methylene resonance. Despite insufficient
details of this experiment being provided this result was also consistent with the linewidth
arising from inhomogeneous broadening. More recently, however, Stover and Frechet
[117] have shown that MAS dramatically reduces the linewidths in 13C spectra of
crosslinked PS gels. MAS is effective in averaging weak dipole-dipole couplings and
broadening due to magnetic field inhomogeneities. In 1992 Brown and co-workers [104]
showed that two peaks were observed in the 1H NMR spectra for the small molecules
both inside and outside the swollen gel. This is indicative of a difference in bulk
susceptibility and hence local field in these two environments. It is likely therefore that
this mechanism also contributes to the NMR linewidths in these heterogeneous systems.
Line broadening due to susceptibility effects would scale linearly with field, as observed
by Errade and co-workers [118], would lead to an inhomogeneous line susceptible to
hole burning [115], and be averaged by MAS [117]. At higher crosslink densities dipolar
coupling would become increasingly important. To demonstrate that this issue is far
from being resolved, a study of linear unswollen polyisoprene by English [119] concluded
that the dominant mechanism of broadening in this case was dipolar in origin resulting
from anisotropic motion of chain segments. Given the above discussion it is considered
that these two effects, namely weak dipolar interactions and local field inhomogeneities,
contribute most significantly to linewidths in the 1H and 13C spectra at moderate crosslink
densities [120].
13.5.3 High-Resolution 13C MAS NMR of Rubbers
It has recently been recognised that high resolution NMR spectra of molecules attached
to swollen, crosslinked polymer beads can be obtained by the method of MAS. A number
of instrument manufacturers are marketing MAS probes for both 1H and 13C observation
designed for highly uniform static magnetic fields. The main potential area of application
of this technology is the field of combinatorial chemistry, in which small molecules are
synthesised on polymer matrices, and are required to be characterised before removal. It
follows that high resolution MAS NMR of swollen polymers will become an important
tool for characterisation of new structures and surface grafting in heterogeneous rubbers.
As an example, Doskocilova and co-workers [121] have recently used swollen-state 1H
and 13C MAS NMR to measure changes in structure of BR on thermal degradation. The
method of cross-polarisation was used to obtain spectra identical to those measured
using single pulse excitation. In highly heterogeneous materials cross-polarisation will
tend to favour signals from more rigid regions (at shorter cross-polarisation contact
times), while the single pulse experiment is biased towards 13C nuclei with shorter T1
relaxation times, that is in this case more mobile carbon nuclei. This result provided
evidence that the material was homogeneous in nature.
510
13.6 Summary
It is clear from the examples given in this review that there are many reasons for examining
the structure of crosslinked rubbers by swollen-state NMR. The first of these is the
increased resolution in the NMR spectrum compared with the molten state, as a result in
the more effective averaging of the dipole-dipole couplings between nuclear spins. With
conventional high-resolution solution NMR equipment it is possible to determine the
crosslink structure and concentrations of crosslinks prior to gelation. The demands on
spectrometer time for these experiments are prodigious since the concentrations of carbons
or proton nuclei participating in crosslinks are by definition less than one per polymer
chain. However, the new information obtained on the mechanisms of crosslinking fully
justifies this investment in time.
At higher crosslink concentrations two approaches can be adopted. Firstly the NMR
experiment can be designed to measure a parameter sensitive to changes in the segmental
motion of the bulk sample. Examples of these are measurements of 1H spin-spin relaxation
times, or widths of lines in the spectra. Since the NMR signal does not arise from the
entire sample under solution-state conditions it is necessary to construct a master curve
relating the NMR parameter to the crosslink density. This method is especially useful for
mixtures of rubbers or for filled rubbers. A second approach is to use the methods of
MAS and high-power proton decoupling to increase the efficiency of averaging of the
residual dipole-dipole couplings, and broadening due to magnetic field inhomogeneities,
and hence obtain a high-resolution spectrum of the whole sample. This is expected to be
an important method in future studies of gel structure.
An additional advantage of the reduced NMR linewidths and rates of transverse relaxation
is the possibility of higher sensitivity in NMR imaging experiments. This advantage is of
course most evident in systems where the transverse relaxation times are reduced to very
short values by high local crosslink densities or the presence of fillers. Examples have
been given where images are obtained of either the plasticised rubber phase or of the
solvent itself. The measurement of diffusion coefficients in the presence of a gradient in
solvent concentration by NMR imaging is also an important experiment with practical
significance, as is the measurement of self-diffusion coefficients by PFG NMR.
An understanding of the effect of crosslinking on local chain motion in polymers has

been developing over the past 20 years. The strong agreement between values of correlation
times for segmental motion of uncrosslinked chains determined by NMR and other
methods, and the clear relationship with free volume theory underlines the importance
of the NMR experiments. It is evident that the presence of permanent crosslinks in polymer
chains leads to a pronounced motional heterogeneity compared with linear chains in
solution or in the bulk state, for which a sound theoretical understanding of the mechanism
511
of chain reorientation has previously been established. The full application of this
knowledge to crosslinked rubbers has yet to be achieved, and promises to shed light on
the effect of crosslinking on segmental chain motion.
References
1. P.J. Flory and J. Rehner, Journal of Chemical Physics, 1943, 11, 512.
2. P.J. Flory and J. Rehner, Journal of Chemical Physics, 1943, 11, 521.
3. P.J. Flory, Journal of Chemical Physics, 1950, 18, 108.
4. P.J. Flory, Journal of Chemical Physics, 1941, 9, 660.
5. M.L. Huggins, Journal of Chemical Physics, 1941, 9, 440.
6. K.C. Valanis and R.F. Landel, Journal of Applied Physics, 1967, 38, 2997.
7. P.J. Flory, Macromolecules, 1979, 12, 119.
8. J.E. Mark in Physical Properties of Polymers, Eds., J.E. Mark, A. Eisenberg,

W.W. Graessley, L. Mandelkern, E.T. Samulski, J.L. Koenig and G.D. Wignall,
ACS, Washington, 2nd Edition, 1993, 1-59.
9. N. Schuld and B.A. Wolf in Polymer Handbook, Eds., J. Brandrup, E.H.

Immergut, E.A. Grulke, A. Abe and D.R. Bloch, Wiley, New York, 1999,
Chapter 7.
10. G.B. McKenna, K.M. Flynn and Y. Chen, Polymer Communications, 1988, 29,
272.
11. G.B. McKenna, K.M. Flynn and Y. Chen, Polymer, 1990, 31, 1937.
12. E.A. Kearsley and L.J. Zapas, Journal of Rheology, 1980, 24, 483.
13. Z. Hrnjak-Murgic, J. Jelencic, M. Bravar and M. Marovic, Journal of Applied

14. S.P. Malone, C. Vosburgh and C. Cohen, Polymer, 1993, 34, 5149.
15. K. Schmidt-Rohr and H.W. Spiess, Multidimensional Solid-state NMR and

Polymers, Academic Press, London, 1994.
512
16. L.E. Nielsen, Journal of Macromolecular Science – Reviews in Macromolecular

Chemistry, 1969, C3, 69.
18. D.J.P. Harrison, W.R. Yates and J.F. Johnson, Journal of Macromolecular Science
C, 1985, 25, 481.
19. D.R. Bauer, Progress in Organic Coatings, 1986, 12, 155.
20. R.A. Kinsey, Rubber Chemistry and Technology, 1990, 63, 407.
21. A.K. Whittaker, Annual Reports on NMR Spectroscopy, 1997, 34, 105.
22. M. Mori and J.L. Koenig, Annual Reports on NMR Spectroscopy, 1997, 34, 231.
23. N. Bloembergen, E.M. Purcell and R.V. Pound, Physics Review, 1948, 73, 679.
24. R. Dejean de la Batie, F. Laupretre and L. Monnerie, Macromolecules, 1988, 21,

2045.
25. C.K. Hall and E. Helfand, Journal of Chemical Physics, 1982, 77, 3275.

2052.

122.
28. F. Laupretre, L. Bokobza and L. Monnerie, Polymer, 1993, 34, 468.
29. F. Laupretre and L. Monnerie, Macromolecules, 1999, 32, 3017.
30. S. Glowinkowski, D.J. Gisser and M.D. Ediger, Macromolecules, 1990, 23, 3520.
31. D.B. Adolf and M.D. Ediger, Macromolecules, 1991, 24, 5834.
32. M.A. Krajewski-Bertrand and F. Laupretre, Macromolecules, 1996, 29, 7616.
33. C. Baysal, B. Erman, I. Bahar, F. Laupretre and L. Monnerie, Macromolecules,

1997, 30, 2058.
34. E.O. Stejskal and J.E. Tanner, Journal of Chemical Physics, 1965, 42, 288.
513
35. P.T. Callaghan, Principles of Nuclear Magnetic Resonance Microscopy, Oxford

University Press, Oxford, 1991.
36. B.D. Boss, E.O. Stejskal and J.D. Ferry, Journal of Physical Chemistry, 1967, 71,
1501.
37. A. Guillermo, M. Todica and J.P. Cohen-Addad, Macromolecules, 1993, 26,

3946.
38. A. Banis, P.T. Inglefield, A.A. Jones and W.Y. Wen, Journal of Polymer Science:
41. S. Schlick, Z. Gao, S. Matsukawa, I. Ando, E. Fead and G. Rossi,

42. A.I. Maklakov, V.A. Sevryugin, V.D. Skirda and N.F. Fatkullin, Polymer Science
USSR, 1984, 26, 2804.
43. V.A. Sevryugin, V.D. Skirda and A.I. Maklakov, Polymer, 1986, 27, 290.
44. V.D. Skirda, N.F. Fatkullin, V.I. Sundukov and A.I. Maklakov, Polymer Science
USSR, 1987, 29, 2229.
45. V.D. Skirda, V.I. Sundukov, A.I. Maklakov, O.E. Zgadzai, I.R. Gafurov and G.I.
Vasiljev, Polymer, 1988, 29, 1294.
46. V.D. Skirda, M.M. Doroginitskii, V.I. Sundukov, A.I. Maklakov, G. Fleischer,
K.G. Häusler and E. Straube, Makromoleculare Chemie, Rapid Communications,
1988, 9, 603.
47. B. Blümich and P. Blümler, Makromoleculare Chemie, 1993, 194, 2133.
48. J.F. Koenig, Macromolecular Symposia, 1994, 86, 283.
49. P. Jezzard, J.J. Attard, T.A. Carpenter and L.D. Hall, Progress in NMR
Spectroscopy, 1991, 23, 1.
514
50. P. Jezzard, C.J. Wiggins, T.A. Carpenter, L.D. Hall, P. Jackson, N.J. Clayden and
N.J. Walton, Advanced Materials, 1992, 4, 82.
51. R.A. Komoroski, Analytical Chemistry, 1993, 65, 1068.
52. L.F. Gladden, Chemical Engineering and Science, 1994, 49, 3339.
53. W. Kuhn, Die Angewandte Chemie, 1990, 29, 1.
54. Magnetic Resonance Microscopy, Methods and Applications in Materials Science,

Agriculture and Biomedicine, Eds., B. Blümich and W. Kuhn, VCH Publishers,
Weinheim, 1992.
55. M.R. Halse, H.J. Rahmann and J.H. Strange, Physica B, 1994, 203, 169.
56. A.G. Webb and L.D. Hall, Polymer Communications, 1990, 31, 422.
57. A.G. Webb and L.D. Hall, Polymer Communications, 1990, 31, 425.
58. A.G. Webb and L.D. Hall, Polymer, 1991, 32, 2926.
59. S.R. Smith and J.L. Koenig, Macromolecules, 1991, 24, 3496.
60. M.A. Rana and J.L. Koenig, Macromolecules, 1994, 27, 3727.
61. M. Mori and J.L. Koenig, Journal of Applied Polymer Science, 1998, 70, 1385.
62. R.S. Clough and J.L. Koenig, Journal of Polymer Science: Polymer Letters
Edition, 1989, 27, 451.
63. S.J. Oh and J.L. Koenig, Polymer, 1999, 40, 4703.
64. R. Folland and A. Charlesby, Radiation Physics and Chemistry, 1977, 10, 61.
65. R. Folland and A. Charlesby, Polymer, 1979, 20, 211.
66. R. Folland and A. Charlesby, Polymer, 1979 20, 207.
67. A. Charlesby, P. Käfer and R. Folland, Radiation Physics and Chemistry, 1978,
11, 83.
68. A. Charlesby and B.J. Bridges, European Polymer Journal, 1981, 17, 645.
69. G.C. Munie, J. Jonas and T.J. Rowland, Journal of Polymer Science: Polymer
515
70. D.R. Brown, G.C. Munie and J. Jonas, Journal of Polymer Science: Polymer
Physics Edition, 1982, 20, 1659.
71. G. Simon, H. Schneider and K-G. Häusler, Progress in Colloid and Polymer
Science, 1988, 78, 30.
72. G. Simon, A. Birnsiel and K-H. Schimmel, Polymer Bulletin, 1989, 21, 235.
73. G. Simon, B. Götschmann, D. Matzen and H. Schneider, Polymer Bulletin, 1989,

21, 475.
74. G. Simon and H. Schneider, Die Makromoleculare Chemie, Macromolecular

Symposia, 1991, 52, 233.
76. C.G. Fry and A.C. Lind, Macromolecules, 1988, 21, 1292.
77. D.S. Bradley, E.D. von MeerWall, G.D. Roberts and J. Kamvouris, Journal of
Polymer Science: Polymer Physics Edition, 1995, 33, 1545.
78. I.I. Nazarova, Yu.A. Ol’Khov and S.M. Baturin, Polymer Science USSR, 1980,
22, 433.
79. I.I. Nazarova, V.B. Nazarov and S.M. Baturin, Polymer Science USSR, 1982, 24,
1967.
80. T.G. Neiss and E.J. Vanderheiden, Macromolecular Symposia, 1994, 86, 117.
81. N. Parizel, G. Meyer and G. Weill, Polymer, 1993, 34, 2495.
82. J.P. Cohen-Addad and C. Schmidt, Journal of Polymer Science: Polymer Physics
Edition, 1987, 25, 487.
83. J.P. Cohen-Addad, E. Soyez, A. Vaillat and J.P. Queslel, Macromolecules, 1992,
25, 1259.
84. K. Fukumori, T. Kurauchi and O. Kamigaito, Polymer, 1990, 31, 713.
85. Yu. Ya. Gotlib, M.I. Lifshits, V.A. Shevelev, I.S. Lishanskii and I.V. Balanina,
Polymer Science USSR, 1976, 18, 2630.
86. W. Kuhn and F. Grün, Kolloid Zeitschrift, 1942, 101, 248.
516
87. V.M. Litvinov, W. Barendswaard and M. Van Duin, Rubber Chemistry and
Technology, 1998, 71, 105.
89. W. Barendswaard, V.M. Litvinov, F. Souren, R.L. Scherrenberg, C. Gondard and

C. Colemonts, Macromolecules, 1999, 32, 167.
90. H. Menge, S. Hotopf, S. Ponitzsch, S. Richter, K.F. Arndt, H. Schneider and U.

Heuert, Polymer, 1999, 40, 5303.
91. C. Schmit and J.P. Cohen-Addad, Macromolecules, 1989, 22, 142.
92. J.P. Cohen-Addad, Journal of Physics, 1982, 43, 1509.
93. J.P. Cohen-Addad and R. Dupeyre, Polymer, 1983, 24, 400.
94. J.P. Cohen-Addad, Polymer, 1983, 24, 1128.
95. J.P. Cohen-Addad and J. Guillermo, Journal of Polymer Science: Polymer Physics
Edition, 1984, 22, 931.
96. M.G. Brereton, Macromolecules, 1990, 23, 1119.
97. M.G. Brereton, Journal of Chemical Physics, 1991, 94, 2136.
99. M.E. Reis, M.G. Brereton, P.G. Klein and P. Dounis, Polymer Gels and
Networks, 1997, 5, 285.
234.
102. P.S. Brown and A.J. Tinker, Kautschuk und Gummi Kunststoffe, 1995, 48, 606.
103. M.D. Ellul, J. Patel and A.J. Tinker, Rubber Chemistry and Technology, 1995,
68, 573.
1992, 65, 744.
517
105. S.A. Groves and A.J. Tinker, Journal of Natural Rubber Research, 1996, 11, 125.
109. S. Cook, Kautschuk und Gummi Kunststoffe, 1999, 52, 350.
110. J.H. O’Donnell and A.K. Whittaker, Polymer, 1992, 33, 62.
111. S. Mohanraj and W.T. Ford, Macromolecules, 1985, 18, 351.
112. J.C. Randall, F.J.Zoepfl and J. Silverman, Makromolekulare Chemie, Rapid

Communications, 1983, 4, 149.
113. F. Horii, Q. Zhu, R. Kitamaru and H. Yamaoka, Macromolecules, 1990, 23, 977.
114. Q. Zhu, F. Horii, R. Kitamaru and H. Yamaoka, Journal of Polymer Science:

Polymer Chemistry Edition, 1990, 28, 2741.
115. A.D. Bain, D.R. Eaton, A.E. Hamielec, M. Mlekuz and B.G. Sayer,
116. G.A. Morris and R. Freeman, Journal of Magnetic Resonance, 1978, 29, 433.
117. H.D.H. Stover and J.M.J. Frechet, Macromolecules, 1991, 24, 883.
118. L.A. Errede, R.A. Newmark and J.R. Hill, Macromolecules, 1986, 19, 651.
119. A.D. English, Macromolecules, 1985, 18, 178.
120. D. Doskocilova and B. Schneider, Pure and Applied Chemistry, 1982, 54, 575.
121. D. Doskocilova, J. Straka and B. Schneider, Polymer, 1993, 34, 437.
518
14
Multidimensional NMR Techniques for the
Characterisation of Viscoelastic Materials
Dan E. Demco, Siegfried Hafner and Hans W. Spiess
14.1 Introduction
An important objective in materials science is the establishment of relationships between

the microscopic structure or molecular dynamics and the resulting macroscopic properties.
Once established, this knowledge then allows the design of improved materials. Thus,
the availability of powerful analytical tools such as nuclear magnetic resonance (NMR)
spectroscopy [1-6] is one of the key issues in polymer science. Its unique chemical selectivity
and high flexibility allows one to study structure, chain conformation and molecular
dynamics in much detail and depth. NMR in its different variants provides information
from the molecular to the macroscopic length scale and on molecular motions from the
1 Hz to 1010 Hz. It can be applied to crystalline as well as to amorphous samples which
is of particular importance for the study of polymers. Moreover, NMR can be conveniently
applied to polymers since they contain predominantly nuclei that are NMR sensitive
such as 1H and 13C.
While well-established for liquid-like samples such as macromolecules in solutions [7], the
applications of NMR to solid or solid-like polymers is more demanding because of the
presence of anisotropic interactions that complicate the analysis of the results. Several
techniques for the removal of these interactions have thus been developed and are nowadays
in a state where they can be routinely applied. Non-averaged anisotropic interactions on
the other hand provide valuable information that is lost in the solution state. Thus, while
it is often necessary to remove the anisotropic interactions, in many cases one would
simultaneously like to preserve them in order to exploit their information content.
This is where two-dimensional (2D) spectroscopy comes into play, for example, by
correlating one dimension where the anisotropic interactions are preserved with a (high-
resolution) dimension where they are removed. That is, both objectives can be achieved
within one two-dimensional experiment. Like conventional one-dimensional (1D) NMR
techniques, 2D techniques were first applied to liquid or solutions, where they provided
invaluable information for the structure assignments in biomolecules [7]. Later, after the
519
experimental difficulties related to the NMR of solids have been overcome, they were
also applied to solid polymers and elastomers [1-6].
From the viewpoint of NMR, elastomers and other viscoelastic polymers above their
glass-transition temperature (Tg) exhibit both, solid-like and liquid-like features. Whereas
the segmental motions give rise to the liquid-like behaviour, the presence of permanent
or non-permanent crosslinks leads to residual dipolar couplings, that are responsible for
the solid-like properties [8]. While this promises that both properties can be exploited,
the application of 2D techniques to viscoelastic materials has to deal with the difficulties
related to both, rigid and mobile samples. Compared with the wealth of applications in
solution and in solid state, it has not been widely applied to viscoelastic polymers, although
it can provide information on such important fields as the chain dynamics in elastomers,
the local structure, residual couplings (induced by chemical crosslinks and topological
constraints), dynamic order parameters, inter-nuclear distances, intermolecular
interactions (which are important, for example, for the miscibility), the effects of fillers
on molecular motions, segmental orientation under mechanical stress and others.
After a brief introduction of the basic tools of NMR in Section 14.2, the 2D techniques
that have been already applied to rubbers or viscoelastic materials (Section 14.3) will be
reviewed. After briefly introducing each of the techniques, a more detailed overview of
the applications and a discussion of some of the highlights will be given. This structure,
where all information on a given method is presented within one section enables the
interested reader to decide more easily which of the techniques might be most useful to
them. NMR imaging which can be considered as a special form of 2D NMR will not be
discussed but the interested reader is referred to the corresponding chapter in this book.
14.2 Basics of NMR in Viscoelastic Polymers
14.2.1 Anisotropic Spin Interactions
The information that can be extracted from solid-state NMR spectra is encoded via spin
interactions such as the chemical shielding, the quadrupolar interaction and the homo-
and hetero-nuclear dipolar interactions [1,9-10]. Some knowledge of the spin interactions
that determine the features of the spectra are thus of prime importance.
A common characteristic of the relevant spin interactions is that they are anisotropic and
can be described by second-rank tensors. The resulting orientation-dependent NMR
frequency is of the following form [1,9]:
520
Multidimensional NMR Techniques for the Characterisation of Viscoelastic Materials
δ
ω(α, β) = ω L + (3 cos2 β − 1 − η sin2 β cos 2α) (14.1)
2
where ωL is the Larmor frequency at the isotropic chemical shift value and the other terms
reflect the deviations due to angular-dependent contributions. δ represents the strength of
the anisotropic interactions, η (0 ≤ η ≤ 1) is the asymmetry parameter which describes the
deviation from axial symmetry and α, β are the polar angles that relate the orientation of
the principle-axes system of the interaction tensor with the main magnetic field.
The most important interaction for chemical information is the chemical shift. It results
from the shielding of the magnetic field at the position of the nucleus by the electrons.
The Larmor frequency is thus shifted by an isotropic contribution ωL and by angular
dependent contributions (polar angles α and β). Assuming an equal probability for all
directions one can calculate the powder average and a broad powder spectrum is obtained
which reflects the chemical-shift anisotropy. For the case of an axially symmetric chemical-
shift tensor, η is zero and the angular-dependent term in Equation 14.1 simplifies to
δ (3cos2β-1)/2. Usually the chemical shift spans about 10 ppm for 1H and 200 ppm for
13
C nuclei.
For abundant nuclei with spin 1/2, the spectrum is often dominated by heteronuclear or
homonuclear dipolar interactions, i.e., the interactions between the magnetic moments
of two neighbouring spins. In this case there is no isotropic contribution and η is zero, so
that Equation 14.1 simplifies correspondingly. For a two-spin system one obtains a spin
Hamiltonian of the form:
⎛ μ ⎞ γ γ h ⎛ 3 cos2 θ − 1⎞ rr
H D = ⎜ 0 ⎟ 1 32 ⎜ ⎟ (3I1z I2z − I1 I2 ) (14.2)
⎝ 4 π ⎠ r12 ⎝ 2 ⎠
where r12 is the magnitude of the r vector connecting the two spins, θ is the angle of this
vector to the magnetic field, the Ii are spin operators and the γi (i=1,2) are the magnetogyric
ratios of the spins. That is, the strength of the resulting line splitting depends strongly on
the distance between the two spins, so that information about distance can be extracted
from such spectra. Homonuclear interaction (equivalent spins with γ1 = γ2 = γ) and
heteronuclear interaction (non-equivalent spins with γ1≠γ2) have r r to be distinguished. In
the latter case, the flip-flop term which is part of the product I1I2 in Equation 14.2 can
be neglected. For a powder sample one has again to take into account all angles β and
thus obtains the so-called Pake spectrum [1,9] with a considerable anisotropic line-
broadening of up to 50 kHz for homonuclear and up to 25 kHz for heteronuclear dipolar
interaction. Since the dipolar coupling is a through-space interaction, however, in principle
521
the sum has to be evaluated over all possible pair interactions. This, and the presence of
molecular motion, leads to considerable complications and is responsible for the
experimental finding that in practice not a Pake spectrum but a relatively structureless
line-shape is obtained.
In case of a deuterated sample (spin 1 case), the spectra are usually dominated by the
quadrupolar interaction, that is, the coupling of the nuclear quadrupole moment with
the electric field gradient of the C-2H bond. For deuterons in C-2H bonds this can lead to
a splitting of about 250 kHz. As in the case of dipolar interaction, a Pake spectrum is
obtained for a powder sample. The z-principal axis of the quadrupolar interaction is
oriented along the bond axis which makes deuteron NMR particularly useful for studies
of segmental orientations and molecular dynamics (reorientation) [1].
In sufficiently mobile, (i.e., liquid-like), systems, the anisotropy is averaged out by the
isotropic thermal motions leaving only the isotropic contributions. As already stated,
viscoelastic polymers represent an intermediate between the two extremes of rigid or
mobile materials and the implications of this will be discussed in 14.2.3 in more detail.
14.2.2 Manipulation of Spin Interactions
The rich information content of solid-state spectra makes them difficult to evaluate, in
particular if more than one of the interactions discussed previously has to be taken into
account. The evaluation of the spectra would often be impossible but NMR methodology
provides the possibility to decouple and recouple spin interactions nearly as desired (see
[11] for a comprehensive introduction). Moreover, these different information sources
can be separated and correlated using the multidimensional techniques discussed
subsequently.
The most prominent example of a technique for decoupling or line-narrowing is magic

angle spinning (MAS) (see also [1,9,11]). Here, the angular dependent part of the
interactions is modulated by rapidly spinning the sample around an axis inclined at an
angle Θ to the magnetic field. If the spinning axes is chosen along the so-called magic-
angle Θm=54.7°, the relevant scaling factor (3cos2Θ -1)/2 becomes zero and the anisotropic
part of the interaction vanishes. An example using an elastomer is shown in Figure 14.1.
While the static spectrum is broadened by (already motionally averaged) dipolar couplings
(Figure 14.1a), these broadening effects are removed in the MAS spectrum (Figure 14.1b).
Often it is necessary to manipulate the spin interactions also by pulse techniques [1,9,11].
These act on the spin operators in the corresponding Hamiltonians (see for example
Equation 14.2) rather than on the geometric part. Depending on the applied pulse
522
Figure 14.1 1D spectra of a typical rubber, styrene-butadiene-rubber (SBR). a) Static

1
H spectrum acquired at a Larmor frequency of 500 MHz. The dipolar coupling is
motionally averaged and different lines can be distinguished although they are still
broadened by the residual dipolar couplings. b) MAS spectrum of the same sample at
a MAS spinning frequency of 15 kHz. The line-broadening due to anisotropic spin
interactions, e.g., residual dipolar couplings, is removed
Figure 14.1b reproduced with permission from R. Graf, D.E. Demco, S. Hafner and H.W.
Spiess, Solid State Nuclear Magnetic Resonance, 1998, 12, 139, Figure 2b. Copyright
1998, Academic Press
523
sequence, a given spin interaction can be switched on and off in order to discriminate
between the different contributions to the desired information. For instance, 13C chemical-
shift values can be determined by selectively irradiating the protons. This so-called
heteronuclear dipolar decoupling removes any influence of the coupling to the protons
from the 13C spectrum. Using long radio-frequency pulses on both sides with well-matched
amplitudes, the flip-flop part of the heteronuclear dipolar interaction can be preserved.
This can be used for the transfer of the proton polarisation to the carbon side (cross
polarisation) [1,9] which is often used to artificially increase the polarisation of the dilute
spins to improve the signal-to-noise ratio. Sometimes it is also applied to edit those of
the dilute spins that are closely coupled to abundant spins.
Undesired homonuclear spin interactions can be also suppressed using suitable multiple-
pulse sequences while still exploiting the information content provided by interactions
that are not affected. Using a combination of MAS and pulse decoupling it is even possible
to reintroduce parts of an interaction that would be averaged out by one of the
manipulation techniques alone (‘recoupling’) [11]. This high flexibility of solid-state NMR
enables one to fully exploit the rich information content provided by the spin interactions.
It becomes particularly powerful if such experiments are combined to multidimensional
NMR techniques as discussed in Section 14.3.
14.2.3 Residual Couplings and Dynamic Order Parameters
In polymers and in particular in elastomers, molecular dynamics plays an important role

and largely determines the mechanical properties. An investigation of segmental motions
is thus of prime importance and NMR is a particularly suitable technique for this purpose.
The measurement of dynamics by NMR takes place via the influence of motions on the
anisotropic spin interactions. Modulation of the spin interactions by molecular motion
is responsible for the relaxation of the spin system back to the equilibrium state. This so-
called longitudinal (T1) relaxation has to be distinguished from transverse (or T2)
relaxation that describes the dephasing of coherences under influence of residual, (i.e.,
partially averaged), spin interactions. Both are thus sensitive probes for molecular motions
in (viscoelastic) polymers as will be briefly discussed in Section 14.2.4. If the dynamics of
individual groups are of interest, the motions can be followed in great detail by two-
dimensional techniques on the time scale of tens of milliseconds [1]. Such investigations
have proved to be extremely valuable for the investigation of polymers around the Tg.
At temperatures of about 50 k above Tg, which is the more interesting regime for rubbery
materials, the motions are rapid enough (tens of kHz) to effectively average the spin
interactions. However, they are not fully anisotropic in high-molecular-mass polymers
524
and leave residual dipolar couplings as a consequence of the anisotropy [8] (see also
Figure 14.1a). Restrictions of the chain motions result on a larger length scale from
entanglements and crosslinks and on a shorter length scale from local chain order due to
the presence of local conformations or stiffening structural elements. For characterising
these effects, an autocorrelation function can be defined which is of the form S(t)
= < (3cos 2 θ (t) - 1) > / 2. Here, θ(τ) is the instantaneous angle relating a vector fixed to
a structural element and <...> represents the average over segmental orientation and
time. This function will not immediately decay to zero but to a certain level which can be
considered as a dynamic order parameter of segmental motion (see Figure 14.2). On a
longer time scale isotropisation can occur if no permanent crosslinks are present and S -> 0.
Since this plateau represents the interesting regime for characterising viscoelastic properties
of a polymer, we will in the following concentrate on multidimensional NMR techniques
for measuring site-selectivity the order parameter or, correspondingly, residual couplings.
First however, a brief overview on 1D NMR investigations serving a similar purpose will
Figure 14.2 Scheme of a two-step decay of the correlation function S of a unit

undergoing anisotropic motion before finally full isotropisation is achieved. The
dashed line indicates the case of a one-step decay by isotropic motion. For the study of
viscoelastic polymers the intermediate plateau that reflects residual couplings is the
most interesting part
525
be given, since they form the basis for the more involved multidimensional NMR studies.
A more detailed review for the particular case of crosslinked elastomers can be found in
reference [12].
14.2.4 One-dimensional NMR Studies of Molecular Motions and Dynamic

Order
As already stated, molecular dynamics in polymers have to be probed on various time

scales from relatively fast segmental motions (with correlation times of the order of 10-10 s)
down to slow motions in the 10-3 s range or even slower. Dynamics in the range of around
10-10 s can be investigated by T1 measurements which are sensitive to motions around the
Larmor frequency. Particularly useful for such investigations is the field cycling technique,
where the magnetic field and thus also the Larmor frequency ω can be varied over several
orders of magnitudes down to the 10 kHz regime. By acquiring T1(ω), molecular dynamics
can be probed over the corresponding range [10] (and references therein).
Slow motions can be studied using so-called longitudinal relaxation in the rotating frame
(or T1ρ). In these experiments a radio frequency field (‘spin lock pulse’) with field strength
of around 10-100 kHz plays the same role as the (much stronger) main magnetic field
does for T1 [10]. For example, the changes in the phase structure under mechanical
deformation of thermoplastic elastomers have been investigated by T1ρ measurements of
13
C nuclei under MAS [13].
A technique for the study of ultra slow molecular motions in viscoelastic materials is
based on the measurement of the stimulated and primary echoes of a three-pulse
experiment [10]. The ratio of the echo amplitudes depends on the characteristic correlation
times, the second moment of the dipolar fluctuations and the average dipolar constant
which is proportional to the order parameter of the system [14,15]. This so-called ‘dipolar-
correlation effect’ (DCE) was used for the study of ultra slow molecular dynamics and
segmental order in polymer melts and networks [16,17]. The dipolar correlations in SBR
networks were found to exist on a time scale exceeding 300 ms [16] and the mean squared
fluctuations of the dipolar coupling constant depends on the number of segments between
crosslinks N, according to N-0.75. This was interpreted by assuming that the meshes of
the network are more stretched than expected for Gaussian random coils [17].
Transverse relaxation (T2) and line shape analysis has also been the topic of many studies
of molecular dynamics in viscoelastic media [12]. It was found to be indicative of
526
heterogeneities of molecular motions [8,18]. Transverse relaxation curves therefore often

turn out to be composed of a superposition of multiexponential and Gaussian decays
[8,19]. Measurement of residual dipolar couplings, molecular motions and other
parameters have been performed by this relaxation technique on different nuclei like 1H,
13
C, 15N, 31P and 29Si and could be related to material properties such as the crosslink
density [8,19]. In order to separate the liquid-like and solid-like contributions of the T2
decay, a linear combination of Hahn echoes and solid echoes have been used [20,21].
While Hahn spin-echo amplitudes are encoded by irreversible dephasing due to molecular
motions and by reversible residual dipolar couplings, only the first effect encodes the
amplitudes of solid echo (actually mixed echo) to first approximation [8,18]. The
appropriate combination of these spin echoes thus gives access to residual dipolar
couplings and the fluctuation rate [20,21]. The method was applied to measure the degree
of segmental order induced in natural rubber upon stretching [20].
A powerful technique for the study of orientation and dynamics in viscoelastic media is
line shape analysis in deuteron NMR spectroscopy [1]. For example, the average
orientation of chain segments in elastomer networks upon macroscopic strain can be
determined by this technique [22-31]. For a non-deformed rubber, a single resonance
line in the deuterium NMR spectrum is observed [26] while the spectrum splits into a
well-defined doublet structure under uniaxial deformation. It was shown that the usual
network constraint on the end-to-end vector determines the deuterium line shape under
deformation, while the interchain (excluded volume) interactions lead to splitting [26-
31]. Deuterium NMR is thus able to monitor the average segmental orientation due to
the crosslinks and mean field separately [31].
The network structure of unfilled and filled elastomers was probed by the quadrupolar
splitting in 2H solid-echo spectra of uniaxially strained samples [23,32] (see Figure 14.3).
The local chain order at a given elongation is larger by a factor of 1.5-2 in the filled
system. A decrease of local chain mobility in the absorption layer is observed under
stress. The same method was applied to investigate molecular dynamic in thermoplastic
elastomer based on hydrogen bonding complexes [33,34].
The 1D NMR techniques described above already allow a detailed investigation of polymer
dynamics but they are mostly not selective in the sense that they do not provide information
on the averaging of particular couplings. For the interpretation of some of the above
results therefore a representative spin-pair along the chain was assumed, thus neglecting
local site-specific motions as well as the geometry of the bonds. This is where
multidimensional NMR techniques come into play.
527
Figure 14.3 Deuterium solid-echo spectra of unfilled (a) and filled (b)
poly[dimethylsiloxane] networks at 305 K with and without mechanical stress as given
by the parameter λ [23]. This example demonstrates the sensitivity of the NMR
lineshape and thus of the spin interactions to internal and external conditions
Reproduced with permission from V.M. Litvinov and H.W. Spiess, Die Makromolekulare
Chemie, 1992, 193, 1181, Figure 8. Copyright 1992, Wiley-VCH
Figure 14.4 (a) Principle of 2D NMR spectroscopy. The coherences are excited by a
preparation sequence and evolve for a time t1. In the following mixing time the system
is allowed to undergo changes after which the resulting signal is read out (direct
detection). By incrementing the evolution time t1 in successive experiments, a 2D data
set is produced. Fourier transformation provides the 2D NMR spectrum. (b) Simple
example for the general scheme in form of a three-pulse sequence. The delays t1, t2 and
tm correspond to the evolution, detection and mixing periods, respectively
528
14.3 Multidimensional NMR Spectroscopy of Viscoelastic Materials
14.3.1 Principle of Multidimensional NMR
In 2D NMR spectroscopy, a two-dimensional data set is acquired as a function of two time

variables t1 and t2 as shown schematically in Figure 14.4 [1, 7]. Figure 14.4a shows the
general case while the three pulse sequence of Figure 14.4b represents a typical example.
A 2D NMR experiment usually is divided in several time periods that follow each other.
The first is the so-called preparation period in which coherences are excited by a suitable
pulse sequence, which in the simplest case is only one radio-frequency pulse (see
Figure 14.4b). Unlike 1D NMR spectroscopy, the excited signal is not directly acquired
but is allowed to evolve in the so-called evolution period under influence of the relevant
spin interactions. The evolution time t1 is incremented in subsequent experiments and
provides the first time dimension of the experiment. After the evolution period (and an
optional mixing time), the remaining signal is directly detected in the detection period
for each time increment thus generating a 2D data set. Two-dimensional Fourier
transformation then gives the 2D spectrum. Optionally a so-called mixing period of
length tm can be inserted between the evolution and detection periods. During tm changes
in the system can occur, for instance, by molecular motions, the action of spin interactions
or by spin manipulation (for instance cross-polarisation to another nucleus).
The different aspects of 2D NMR spectroscopy are reflected in the different variants that
can be distinguished. One variant, separation spectroscopy, is used to separate different
interactions taking advantage of the spin manipulation techniques [1,7]. For instance,
during the evolution period the spin manipulation can be made such that only the isotropic
chemical shift is acquired while in the detection period the full spectrum is acquired.
This leaves the (motionally averaged) anisotropy to be studied site-selectively. Other 2D
NMR techniques, so-called correlation techniques [1,7], aim at obtaining new information
by correlating different interactions. For example, the 13C chemical shift anisotropy can
be correlated with the heteronuclear dipolar powder pattern in order to obtain information
on the relative orientation of the two tensors [7]. Considering the manifold of spin
manipulation techniques there is a wealth of such 2D NMR techniques that can be derived
for different purposes. Finally, introducing a mixing time tm, 2D exchange spectroscopy
can be performed. The most important application of such exchange techniques with
respect to polymer investigations is the study of slow molecular dynamics [1]. In these
experiments, reorientations due to molecular dynamics are allowed to take place during
the mixing time tm and lead to characteristic off-diagonal patterns in the resulting 2D
spectra. If the mixing time is increased in a series of 2D experiments, slow dynamics in
the range of milliseconds to seconds can be investigated in detail [1].
529
14.3.2 Two-dimensional NMR Techniques and Applications to Viscoelastic

Polymers
The investigation of molecular dynamics and local segmental orientation by 2D NMR
spectroscopy has a number of advantages [1,7] particularly when dealing with complicated
spin systems such as viscoelastic polymers. Next, examples of such investigations will be
discussed, concentrating on techniques that are particularly sensitive to the viscoelastic
regime (see Figure 14.2), that is, NMR techniques for measuring residual dipolar couplings
and dynamic order parameters. Readers interested in a more general overview of
multidimensional NMR of polymers are referred to [1].
14.3.2.1 Dynamic Order Parameters from 2D 1H Magnetisation Exchange

Spectroscopy
Two-dimensional 1H magnetisation-exchange spectroscopy and its reduced one-dimensional

variant have been used to probe residual dipole-dipole couplings between different functional
groups of an SBR sample, namely between the CH- and CH2-groups of the butadiene units
[35,36]. The intergroup residual dipolar couplings can then be correlated with the shear
modulus which is an independent measure of the crosslink density [36]. Hence, the dynamic
local-order parameter associated with the partially averaged dipolar coupling can be
evaluated taking the average over the relevant dynamic processes.
• Principle of the method
The basic radio frequency pulse sequence is depicted in Figure 14.4.b. Following the
general scheme of 2D NMR spectroscopy, an evolution period of duration t1 follows the
excitation pulse. This interval ends with a flip-back pulse and the system is allowed to
exchange during the mixing period tm. After a third pulse the remaining signal is acquired
during the detection interval t2 [36]. The intensity of the crosspeaks in the resulting 2D
spectra then reflects the degree of exchange between the corresponding groups. Since full
2D NMR spectroscopy is time consuming, for quantitative evaluations a reduced one-
dimensional form has also been used. In this case the time t1 is fixed and adjusted such
that it acts as a chemical-shift filter [1] for one of the two butadiene lines in the static
SBR spectrum. The reappearance of the filtered line with increasing mixing time is then
a measure of the magnetisation exchange between the protons of the CH and CH2 groups.
Note, that unlike the 500 MHz spectrum in Figure 14.1, the styrene line and the CH line
could not be distinguished in these experiments so that only the CH and CH2 lines of the
butadiene units are resolved. They are separated by about 3.35 ppm, corresponding to
1 kHz at the Larmor frequency of 300 MHz.
530
The theory needed for the evaluation of the exchange process was also developed [36].
In the limit of short mixing times tm, the initial decay rates for the CH line were found to
depend only on the intergroup residual dipolar coupling D CH − CH 2 according to
I z (t m )
1
(D )
2
CH − CH 2
= 1− t 2m (14.3)
2
where I z ( t m ) is the observable NMR for the CH protons (for the CH2 protons the
factor 1/2 must be substituted by 1/4 in Equation 14.3). The symbol (...) represents the
statistical ensemble average of the space part of the dipolar coupling in a disordered
elastomer. Note, that from such measurements, the intergroup dipolar coupling D CH − CH 2
can be determined which is largely parallel to the segment axis. That is, it is much more
informative for the detection of chain order and chain motion than intragroup couplings
would be.
For short mixing times, the measurable exchange of longitudinal proton magnetisation
is predominantly determined by the protons belonging to nearest-neighbour CH- and
CH2-groups. At longer mixing times on the other hand, a much larger spin system along
the polymer chain must be considered, in principle. In reference [37], however, it was
shown that even in this case the intergroup residual dipolar coupling can be measured
from the initial-time regime of the magnetisation-exchange decay curve.
• Two-dimensional 1H magnetisation-exchange processes in crosslinked SBR
From the point of view of molecular motion, crosslinked elastomers are highly
heterogeneous systems and a solid-like behaviour is present together with slow and fast
motions of both, chain segments and functional groups. Hence, the magnetisation-
exchange process is also expected to be heterogeneous, that is, driven by different
mechanisms.
To investigate this heterogeneity, a series of static 2D NMR experiments were performed

on SBR by increasing mixing time [36]. As already discussed previously, the aromatic
and the olefinic protons are not resolved under the applied experimental conditions as
can be seen in the 2D spectra shown in Figure 14.5.
A remarkable feature of the diagonal 2D spectrum for short mixing times is, that the
width of the diagonal peaks perpendicular to the (ω1,ω2)-diagonal is much smaller than
the widths along the diagonal. This indicates, that the 1 H NMR spectrum is
heterogeneously broadened [36]. The width of each line along the diagonal reflects both,
531
Figure 14.5 Two-dimensional 1H magnetisation-exchange spectra of a SBR sample

recorded with the three-pulse sequence of Figure 14.4b. Different mixing times have
been used: tm = 2.5 ms, (a) 5 ms, (b) 15 ms, (c) and 30 ms (d). The 2D surface
representation for tm=30 ms in (e) shows that all cross-peaks are positive [36]
Reproduced with permission from D.E. Demco, S. Hafner, C. Fülber, R. Graf and H.W.
Spiess, Journal of Chemical Physics, 1996, 105, 11285, Figure 4. Copyright 1996,
American Institute of Physics
532
the solid- and the liquid-like contributions, whereas the width perpendicular to the
diagonal reflects selectively the liquid-like linewidth only. This feature is very pronounced
for the CH diagonal peak and less for the CH2 diagonal peak because of the strong
proton-proton coupling within the methylene group. A similar result was already found
in a 13C magnetisation-exchange experiment on natural rubber aimed to detect chain
diffusion [38]. While no chain diffusion was detected, the lines were also found to be
spread out along the diagonal.
In the 1H magnetisation-exchange experiment on SBR not much cross-peak signal is

found for mixing times shorter than 2.5 ms (see Figure 14.5a). This indicates, that within
this time scale, the CH and CH2 functional groups can be considered in reasonable
approximation to be isolated. The inter-group residual dipolar coupling leads to
magnetisation exchange for longer mixing times. Intense positive cross-peaks are thus
present for tm = 30 ms (see Figures 14.5d and 14.5e). Such positive cross-peaks can arise
for two reasons: from the exchange process mediated by residual dipolar couplings and
from cross-relaxation induced by segmental motions with correlation times τc longer
than ω0-1 [7] (see also Section 14.3.2.2).
• Proton intergroup residual dipolar couplings
The residual dipolar couplings between the protons of the CH group and the CH2 group
were determined from the tm-dependence of the peak intensities. The decay of the
longitudinal magnetisation M(tm) of the diagonal signals (normalised to the value M0
that corresponds to zero mixing time) is recorded for short mixing times for the CH
group. According to Equation 14.3, the magnetisation-exchange dynamics should show
an initial quadratic dependence on tm. This dependence is indeed found in the experiments
and is shown in Figure 14.6a for SBR as an example.
From the slope of the initial magnetisation decay, the values of the effective residual
1/2
( )
2
CH − CH 2 CH − CH 2
dipolar couplings Deff ≡ Deff could be evaluated (see Figure 14.6a) and
it could be shown that these couplings scale with the crosslink density [36]. The correlation
CH − CH 2
of the coupling constant Deff with the shear modulus G as an independent measure
of the crosslink density demonstrates this behaviour (Figure 14.6b). Note, however, that
the line in Figure 14.6b does not cross the origin but provides a significant value for
CH − CH 2
Deff at G=0. This is not in contradiction with the applied model, but reflects the
influence of local chain order (physical crosslinks) [36]. From extrapolating to G=0 in
Figure 14.6b, the inter-group residual coupling of an uncrosslinked SBR melt was
CH − CH 2
determined to be Deff ≈533 Hz. Using average internuclear distances and bond angles
533
Figure 14.6 (a) The initial decays of the proton magnetisation for the exchange process
CH® CH2. The initial part of the magnetisation-exchange data (starting with tm=100
s) shows the predicted dependence in t 2m . The effective CH-CH2 intergroup dipolar
CH − CH 2
coupling constant Deff can be evaluated by fitting the theoretical curve. (b)
CH − CH 2
Deff versus the shear modulus G for a series of differently crosslinked SBR
samples [36]
Reproduced with permission from D.E. Demco, S. Hafner, C. Fülber, R. Graf and H.W.
Spiess, Journal of Chemical Physics, 1996, 105, 11285, Figure 6. Copyright 1996,
American Institute of Physics
534
for SBR, one can extract the value for the dynamic-order parameter P2 H − H2
of the
r CH − CH 2 − CH 2
CH-CH2 intergroup linkage rCH − CH 2 according to Deff = DCH
rigid P2 H − H2
. This
yields P2 ≈ 0.1 and suggests an even higher value for the dynamic-order parameter
H − H2
r r
for the corresponding carbon-carbon bond P2 C − C ≈ 0.13, since rCH − CH 2 and rC − C form
an angle of 24°. Thus, even in polymer melts the chain motion can be largely anisotropic,
as shown previously for poly(methacrylates) [39].
Magnetisation-exchange NMR spectroscopy offers a convenient means for measuring

the residual dipolar couplings between functional groups along the polymer chains. The
intergroup dipolar coupling is remarkably high even in uncrosslinked melts representing
relatively high local segmental order. In crosslinked samples, it scales with the crosslink
density as predicted by the scale-invariant model of residual dipolar couplings [8],
emphasising its relation to viscoelastic properties of the elastomers.
14.3.2.2 Molecular Motions from 2D NOESY-MAS Spectroscopy
Previously the coherent magnetisation-exchange process by 1H non-vanishing average

dipolar couplings was discussed. Under fast MAS conditions, however, the residual
couplings responsible for the exchange are largely refocused for full rotor periods and
cross-relaxation by the fluctuating part of the dipolar interaction predominates. This so-
called Nuclear Overhauser effect (NOE) [7] takes place if the spin system is not in internal
equilibrium. It corresponds to relaxation between dipolar coupled nuclei such as 13C and
1
H or between different sites (lines) of the same type of nucleus. NOE spectroscopy
(NOESY) is well-known in liquid-state NMR where it is one of the standard 2D techniques
to elucidate and assign structures of macromolecules in solution [7].
Following the success of the technique in the solution-state [40], NOESY has been also
applied to polymers or viscoelastic materials [35,36,41-47]. As in solution experiments,
one takes advantage of the fluctuating part of the dipolar coupling to extract useful
structural and dynamical information. Using this method, NOE factors have been
measured in polyisoprene over a range of temperatures with static 13C NOE spectroscopy
[41]. A molecular-weight independent change of regime was observed at around 60 °C
for the backbone motion reflecting a loss of motional cooperativity with increasing
temperature. Also the temperature-independent correlation time of the internal rotation
of the methyl group could be characterised.
13
C NOE spectroscopy under MAS was used for probing polymer miscibility in polymer
blends, polystyrene/polyvinyl methyl ether (PS/PVME) [42]. This study takes advantage
of the fact that crosspeaks appear only between spins that are neighbours of each other,
535
thus establishing NOE as a probe for the degree of mixture on the molecular level.
Additional information on the molecular structure of the blend could be obtained from
the NOE growth rates. The results suggest that there exists a specific interaction between
the phenyl ring of the PS and the PVME methyl group. In reference [43] the same technique
was applied for investigating methyl groups as a source of cross-relaxation in solid
polymers such as polycarbonate or polystyrene.
Static 1H 2D NOE spectroscopy was applied in a first experiment showing that the
technique can be used to measure inter-chain interactions [44]. This work was then
continued by applying the technique under MAS to investigate the inter-molecular
interactions responsible for the miscibility in polybutadiene/polyisoprene blends above
the Tg [45]. It was shown that intermolecular association can be probed by this technique
and the results reveal the existence of weak intermolecular interactions between the
polyisoprene methyl group and the vinyl side chain of the polybutadiene.
• 1H NOESY MAS applied to the investigation of molecular motions

Recently 1H NOESY MAS was used to study molecular motions in technically relevant
materials such as rubbers [46, 47]. For the evaluation of these parameters, it is necessary
to understand the cross-relaxation process in the presence of anisotropic motions and
under sample spinning. Such a treatment is provided in [47] and the cross-relaxation
rates were found to weakly depend on fast motions in the Larmor-frequency range and
strongly on slow motions of the order of the spinning frequency νR. Explicit expressions
for the νR dependent cross-relaxation rates were derived for different motional models.
Examples explicitly discussed were based on a heterogeneous distribution of correlation
times [1,8,48] or on a multi-step process in the most simple case assuming a bimodal
distribution of correlation times [49-51].
The derived relationships were tested experimentally on a crosslinking series of SBR and
the cross-relaxation was studied as a function of the rotor frequency νR. As an illustration
of such measurements, Figure 14.7a shows a surface representation of the gradient-selected
2D NOESY spectrum of an SBR sample acquired with a mixing time of 2.7 ms at a
rotor-frequency of 8 kHz. The lines corresponding to the aliphatic, the olefinic and the
aromatic protons are well resolved and can be assigned as indicated. Already at short
mixing times pronounced cross-peaks are visible, in particular between the olefinic CH
and the CH2 group.
Figure 14.7b shows the decay curve of the diagonal peaks for mixing times up to 3 ms.
Unlike the magnetisation-exchange case [36,37], the decay is found to be approximately
linear in tm in the short mixing time regime and exponential or biexponential for long
536
Figure 14.7 (a) 2D 1H-NOESY surface spectrum of SBR acquired for a mixing time of
2.7 ms and under 8 kHz MAS [47]. Already at this short mixing time pronounced
crosspeaks are visible. (b) Short time decay of the crosspeak intensity evaluated from a
series of 2D spectra. Such curves can be analysed, providing information on
internuclear distances and molecular dynamics (correlation times)
Reproduced with permission from T. Fritzhanns, D.E. Demco, S. Hafner and H.W. Spiess,
Molecular Physics, 1999, 97, 931, Figures 2c and 3a. Copyright 1999, Taylor & Francis
537
times [47]. A similar curve can be plotted for the increase of the cross-peaks. To investigate
the effect of the sample spinning on the cross-relaxation rates, series of 2D NOESY MAS
spectra were acquired as a function of the rotor frequency. The rates evaluated from
these series were found to decrease linearly with νR which is an indication that cross-
relaxation in elastomers is dominated by slow motions in the 10 kHz regime rather than
by motions in the Larmor-frequency range as in case of liquids. The cross-relaxation
rates at room temperature were found to depend only moderately on the crosslink density
which is another indication that cross-relaxation is dominated by the α process.
The experimental rates were analysed in terms of the explicit expressions derived by the
theoretical treatment [47]. As expected for a statistical copolymer, segmental motions in
the SBR samples cannot be described by a single correlation time [1]. From T1 data that
were also measured the correlation time should be in the range of 10-8 s while the νR
dependence of the cross-relaxation rates requires a correlation time of 10-5 s. So it is
clear that at least two different correlation times are necessary to account for the
experimental findings. The relaxation dispersion data T1(ω) (ω being the Larmor
frequency) [52], however, show no discontinuity that would indicate the presence of two
distinct motional processes. Thus, the data were analysed in terms of a broad distribution
of correlation times for almost isotropic segmental motions (α relaxation). With a simple
log-Gaussian distribution function of reasonable parameters [47,48] it was found that
one can account for both, the T1 values and the νR-dependent cross-relaxation rates.
From the centre of gravity of the distribution, the Tg could be estimated as Tg = –43 °C
using the well-known Williams-Landel-Ferry (WLF) equation [1]. This value compares
favourably with the known value of Tg = –53 °C for uncrosslinked SBR-1500.
14.3.2.3 Selective Residual Dipolar Couplings by 1H Multiple-Quantum NMR

Spectroscopy
Multiple-quantum (MQ) NMR spectroscopy is well-established for structural studies of

liquids and highly mobile solutes in liquid crystals [7,53,54]. In recent years there has been
a sustained effort to obtain homonuclear [55-63] and heteronuclear [64-66] high-resolution
MQ NMR spectra also for organic solids using fast MAS to increase resolution and
sensitivity. Such MQ spectra proved to be valuable tools for determining dipolar
connectivities between spin-1/2 nuclei [56-58, 60-63]. More quantitatively, dipolar couplings,
internuclear distances and molecular torsion angles can be measured by these techniques
[56-63]. For viscoelastic materials it was recently shown, that 1H high-resolution MQ
NMR spectroscopy offers the possibility to measure site-selective residual dipolar couplings
between all resolved protons [60, 61]. Thus the MQ technique is an attractive tool for
studying structure and dynamics in polymer melts [60] and elastomers [61-63].
538
Figure 14.8 (a) General scheme of two-dimensional multiple-quantum (MQ)

spectroscopy. During both intervals of free precession (MQ coherences during t1 and
SQ coherences during t2) dipolar decoupling can be achieved by MAS. Possible pulse
sequences used for excitation/reconversion of MQ coherences are: (b) broadband
BABA [57] and (c) C7 [67]
Reproduced with permission from R. Graf, D.E. Demco, S. Hafner and H.W. Spiess, Solid
State Nuclear Magnetic Resonance, 1998, 12, 139, Figure 1. Copyright 1998, Academic
Press
• Principles of the method
The MQ experiment [53, 54] follows the general scheme of 2D NMR spectroscopy is
shown in Figure 14.8a.
In the beginning, MQ coherences are excited (preparation period) and evolve for an
evolution time t1. At the end of the evolution period, the MQ coherences are reconverted
to detectable single-quantum (SQ) coherences (mixing period) which then are detected.
539
The experiment is performed under fast MAS conditions, that is, the rotor frequency ωR
is larger than the anisotropic spin interactions. Hence, an efficient averaging of the
anisotropic spin interactions takes place during the evolution and detection periods leading
to well-resolved, nearly liquid-like NMR spectra. On the other hand, during the excitation
and reconversion periods of the experiment dipolar interactions are required for the
generation (and reconversion) of the MQ coherences. The relevant parts of the dipolar
Hamiltonian are thus reintroduced by broadband dipolar-recoupling pulse sequences
such as back-to-back (BABA) [57] (Figure 14.8b), or C7 [67] (Figure 14.8c). Note, that
only those spins contribute to the double quantum (DQ) signal that are relatively strongly
coupled to each other.
• Dipolar connectivities from the high-resolution 1H DQ MAS spectrum
Figure 14.9b shows the 1H DQ spectrum of a crosslinked SBR sample which has been
acquired at a spinning frequency of 10 kHz [61].
Figure 14.9 Figure 14.9 1H DQ MAS spectrum of a SBR sample with a low value of
crosslink density (0.8 phr-0.8 phr sulfur-accelerator) as shown in b [61]. The rotor
frequency was 10 kHz and the t1 increment was 15 μs. The diagonal peaks and the
cross peaks of the functional groups are assigned as indicated (compare (a) and (b))
State Nuclear Magnetic Resonance, 1998, 12, 139, Figure 3. Copyright 1998, Academic Press
540
On top of the 2D spectrum, the single-quantum projection is shown, which corresponds

to the MAS spectrum in Figure 14.1 apart from some deviations caused by DQ-filter
effects. The peaks in the 2D DQ spectrum correspond to double-quantum coherences
between two spins which must be relatively close neighbours in space in order to contribute
significantly to the peak intensity as follows from the strong distance sensitivity of the
dipolar coupling, Equation 14.2. From the existence of the corresponding peaks therefore
through-space dipolar connectivities can be easily established.
The assignment of the peaks is shown along the ω1 dimension by pairs of letters a-f (see
Figure 14.9a) which indicate the functional groups that are involved in the generation of
the corresponding DQ peaks. From a simple qualitative inspection of the DQ spectrum
it can be seen that there are dipolar connectivities between practically all functional
groups. The strongest DQ signals are found between protons of the polybutadiene groups,
but also DQ peaks of considerable intensity are visible between the polybutadiene protons
and the aromatic protons of the styrene units. This indicates a good mixture of the
different functional groups on the nanometer length scale as is found for instance in the
case of a statistical copolymer.
• Site-selective double-quantum build-up curves and correlation with

crosslink density
A more quantitative evaluation of the DQ peak intensities is possible by acquiring the
experimental DQ build-up curves for each of the peaks performing a series of DQ
experiments with increasing excitation time. As an example, the build-up curves for the
three main contributions are shown in Figure 14.10 for two samples with different
crosslink densities [61].
A polynomial fit of the experimental DQ built-up curves using a relationship derived in

the spin-pair approximation allows the ratio of the corresponding residual dipolar
(
couplings to be determined, i.e., D CH 2 SCH
s
2
)(
: D CH 2 − CH SCH
s
2 − CH
)(
: D CH − CH SCH
s
− CH
),
where D represents a pre-averaged dipolar coupling constant and Sijs the corresponding
ij
2 − CH
scaled dynamic-order parameter [60, 61]. The order parameter SCH
s (calibrated using
CH 2 − CH
a 1D MAS sideband pattern) is found to be close to the value Ss = 0.1 that has
been estimated for SBR samples by the magnetisation-exchange experiments described
above [36]. The order parameter corresponding to the CH-CH coupling is even higher
[61]. This coupling provides the best measure for the chain dynamics since it is
predominantly aligned along the segmental axis.
For investigating a series of samples with different crosslink densities, however, the
procedure described previously is very time-consuming. In this case, a somewhat less
541
Figure 14.10 Site-selective 1H DQ build-up curves for low (a) and high (b) crosslinked SBR
samples [61], respectively. The DQ signals correspond to the -CH2- (■), CH2-CH- (▲) and
CH=CH coherences (●). The vertical dashed lines mark the excitation times for the
maximum signal for the methylene protons, which differs for the two cases. More
quantitatively, the build-up curves can be evaluated with respect to residual dipolar couplings
State Nuclear Magnetic Resonance, 1998, 12, 139, Figure 4. Copyright 1998, Academic Press
542
detailed 1D measurement of the integral DQ-filtered signal can be used to establish the
relationship between the total DQ intensity and the crosslink density. The square root of
the DQ intensity is found to be an approximately linear function of the crosslink density
but, more important, the topological constraints that are already present in the melt are
found to dominate the residual couplings (see also [60]). Thus the technique is not so
much a sensitive measure of the crosslink density but is rather a sensitive and selective
probe for local dynamic chain order.
The strongest residual dipolar couplings were also edited in a non-rotating crosslinked
poly(isoprene) series by exciting DQ and triple-quantum coherences in the short time
regime [62,63]. From this, the dynamic order parameters of the methylene and methyl
groups were estimated and correlated with the crosslink density. Essentially the same
behaviour was found as for SBR.
14.3.2.4 Residual Local Dipolar Fields by Heteronuclear Correlation

Experiments
The NMR spectrum of rare spins, (e.g., 13C), contains extremely useful information
because of the usually high chemical-shift dispersion and the possibility to easily eliminate
undesired spin interactions. It would thus be advantageous to combine this high-resolution
dimension with 1H spectroscopy in the form of a heteronuclear correlation experiment.
Next, the so-called wide-line separation (WISE) experiment [1,68-71], in which 1H
broadline information is correlated with the 13C spectroscopic information will be
discussed. There are however also other possibilities for including heteronuclear
information in a 2D experiment, for instance, 2D heteronuclear J-resolved spectroscopy
which was applied for the investigation of filled natural rubber [38]. It was concluded
that there must be a high degree of motion to allow the scalar 13C-1H couplings to be
revealed by MAS alone. The 13C linewidth was found to be determined by susceptibility
effects due to the presence of the filler.
While the (heteronuclear and homonuclear) residual dipolar couplings have to be eliminated
in such J-resolved experiments, they are the main source of information in the WISE
experiment described in the next section. For the investigation of viscoelastic materials the
experiment thus must be performed under static conditions or under slow MAS.
The 2D WISE experiment [68, 71] is based on the simple pulse scheme presented in
Figure 14.11.
543
Figure 14.11 Basic 2D pulse sequence used for the measurement of residual dipolar
local fields. After the excitation pulse the spins are allowed to evolve for some time t1
(indirect dimension) under influence of the relevant spin interactions before cross-
polarisation takes place. The direct detection during time t2 then takes place on the
13
C side typically under proton dipolar decoupling (DD). The basic scheme can be
extended by various spin manipulation techniques (not shown) during time t1. For
instance, the heteronuclear dipolar contribution can be removed by a decoupling pulse
on the 13C side [73]
After the excitation of the protons by a 90° pulse, a variable period t1 follows during
which a decay of the magnetisation by homonuclear and heteronuclear dipolar couplings
takes place. Optionally, the heteronuclear couplings can be eliminated during this time
by S-spin decoupling using a suitable pulse sequence applied on the S-spin side during
the t1 period [68]. At the end of this period, a rare spin magnetisation is created by cross-
polarisation. The duration of the cross-polarisation process is chosen short enough that
spin diffusion between protons can be neglected. Finally, the 13C signal is recorded in the
t2 dimension in the presence of proton dipolar decoupling (DD). The resulting 2D data
set can then be Fourier transformed in both dimensions (resulting in a 2D WISE spectrum)
or only in the 13C dimension. In the latter case the information is obtained in form of the
1
H decay for each resolved 13C line (13C edited 1H relaxation). The experiment is sometimes
performed under slow MAS conditions to obtain well-resolved 13C lines. For viscoelastic
materials, the spinning must be very slow in order to not drastically influence the proton
lines which contain the information on the heteronuclear and homonuclear residual
couplings. As an alternative, one could also generate a sideband pattern by MAS and
evaluate it with respect to the residual couplings (see below). In any case, the experiment
544
provides a direct visualisation of the local dipolar field for functional groups with different
molecular mobilities.
One should, however, be aware that it is often not only the 1H-1H dipolar coupling but
also the 1H-13C heteronuclear dipolar coupling that affects the magnetisation decay of
those protons that are close enough to the 13C nuclei to cross polarise them [35,68,72-
74]. This effect takes place because the 1H-13C internuclear distance is smaller than the
1
H-1H distance, for instance, in a methylene group. The heteronuclear dipolar coupling
thus often is of the same order of magnitude or even stronger than the homonuclear
dipolar coupling [68,74].
• Residual heteronuclear local dipolar couplings by WISE-MAS experiments
For elastomers and rubbery-like materials well above the Tg, the high molecular mobility
reduces the dipolar couplings dramatically. The WISE experiment allows one to investigate
site-selectively residual dipolar interactions and thus molecular dynamics by editing the
corresponding proton slices of the 2D data set.
In references [35,74] 13C-edited 1H transverse relaxation was investigated in a SBR

crosslinking series. A short contact time was used to avoid 1H spin diffusion. As a
consequence of this, only 1H atoms directly bonded to 13C atoms are observed and the
1
H transverse relaxation is thus found to be mainly governed by 1H-13C heteronuclear
couplings which makes analysis simple. The 13C-edited 1H transverse relaxation could
be fitted with only one adjustable parameter, the effective number of statistical segments
[35]. This was invoked as a justification for the heterogeneous model used to describe
transverse relaxation. Moreover, the effect of crosslinking could be investigated and it
was shown that it affects the dynamics of all functional groups to the same extent [74].
A similar experiment has been performed [68] but under MAS conditions. For a series of
crosslinked natural rubber samples (A - F1), 13C edited 1H spinning sidebands have been
extracted from the 2D spectrum. These sideband pattern are encoded by the residual
dipolar couplings of the corresponding functional groups and are presented in
Figure 14.12.
For the experimental conditions given, the spectra are found to be dominated by
heteronuclear dipolar interactions which leads to relatively narrow, well-separated dipolar
spinning sidebands. The heteronuclear dipolar couplings could be evaluated by simulating
the spectra on this basis (see right-hand column in Figure 14.12). They were found to be
between 0.9 kHz and 1.5 kHz for the samples of the series (the inter-crosslink masses Mc
were between 6700-11000 g/mol). A practically identical envelope of the spinning-
545
Figure 14.12 Slices from a 2D experiment corresponding to the pulse sequence of

Figure 14.11 performed on a series of crosslinked natural rubber samples A- F1
(defined in [68]) under MAS. The slices reflecting proton sideband pattern for the
different functional groups are encoded by the 1H residual dipolar couplings. The
distinct features of dipolar slices prove that the different functional groups may be
considered as relatively isolated groups of spins on the time scale of the evolution and
cross polarisation
Reproduced with permission from C. Malveau, P. Tekely and D. Canet, Solid State NMR,
1997, 7, 271, Figure 3. Copyright 1997, Academic Press
546
sidebands intensities in the indirect dimension is found for all functional groups. This
indicates that the effective 1H- 13C vectors of different segments are affected in
approximately the same way by internal motions.
Spin diffusion as an alternative explanation of this effect could be excluded by a similar

experiment with additional 13C decoupling during the evolution period t1 which eliminates
the encoding by the heteronuclear dipolar couplings [68]. Distinct features are then visible
for the spinning sideband patterns of the CH3, CH2 and CH groups (not shown) proving
that different functional groups are relatively isolated from each other on the time scale
of the experiment as a consequence of the reduction of homonuclear dipolar interactions
by high segmental mobility and MAS. Therefore, the indirect detection of the spinning
sideband patterns via cross polarisation 13 C spectroscopy allows for selective
measurements of residual dipolar interactions on different functional groups [68].
• Investigation of phase-separated systems using WISE
Apart from the investigation of residual couplings described previously, WISE has also
been applied for the investigation of phase-separated systems, for instance co-polymers,
consisting of a rubbery phase and a more rigid phase [75-77]. While the different phases
can be discriminated in the 13C dimension by their chemical composition, the dynamics
of each of them can be studied by the corresponding 1H slice.
A model polyurethane consisting of monodisperse hard segments (45% weight) and soft
segments has been investigated [75] and it could be shown that both phases are completely
separated from each other. The dynamics and morphology of a poly(isobutylene-co-p-
methylstyrene) copolymer were investigated by various solid-state NMR techniques
including WISE [76]. WISE was used to investigate dynamics by determining the transverse
relaxation for polyisobutylene and methylstyrene blocks separately and differences
between the two were found. Domain sizes were investigated by another 2D heteronuclear
correlation experiment, where pulsed spin decoupling was used on the proton side and
an additional time τ was introduced in order to allow spin diffusion to take place. The
spin diffusion could be followed by the increase of the crosspeak intensities in the resulting
2D heteronuclear correlation spectra with increasing time τ. A domain size of the p-
methylstyrene blocks of around 3 nm could be estimated. Two commercial block-
copolymers have been investigated [77] using among other techniques WISE. The WISE
spectra showed that the mobilities of the polystyrene and rubber phases of the copolymers
are very different and that the interfacial region between the two must be very small. The
size of this region could be determined to be 2 nm using a spin diffusion experiment.
547
14.3.2.5 Deuterium NMR Studies on Thermoplastic Elastomers
For the investigation of the molecular dynamics in polymers, deuteron solid-state nuclear
magnetic resonance (2D-NMR) spectroscopy has been shown to be a powerful method
[1]. In the field of viscoelastic polymers, segmental dynamics of poly(urethanes) has been
studied intensively by 2D-NMR [78, 79]. In addition to 1D NMR spectroscopy, 2D
NMR exchange spectroscopy was used to extend the time scale of molecular dynamics
up to the order of milliseconds or even seconds. In combination with line-shape simulation,
this technique allows one to obtain correlation times and correlation-time distributions
of the molecular mobility as well as detailed information about the geometry of the
motional process [1].
The scheme typically used for such 2D-exchange NMR experiments corresponds to that
given in Figure 14.4b, however, for deuterons, often a solid-echo sequence is used for
detection instead of the last pulse. During the mixing time tm of the experiment molecular
reorientations are allowed to take place. If the molecular orientation of a C-2H bond has
changed due to slow molecular motions, the signal continues to evolve with a new
frequency. For reorientation about a well-defined angle θ, the 2D exchange spectrum
exhibits characteristic ridges in the form of ellipses that can be analysed for motion (for
details about the analysis see [1]).
• Local motions and segmental orientation in supramolecular hydrogen

bond assemblies
In the field of viscoelastic materials, the technique has been mainly applied to thermoplastic
elastomers. One interesting elastomer belonging to this class consist of polybutadiene
chains functionalised by 4-(3,5-dioxo-1,2,4-triazolidin-4-yl) benzoic acid (U4A) units
which act as effective junctions zones. The molecular dynamics of the phenyl rings of the
U4A units has been probed by 1D and 2D 2H-NMR [80]. In this system, there are three
spatially separated environments which are reflected in the mobility of the polar units.
Phenyl rings which are incorporated in the structure are either rigid or undergo 180°
phenyl flips. The small fraction of free functional groups move isotropically and their
mobility is coupled to the dynamics of the polymer matrix. The 1D 2H-NMR spectra can
be described quantitatively assuming a distribution of correlation times over 2-3 decades
and the geometry of the motional processes is defined by the environment in the clusters
up to the order-disorder transition temperature. 2D spectra of the model compounds
(see Figure 14.13) show an elliptical exchange patterns, indicating well-defined slow
180° phenyl flips on a time scale of 100 ms up to 3 s.
548
Figure 14.13 2D deuterium exchange spectra of a thermoplastic elastomer (tBu-dU4A)

[80]. Experimental results at 27 °C with mixing times of (a) tm=1 ms, (b) tm=500 ms
and (c) tm=3 s indicating the presence of slow 180° phenyl flips
Reproduced with permission from A. Dardin, C. Boeffel, H.W.Spiess, R. Stadler and E.T.
Samulski, Acta Polymerica, 1995, 46, 291, Figure 12. Copyright 1995, Wiley-VCH
549
14.5 Conclusion
Multidimensional NMR spectroscopy proves to be a powerful method to reveal structural

and dynamical information at the molecular level in elastomers. Residual dipolar couplings
can be measured site-selectively and correlated with the crosslink density and mechanical
stress. The local segmental order and information on local molecular motions can be
also obtained with newly developed 2D NMR methods. The information at the molecular
level can be correlated with macroscopic properties of elastomers and provides the basis
for a better design of material properties for specific applications.
Acknowledgements
Useful discussions with Dr. G. Heinrich, Dr. H. Dumler (both from Continental AG,
Hannover), with Professor B. Blümich, D. Canet, P. Tekely, and with our colleagues T.
Fritzhanns, C. Fülber, R. Graf are gratefully acknowledged. DED thanks the Deutsche
Forschungsgemeinschaft for a Mercator Guest Professor grant and SH for financial
support of part of the work reviewed above.
References
1. K. Schmidt-Rohr and H. W. Spiess, Multidimensional Solid-State NMR and

Polymers, Academic Press, New York, 1994.
2. F. Hori in Solid State NMR of Polymers, Eds., I. Ando and T. Asakura, Elsevier,
Amsterdam, 1998.
3. H.W. Beckham and H.W. Spiess, NMR Basic Principles and Progress, 1994, 32,
163.
4. H.W. Spiess, Annual Reports on NMR Spectroscopy, 1997, 34, 1.
5. P.A. Mirau, S.A. Heffner, G. Koegler and F. Bovey, Polymer International, 1991,
26, 29.
6. B.F. Chmelka, K. Schmidt-Rohr and H.W. Spiess in Nuclear Magnetic Resonance

Probes of Molecular Dynamics, Ed., R. Tycko, Kluwer Academic Publishers,
Dordrecht, The Netherlands, 1994, p.113.
7. R.R. Ernst, G. Bodenhausen and A. Wokaun, Principles of Nuclear Magnetic

Resonance in One and Two Dimensions, Clarendon Press, Oxford, 1987.
550
8. J-P. Cohen Addad, Progress in NMR Spectroscopy, 1994, 25, 1.
9. M. Mehring, Principles of High-Resolution NMR Spectroscopy in Solids, 2nd

Edition, Springer, Berlin, 1983.
10. R. Kimmich, NMR: Tomography, Diffusiometry, Relaxometry, Springer, Berlin,

1997.
11. S. Hafner and H.W. Spiess, Concepts in Magnetic Resonance, 1998, 10, 99.
12. A.K. Whittaker, Annual Reports on NMR Spectroscopy, 1997, 34, 106.
13. A. Schmidt, W. S. Veeman, V. M. Litvinov and W. Gabrielse, Macromolecules,

1998, 31, 1652.
14. R. Kimmich, E. Fischer, P. Callaghan and N. Fatkullin, Journal of Magnetic

Resonance A, 1995, 117, 53.
15. F. Grinberg and R. Kimmich, Journal of Chemical Physics, 1996, 105, 3301.
16. E. Fischer, F. Grinberg, R. Kimmich and S. Hafner, Journal of Chemical Physics,

1998, 109, 846.
17. F. Grinberg, M. Garbaczyk and W. Kuhn, Journal of Chemical Physics, 1999,

111, 11222.
18. J. Collignon, H. Sillescu and H.W. Spiess, Colloid and Polymer Science, 1981,
259, 220.
19. V.D. Fedotov and H. Schneider, Structure and Dynamics of Bulk Polymers by
NMR Methods, Springer, Volume 21, Berlin, 1989.
20. P.T. Callaghan and E.T. Samulski, Macromolecules, 1997, 30, 113.
21. R.C. Ball, P.T. Callaghan and E.T. Samulski, Journal of Chemical Physics, 1997,
106, 7352.
22. B. Deloche, A. Dubault and D. Durand, Journal of Polymer Science: Part B

23. V.M. Litvinov and H.W. Spiess, Die Makromolekulare Chemie, 1992, 193, 1181.
24. P. Sotta and B. Deloche, Journal of Chemical Physics, 1994, 100, 4591.
551
26. P. Sotta, Macromolecules, 1998, 31, 3872.
27. M. Warner, P.T. Callaghan and E.T. Samulski, Macromolecules, 1997, 30, 4733.
28. P. Sotta, B. Deloche, J. Herz, A. Lapp, D. Durand and J.C. Rabadeux,

29. M.G. Brereton and M.E. Ries, Macromolecules, 1996, 29, 2644.
31. M.E. Ries, M.G. Brereton, P.K. Klein, I.M. Ward, P. Ekanayake, H. Menge and
H. Schneider, Macromolecules, 1999, 32, 4961.
32. V.M. Litvinov and H.W. Spiess, Die Makromolekulare Chemie, 1991, 192, 3005.
33. A. Dardin, R. Stadler, C. Boeffel and H.W. Spiess, Die Makromolekulare Chemie,
1993, 194, 3467.
34. A. Dardin, H.W. Spiess, R. Stadler and E.T. Samulski, Polymer Gels and
Networks, 1997, 5, 37.
1996, 29, 6222.
36. D.E. Demco, S. Hafner, C. Fülber, R. Graf and H.W. Spiess, Journal of Chemical
Physics, 1996, 105, 11285.
37. L. Gasper, D.E. Demco and B. Blümich, Solid State Nuclear Magnetic Resonance,
1999, 14, 105.
38. A.P.M. Kentgens, W.S. Veeman and J. van Bree, Macromolecules, 1987, 20,
1234.
39. A.S. Kulik, D. Radloff and H.W. Spiess, Macromolecules, 1994, 27, 433.
40. D. Canet, Nuclear Magnetic Resonance, Concepts and Methods, John Wiley
Publishers, Chichester, UK, 1996.
41. J. Denault and J. Prud’homme, Macromolecules, 1989, 22, 1307.
42. J.L. White and P. Mirau, Macromolecules, 1993, 26, 3049.
552
43. J.L. White, Solid State NMR, 1997, 10, 79.
44. P. Mirau, P., H. Tanaka and F. Bovey, Macromolecules, 1988, 21, 2929.
45. S.A. Heffner and P. Mirau, Macromolecules, 1994, 27, 7283.
46. T. Fritzhanns, S. Hafner, D.E. Demco, H.W. Spiess and F.H. Laukien, Journal of
Magnetic Resonance, 1998, 134, 355.
47. T. Fritzhanns, D.E. Demco, S. Hafner and H.W. Spiess, Molecular Physics, 1999,
97, 931.
48. E. Fischer, R. Kimmich, U. Beginn, M. Möller and N. Fatkulin, Physical Review

E, 1999, 59, 4079.
49. G.Lipardi and A. Szabo, Journal of the American Chemical Society, 1982, 104,
4546.
50. G. Lipardi and A. Szabo, Journal of the American Chemical Society, 1982, 104,
4559.
52. E. Fischer, F. Grinberg, R. Kimmich and S. Hafner, Journal of Chemical Physics,

1998, 109, 846.
53. M. Munowitz and A. Pines, Advances in Chemical Physics, 1987, 66, 1.
54. D. P. Weitekamp, Advances in Magnetic Resonance, 1993, 11, 111.
55. H. Geen, J.J. Titman, J. Gottwald and H.W. Spiess, Chemical Physics Letters,
1994, 227, 79.
56. J. Gottwald, D.E. Demco, R. Graf and H.W. Spiess, Chemical Physics Letters,
1995, 243, 314.
57. M. Feike, D.E. Demco, R. Graf, J. Gottwald, S. Hafner and H.W. Spiess, Journal
of Magnetic Resonance, 1996, 122, 241.
58. R. Graf, D.E. Demco, J. Gottwald, S. Hafner and H.W. Spiess, Journal of
Chemical Physics, 1997, 106, 885.
59. X. Feng, P.J.E. Verdegem, Y.K Lee, D. Sandström, M. Edén, P. Bovee-Geurts, W.J.
de Grip, J. Lugtenburg, H.J.M. de Grott and M.H. Levitt, Journal of the
American Chemical Society, 1997, 119, 6853.
553
60. R. Graf, A. Heuer and H.W. Spiess, Physical Review Letters, 1998, 80, 5738.
61. R. Graf, D.E. Demco, S. Hafner and H.W. Spiess, Solid State Nuclear Magnetic
Resonance, 1998, 12, 139.
62. M. Schneider, L. Gasper, D.E. Demco and B. Blümich, Journal of Chemical

Physics, 1999, 111, 402.
63. M.Schneider, D.E. Demco and B. Blümich, Journal of Magnetic Resonance, 1999,
140, 432.
64. W. Sommer, J. Gottwald, D.E. Demco and H.W. Spiess, Journal of Magnetic
Resonance A, 1995, 112, 131.
65. M. Hong, J.D. Gross and R.G. Griffin, Journal of Physical Chemistry B, 1997,
101, 5869.
66. K. Saalwächter, R. Graf, D.E. Demco and H.W. Spiess, Journal of Magnetic
Resonance, 1999, 139, 287.
67. Y.K. Lee, N.D. Kurur, M. Helmle, O.G. Johannessen, N.C. Nielsen and M.H.
Levitt, Chemical Physics Letters, 1995, 242, 304.
68. C. Malveau, P. Tekely and D. Canet, Solid State NMR, 1997, 7, 271.
69. P. Tekely, D. Nicole, J. Brondeau and J.J. Delpuech, Journal of Physical

Chemistry, 1986, 90, 5608.
70. N. Zumbulyadis, Physics Review B, 1986, 33, 6495.
71. K. Schmidt-Rohr, J. Clauss and H.W. Spiess, Macromolecules, 1992, 25, 3273.
72. P. Tekely, P. Palmas and P. Mutzenhardt, Macromolecules, 1993, 26, 7363.
73. P. Palmas, P. Tekely and D. Canet, Solid State Nuclear Magnetic Resonance,
1995, 4, 105.
74. C. Fülber, D.E. Demco, O. Weintraub and B. Blümich, Macromolecular

Chemistry and Physics, 1996, 197, 581.
75. H-J. Tao, D.M. Rice, W.J. MacKnight and S.L. Hsu, Macromolecules, 1995, 28,
4036.
76. J.L. White, A.J. Dias and J.R. Ashbaugh, Macromolecules, 1980, 31, 1880.
554
77. H. Yu, J. Wang, A. Natanson and M.A. Singh, Macromolecules 1999, 32, 4365.
78. A.D. Meltzer, H.W.Spiess, C.D. Eisenbach and H. Hayen, Makromolekulare

Chemie, Rapid Communications, 1991, 12, 261.
79. J.A. Kornfield, H.W. Spiess, H. Nefzger, H. Hayen and C.D. Eisenbach,
80. A. Dardin, C. Boeffel, H.W. Spiess, R. Stadler and E.T. Samulski, Acta
Polymerica, 1995, 46, 291.
555
556
Deuterium NMR in Rubbery Materials
15
Bertrand Deloche and Paul Sotta
15.1 Introduction
Deuterium NMR (2H NMR) has been used extensively to study and characterise molecular
properties in various complex fluids such as liquid crystals [1, 2, 3] and in polymer
systems [4].
Rubber materials are soft, elastic solids, made of mobile, flexible polymer chains (with a
glass transition temperature (Tg) typically lower than 0 °C) which are linked together to
form a three-dimensional network. They are characterised by a low, frequency independent
elastic modulus (of the order 105 to 106 Pa) and usually by a large maximum reversible
deformation (up to a few hundred per cent). Rubber elasticity is based on the properties
of crosslinked polymer chains: at large spatial scales, the presence of crosslinks ensures
the reversibility of the deformation, while at short scales, mobile polymer chains behave
as molecular, entropic springs.
There are different ways to realise rubber materials with ‘customised’ properties. The
main requirements are usually: chemical and mechanical stability over a certain
temperature range; a given value of the elastic modulus; a high maximum deformation
and eventually, biocompatibility.
In order to characterise the molecular behaviour in such materials, various techniques

have been developed in the past 25 years. The aim of these investigations has generally
been to relate the macroscopic, thermomechanical behaviour, to molecular properties, in
order to improve the molecular design and to engineer materials with better performances.
From a physical point of view, it is easy to understand that the local molecular orientation
in rubber materials is related in a quite direct way to the polymer chain extension, which
in turn gives the amount of (entropic) elastic energy stored in the rubber. Therefore, it
has been essential to implement experimental approaches giving access to chain ordering
in strained rubbers, in order to investigate the way in which the constraints are transmitted
at a molecular scale.
557
Amongst the various techniques sensitive to the molecular behaviour, deuterium (2H)
nuclear magnetic resonance (NMR) has some properties which make it particularly
specific:
2
• H NMR provides a direct, absolute measurement of local chain extension, to which
the modulus is directly related, as mentioned above. In principle, the full distribution
of constraints in polymer chains may be obtained by analysing 2H NMR spectra. This
property has been exploited to image the spatial distribution of constraints in a (non-
uniformly) stretched rubber (see the corresponding chapter in this book);
2
• H NMR is sensitive to a very low degree of local anisotropy (or constraint);
• by selective deuteration, some contrast between various molecular groups or portions

of polymer chains may be introduced in to the NMR response;
2
• H NMR may be sensitive to the presence of defects or heterogeneities at various
spatial scales, which are often believed to play a major role in limiting or enhancing
the elastic properties of the materials.
The chapter is organised as follows:
The basics of 2H NMR are summarised in Section 15.2. Its potential for the
characterisation of molecular motions, are reviewed simply.
Some applications are listed to illustrate these potentials. They may be classified in various
ways. The most direct approach consists in working with labelled (deuterated), model
systems. ‘Real’ materials (designed for industrial applications and processed at a large
scale) are often multicomponent, complex systems, which may be relatively ill defined at
the molecular scale. Thus, working with chemically well defined, labelled materials (which,
however, often have relatively poor mechanical properties by themselves) is a way to
isolate and study the various parameters which play a role in rubber properties. Studies
are done both in the relaxed state and in constrained (uniaxially deformed) states. This
approach is illustrated in Section 15.3. Examples of studies performed in model, single
component networks are presented. However, even in this case, the sensitivity of the
method is such that it may detect the presence of a few percent of molecular defects.
The necessity to label (deuterate) the systems under study may appear as a practical
drawback. It often prevents working with real (industrially designed) materials. A way
to circumvent this problem is to use labelled molecules as 2H NMR probes whose
behaviour reflects that of the rubber matrix. These molecules may be solvent molecules
or even polymer chains chemically identical to the rubber matrix itself. This approach is
described in Section 15.4 and is illustrated by a few examples.
558
In composite, multicomponent systems, 2H NMR is particularly suited to investigate

interfacial properties. This is illustrated in filled rubbers, semicrystalline and thermoplastic
elastomers, based on copolymers (Section 15.5).
15.2 2H NMR Background
The basic concepts of deuterium (2H) NMR and its application to anisotropic fluids
have been described in numerous references [1, 4-7]. This technique has some specificities
which make it particularly relevant to investigate fluid, partially ordered systems:
• The necessity to label (deuterate) the sample may be turned into an advantage, allowing
investigation of specific components of a sample;
• The predominant interaction for a 2H spin system is the quadrupolar interaction,

which couples the electric quadrupole moment of the 2H nucleus to its electronic
surrounding. This interaction is a second-rank tensor HQ which lies approximately
along the C-2H bond in organic molecules. Thus, in practice, 2H nuclei may be
considered to be isolated. It shows that the 2H NMR formalism is similar to that of
an isolated proton pair [8];
• The quadrupolar interaction constant νQ is large, of the order 250 kHz, which makes
2
H NMR very sensitive to the degree of molecular mobility.
In a solid system, the resonance spectrum spreads over a wide frequency range (± 250 kHz),
with a structure related to the static orientational distribution of C-2H bonds in the
system (powder spectrum).
In a fluid system, the spectrum structure is affected by molecular motions. In the limit of
fast molecular motions (νQτ << 1), the lineshape becomes insensitive to the type of
individual motions, this is the so-called motional narrowing regime. What is observed in
this regime is the time average of the quadrupolar interaction, i.e., of the tensor HQ.
For fast, isotropic motions (a simple liquid), the quadrupolar interaction is time-averaged
to zero, which leads to a single resonance line at the Larmor frequency, with a linewidth
1/T2 (of a few Hz usually) related essentially to the correlation time τ of the motions.
In case of anisotropic motions (an anisotropic fluid), the quadrupolar interaction is

partially time-averaged to a residual tensor. This residual interaction Δ modulates the
time relaxation function via a reversible dephasing: M(t)≈exp[–t/T2]cosΔt. For motions
uniaxial around a symmetry axis, the effect on the spectrum is to split the resonance line
into a symmetric doublet with a splitting (in frequency units):
559
3 cos2 Ω − 1
Δ = νQ S (15.1)
2
Two space variables appear in Equation 15.1:
• Ω is the angle between the magnetic field B0 (which provides the reference direction in
a NMR experiment) and the symmetry axis of the molecular motions. Thus, the
angular factor (3cos2 Ω-1)/2 is related to the direction of orientation.
• νQS is the magnitude of the residual tensor. The averaging factor S is the orientational
order parameter, i.e., the degree of motional anisotropy, of a given C-2H bond (with
respect to its average direction of orientation):
3 cos2 θ(t ) − 1
S= (15.2)
2
θ(t) is the instantaneous angle with respect to the average direction of orientation and the
overbar denotes a temporal averaging. Most often, it is more relevant to consider the order
parameter of a molecular symmetry axis instead of a C-2H bond itself, which implies
taking into account preaveraging effects due to internal molecular motions [9, 10].
The two factors given previously may be separated experimentally by changing the
orientation of the sample compared to the magnetic field B0. For a current spectral
resolution of a few Hz, order parameters S (as defined in equation 15.2) as small as 10-4
may be measured, given the large value of the bare interaction constant νQ. Note that
only the absolute value of S is measured.
Thus, one specificity of 2H NMR lies in the ability to observe and to measure directly
some local molecular orientation with a high sensitivity. In the case of a non uniform
system, i.e., containing C-2H bonds with distinct average axis and/or orientation degrees,
two steps have to be clearly distinguished in analysing 2H NMR spectra. First, temporal
averaging gives a doublet for each particular bond in the system, according to
Equations 15.1 and 15.2. Then, the overall frequency spectrum F is obtained directly as
the superposition of all the doublets, with various splittings and relative weights. This is
written simply in terms of the distribution ρ(Δ) of the residual interaction Δ (expressed in
frequency units):
F( Δ ) ≈ ρ( Δ ) + ρ( − Δ )
Thus, in principle, the full distribution of local molecular orientations may be extracted
from the analysis of the lineshape. However, in disordered systems, it is often difficult to
560
separate angular factors and degrees of order. Additionally, measurements of T1, T2 and
other relaxation times allow to investigate the correlation times of the motions.
15.3 Model, Labelled Rubber Systems
15.3.1 General Presentation: End-linked versus Randomly Crosslinked Net-

works
End-linked networks have an elastic chain length N between junctions which is in principle
entirely determined by the precursor chain length distribution, provided that precursor
chains are smaller than the entanglement critical molecular weight Mc. End-linked
poly(dimethylsiloxane) (PDMS) networks, prepared by hydrosililation reactions [11, 12]
with precursor chain molecular weight Mn from 3000 to 25000 g.mol-1, have been studied.
Networks with unimodal and bimodal distributions of precursor chains were synthesised
using various crosslinking agents with branched or cyclic structures and with
functionalities of 4 or 6. However, some of these crosslinking agents are poorly compatible
with PDMS precursor chains. As a result, the homogeneity of the crosslinking density in
these networks may be somewhat controversial [13]. Besides, the density of trapped
entanglements is related to the concentration vc at which the network was formed [14],
and this latter parameter has been varied as well to some extent. A fraction (10% to
20%) of precursor chains are perdeuterated. Note that in PDMS, all 2H nuclei belong to
rotating methyl groups and are completely equivalent, which makes the 2H NMR spectra
analysis very easy. Even in such well defined networks, some defects are present. The
fraction of dangling chains may be estimated from the quantity of unreacted precursor
chains [15]. Recently, networks with a controlled fraction of selectively deuterated dangling
chains have been investigated [16, 17]. Some chains forming loops may not participate
in the elastic behaviour either.
Deuterated poly(1,4-butadiene) (PB) networks have also been investigated [18-22]. The
linear precursor chains (Mn from 75000 to 116000 g.mol-1) are homogeneously, selectively
deuterated in the 1,4 positions, and randomly crosslinked with different types or amounts
of crosslinkers. Star precursor chains deuterated at the center have also been used.
15.3.2 Pseudosolid Behaviour
Figure 15.1a shows a typical 2H NMR spectrum obtained in a deuterated PDMS network
at room temperature, in the absence of an external constraint [23]. The half-height
linewidth is very small (about 60 Hz) compared to the static interaction νQ. This means
561
Figure 15.1 a: 2H NMR spectrum of perdeuterated crosslinked chains in a PDMS

network (average molecular weight between junctions Mn = 10500 g.mol-1, i.e.,
140 monomers) at room temperature. The inset shows a pseudo-solid echo obtained
after a [π/2,π/2] pulse sequence. b: 2H NMR spectrum of perdeuterated molten PDMS
chains (Mn = 10500 g.mol-1) at room temperature. The linewidth is limited here by the
experimental resolution
Reprinted with permission from P. Sotta and B. Deloche, Macromolecules, 1990, 23,
1999, Figure 1. Copyright 1990, American Chemical Society
that a strong motional averaging is achieved, or in other words, that 2H NMR is sensitive
to the high mobility of network chains. Note that the Tg of PDMS is –120 °C, which is
far below room temperature.
On the other hand, the spectrum shown in Figure 15.1a is not a very narrow, liquid-like
line as observed in molten PDMS chains identical to those in the network (half-height
562
linewidth about 4 Hz) (Figure 15.1b). Thus, the motional averaging is not complete in
the network. Due to the junction points, chains may not reorient completely at the time
scale of the 2H NMR experiment, νQ-1. This indicates that 2H NMR is sensitive as well to
the constraints imposed on network chains by the junctions. This is demonstrated most
clearly by the response to an appropriate pulse sequence, which exhibits a well-shaped
‘pseudosolid’ spin echo (insert in Figure 15.1a). This pseudosolid echo is due to a reversible
contribution in the nuclear magnetic relaxation, which demonstrates the presence of a
permanent residual anisotropy at the molecular level [8, 24]. In other words, some (small)
degree of constraint (elongation) is stored in the network chains, even though the system
is, of course, overall isotropic. It is important to emphasise that the degree of orientation
is very small, as compared for example to the case of some liquid crystal system. The
ratio of the linewidth (about 60 Hz) to the quadrupolar constant νQ (about 250 kHz)
gives a degree of local orientation of the order 10-4 at most. On the other hand, the
spectrum is not structured as in a solid sample. It exhibits one broadened line, instead of
the typical line shape with a doublet (powder spectrum) observed in a solid sample. This
means that the degree of residual orientation is distributed both in direction and in
magnitude, because the network is disordered.
15.3.3 Dangling Chains
A more precise analysis of the line shape shows that it contains two components with
different behaviour and characteristic widths (Figure 15.2) [16, 25]. There is first a narrow,
liquid-like component. It is associated with chain segments which are not subject to the
motional restrictions giving rise to the pseudo-solid behaviour described in Section 15.3.2.
This part of the signal is generally attributed to dangling chains. This has been tested
using end-linked PDMS networks with a controlled amount of selectively deuterated
monofunctional precursor chains (which form dangling chains) [16], and also in PB
elastomers [26]. The relative amount of liquid-like signal correlates very well to the
fraction of dangling chains. The broader component is associated with elastic chains,
i.e., to chains connected to the network at both ends, which have a pseudo-solid behaviour.
Thus, in samples in which the number of dangling chains is not controlled a priori, it
may be estimated by a precise analysis of the line shape (or equivalently, the relaxation
function) [25].
15.3.4 Time Scale of Motions
NMR studies in polymer systems often present some difficulties, because polymer chains
may have motions at various spatial scales, with correlation times ranging in very large
domains. A model based on different correlation times in various parts of the networks has
563
Figure 15.2 2H NMR relaxation function obtained in a PDMS model network in the
relaxed state, showing the two components: the fast relaxing component (the curve
with alternatively long and short dashes) is ‘pseudo-solid’ and belongs to elastic chains
(which have restricted motions), the slowly relaxing, exponential component (dashed
line), is liquid-like and belongs to dangling chains. The total signal (black points) is a
superposition of both contributions, which gives the fraction of dangling chains
Reprinted with permission from K. McLoughlin, C. Szeto, T.M. Duncan and C. Cohen,
been proposed to interpret the line shape [27]. The situation becomes simpler when the
different correlation times are well separated. Molten PDMS chains are extremely mobile
at room temperature. Chains of length 10500 g.mol-1, or shorter, are completely liquid (see
Figure 15.1b). The quadrupolar interactions are already averaged by local, intrachain
motions, which have correlation times of the order 10-9 seconds [28]. In the limit of fast
motions, the types of motions do not influence the lineshape, only correlation times do.
In polymer networks, slow motions, which may occur at larger spatial scales, are frozen.
This has been demonstrated in the PDMS model networks described in Section 15.3.1,
by measuring the relaxation of the so-called quadrupolar magnetic order in the 2H spin
system [29]. This relaxation (observed with an appropriate pulse sequence) is sensitive
564
to slow variations of the average nuclear interaction. It involves a relaxation time T1Q
equal to 5/3 of the usual longitudinal relaxation time T1 [3] and the correlation function
<sinΔ0tsinΔtt>, which correlates the residual interaction Δ0 at t = 0 and Δt at t = τ (τ may
be as large as 0.5 s here) [30]. The exponential decay which is shown in Figure 15.3
Figure 15.3 The decay of the so-called quadrupolar order as a function of the time
interval t2 (quadrupolar order is created by a pulse at time t1, then relaxes during a
time interval t2 before being observed via a third pulse). Data from three differently
crosslinked PDMS networks are presented (Mn = 9700 g.mol-1; o,
tetraallyloxyethylene-crosslinked (tetrafunctional); + crosslinked with bis(allyloxy)-3-
[bis(allyloxymethyl]-2, 2-propylene oxide (hexafunctional); Δ tetravinylcyclosiloxane-
crosslinked]. The longitudinal relaxation is plotted for comparison. Experimentally,
the quadrupolar relaxation time T1Q equals about 1.33 T1Z
Reproduced with permission from P. Sotta and B. Deloche, Journal of Chemical Physics,
1994, 100, 4591, Figure 4. Copyright 1994, American Institute of Physics
565
actually demonstrates that network chain end-to-end vectors have no slow reorientations.
This is related to the fact that network junctions have fixed average positions in space
(even though they may fluctuate rapidly around these average positions).
15.3.5 Uniaxially Deformed Model Networks
Figure 15.4 shows 2H NMR spectra obtained in a uniaxially deformed PDMS network,
for different values of the angle Ω between the applied force F and the magnetic field B0
[31, 32]. The line shape is drastically affected compared to that observed in a relaxed
sample: a permanent, well-resolved doublet structure appears, both in stretched and
Figure 15.4 2H NMR spectra in a uniaxially compressed PDMS network (Mn =

10500 g.mol-1, deformation ratio λ = 0.5), for different values of the angle Ω between
the applied uniaxial force and the magnetic field B0
Reproduced with permission from P. Sotta and B. Deloche, Macromolecules, 1990, 23,
566
compressed samples. The doublet splitting Δ, measured as a function of the angle Ω,

reproduces with great accuracy the curve (3cos2 Ω-1)/2, in agreement with Equation 15.1.
This demonstrates that the quadrupolar interactions are time-averaged along the direction
of the applied force F, which means that the chain segment motions associated with the
doublet are uniaxial, with a symmetry axis along this direction. Specifically, for Ω = 54.5°,
the doublet vanishes and the spectrum is similar to the one obtained at λ = 1.
At this point, it is important to emphasise the implications of this behaviour, as regards

the physical origin of elasticity in these systems. Even though the way the anisotropy
appears in the stretched system may seem rather trivial, it is not. The appearance of the
doublet is not simply related to the induced overall anisotropy in a stretched rubber
network. In classical, independent chain models of rubbers, elasticity is an entropic process
only related to conformational restrictions in network chains, which are fixed at junction
points [32, 33, 34]. The macroscopic constraint is transmitted through the junctions via
the anisotropy induced in the static end-to-end vector distribution P(R). Due to end-to-
end stretching along its end-to-end vector R, each chain is oriented along R. As a result,
the average quadrupolar interaction for one chain of N segments (x, y, z are the
components of the reduced vector R/N1/2a, with z axis along B0) is [23]:
Δ 0 ( R ) ≈ νQ
κ
N
(
2z2 − x2 − y2 ) (15.3)
κ is a numerical factor which accounts essentially for chain flexibility. The 2H NMR
spectrum might be simulated by superposing the contributions of all chains, considered
independently [35]. If the macroscopic deformation is transmitted affinely to crosslink
junctions, P(R) may be written in the deformed network (the strain being along z):
⎡ 3λx2 ⎤ ⎡ 3λy2 ⎤ ⎡ 3z2 ⎤

P( R ) ≈ exp ⎢− ⎥ exp ⎢− ⎥ exp ⎢− 2 ⎥
⎢⎣ 2 ⎥⎦ ⎢⎣ 2 ⎥⎦ ⎢⎣ 2λ ⎥⎦
where λ is the deformation ratio. A spectrum calculated in this way is shown in Figure 15.5,
both in relaxed and elongated states. The most striking feature is that no doublet appears
here in the elongated state [23], even though of course there is a non-zero, average overall
orientation <S>o in the deformed network (the so-called Kuhn and Grün orientation
function) [33]:
S 0
= νQ
N
(
κ 2
λ − λ−1 )
The fact that classical theories fail in reproducing the existence of a doublet structure has
been also pointed out recently [36, 37].
567
Figure 15.5 Spectra simulated with a ‘phantom’ network, in which each chain is
elongated, and thus oriented, along its end-to-end vector. A gaussian distribution of
the end-to-end vectors is assumed. a: relaxed (isotropic) state; b: uniaxially elongated
state (deformation ratio λ = 2.5). Affine displacement of junctions is assumed. No
doublet appears
Reproduced with permission from P. Sotta and B. Deloche, Macromolecules, 1990, 23,
In fact, the dynamic uniaxiality reflected in the observed doublet (together with its Ω-
dependence) has a completely different origin. It comes from interactions between chain
segments, which, even though they are weak, have a collective effect in the deformed
sample so that most of the segments exhibit the same uniaxial fluctuations along the
applied force F, as observed experimentally. A mean field model for this effect was proposed
[23]. In this model, the residual quadrupolar interaction in one network chain contains,
in addition to the term Δ0 associated to chain elongation along the end-to-end vector
(Equation 15.3), a second contribution, written as a mean-field term which represents a
uniform, uniaxial contribution to orientation. This uniaxial contribution is responsible
for the splitting of the spectrum in two lines. This may be written as:
3 cos2 Ω − 1
Δ ≈ Δ 0 ( R ) + νQ u S (15.4)
2
568
In the model, the uniform contribution (and thus, the doublet splitting) is proportional
to the overall average orientation <S>. The interaction parameter u characterises the
strength of orientational interactions between segments (0 < u < 1). Thus, for a given
deformation ratio λ, the spectrum contains one constant splitting and a distribution of
additional shifts, which is clearly seen in Figure 15.4.
This mean-field effect is also clearly demonstrated by the observed orientation of free
(unattached) chains chemically identical to network chains, which are introduced during
or after network formation and free to diffuse in it (Figure 15.6) [20, 38, 39]. According
Figure 15.6 2H NMR spectra of perdeuterated free chains (Mn = 10500 g.mol-1)
diffusing in a uniaxially constrained PDMS network (Mn = 10500 g.mol-1), for
different elongation ratios λ. At λ = 1, the free chains probe an overall isotropic
system, and the doublet vanishes. The fraction of free chains is 9%
Reproduced with permission from P. Sotta, B. Deloche, J. Herz, A. Lapp, D. Durand and
J.C. Rabadeux, Macromolecules, 1987, 20, 2769, Figure 2. Copyright 1987, American
Chemical Society
569
to the previous model, the free chains are sensitive only to the uniaxial contribution in
the stretched network (second term in Equation 15.4). Therefore, under identical
conditions (same network characteristics, same elongation), they do present a permanent
doublet with the same splitting (and the same Ω-dependence) as network chains, but
without line broadening related to end-to-end chain elongation (first term in
Equation 15.4). The doublet splitting is of course zero in the relaxed state (λ = 1), because
the average orientation then vanishes and it increases with increasing deformation
(elongation) ratios above [18, 17, 31]. In Figure 15.7, the splitting Δ is plotted as a
Figure 15.7 Variation of the splitting Δ in an elongated PDMS model network

containing a small amount (9%) of free PDMS chains, as a function of (λ2-λ-1), with λ
the elongation ratio. A pair of identical network/free chains systems are used: ▲:
deuterated network chains (Mn = 10500 g.mol-1); o: deuterated free chains
(Mn = 10500 g.mol-1). S is the chain segment order parameter
Reproduced with permission from P. Sotta, B. Deloche, J. Herz, A. Lapp, D. Durand and
J.C. Rabadeux, Macromolecules, 1987, 20, 2769, Figure 7. Copyright 1987, American
Chemical Society
570
function of the strain function (λ2-λ-1) for both network and free chains. Taking into
account the internal motions and structure of the PDMS monomer [9, 10], the chain
segment order parameter S is equal to a few 10-3 for λ≤2 [31, 39].
In the presence of the mean field effect introduced before, the overall average orientation
is enhanced with respect to the hypothetical value <S>0 in the absence of interactions:
1
S = S (15.5)
1− u 0
The factor 1/(1-u) is greater than 1 (u<1). Thus, according to Equation 15.4, the splitting
Δ may be written:
Δ = νQ
κ u 3 cos2 Ω − 1 2
N 1− u 2
(
λ − λ−1 ) (15.6)
Equation 15.6 shows that the splitting is indeed proportional to the strain function (λ2-λ−1).
Thus, the induced orientational order parameter (see Equation 15.1) may be characterised
by the slope P of the doublet splitting Δ versus (λ2-λ-1). Experimentally, in PDMS model
networks, u amounts to a few 10-1 [23, 25].
Note that optical techniques such as the stress-induced birefringence [40, 41], infrared
dichroism [42, 43] or fluorescence polarisation [44, 45] are not able to separate the two
contributions in Equation 15.4, because they are only sensitive to the overall (macroscopic)
symmetry of the system, (i.e., to <S>), whereas 2H NMR is sensitive to the local symmetry
of molecular motions (the doublet is u<S>). In particular, the fact that the doublet is the
same for network and guest chains has sometimes been misinterpreted. Strictly speaking,
it does not prove that interchain interactions are strong, (i.e., u close to 1) [46], but that
there is a mean field effect in the deformed network. This result may be extrapolated to
network dangling chains as well. They exhibit the same orientational behaviour as free
chains when the network is deformed [39]. Besides, it has been claimed that the doublet
structure in the network itself may be dominated by the contribution from dangling
chains [47]. This contribution has been quantified in recent 2H NMR experiments in
PDMS end-linked networks with a controlled amount of selectively labelled dangling
chains [17]. It is clear that both network and dangling chains contribute to the observed
doublet [48].
The doublet appears in various other polymer matrices, which shows that the uniaxial
ordering of chain segments is perhaps a quite general phenomenon. For instance, PB
networks with chains deuterated in 1,4 positions (Figure 15.8) have been studied [18-
21]. Their behaviour is essentially the same, even though the linewidth is much broader
571
at room temperature than in PDMS. This is due to a chain mobility lower and a critical
entanglement molecular weight smaller than in PDMS, which leads to a much higher
viscosity. The increasing line width as λ increases is attributed here to the broad distribution
of chain lengths between junctions, because shorter chains are more oriented than longer
ones, and thus produce a larger doublet. Spectra can be fitted with a most probable
distribution of chain lengths. Other systems are described in Section 15.5.
Figure 15.8 2H NMR spectra obtained in a randomly crosslinked, deuterated PB

network. Precursor chain molecular weight 115700 g.mol-1, 1.2% crosslink agent, the
average molecular weight between junctions is 27400 g.mol-1 (as determined by
swelling experiments) or 11600 g.mol-1 (elastic measurements). The smooth curves are
fits with a most probable distribution of chain lengths with number average molecular
weight Mc = 11600 and with non-Gaussian chain statistics
Reproduced with permission from W. Gronski, R. Stadler and M.M. Jacobi,
572
15.3.6 Physical Origin of the Induced Orientation
As mentioned in Section 15.3.5, the doublet has been interpreted by short-range interchain
interactions. Numerical simulations using a Monte Carlo method have been performed
to confirm this interpretation [49-51]. In these simulations, chains are generated as self
avoiding walks on a cubic lattice, placed in a box. In order to represent network chains,
extremities of chains are affixed to the lattice. Interactions are introduced only via short-
range excluded volume (which prevents double occupancy of a lattice site). The
deformation is transmitted exclusively via the end-to-end vector distribution. The full
distribution of orientation, (i.e., the distribution for the average orientation of each
segment in the system), is sampled during the simulation. Figure 15.9 shows the simulated
spectrum in an elongated network. It does present a doublet structure qualitatively similar
to that observed experimentally, that varies linearly with (λ2-λ-1) and has the Ω-dependence
in Equation 15.5 [49]. One advantage of numerical simulations is that a given interaction
may be switched off easily: interchain (excluded-volume) interactions may be suppressed,
Figure 15.9 Spectrum for an elongated network (λ = 3.41, N = 51, Ω = 0°) obtained by
Monte-Carlo simulation. The abscissa scale denotes the reduced interaction Δ/νQ, i.e.
the orientational order parameter S. Ordinates are in linear arbitrary units
Reproduced with permission from P. Sotta, Macromolecules, 1998, 31, 8417, Figure 1.
573
and the spectrum obtained in a deformed network in this case does not present a doublet
[49]. These simulations demonstrate that a purely isotropic interaction (excluded volume)
can generate a uniaxial, (i.e., collective), orientational mean-field in a uniaxially stretched
system of chains. A model, based on translational entropy, has been proposed [50]. Other
interpretations have been given. It was suggested that junction fluctuations may give rise
to such a uniaxial contribution in stretched samples [52]. However, this does not explain
the free chain orientation. Another model based on anisotropic screening of excluded
volume interactions has been proposed [21, 37].
To emphasise the role of orientational interactions in the appearance of the doublet in

2
H NMR spectra, model PDMS networks have been swollen to various extents with a
good solvent of PDMS [53]. Figure 15.10 shows the slopes P=Δ/(λ2-λ-1) obtained in a
PDMS network swollen with chloroform. The swelling process results in stretching
Figure 15.10 Slopes P=Δ/(λ2-λ-1) obtained on deuterated PDMS networks

(Mn=10500 g.mol-1), swollen with chloroform, versus the polymer network volume
fraction Φ (the slopes are normalised to the slope at Φ=1). Data from two networks
with different characteristics are shown
Reproduced with permission from P. Sotta, B. Deloche and J. Herz, Polymer, 1988, 29,
1171, Figure 2b. Copyright 1988, Elsevier Science
574
network chains, and thus in enhancing the local average degree of orientation. This
would lead to an increase of P, which is not observed. Instead, a decrease of the slope P
is observed. This has been interpreted within the mean-field model mentioned in
Section 15.3.5, as the effect of a weakening of interchain interactions, due to the presence
of solvent molecules which decreases the local density of chain segments [53].
There is another interesting illustration of the mean-field effect presented previously.

End-linked PDMS networks with a bimodal distribution of precursor chains (Mn = 25000
and 3000 g.mol-1) with various fractions ΦL of long chains, have been prepared. Short or
long chains may be observed separately, due to selective deuteration. It is remarkable
that short and long chains exhibit the same doublet in a given network [54].This shows
that the doublet originates in a collective effect, which affects both type of chains in the
same way, with the same magnitude. The doublet splitting obeys an ‘ideal mixing law’:
Δ=ΦLΔL+(1-ΦL)ΔS; it is an average of that ΔL observed in the network with 100% long
chains, and that ΔS in the network with 100% short chains.
15.3.7 Correlation to Elastic Properties
In this section it is shown that 2H NMR is an excellent and quite direct way to characterise
elastic properties of a rubber material. It has been mentioned in Section 15.3.5, that this
characterisation may be performed in two ways. First, analysing the line shape, (i.e.,
measuring the linewidth), in the relaxed state gives the magnitude of local constraints
stored in elastic chains, and this in turn is directly related to macroscopic elastic properties
of networks. Alternatively, the slope P=Δ/(λ2-λ-1) is a function of parameters which are
involved in elasticity as well. Both quantities contain the dominant factor 1/N, where N
is the average length of elastic chains, and they should be correlated in a series of samples
with various characteristics. This has been checked. However, due to the presence of
dangling chains and perhaps other defects such as loops, it was shown that the line shape
has to be analysed very carefully to establish such a correlation [25]. Indeed, the line
shape allows the probing of the full distribution of local constraints (which may be quite
inhomogeneous), whereas the doublet splitting only reflects a mean-field average in the
system (Section 15.3.5).
To probe the 1/N dependence in the slope P, experiments were carried out on model end-
linked networks with different precursor chain lengths (average molecular weight Mn
varying from 3100 to 23000 g.mol-1) [55, 56]. As expected, the slope P decreases when
N increases. However, it does not follow a 1/N variation: P varies only by a factor of
about 3 as N varies by a factor of 7. In an end-linked network, N should not be simply
assimilated with the length of precursor chains, since trapped entanglements may play
the role of effective crosslinks (often denoted as physical crosslinks), in addition to chemical
575
crosslinks [57]. Therefore, some effective mesh size Neff should be introduced to take the
actual density of crosslinks into account [58]. A way to vary the number of trapped
entanglements is to perform the crosslinking reaction at different polymer concentrations
vc [14]. Indeed, network elastic properties depend on vc : the higher vc, the higher the
elastic modulus and the polymer volume fraction at swelling equilibrium Φe [59, 60].
Amongst the various PDMS networks studied, those which differ by vc only do exhibit
different slopes P [61, 62]. In Figure 15.11 the slope P is plotted versus Φe [57] for a
series of networks with various crosslink and/or trapped entanglement densities. At
equilibrium swelling in a good solvent, each chain of length N (between two adjacent
junctions) occupies a volume of the order R3 = (aN3/5)3, which leads to Φe = N-4/5 [63]
and thus P = Φe5/4. Though the dispersion of points in Figure 15.11 is too wide to test
this power law dependence, there is a clear relationship between both quantities [25,
57]. On the other hand, Figure 15.12 shows the dependence of the slope P on the elastic
modulus in a series of PDMS networks with various precursor chain lengths and elastic
Figure 15.11 Slopes P=Δ/(λ2-λ−1) as a function of the polymer volume fraction at

swelling equilibrium in cyclohexane Φe in a series of end-linked PDMS networks with
various lengths between junctions (A, A′: Mn = 25000 g.mol-1; B, B′: Mn =
10500 g.mol-1; C: Mn = 3100 g.mol-1) and polymer concentration during crosslinking
in toluene (A, B, C: vc = 0.7; A’, B’: vc =1)
Reproduced with permission from A. Dubault, B. Deloche and J. Herz, Progress in
Colloid and Polymer Science, 1987, 75, 45, Figure 3. Copyright 1987, Spriger-Verlag
GmbH & Co KG
576
Figure 15.12 Slopes P=Δ/(λ2-λ-1) plotted versus the effective molecular weight between
crosslinks Mc and the elastic modulus Ge, obtained from swelling measurements, in
series of PDMS networks with various precursor chain lengths and elastic chain
fractions: ● network chains only are deuterated (observed); O dangling chains only
are deuterated; ❑ free probe chains only are deuterated
Reproduced with permission from K. McLoughlin, J.K. Waldbieser, C. Cohen and T.M. Duncan,
chain fractions [17]. There is a linear correlation between both quantities, as it is expected
since both quantities are proportional to 1/Neff.
In PB networks, the variation of the line shape as a function of the applied stress was
interpreted in terms of a chain length distribution. Shorter chains may be more oriented
than longer ones, at a given elongation [18], which may lead to a non-affine behaviour
at the chain scale. The question of the spatial scale to which the deformation is affinely
transmitted, has been investigated intensively by small angle neutron scattering [64].
However, it may happen as well that the chain portions close to junction points are more
oriented (have a more restricted mobility) than those in the middle of the chains [19].
In Figure 15.13, the slopes P=Δ/(λ2-λ-1), measured in end-linked PDMS networks with
bimodal distributions of precursor chain lengths and various types of crosslinking agents,
577
Figure 15.13 Slopes P=Δ/(λ2-λ-1) versus the elastic modulus 2(C1+C2) (C1 and C2 are
the Mooney-Rivlin coefficients), in bimodal PDMS networks composed of a mixture
of long (Mn = 25000 g.mol-1) and short chains (Mn = 3000 g.mol-1). Either long
(capital letters) or short (small letters) chains are deuterated. The long chain volume
fractions are 15% (d, D), 50% (m, M) and 75% (s, S)
Reproduced with permission from B. Chapellier, B. Deloche and R. Oeser, Journal de
Physique II (Paris), 1993, 3, 1619, Figure 5. Copyright 1993, EDP Sciences
578
are plotted versus the elastic modulus (expressed in terms of the Mooney-Rivlin coefficients
C1 and C2 as 2(C1+C2) [33]), which have been measured independently for the whole series
of samples. The variation of the slope P is indeed perfectly linear in the elastic modulus,
which corroborates very nicely the mean-field model presented in Section 15.3.5 [54].
The results reported in Figures 15.11, 15.12 and 15.13 show relationships between
microscopic data (molecular orientation) and macroscopic ones (elastic modulus). This
suggests that the orientational interactions introduced in Section 15.3.6 to account for
the 2H NMR doublet induced under stress, play a role in rubber elasticity. Indeed, the
force to apply to realise a given deformation is expected to be lower in the presence of
interactions which tend to orient chain segments along the force direction. An elementary,
phenomenological model of rubber elasticity including short-range interactions has been
proposed [65]. It predicts the λ-dependence of the reduced force (and of the dilation
modulus) and reproduces fairly accurately the elastic response of PDMS networks [66].
15.4 Using Deuterated Probes

The necessity to label (deuterate) the systems under study may often appear as a
disadvantage, since it prevents working with real (industrially designed) materials. A
way to circumvent this drawback is to use labelled (deuterated) molecules as 2H NMR
probes of material properties [67].
The uniaxial contribution to the stress-induced orientation (the second term in

Equation 15.4) has been attributed mainly to orientational interactions between chain
segments. When guest molecules are introduced in a rubber matrix, these interactions
take place between network chain segments and guest molecules as well. This effect has
been recognised earlier, and may be used to study indirectly the behaviour of the matrix.
Two types of guest molecules have been used for this purpose.
One possibility is to use free polymer chains, chemically identical to the network chains. In
order to diffuse in the network easily, the free chains should not be too long, typically
significantly smaller than the mesh size of the network under study. The advantage here is
that guest molecules are locally identical to network chains and thus behave as ideal, non-
perturbative NMR probes. Indeed, it was shown that free chain segment exhibit the same
degree of uniaxial order as network chain segments, with the same strain dependence (see
Figure 15.7) [39]. These properties remain valid when oligomers are used as probe molecules,
which suggests that orientational couplings between chain segments are short range. The
presence of probe chains, identical to network chains does not affect the local chain segment
density. Thus, the induced order remains unaffected up to a quite large fraction of free
chains or oligomers (about 25%) [53]. End-linked PDMS [38, 39], poly(urethane) (PU)
networks [39], as well as PB networks [20] have been investigated in this way.
579
Another possibility is to use deuterated molecules of a good solvent for the rubber matrix
under investigation. The idea here is to introduce a relatively low fraction of solvent, so
that the matrix properties are not affected significantly. The appearance of a well resolved
doublet Δ in the NMR spectra of solvent molecules as the rubber is uniaxially stretched,
shows that the solvent molecules are also sensitive to the uniaxial anisotropy induced in
the deformed rubber [67]. This phenomenon is quite general and it has been observed
with a number of polar or non-polar solvents (short alkane molecules, benzene,
chloroform, cyclohexane, toluene, water). Indeed, it was shown that the induced
orientational order, as measured in the doublet splitting, is extremely sensitive to the
presence of even a small fraction of solvent: for example, in the system PDMS network/
chloroform, the orientational order induced on solvent molecules drops by about 20%,
when the solvent volume fraction varies from 0% to 10%. The solvent degree of order is
always smaller than the chain segment order [9]. However, due to the high mobility of
solvent molecules, the linewidth observed in the solvent spectra is always very small,
which allows to observe resolved doublets even with very small splittings (typically a few
Hz) at small deformations (smaller than 5%, i.e., λ≤1.05) (Figure 15.14). Note also that,
due to this mobility, only an average of the local orientation in the network is reflected
by the solvent. In the low deformation limit, it has been shown that the solvent doublet
is proportional to that of network chains (and thus exhibits the λ2-λ-1 dependence), with
a proportionality constant which depends on the pair polymer/solvent only [9, 53]. Solvent
molecules are thus very sensitive probes of the rubber matrix behaviour, and a number
of studies dealing with network parameters such as the density of crosslinks or trapped
entanglements have been achieved in this indirect approach [55, 61].
Anisotropic swelling of a rubber sample is another way to obtain uniaxial deformations.

This may be achieved by swelling a cylindrical test sample, while maintaining its fixed
diameter (sample confined in a tube). Such confined swelling experiments are convenient
to deform uniaxially fragile systems as gels or physical networks. A deuterated solvent
may be used as both the constraint agent and the 2H NMR probe allowing measurement
of the induced uniaxial anisotropy. This experiment has been performed in various types
of rubbers [68, 69]. It may be used to monitor the change in anisotropy as the constraint
is released and its temporal relaxation as equilibrium is approached.
The orientation of the swelling agent (solvent or free chains) has to be taken into account
in the analysis of the stress-optical behaviour of swollen networks. Specifically, the segment
polarisability (relative to the network chains or to the diluent chains), as currently derived
from stress-optical coefficients [33], may not be representative of intrinsic properties of
isolated chains. Short-range orientational interactions between the probe molecules and
network chains (and between the chains of the matrix itself) must be considered in the
interpretation of opticoelastic properties of swollen (and dry) rubbers [67].
580
Figure 15.14 2H NMR spectra of deuterated benzene molecules diffusing in a uniaxially

stretched PB network for various elongation ratios. The solvent volume fraction is 9%
Reproduced with permission from A. Dubault, B. Deloche and J. Herz, Polymer 1984,
25, 1405, Figure 1. Copyright 1984, Elsevier Science
581
15.5 Filled, Composite and Thermoplastic Elastomers
15.5.1 Filled Elastomers
15.5.1.1 Carbon Black Filled Vulcanised Natural Rubbers
Deuterium has a quite low natural abundance (1.5 x 10-2 %). However, the main problem
to obtaining measurable 2H NMR signals is related mainly to the rather broad linewidth in
systems with interesting rubbery properties, (i.e., with enough local constraints stored in
elastic chains, thus, large residual quadrupolar interactions), rather than to the small natural
abundance itself. This can be circumvented by using the magic angle spinning (MAS)
technique, which allows concentration of the widely spread spectral intensity into a family
of sharp spinning sidebands. This has been used in a series of filled and unfilled sulfur-
vulcanised natural rubber (cis-polyisoprene) samples [70]. Figure 15.15 show a series of
natural abundance 2H spectra obtained under MAS at a spinning speed of 0.5 kHz, in
natural rubbers with various proportions of vulcanisers and carbon black fillers. The spectra
are fitted with a two-component model which includes a ‘mobile’ (narrower) component
and a ‘rigid’ (broader) one. The residual interactions obtained from best fits of the spectra
varies from 0.2 to 0.6 kHz (mobile component) and from 1.2 to 1.7 kHz (rigid component).
Thus, this experiment demonstrates that it is possible in principle to estimate the magnitude
of local chain orientation using natural abundance 2H NMR.
15.5.1.2 Carbon Black Filled Diene Rubbers
1
H and 2H NMR have been used in styrene-butadiene rubber (SBR) with and without
carbon-black fillers to estimate the values of some network parameters, namely the average
network chain length N. The values obtained from both approaches were checked to
make sure that they were consistent with each other and with the results of other methods
[71, 72, 73]. To this purpose, a series of samples with various filler contents and/or
crosslink densities were swollen with deuterated benzene. The slopes P=Δ/(λ2-λ-1) obtained
on deuterated benzene in uniaxially stretched samples were measured. The slopes increase
significantly with the filler content, which suggests that filler particles act as effective
junction points [72, 73].
Deuterated PB networks filled with carbon black have been investigated recently [74].
The 2H NMR lineshape is different from that in unfilled elastomers: an asymmetric
doublet is observed as the sample is uniaxially stretched (λ=1.8)). This asymmetry is
related to the presence of carbon black fillers, which induce magnetic inhomogeneities.
582
Figure 15.15 Natural abundance 2H MAS spectra observed in a series of vulcanised

natural rubbers with various vulcaniser (sulfur: 1 or 3 wt%) and/or filler (carbon
black: 0 or 40 wt%) contents. The spinning speed is 0.5 kHz. The number of scans is
about 300000. Spectra are simulated with two components (a ‘mobile’ and a ‘rigid’
one) with various residual quadrupolar interactions
Reproduced with permission from C. Malveau, P. Tekely and D. Canet, Solid State
Nuclear Magnetic Resonance, 1997, 7, 271, Figure 5. Copyright 1997, Elsevier Science
583
This effect allows to quantify the fraction of polymer in close contact with filler particles.
It is strongly suggested that that fraction is submitted to local deformations different
from that within the bulk matrix.
15.5.1.3 Silica Filled Silicon Rubbers
In composite systems, 2H NMR is particularly suited to investigate interfacial properties.

Indeed, isolated nuclei are observed, which potentially allows spatially selective
information to be obtained. It has been used to investigate polymer chain mobility at the
polymer-filler interface, mainly in filled silicon (in particular PDMS) networks. The chain
mobility differs considerably at the polymer-filler interface, and this may be interpreted
in terms of an adsorbed polymer layer at the filler surface. T1 relaxation measurements
allowed to determine the fraction of chain units involved in the adsorption layer, or
equivalently, the thickness of the layer [75, 76, 77]. The molecular mobility and the
thickness of the adsorption layer are very sensitive to the type of filler surface [78].
It was shown that the stress-induced orientational order is larger in a filled network than
in an unfilled one [78]. Two effects explain this observation: first, adsorption of network
chains on filler particles leads to an increase of the effective crosslink density, and secondly,
the microscopic deformation ratio differs from the macroscopic one, since part of the
volume is occupied by solid filler particles. An important question for understanding the
elastic properties of filled elastomeric systems, is to know to what extent the adsorption
layer is affected by an external stress. Long-time elastic relaxation and/or non-linearity
in the elastic behaviour (Mullins effect, Payne effect) may be related to this question
[79]. Just above the melting temperature Tm, it has been shown that local chain mobility
in the adsorption layer decreases under stress, which may allow some elastic energy to be
dissipated, (i.e., to relax). This may provide a mechanism for the reinforcement of filled
PDMS networks [78].
Silica hydrolysis-condensation of alkoxysilane-terminated chains of various polymers

(PB, poly(propylene)-glycol, poly(ethylene)-glycol) leads to the formation of silica nodules.
Hybrid (organic-inorganic) nanocomposite rubbery materials are obtained in this way.
These materials have been swollen by various deuterated solvents, in order to probe the
order induced in the polymer matrix under stress [80].
15.5.2 Semi-crystalline Polymers
In semi-crystalline polymers, crystallites may act as effective crosslinks, which gives rise
to rubbery properties above Tg. In systems with long chains, the presence of crystallites
584
allows to relax the constraints in the amorphous phase, while the chains may be quite
strongly oriented as regards their overall configuration.
Deuterated n-octane was used as a probe of the amorphous regions in a stretched PDMS
network cooled down below Tm (octane molecules do not penetrate into crystallites). It
was observed that the local chain orientation in the amorphous regions decreases as the
system crystallises [81].
Drawn isotactic polypropylene (iPP) fibres have been studied by using deuterated n-
decane as a 2H NMR probe of the chain deformation ratio in the amorphous regions. It
is observed that the slope P=Δ/(λ2-λ-1) (defined in the limit of low deformations) indeed
depends on the annealing temperature [82]. Thus, annealing above the melting temperature
Tm of the crystallites allows chains to relax to some extent. Then, the local deformation
ratio in the amorphous phase λl becomes lower than the macroscopic one λ and depends
on the annealing temperature, i.e., on the amount of chain relaxation. Therefore, such
systems have strongly non-affine deformation at the chain scale.
Fully deuterated linear poly(ethylene) (PE) has been investigated also via various 2H
NMR techniques below Tm, i.e. in the semi crystalline state [83, 84]. The crystallinity
ratio was measured as a function of temperature and it was shown that the motions are
highly restricted in the amorphous regions of PE. It was shown that the onset of β and α
transitions (at which mobility appears in the crystalline phase) may be observed by 2H
NMR on raising the temperature. This onset of local motions in the crystalline phase is
related to the chain relaxation process quoted previously.
15.5.3 Thermoplastic Elastomers
Block copolymers allow to realise thermoplastic elastomers, by combining a central, soft

block with rigid (crystalline and/or glassy) blocks at chain extremities. The rigid segments
segregate into hard microdomains (or nodules) which act both as crosslink points with
high functionality and as solid, reinforcing particles.
15.5.3.1 Polyamide–Polyether Block Copolymers
Polyamide–polyether (PA-PEG) block copolymers have been studied using deuterated

water (2H2O) molecules as 2H NMR probes [85, 86, 87]. Well resolved doublets are
observed on 2H2O on stretching the samples at room temperature. At a given elongation
ratio λ, the splitting depends on the water concentration c. The linewidth is that of a
liquid and becomes comparable to that in bulk water for c≈0.6. This confirms that water
585
probes the mobile, elastomeric part of the sample only (PEG block). Moreover, the
linewidth follows the Williams-Landel-Ferry (WLF) equation and is independent of the
stretching ratio and the size of the hard PA block.
In these systems, the 2H 2O splitting is not linear in the strain function (λ2-λ -1)
(Figure 15.16). This indicates that the microscopic deformation ratio λl (at the chain
Figure 15.16 2H NMR splittings in deuterated water molecules in a PA-PEG block-

copolymer versus the stretching functions λ2-λ-1 (a) or λ*2-λ*-1 (b) (see Section
15.5.3.1), at various water concentrations c (the points at c = 0% are deduced from
the measured splittings Δ(c) by extrapolation)
Adapted with permission from J. Rault, C. Macé, P. Judeinstein and J. Courtieu, Journal of
Macromolecular Science B, 1996, 35, 115, Figure 6. Copyright 1996, Marcel Dekker, Inc
586
scale) is different from the macroscopic one λ, or in other words, that the deformation is
not affine at the chain scale. This is a quite general property in thermoplastic elastomers.
As PA-PEG elastomers are stretched (at room temperature) and then annealed just above
Tm, they do not relax to their initial dimension. This indicates that flow occurs during
deformation, because PA segments recrystallise with the chain axis parallel to the stretch
direction. However, the splitting variation tends to be linear in (λ*2-λ*-1), in which λ* is
the deformation ratio referred to the relaxed dimension measured after annealing
(Figure 15.16) [87]. This indicates that the local deformation ratio λl is equal to λ*. Note
additionally that 2H2O has be used as well to probe PA-PEG block-copolymers doped
with several caesium salts [88].
15.5.3.2 Polyurethane Elastomers
PU elastomers based on a polyester diol as the soft block and isocyanate groups coupled
via a chain extender as the rigid block, have been investigated. The role of the chain
extender is to induce some mobility and disorder within the hard microdomains. The
elastomers are selectively deuterated in the rigid block. 2H NMR spectra show the
coexistence of a solid and a mobile fraction [89]. The solid fraction shows no onset of
fast molecular motions at all, while the mobile fraction increases with temperature. These
results suggest that some of the hard microdomains (perhaps the most disordered ones)
melt successively when temperature is raised.
PU systems with hard blocks made of piperazine, have been also studied [90, 91, 92]. In
systems deuterated in the segments of the soft blocks, 2H NMR results show that a
fraction of the soft segments have a restricted mobility, due to the connectivity to hard
blocks, rather than to the adsorption on those blocks acting as filler particles [93]. To
study the interface in a specific way, systems with hard blocks selectively deuterated at
different positions have been synthetised. A reduction of quadrupolar interaction according
to the position was observed, corresponding to an increase of the mobility in hard segments
on approaching the hard-soft interface.
15.5.3.3 H-bonded Polybutadiene Networks
New crosslinked elastomers may be formed by self-aggregation of substituted functional

units able to form directed hydrogen-bonds. 1,4-PB has been randomly substituted with
about 4 mol% urazole units [94]. The urazole units form hydrogen-bonded
supramolecular, plate-like aggregates which act as effective crosslinking zones, thus
creating a thermoplastic elastomer [95, 96, 97]. The system is deuterated either in the PB
backbone or in the urazole units, which allows to investigate the molecular dynamics in
587
either part of the elastomer. The 2H NMR spectra in the relaxed state show motional
heterogeneities along the PB chains as well as in the junction domains. About 97% of the
urazole units are included in the aggregates, in which their mobility is completely frozen.
There are however some phenyl ring flips at the boundary of the domains. The 3%
urazole units which are not included in the aggregates undergo fast isotropic motions.
Part of the PB segments in the elastic chains are restricted in their mobility: a layer of
about two monomers units adjacent to the aggregates is immobilised.
The stress-induced quadrupolar doublets have been observed in samples deuterated in

PB chains [98]. At high elongation ratios, the doublet splittings are not proportional to
the strain function (λ2-λ-1): as λ increases, the splittings first increase rapidly, then
become much less dependent on λ or even nearly constant in some cases. There is also
a huge increase in linewidth on increasing λ. These features have been tentatively
interpreted as a very broad distribution of chain segment orientation in the highly
stretched samples [96].
15.6 Conclusion
The results discussed in this chapter demonstrate that 2H NMR is a powerful technique
for investigating microscopic properties in rubber networks. Most of the experiments
described here are easy to handle on standard NMR equipment. Due to the absence of
interactions between 2H nuclei, spectra and line shapes are easy to interpret and give
quite direct information, at least in the first step of analysis, which is that generally
required to correlate microscopic to macroscopic properties in these systems. Additionally,
in contrast to optical techniques (as birefringence, infrared dichroism, fluorescence
polarisation) the information which is obtained is very specific, because spatial and
temporal averaging processes are clearly distinguishable in NMR.
These basic properties have been exploited in the experiments described here. Various
time scales have been investigated. The static or dynamic nature of the observed local
anisotropy has been specified. It has been demonstrated that 2H NMR is sensitive to very
low degrees or very small variations in the magnitude of the elastic constraints stored in
elastic chains.
The reliability and sensitivity of the probe method has been emphasised. It circumvents
almost completely the perturbations inherent to some other probe techniques (electron
spin resonance, fluorescence). In particular, free chains appear to be ideal, non-perturbative
NMR probes for testing chain segment orientation in strained rubbers. The solvent probe
method is easy to handle and unexpensive, since it does not require the synthesis of
588
labelled polymers. In view of industrial applications, the technique may be used in non-
transparent samples, such as multi-component or filled elastomers. This technique should
be widely used to characterise rubbery materials, as regards for example the presence of
residual constraints or deformations resulting from material processing and/or from
intensive mechanical deformations.
One of the most relevant physical results discussed here is that the stress-induced
orientation in rubbers is not a single chain process, but involves cooperative interactions
between chain segments. Indeed, the 2H NMR response of elastic chains in a uniaxially
strained rubber is similar to that of a uniaxial fluid. This basic result is in contrast to
various theoretical descriptions wherein the physical effects of the chains are mainly
restricted to their sole action on the junctions. These local interactions between chains
may affect macroscopic elastomer properties (elasticity, optoelasticity, resilience) and
thus must be taken into account in order to interpret these properties at the microscopic
level. For example, a way to change (and perhaps to enhance) drastically the properties
of rubbers would be to enhance the orientational interactions by introducing liquid
crystalline (nematogenic) groups along the chains [99, 100]. The collective, dynamical
behaviour of chain segments observed in strained rubbers is a general property of molten
polymer systems in which the chains are constrained or confined, such as liquid crystal
polymers [101], block copolymers [102], grafted polymer brushes [10,103], coated
polymer films [104] and melts under shear flow [105].
More specific 2H NMR investigations, perhaps using selective deuteration of network

chains, may allow to obtain a very precise description of rubbery materials at a molecular
scale, as regards the characterisation of the heterogeneity in the local order or the dynamics.
The distribution of local order (or of the local degree of constraint) is probably related to
the heterogeneity of the crosslink density, or equivalently to the distribution of elastic
chain lengths. Some degree of motional heterogeneity may be associated to the presence
of junctions in model (single component) networks. On the other hand, a strong gradient
of chain segment mobility is related to the presence of polymer-solid interfaces in filled
and/or thermoplastic (multicomponent) elastomers. The mechanical properties of these
rubbery materials depend in a crucial way on both properties, namely, the repartition of
(effective) crosslinks and/or the nature of the interfaces in filled systems.
Note finally that, due to its high specificity and sensitivity to the local degree of constraints
in a rubbery material, 2H NMR has been used to obtain a spatial map of the constraints
in a material by imaging techniques [106, 107]. These experiments are described in another
chapter of this book. 2H NMR has been used as well to investigate other related polymer
systems, such as gels [108], thermosets [109], fibres [110, 111, 112, 113] and various
solid (glassy and/or semi crystalline) polymers [114].
589
References
1. J. Seelig, Quarterly Reviews of Biophysics, 1977, 10, 353.
2. J. Charvolin and B. Deloche in The Molecular Physics of Liquid Crystals, Eds.,

G.P. Luckhurst and W.G. Gray, Academic Press, London, UK, 1979.
3. J.H. Davis, Biochimica et Biophysica Acta, 1983, 737, 117.
4. H.W. Spiess, Advances in Polymer Science, 1985, 66, 23.
5. M.H. Cohen and F. Reiff in Solid State Physics, Eds., F. Seitz and D. Turnbull,
Academic Press London, 1975, Volume 5, 321.
6. H.W. Spiess, Colloid and Polymer Science, 1983, 261, 193.
7. E.T. Samulski, Polymer, 1985, 22, 177.
8. J.P. Cohen Addad, Progress in NMR Spectroscopy, 1993, 25, 1.
9. H. Toriumi, B. Deloche, E.T. Samulski and J. Herz, Macromolecules, 1985, 18,

305.
10. M. Zeghal, B. Deloche and P. Auroy, Macromolecules 1999, 32, 4947.
11. J. Herz, A. Belkebir and P. Rempp, European Polymer Journal, 1973, 9, 1165.
12. M. Beltzung, C. Picot, P. Rempp and J. Herz, Macromolecules, 1982, 17, 1594.
13. J. Bastide and L. Leibler, Macromolecules, 1988, 21, 2647.
14. S.J. Candau, A. Peters and J. Herz, Polymer, 1981, 22, 1504.
15. D.R. Miller and C.W. Macosko, Macromolecules, 1976, 9, 206.
16. K. McLoughlin, C. Szeto, T.M. Duncan and C. Cohen, Macromolecules, 1996,

29, 5475.
17. K. McLoughlin, J.K. Waldbieser, C. Cohen and T.M. Duncan, Macromolecules,

1997, 30, 1044.
18. W. Gronski, R. Stadler and M.M. Jacobi, Macromolecules, 1984, 17, 741.
19. W. Gronski, D. Emeis, A. Bruederlin, M.M. Jacobi and R. Stadler, British

Polymer Journal, 1985, 17, 103.
590
20. M.M. Jacobi, R. Stadler and W. Gronski, Macromolecules, 1986, 19, 2887.
21. M.E. Ries, M.G. Brereton, P.G. Klein, I.M. Ward, P. Ekanayake, H. Menge and
H. Schneider, Macromolecules, 1999, 32, 4961.
23. P. Sotta and B. Deloche, Macromolecules, 1990, 23, 1999.
24. J.P. Cohen Addad in Physical Properties of Polymeric Gels, Ed., J.P. Cohen
Addad, John Wiley & Sons, Chichester, UK, 1996, p.39.
26. H. Menge, S. Hotopf, U. Heuert and H. Schneider, Polymer, 2000, 41, 3019.
27. C.D. Poon and E.T. Samulski, Journal of Non-Crystalline Solids, 1991, 131, 509.
28. J. Hirschinger and A.D. English, Journal of Magnetic Resonance, 1989, 85, 542.
29. P. Sotta and B. Deloche, Journal of Chemical Physics, 1994, 100, 4591.
30. H.W. Spiess, Journal of Chemical Physics, 1980, 72, 6755.
31. B. Deloche, M. Beltzung and J. Herz, Journal of Physics Letters, 1982, 43, L763.
32. P.J. Flory, Statistical Mechanics of Chain Molecules, Interscience, New York, NY,
USA, 1969.
33. L.R.G. Treloar, The Physics of Rubber Elasticity, Clarendon Press, Oxford, UK,
1975.
34. J.E. Mark and B. Erman, Rubber Like Elasticity, a Molecular Primer, Wiley-
Interscience, New York, NY, USA, 1988.
35. M.I. Lifshits, Polymer, 1987, 28, 454.
36. M. Warner, P. T. Callaghan and E. T. Samulski, Macromolecules 1997, 30, 4733.
38. B. Deloche, A. Dubault, J. Herz and A. Lapp, Europhysics Letters, 1986, 1, 629.
39. P. Sotta, B. Deloche, J. Herz, A. Lapp, D. Durand and J.C. Rabadeux,

591
40. B. Erman and P.J Flory, Macromolecules, 1983, 16, 1601.
41. K. Nagai and T. Ishikawa, Journal of Chemical Physics, 1966, 45, 3128.
42. H.W. Siesler, Advances in Polymer Science, 1984, 65, 1.
43. B. Jasse and J.L. Koenig, Journal of Polymer Science, Polymer Physics Edition,
1979, 17, 799.
44. L. Monnerie in Static and Dynamic Properties of the Polymeric Solid-State, Eds.,
R.A. Pethrick and R.W. Richards, D. Reidel Publishing, Dordrecht, The
Netherlands, 1982.
45. J.P. Jarry, B. Erman and L. Monnerie, Macromolecules, 1986, 19, 2750.
46. C.M. Ylitalo, J.A. Zawada, G. Fuller, V. Abetz and R. Stadler, Polymer, 1992, 33,
2949.
47. J. Kornfield, G-C. Chung and S.D. Smith, Macromolecules, 1992, 25, 4442.
48. B. Deloche, Die Makromolekulare Chemie - Macromolecular Symposia, 1993,

72, 29.
49. M. Depner, B. Deloche and P. Sotta, Macromolecules, 1994, 27, 5192.
50. P. Sotta, P.G. Higgs, M. Depner and B. Deloche, Macromolecules, 1995, 28, 7208.
53. P. Sotta, B. Deloche and J. Herz, Polymer, 1988, 29, 1171.
54. B. Chapellier, B. Deloche and R. Oeser, Journal de Physique II (Paris), 1993, 3,

1619.
55. A. Dubault, B. Deloche and J. Herz, Polymer 1984, 25, 1405.
56. D. Yang, F. Li and Z. Qiu, Journal of Macromolecular Science A, 1990, 27, 149.
57. A. Dubault, B. Deloche and J. Herz, Progress in Colloid and Polymer Science,
1987, 75, 45.
1996, 29, 6222.
592
59. M. Beltzung, C. Picot and J. Herz, Macromolecules, 1984, 17, 663.
60. V.G. Vasiliev, L.Z. Rogovina and G.L. Slonimsky, Polymer, 1985, 26, 1667.
61. A. Dubault, B. Deloche and J. Herz, Macromolecules, 1987, 20, 2096.
62. A. Dubault, B. Deloche and J. Herz, Macromolecules, 1990, 23, 2823.
63. P.G. de Gennes, Scaling Concepts in Polymer Physics, Cornell University Press,
Ithaca, NY, USA, 1979.
64. J. Bastide and S.J. Candau in Physical Properties of Polymeric Gels, Ed., J.P.
Cohen Addad, John Wiley & Sons, Chichester, UK, 1996, p.143.
66. H. Pak and P.J. Flory, Journal of Polymer Science: Polymer Physics Edition,
1979, 17, 1845.
68. B. Deloche, A. Dubault and D. Durand, Journal of Polymer Science, Part B:

69. Y. Rabin and E.T. Samulski, Macromolecules, 1992, 25, 2985.
70. C. Malveau, P. Tekely and D. Canet, Solid State Nuclear Magnetic Resonance,
1997, 7, 271.
71. K. Baumann and W. Gronski, Kautschuk und Gummi Kunststoffe, 1989, 42, 383.
72. G. Simon, Polymer Bulletin (Berlin), 1991, 25, 365.
73. G. Simon and H. Schneider, Die Makromolekulare Chemie - Macromolecular.

Symposia, 1991, 52, 233.
74. P. Ekanayake, H. Menge, H. Schneider, M.E. Ries, M.G. Brereton and P.G. Klein,
75. V.M. Litvinov and A.A. Zhadanov, Doklady Physical Chemistry, 1985, 283, 811.
76. V.M. Litvinov, Polymer Science USSR, 1988, 30, 2250.
77. V.M. Litvinov and H.W. Spiess, Makromolekulare Chemie, 1991, 192, 3005.
593
78. V.M. Litvinov and H.W. Spiess, Makromolekulare Chemie, 1992, 193, 1181.
79. G. Kraus, Journal of Applied Polymer Science, Applied Polymer Science

Symposium, 1984, 39, 75.
80. V. Dessolle, E. Lafontaine, J.P. Bayle and P. Judeinstein in Better Ceramics

Through Chemistry VII, Materials Research Society Symposium Proceedings,
1996, 435, 475.
81. S. Valic, P. Sotta and B. Deloche, Polymer, 1999, 40, 989.
82. M. Botev, R. Neffati and J. Rault, Polymer, 1999, 40, 5227.
83. K. Rosenke, H. Sillescu and H.W. Spiess, Polymer, 1980, 21,757.
84. D. Hentschel, H. Sillescu and H.W Spiess, Polymer, 1984, 25, 1078.
85. E. Okoroafor and J. Rault, Journal of Polymer Science: Polymer Physics Edition,
1991, 29, 1427.
86. J. Rault and H.M. Le Huy, Journal of Macromolecular Science B, 1996, 35, 89.
87. J. Rault, C. Macé, P. Judeinstein and J. Courtieu, Journal of Macromolecular

Science B, 1996, 35, 115.
88. V. Dessolle, J.P. Bayle, J. Courtieu and P. Judeinstein, Journal of Physical

Chemistry, 1999, 103, 2653.
89. N.J. Clayden, C. Nijs and G. Eeckhaut, Macromolecules, 1998, 31, 7820.
90. A.D. Meltzer, H.W. Spiess, C.D. Eisenbach and H. Hayen, Makromolekulare
Chemie, Rapid Communications, 1991, 12, 261.
91. J.A. Kornfield, H.W. Spiess, H. Nefzger, H. Hayen and H. Eisenbach,

92. A.D. Meltzer, H.W. Spiess, C.D. Eisenbach and H. Hayen, Macromolecules,
1992, 25, 993.
93. C.D. Eisenbach in Integration of Fundamental Polymer Science and Technology,

Ed., L.A. Kleintjens and P.J. Lemstra, Elsevier Applied Science Publishers,
Amsterdam, The Netherlands, 1986, p.317.
94. C. Hilger and R. Stadler, Makromolekulare Chemie, 1990, 191, 1347.
594
95. A. Dardin, C. Boeffel, H.W. Spiess, R. Stadler, V. Abetz and E.T. Samulski,
Polymer Materials Science and Engineering, 1994, 71, 248.
96. A. Dardin, C. Boeffel, H.W. Spiess, R. Stadler and E.T. Samulski, ACS
Symposium Series, 1995, 597, 190.
97. A. Dardin, C. Boeffel, H.W. Spiess, R. Stadler and E.T. Samulski, Acta
Polymerica, 1995, 46, 291.
98. V. Abetz, A. Dardin, R. Stadler, J. Hellmann, E.T. Samulski and H.W. Spiess,
Colloid and Polymer Science, 1996, 274, 723.
99. P.G. de Gennes, CR Academy of Science Paris, Series B, 1975, 281, 101.
100. P.G. de Gennes, CR Academy of Science Paris, Series II B, 1997, 324, 343.
101. P. Esnault, D. Galland, F. Volino and R. B. Blumstein, Macromolecules 1989, 22,

3734.
102. S. Valic, B. Deloche, Y. Gallot and A. Skoulios, Polymer, 1995, 36, 3041.
103. M. Zeghal, B. Deloche, P-A. Albouy and P. Auroy, Physical Review E, 1997, 56,
5603.
104. S. Rivillon, P. Auroy and B. Deloche, Physical Review Letters, 2000, 84, 499.
105. P.T. Callaghan, M.L. Kilfoil and E.T. Samulski, Physical Review Letters, 1998,
81, 4524.
106. M. Klinkenberg, P. Blümler and B. Blümich, Journal of Magnetic Resonance A,

1996, 119, 197.
108. M. Corti, L. Pavesi, A. Rigamonti and F. Tabak, Physical Review A, 1991, 43,
6887.
109. S. Jahromi, Macromolecules, 1994, 27, 2804.
110. R.E. Diehl, Journal of Chemical Physics, 1968, 48, 831.
111. M. Botev, P. Judeinstein, R. Neffati and J. Rault, Macromolecules, 1996, 29,

8538.
595
112. J.L. Hutchison, N.S. Murthy and E.T. Samulski, Macromolecules, 1996, 29,
5551.
113. L.S. Loo, R.F. Cohen and K.K. Gleason, Science 2000, 288, 119.

Polymers, Academic Press, London, UK, 1994.
596
Abbreviations
Abbreviations
1D Two-dimensional
1Q Single-quantum
2D One-dimensional
2
D-NMR Deuteron solid-state NMR
2
H 2O Deuterated water
2Q Double-quantum
3Q Triple-quantum
9BEN 9 borobicyclo [3, 3, 1] nonane
A1 Amide 1
ABS Acrylonitrile-butadiene-styrene terpolymer
AC Accelerator
ACM Acrylate rubber
ACN Acrylonitrile
AFM Atomic Force Microscopy
Al-CSM Aluminium salt of chlorosulfonate polyethylene
ALMA-d5 Allyl-d5 methacrylate
Al-m-EPDM Aluminium salt of maleated EPDM
AN Acrylonitrile
AO 1-Allyl oxyoctane
APT Attached Proton Test
ASTM American Standards for Testing and Materials
ATR Attenuated total reflectance spectroscopy
ATR-IR Attenuated total reflectance-IR spectroscopy
B/S Butadiene/styrene
BABA Back-to-back
597
BDO 1, 3-Butadiene
BET Brunauer, Emmett and Teller
BHT Butylated hydroxytoluene
BPO Benzoyl peroxide
BR Polybutadiene rubber
Bt Benzothiazole
CB Carbon black
CBR Cyclised polybutadiene rubber
CBS N-cyclohexyl-2-benzothiazole sulfenamide
CC Crosslinks
CHD Cis-1,4 hexadiene
CIIR Chlorobutyl rubber
CK Co-operative kinematics
CM Chlorinated polyethylene
CNR Chlorinated natural rubber
COLOC Chemical shift correlation
COSY Correlation Spectroscopy
CP Cross Polarisation
CPMG Carr, Purcell, Meiboom and Gill
CR Chloroprene rubber
CRAMPS Combined Rotation and Multiple Pulse Spectroscopy
CSDPF Carbon-silica dual phase fillers
CSM Chlorosulfonated PE
CT Cold trap
DAIP Diallylisophthalate
DANTE Delays alternating with nutations for tailored excitations
DAO 1,8-diallyl oxyoctane
DAP Diallylphthalate
DATP Diallylterephthalate
DBM Dibutyl maleate
DBTDL Bibutyl tin dilaurate
DBTM Dibutyl tin maleate
DCDPH Selectively hydrogenated DCPD
DCE Dipolar correlation effect
598
Abbreviations
DCP Dicumyl peroxide

DCPD Dicyclopentadiene
DD Dipolar decoupling
DEPT Distortionless enhancement by polarisation transfer
DETA Dielectric thermal analysis
DIN Deutsches Institut für Normung
DMA Dynamic mechanical analysis
DMS Dimethyl siloxane
DOS Disk operating system
DPNR Deproteinised NR
DQ Double quantum
DRIFTS Diffuse reflectance IR-FT Spectroscopy
DSC Dynamic scanning calorimetry
DTA Differential thermal analysis
DTG Differential thermal gravimetry
DTGS Deuterated triglycine sulfate
EAA Polyethylene-co-acrylic acid
EBA Polyethylene-co-butyl acrylate
ECO Epichlorohydrin
EGP Evolved gas profile
EHA 2-Ethyl hexyl acrylate
EN Chain entanglement
EN/TE Temporary chain entanglements
EN/TR Trapped chain entanglements
ENB 5-Ethylidene-2-norbornene
ENBH 2-Ethylidene norbonane
ENR Epoxidised natural rubber
ENVP EPDM-g-NVP
EP Ethylene-propylene copolymer(s)
EPDM Ethylene-propylene diene terpolymer(s)
EPM Ethylene-propylene copolymer(s)
EPR Ethylene-propylene rubber
ER Polyurethane-epoxide resin
ESBS Epoxidised butadiene-styrene block copolymer
599
ESO Epoxidised soya bean oil

EV Efficient vulcanisation
EVA Ethylene-vinyl acetate copolymer
EVS Polyethylene-co-vinyl trimethoxy silane
FDP Farnesyl diphosphate
FGP Functional gas profiles
FID Flame ionisation detection
FIND Free induction decay(s)
FRLS Fire-resistant low smoke
FT Fourier transform
FT-IR Fourier transform-infrared spectroscopy
GC Gas chromatography
GHPD Gated high power decoupling
GMA Glycidyl methacrylate
GPC Gel permeation chromatography
HBR Hydrogenated butadiene rubber
HCR Hydrogenated CR
HD 1,4-hexadiene
HEMA 2-hydroxy ethyl methacrylate
HENR Hydrogenated epoxidised natural rubber
HETCOR Heteronuclear chemical shift correlation
HMQC Heteronuclaer multiple-quantum coherence
HNBR Hydrogenated nitrile rubber
HNR Hydrogenated natural rubber
HOFR Heat, oil and fire resistant
HPLC High pressure liquid chromatography
HPV High pressure vulcanisation
HQEE Hydroquinone bis (b-hydroxyethyl) ether
HRTG-EI/SI QMSHigh resolution thermogravimetry–electron impact/soft ionisation
quadrupole mass spectrometry
HSBR Hydrogenated styrene butadiene rubber
HTNR Hydroxyterminated liquid NR
HXSBR Hydrogenated carboxylated styrene butadiene rubber
ICTA International Confederation for Thermal Analysis
600
Abbreviations
IDP Isopentenyl diphosphate

IIR Polyisoprene rubber
IIRP International Institute of Synthetic Rubber Producers
IPN Interpenetrating network(s)
IPNF Interpenetrating network foam
iPP Isotactic polypropylene
IR Infra red
ISAF Intermediate super abrasion furnace
ITPE Ionic thermoplastic elastomer
LDPE Low density linear polyethylene
LDPE-g-PNVP LDPE grafted with 2-N-vinyl pyrrolidone
LPA Copolymer of polyamide
MA Maleic anhydride
MAS Magic angle spinning
MBS 2(4-Morpholinothio) benzathiazole
MBT Mercaptobenzothiazole
MDI Methylene bis-4-phenol isocyanate
MDPE Medium density polyethylene
MDSC Modulated differential scanning calorimetry
MEPDM Maleated EPDM
MMA Methyl methacrylate
Mn Molecular weight
MNB 5-Methylidene-2-norbonene
MOR N-oxydiethylene-2-benzothiazole sulfenamide
MP Melting point
MPTES Mercaptopropyl triethoxysilane
MQ Multiple quantum
MRI Magnetic resonance imaging
MS Mass spectrometry
MSA Maleic anhydride grafted polypropylene
m-SEBS Maleated styrene-ethylene-butylene-styrene block copolymer
MS-PCA Mass spectrometry with principal component analysis
MW Molecular weight
Na-m-SEBS Sodium salt of maleated styrene ethylene butylene styrene
601
Na-s-m-SEBS Sodium salt of sulfonated maleated SEBS

NBR Nitrile-butadiene rubber
NIST National Institute of Science and Technology
NMA N-methylol acrylamide
NMR Nuclear magnetic resonance
NMR-MOUSE NMR-mobile universal surface explorer
NOE Nuclear Overhauser effect
NOESY NOE spectroscopy
NR Natural rubber(s)
NVP N-vinyl pyrrolidine
OIT Oxidative induction time
OPD Optical path difference
PA Photoacoustic
PA-12 Polyamide-12
PA-FTIR Photoacoustic-Fourier transform-Infrared spectroscopy
PA-PEG Polyamide–polyethylene glycol
PAS Photoacoustic spectroscopy
PB Poly(1,4-butadiene)
PBT Polybutylene terephthalate
PC Polycarbonate
PCA Principal component analysis
PCB Polychlorinated biphenyl(s)
PCL Polycaprolactone
PCP Polchloroprene
PDMS Polydimethylsiloxane
PDSC Pressure differential scanning calorimetry
PDXL Polydioxolane
PE Polyethylene
PED Deuterated PE
PEEK Polyether ketone
PEG Polyethylene glycol
PEGDA Poly(ethylene glycol)diacrylate
PEgDBM PE grafted to dibutyl maleate
PEH Normal PE
602
Abbreviations
PEO Polyethylene oxide

PEXa Peroxide crosslinking system
PEXb Silane crosslinking system
PEXc Electron beam crosslinking system
PF Phenolic formaldehyde
PFG Pulsed field gradients
PFGE Pulsed field gradient echo spectroscopy
PFPA Pentafluorophenyl acrylate
PFPMA Pentafluoro phenyl methacrylate
phr Parts per hundred rubber
PIB Polyisobutylene
PMDETA Penta-methyl diethylene triamine
PMMA Polymethylmethacrylate
PP Polyopropylene
PPA Polypentenamer
PPL Poly-b-propiolactone
ppm Parts per million
PPO Poly(propylene oxide)
PR Alkyl phenol formaldehyde resin
PS Polystyrene
PTA Pulsed thermal analysis
PTFE Polytetrafluoroethylene
PTMG Poly(tetramethylene ether glycol)
PTMO Polytetramethyleneoxide
PTOP Para-tert-octyl phenolic
PU Polyurethane(s)
PUA Polyurethane acrylate
PUF Polyurethane foam
PVC Polyvinyl chloride
PVME Polyvinyl methyl ether
PyGC-FID pyrolysis gas chromatography – FID
QMS Quadrupole mass spectrometry
R Dichroic ratio
rf Radiofrequency
603
RG Rosencwaig and Gersho theory

RX Receiver signal
S/N Signal-to-noise ratio
SAN Styrene acrylonitrile
SBR Styrene-butadiene rubber(s)
SBS Styrene-butadiene-styrene rubber
SEBS Styrene-ethylene-butylene-styrene block co-polymer
SEC Size exclusion chromatography
SEM Scanning electron microscopy
SEV Semi-efficient vulcanisation
SGP Specific gas profiles
SI QMS Soft ionisation quadrupole mass spectrometry
SIS Styrene-butadiene-styrene
SPRITE Single-point ramped imaging with T1 enhancement
SQ Single quantum
SRF Semi-reinforcing furnace carbon black
STA Simultaneous thermal analysis
TAC Triallylcyanurate
TBBS N-tertiary butyl-2 benzothiazole sulfenamide
TBSI N-tertiary butyl-2 benzothiazole sulfenimide
TBTCl Tributyltin chloride
TBTLi Tributyltin lithium
TD Thermal desorption
TDI Toluene diisocyanate
TeCEA Tetrachloroethyl acrylate
TESBT Triethoxysilyl benzothiazole
TESPT Bis(triethoxysilyl propyl) tetrasulfide
Tg Glass transition temperature
TG Thermogravimetry
TGA Thermogravimetric analysis
THF Tetrahydrofuran
Tm Melting temperature
TMA Thermomechanical analysis
TMBA Thermo molecular beam analysis
604
Abbreviations
TMPS Poly(tetramethyl-p-silphenylene-siloxane)
TMPTA Trimethylol propane triacrylate
TMPTMA Trimethylol propane trimethacrylate
TMTD Tetramethyl thiuram disulphide
TOCSY Total correlation spectroscopy
ToFMS Time-of-flight MS
TPE Thermoplastic elastomer(s)
TPEE Thermoplastic polyester elastomers
TPGDA Tripropylene glycol diacrylate
TPP Triphenyl phosphate
TPU Thermoplastic polyurethane
TRIM Trimethylolpropanetrimethacrylate
TVA Thermal volatilisation analysis
TX Transmitter signal
U4A 4-(3,5-Dioxo-1,2,4-triazolidin-4-yl) benzoic acid
UV Ultraviolet
VAC Vinyl acetate
VI Video imaging
VNB 5-Vinyl-2-norbornene triallylcyanurate
VTEO Vinyl triethyl silane
VTMO Vinyl trimethyl silane
WISE Wide line separation
WLF Williams-Landel-Ferry
Xe Xenon
XNBR Carboxylated nitrile rubber
XPS X-ray photo-electron spectroscopy
XRD X-ray diffraction
XSBR Carboxylated styrene butadiene rubber
XSBR Hydogenated, carboxylated SBR
ZDMC Zinc dimethyl dithiocarbamate
Zn-EMA Zinc salt of ethylene-methacrylic acid copolymer
Zn-m-EPDM Zinc salt of maleated ethylene-propylene diene monomer rubber
Zn-m-HDPE Zinc salt of maleated high density polyethylene
Zn-PEA Zinc salt of polyethylene-co-acrylic acid
605
Zn-PPA Zinc salt of polypropylene-co-acrylic acid

Zn-s-EPDM Zinc salt of sulfonated EPDM
Zn-s-m-EPDM Zinc salts of sulfonated maleated EPDM
Zn-XNBR Zinc salt of carboxylated nitrile rubber
606
Contributors
Jean Pierre Cohen Addad

Université Joseph Fourier - Grenoble, Laboratoire de Spectrometrie Physique, BP 87,
38402 St Martin d’Heres Cedex, Grenoble I, France
Jan C.J. Bart

DSM Research, PO Box 18, 6160 MD Geleen, The Netherlands
Bernhard Blumich
Institut für Technische Chemie und Makromolekulare Chemie, RWTH, Worringerweg
1, D-52056, Aachen, Germany
Prajna P. De
Rubber Technology Centre, Indian Institute of Technology, Kharagpur 721302, West
Bengal, India
Bertrand Deloche
Laboratoire de Physique des Solides (CNRS UMR 8502), Université de Paris-Sud, 91405
Orsay Cedex, France
Dan E. Demco
Institut für Technische Chemie und Makromolekulare Chemie, RWTH Aachen,
Worringerweg 1, D-52056, Aachen, Germany
Herman G. Dikland
DSM Elastomers, Global R&D, PO Box 1130, 6160 BC Geleen, The Netherlands
Martin van Duin

Eng Aik Hwee

Ansell Shah Alam Sdn Bhd, Lot 16, Persiaran Perusahaan, Seksyen 23, 40000 Shah
Alam, Selangor Darul Ehsan, Malaysia
607
Siegfried Hafner
Varian Deutschland GmbH, Alsfelder Strasse, 64289 Darmstadt, Germany
Jack L. Koenig
Case Western Reserve University, Department of Macromolecular Science, Kent Hale
Smith Building – Room 212, 10900 Euclid Avenue, Cleveland, OH 44106-7202, USA
Victor M. Litvinov
DSM Research BV, PO Box 18, 6160 MD Geleen, The Netherlands
Dallas D. Parker
Case Western Reserve University, Department of Macromolecular Science, Kent Hale
Smith Building – Room 212, 10900 Euclid Avenue, Cleveland, OH 44106-7202, USA
James R Parker
PPG Industries Inc., Chemicals Group - Technical Centre, 440 College Park Drive,
Monroeville, PA 15146, USA
Charles Raemaekers
Jitladda Sakdapipanich
Department of Chemistry, Faculty of Science, Mahidol University, Rama VI Road,
Bangkok 10400, Thailand
Nikhil K. Singha
Department of Chemical Engineering, Division of Polymer Chemistry, Eindhoven
University of Technology, PO Box 513, 5612 MB Eindhoven, The Netherlands
Paul Sotta
Laboratoire de Physique des Solides (CNRS UMR 8502), Université Paris-Sud, 91405
Orsay Cedex, France
Hans W. Spiess
Max-Planck Institut für Polymerforschung, Postfach 3148, D - 55021 Mainz, Germany
608
Contributors
Yasuyuki Tanaka
Department of Chemistry, Faculty of Science, Mahidol University, Rama VI Road,
Bangkok 10400, Thailand
Wiebren S. Veeman
Institute of Chemistry, Gerhard-Mercator-University, Duisburg, Germany
Andrew K. Whittaker
Centre for Magnetic Resonance, University of Queensland, QLD 4072, Australia
609
610
Index
A Anisotropic interactions 322, 519–522,

524, 529, 558, 559–560
ABS (acrylonitrile–butadiene–styrene) 21– Anti-Markovnikov rule 142–143
22, 26 AO (1-allyloxyoctane) 230–232
Absorptivity 82 APT (attached proton test) 329
Accelerators ASTM standards
IR spectra 24–25 D297-93 14
sulfur vulcanisation 215, 322–323, D3677 83
328–329, 330–333, 336, 506–507 E1131 14
Acid-catalysed cyclisation 420–421 PA–FTIR 64
Acrylonitrile, determination in NBR 61– temperature calibration 13
62, 88–90
ATR see attenuated total reflectance
Acrylonitrile–butadiene–styrene (ABS), (ATR) spectroscopy
thermal degradation 21–22, 26
Attached proton test (APT) 329
Activators, sulfur vulcanisation 215, 323
Attenuated total reflectance (ATR)
Additives 27 spectroscopy 81, 104–108, 109, 216,
Adhesion 111–113 227–228
Adsorption
filler surface 354, 371–372, 375–377,
378–379 B
mineral aggregates 293–294, 315–317
Balata
Aflas FA-150C rubber, thermal
degradation 21–22 cyclisation 142
Ageing IR spectra 88
elastomers 264–265, 274–275, 348 isomers 88, 403, 431
NMR imaging 264–265, 274–275 BDO (1,4-butanediol polyurethanes)
189–192
see also photoageing; thermo-
oxidative degradation Beer–Lambert’s law 82
Aggregated polymers 294 Bell, Alexander Graham 49
Alkyl phenol formaldehyde resin (PR), IR Bending vibrations 79, 170–172
spectra 94–95 BHT (butylated hydroxytoluene) 237
Alkylthiols, EPDM crosslinking 222–223 Biocompatibility 557
1-Allyloxyoctane (AO) 230–232 Biosynthesis 402, 403
Aluminium, polyethylene laminates 178– Blends
13
179 C NMR 340
611
crosslink density 502–510 hydrogenation 129–131

1
H NMR transverse magnetisation hydrosilylation 142–144
relaxation 384 isomers 83, 88, 90, 139, 212–213
interfaces 271, 354 NMR imaging 500–501
ionomeric 151–153, 154–155 peroxide-curing 227, 338–339
IR spectra 92–98, 111–113, 180, 182– sulfur vulcanisation 212–213, 333–338
183 swollen 500–501
mixing process 264 1,4-butanediol (BDO) polyurethanes
NMR imaging 264, 271, 272, 338, 189–192
384 Butylated hydroxytoluene (BHT) 237
orientation effects 71 N-t-butyl-2-benzothiazole sulfenamide
polypropylene (TBBS) 330–333, 336–337
EPDM 482–485 Butyl rubber
EPM 465–474, 482 branch-point 426–427
PVC 180, 182–187 1
H NMR 426–427
self-crosslinking 96–98, 183, 186 IR spectra 94–95, 127
thermoplastic elastomers 188, 189 unsaturation 405, 406
Block copolymers
EPM–polypropylene 465–474
IR spectra 98, 100 C
network structure 293 Cable sheathing compounds 172, 175
NMR 416, 419 Calibration
PBT/PTMO 486–487 IR spectroscopy 82
polyamide–polyether 585–587 thermogravimetric analysis 13–14
see also copolymers Carbon black
Bloom (surface layer) 66–68 IR spectra 198–199
Bohr equation 77 properties 465
Bond strength 80 Carbon-black-filled rubbers 81
BR see butadiene rubber ATR–FTIR spectra 105–106, 107
Branch-points 168–169, 426–429, 439 chain adsorption 317, 371–372
Bromination, IR spectra 137 13
C NMR 341–347
Butadiene rubber (BR) chlorosulfonated polyethylene 58–60
chain segment diffusion 499 crosslink density 507
crosslinking 227, 335–336, 337, 339 diene rubbers 582, 584
crosslink density 335–336, 506–507 DRIFTS 104, 105
cyclohexane self-diffusion 497–498 EPDM 369–370, 372, 373, 463, 465,
IR spectra 126, 127 474–480
chlorination 135–138 free radicals 368, 369
cyclisation 140–142 1
H NMR transverse magnetisation
epoxidation 146 relaxation 265–266, 369–374, 475
hydrocarboxylation 146 IR spectra 198–199
612
Index
NBR 504 end-linked networks 299, 561, 563,

NMR-MOUSE 278–280 575, 577–579
NR 341–344, 582, 583 entanglements 309–315, 354, 360,
PA–FTIR spectra 57–60 366–367, 377–378, 505–506, 561
polymer-filler interactions 474–475 extension 557–558
SBR 24–25, 57–58, 372, 374 free chains 570–571, 579
sulfur vulcanisation 336–337, 341– linkage effect 302
347, 582, 583 long-chain dynamics 313–315
Carbon–carbon unsaturation 126, 128– molecular mass distribution 364–366
129 network chain density 333, 334, 344,
Carbon–silica dual phase fillers (CSDPF) 346–347, 355–360, 367
108 polymeric networks 293–307
Carboxylated nitrile rubber (XNBR) scission 424–426
IR spectra 85, 96–97, 99, 194–198 short-chain dynamics 312
self-crosslinking blends 96–98, 99 strand length dependence 301–302
Carboxylated styrene–butadiene rubber swelling effects 302, 303, 304, 306,
(XSBR), IR spectra 131, 133 344, 346–347
Carboxylate ionomers, FTIR spectra 147– telechelic chains 299, 423–424, 439
149 terminal groups 424–426, 436–437,
Carboxynitroso rubber, thermal 438
decomposition 28 thermal behaviour 301–302
Chain dynamics 312–315, 525 see also molecular structure
1
H NMR transverse magnetisation Characterisation
relaxation 313, 355–360, 361, 371, chemical characterisation 321–400
374–376, 379 chemically modified rubbers 125–165
2
H NMR 562–566 elastomers 1–47
polymer solutions 494–496 rubbery materials 125–165
segmental mobility 495–496, 498– chemical and physical networks 353–
499, 511–512, 520, 524–525, 530, 400
538, 548–549 molecular behaviour 557–558
time scale of motions 563–566 soft polymeric matter 291–320
Chain structure structural 17
adsorption 293–294, 315–317, 354, viscoelastic materials 519–555
371–372, 375–377, 378–379 vulcanisates 321–352
branch-points 168–169, 426–429, 439 Chemical analysis
conformation 504 chemical composition distribution
crosslinking 302–306, 493 410–413
dangling chains 563, 564, 571 elastomers 402–422
disentanglement 361 (multi) hyphenated TGA 17
elongation 296–298, 567–570, 573 network structure 354–355
end-groups 404, 405, 422–429, 436– Chemical degradation 442
437, 438 Chemically modified rubbers
613
IR spectroscopy characterisation 125– CK (co-operative kinematics theory) 494–

165 496
13
NMR 419–422 C NMR spectroscopy
Chemical networks see crosslinking binary blends 340
Chemical shift carbon-black-filled rubbers 341–347
13
C NMR spectroscopy 325–326, 328, chemical shift 325–326, 328, 524, 543
524, 543 crosslinking 239, 329–330
Fourier transform NMR 508 crosslink density 329–330, 353, 508–
NMR imaging 270–271, 277, 521 510
129
Xe NMR spectroscopy 457–458, elastomers 402–403
469–470, 476–480 EPM 405–408, 413–414, 508–509
Chemical structure EVA 428–429
NMR 324–326 1
H–13C heteronuclear coupling 543–
elastomers 401–404 547
NR 433–436 high-pressure vulcanisation 339
cis-polyisoprene vulcanisate 330–333 high-resolution 510
Chlorination latex state 446–448
BR 135–138 linewidths 508–510
chlorocarbenes 136 magic-angle spinning 329–330, 336,
IR spectra 135–138 510, 535–536
Chlorocarbenes, chlorination 136 NBR 416, 417–418
Chloroprene rubber (CR) NR 433–434
crystallinity 91 PB 414–415, 495–496
hydrogenation 134 peroxide vulcanisates 338–339
IR spectra 83, 85, 86, 90, 126, 127, polyisoprenes 340, 344, 347, 494–495
134 naturally occurring 430–432
isomers 90–91 polyurethane–urea 416, 419
Chlorosulfonated polyethylene (CSM) SBR 408–410
IR spectra 85, 96–97, 99 silica-filled rubbers 347
PA–FTIR spectra 58–60, 102–104 silica-filled silicone rubbers 376
rubber–filler interaction 102–104 sulfur vulcanisates 322
self-crosslinking blends 96–98, 99 BR 333–338
Cchromatography see high-pressure NR 324–333
liquid chromatography SBR 338
Cis–trans isomerisation swollen state 444–446, 508–510
BR 139, 212–213, 334–335, 336, terminal groups 424–425
338–339 thermo-oxidative degradation 348
IR spectra 139, 212, 214 WISE experiment 543–547
polyisoprene 323–324, 328, 338–339, Cohen–Addad–Sotta model 258–259
341, 343, 347, 403, 429–435 Cold traps 3, 4–5, 12
see also isomers Combinatorial chemistry 510
614
Index
Combined rotation and multiple-pulse inter-crosslink components 505

spectroscopy (CRAMPS) 322 molecular structure alteration 230
Compatibilisation, PCP/EPDMgDBM 92–94 network structure 292–294, 296–307,
Composites see carbon-black-filled rubbers 345–347
1
Composition see chemical analysis H NMR 220, 223–224, 239
Concentration, IR spectroscopy 82 NMR
Conveyor belts, quality control 281 imaging models 257–259
Co-operative kinematics (CK) theory reviews 493
494–496 NR 226, 326
Copolymers optical spectroscopy 207–246
high-resolution NMR 401–448 peroxide-curing 225–227
chemical composition 402–422 co-agents 229–237, 238
sequence distribution 402, 413–419, polydiene rubbers 210–211, 215–216
420, 444, 446 polyethylenes 177–178, 179
IR spectra 92–96 silane-grafted 174, 175
quantitative analysis 88–90 radiation 156, 177–178
phase-separated systems 547 randomly crosslinked chains 302–306,
see also block copolymers 561
Coupling agents 104, 108, 110 rubbers 209–210
Covulcanisation 269–271 SBR 531–533, 541–542, 545
CR see chloroprene rubber self-crosslinking blends 96–98, 183,
CRAMPS (combined rotation and 186
multiple-pulse spectroscopy) 322 sulfur vulcanisation 323
Crosslinking swelling 491–493
BR 227, 335–336, 337 thermoplastic rubbers 168
characterisation limits 321 see also network structure
13
C NMR 239, 545 Crystallinity
crosslink density 234, 271–273 chloroprene rubber 91
blends 502–510 EPM 307–308
BR 335–336, 506–507 ethylene–octene copolymer 382–383
13
C NMR 329–330, 446, 508–510 polyethylene fibres 459, 460
double-quantum NMR 541–543
PVC 183
EPDM 224
EPM 508–509 see also semi-crystalline materials
1
H NMR 353, 446, 506–509, 509– Crystallisation
510 crystallites 354, 584–585
double-quantum MAS 541–543 kinetics 308–309
NMR-MOUSE 279–280 NMR 307–309
EPDM 208–209, 216, 222–225, 237– PE 169–172
239 CSDPF (carbon–silica dual phase fillers)
properties relationships 224–225 108
filled rubbers 377 CSM see chlorosulfonated polyethylene
615
Cumyloxy fragments 30, 31 Depth profiling 66–70

Curing see also thermal diffusivity
conditions 353, 361, 386 Desulfuration 323
crosslink density 508 Deuterium NMR see 2H NMR
curing points 421 Dialkenylsulfides, Fourier transform
(multi) hyphenated TGA 18 Raman spectra 217–219, 223
network topology 353, 361 Diallylisophthalate (DAIP) 234–235
physical properties 353, 354 1,8-diallyloxyoctane (DAO) 230–232
Cyclisation Diallylphthalate (DAP) 234–235
acid-catalysed 420–421 Diallylterephthalate (DATP) 229, 232–235
BR 140–142 Dibutyl maleate (DBM)
IR spectra 137, 139–142 EPDMgDBM 92–94
Cycloaliphatic epoxide–ENR–GMA, grafting 92–94, 172, 174
FTIR spectra 100–102 PEgDBM 92–93
Cyclohexane Dicumylperoxide (DCP) 226–227, 232,
self-diffusion in BR 497–498 233, 338–339
swollen rubbers 266–267, 497–498, Dicyclopentadiene (DCPD) 207, 228–229
501 Diene rubbers
Cyclopolymerisation, peroxide-curing co- carbon-black-filled 582, 584
agents 235 crosslinking 209–210
cyclisation 140
epoxidation 146
D hydrosilylation 142, 143
Defects IR spectra 125–126, 127
elastomers 264–269 microstructure 126
1
H NMR transverse magnetisation oxidation 146
relaxation 265–269, 360–366 structure 401, 402–403
Deformation sulfur vulcanisation 323–324
low-density polyethylene 170–172 telechelic 423–424, 439
polymers 167–168 Differential scanning calorimetry (DSC)
rubbery materials 167–168, 557–558, and thermogravimetric analysis 1, 2–
567–572, 580, 581 3, 5–7, 10–11, 15–17, 29–30
sample deformation in NMR imaging mass spectrometry 2–3, 11, 15–17,
275–276 30, 31
uniaxially deformed model networks Differential thermal analysis (DTA)
566–572, 580, 581 and thermogravimetric analysis 2–3,
Degradation see chemical degradation; 5–6, 10–11, 15–17, 30–33
thermal degradation mass spectrometry 2–3, 6, 11, 15–17,
Delays alternating with nutations for 33–34
tailored excitations (DANTE) 509–510 Differential thermal gravimetry (DTG)
Deproteinisation 433–435 DTG curves 9, 14, 15
616
Index
and thermogravimetric analysis 2–3, network structure 354, 377–378

9, 14–16, 36 NMR 304–305
Diffuse reflectance infrared Fourier PDMS 305, 567, 576–579
transform spectroscopy (DRIFTS) 81, rubbery materials 557–558, 567–568
104, 105 soft polymeric matter 291–292, 295,
see also Fourier transform infrared 296, 303
spectroscopy vulcanisation process 321
4,4′-diphenylene methane diisocyanate Elastomer products 265–269
(MDI), polyurethanes 189–190
Elastomers
Dipolar-correlation effect (DCE) 526
ageing 264–265, 274–275, 348
Dipolar decoupling (DD) 322, 522–524,
branch-points 426–429, 439
544, 547
chemical composition 402–422
Dipole–dipole interactions
13
copolymers 401
C NMR 322
defects 264–269
decoupling 322, 522–524, 544, 547
end-groups 404, 405, 422–426, 436–
double-quantum filtering 540–543
1
437, 438
H magnetisation-exchange
filled 582–584
spectroscopy 530
1
heterogeneities 264–269
relaxation 257–258, 281, 361 high-resolution NMR 401–456
mechanical load 265
multi-quantum filtering 259–264,
538–543 mixing process 264
NMR imaging 272, 297, 315, 521– (multi) hyphenated thermogravimetric
524 analysis techniques 1–47
PE 174 NMR imaging 247–289
models of transverse relaxation 257–
soft polymeric matter 306, 310, 315
259
solid-like properties 257, 520
peroxide-curing 225–227
Disproportionation 424–425
reinforcement 315
Distortionless enhancement by
rubber–filler interaction 102–108
polarisation transfer (DEPT) 328, 331,
334 semi-crystalline 381–383
Double-quantum NMR 540–543 structure determination 401–404
DSC see differential scanning calorimetry sulfur vulcanisation 209–210
DTA see differential thermal analysis thermoplastic 168, 188–199, 548–
549, 585–588
Dynamic mechanical analysis (DMA)
358–9 vulcanisation 264
129
Xe NMR spectroscopy 457–489
Electromagnetic spectrum 77–78
E
Electron beam irradiation, IR spectra
Einstein relationship 468 156–157, 175–176
Elasticity Elongation
2
H NMR 557–558, 567–568, 575–579 chain structure 296–298, 567–570, 573
617
PE films 169–170 373, 463, 465, 474–480

see also stretching chain length distribution function 365
Emulsions, SBR 62–63 crosslinking 208–209, 216, 222–225,
End-groups, NMR 404, 405, 422–429, 237–239
1
436–437, 438 H NMR transverse magnetisation
ENR see epoxidised natural rubber relaxation 386–387
Entropy reduction 296–297, 345 IR spectra 84, 89, 93
EPDM see ethylene–propylene–diene grafting 92–94, 95, 156–157
terpolymer ionomeric blends 151–153, 154–155
EPM see ethylene–propylene copolymers termonomer content 90, 91
Epoxidation, IR spectra 146 mechanical properties 224–225, 238
Epoxidised natural rubber (ENR) network structure 223–225
hydrogenation 132, 134 NMR imaging 254–255
IR spectra 85, 100–102, 104–108, 134 oil-extended 366–367
see also natural rubber peroxide-curing 209, 227–237
co-agents 230, 238
Epoxidised soyabean oil (ESO) 187
third monomer effects 227–229, 238
Epoxy titration method 405–406
polypropylene blends 482–485
Ethylene–methacrylic acid copolymer,
production 207–208
zinc salt 151–153, 154–155
properties 208
Ethylene–octene copolymer, crystallinity
382–383 SEC–NMR 410–413
Ethylene–propylene copolymers (EPM) sulfur vulcanisation 208–209, 216–
13 225, 238, 365, 386–387, 422
C NMR 405–408, 413–414, 508–
5-ethylidene-2-norbornene conver-
509
sion 219–221, 224, 238
crosslink density 508–509
thermal decomposition 23–24, 30, 31
crystallinity 307–308 129
Xe NMR 458, 463–464, 474–480,
IR spectra 89, 90, 180, 181
482–486
NMR 325, 405–410, 413–414 Ethylene vinyl acetate (EVA)
PA–FTIR spectra 64–65 branch-points 428–429
peroxide-curing 230–231, 232–234,
electron beam irradiation 156–157
235, 236
IR spectra 172, 173
polypropylene blends 465–474, 482
NMR 416, 419, 420, 428–429
properties 207
polyethylene copolymer 172, 173
propylene determination 90, 405–410,
2-ethylhexyl acrylate (EHA), 1H NMR
414, 469
transverse magnetisation relaxation 358–
sequence distribution 413–414, 424– 359, 361–363
425
129
5-ethylidene-2-norbornene (ENB) 207,
Xe NMR 465–474 216–217, 219–221, 224, 238, 410–413
Ethylene–propylene–diene terpolymer
EVA see ethylene vinyl acetate
(EPDM)
Evolved gas analysis, (multi) hyphenated
carbon-black-filled 369–370, 372, TGA 15, 17, 18
618
Index
Evolved gas profiles (EGP) 9, 21 see also diffuse reflectance infrared

Extinction coefficient 82 Fourier transform spectroscopy ;
infrared spectroscopy
Fourier transform NMR 248–251, 508,
F 528–529, 544
Far infrared region 78 Fourier transform Raman spectroscopy 211
Fibres, PE 459, 460 dialkenylsulfides 217–219
Filled rubbers sulfur vulcanisation 217–223, 224
chain grafting 379–381 see also Raman spectroscopy
elastomers 582–584 free radicals
filler clusters 265–266 carbon-black-filled rubbers 368, 369
1
H NMR transverse magnetisation macroradical side reaction suppression
relaxation 368–379 230–232
2
H NMR 582–584 sulfur vulcanisation 213, 335
network structure 377–378 Frequencies 77–78
rubber–filler interaction 102–108, Larmour frequencies 248–249, 295,
368, 371, 372, 374, 376, 380, 474– 297, 462, 521, 523, 526
475 stretching 79–80
see also carbon-black-filled rubbers; vibrational 78–80
silica-filled rubbers; silica-filled silicon FTIR see Fourier transform infrared
rubbers spectroscopy; photoacoustic Fourier
Films transform infrared spectroscopy
Kapton 70 Functional groups
PE 68–69, 169–170, 174–178 analysis 80
Fire resistant compounds 92–93 formation during milling 109, 111
Food industry, radiation processing 174
Fourier transform infrared spectroscopy
(FTIR) 81, 211 G
carboxylate ionomers 147–149 γ-irradiation
peroxide-curing co-agents 232, 233, low-density polyethylene film 176–178
234–236, 237, 238
NR 339
photoacoustic 49–76
PIB 426
polyethylenes
polyisoprene 503–504
aluminium laminates 178–179
low-density 170–172, 176–178 polymers 153
medium-density 178, 179 vulcanisation 339
PVC blends 181–182 Gas chromatography (GC)
sulfonated ionomers 149–151 and thermogravimetric analysis 3–5,
sulfur-vulcanised NR 214 11–12, 15–17
mass spectrometry 3–4, 11, 15–17,
and thermogravimetric analysis 2–4,
34–35
6–10, 11–12, 15–18, 19–25, 34–35
Gaskets, NMR imaging 265–266
619
Gas phase, 129Xe NMR spectroscopy 458– Hamiltonians 521, 522, 540
459 Helium atmosphere, photoacoustic
Gated high-power decoupling (GHPD) Fourier transform infrared spectroscopy
334 53–54
GC see gas chromatography Herschel, Sir William 77
Gels 291, 292, 299, 301–305, 310, 447 Heterogeneity
gelation 302–304, 307 elastomers 264–269
1
stretched 306 H NMR transverse magnetisation
swelling 491–492 relaxation 265–269, 360–366
Glass-to-rubber transition temperature molecular-scale 360
(Tg) 168, 279–280, 493, 520, 524–525, morphological 361
557, 562 network structure 353, 355, 360–366
Glassy materials, log (shear) modulus 167 spatial 273, 353, 361
Glycidyl methacrylate (GMA) 100–102 Heteronuclear correlation experiments
Goodyear 209 543–547
Gotlib–Fedotov–Schneider model 258, Hevea rubber
259 cyclisation 142
Gradient echoes isomers 88, 403
NMR imaging 253–255, 265, 271, NMR 436
462–463, 496–497, 498–500 High-density polyethylene (HDPE), IR
129
Xe NMR spectroscopy 462–463 spectra 194–198
Grafting High-performance liquid chromatography
dibutyl maleate 92–94, 172, 174 404
EPDM 92–94, 95, 156–157 High-pressure liquid chromatography
filler surface 379–381 331–332
IR spectra 92–96, 156–157, 172, 174 High-pressure vulcanisation (HPV) 339
NMR 421 High-resolution NMR
13
polyethylene 92–94, 172, 174, 175– C MAS NMR 510
176, 178 elastomer structural characterisation
PVC 188 401–456
radiation induced 178 applications 436–448
vinyl silanes 174 multinuclear 436–443
Guayule rubber, isomers 88 vulcanisate chemical characterisation
Gutta percha, isomers 88, 403, 429, 431 321–352
1
H magnetisation-exchange spectroscopy
529, 530–535
H 1
H NMR relaxation
Hahn-echo imaging 253–257, 258–259, butyl rubber 426–427
262, 265, 266–267, 268, 269–271, 274– CRAMPS techniques 322
275, 277–280, 282, 503, 527 crosslinking 220, 223–224, 239, 299–
Halogenation, IR spectra 135–138 306
620
Index
crosslink density 353, 506–508, 509– network density 356–360, 367

510 real-time experiments 386–387
double-quantum spectroscopy 540– NR 267–268, 279
543 quality control 281, 387–388
elastomers 402–403 real-time experiments 385–387
EPDM 410–413 residual dipolar coupling 524–526,
1
H–13C heteronuclear coupling 543– 527
547 rubbery material characterisation
linewidths 506–508, 509–510 353–400
multi-quantum spectroscopy 259–264, SBR 267–268, 279–280
538–543
semi-crystalline elastomers 381–383
nanometer scale probe 291
solvent effects 361–363, 499
NOESY–MAS spectroscopy 535–538
swollen state 354, 361–363, 378,
NR 433, 435
500–506
PB 314
temperature dependence 276, 356–
PDMS rubber bands 262–263 357, 375, 376
pseudo-solid spin-echoes 298–299, theory 248–249, 252, 255–259, 295,
310 297, 299
water/ice 321–322
thermo-oxidative ageing 274–275
WISE experiment 543–547
1
viscoelastic materials
H NMR transverse magnetisation ionic 383–384
relaxation molecular dynamics 526–527
blends 384 viscoelastic properties 365–366
chain dynamics 313, 355–360, 361, vulcanisation process 273–274
371, 374–376, 379 see also NMR imaging; NMR-
crosslink density 257–259, 502–506 MOUSE
decay curves 363–364, 366, 382, 502– 2
H NMR
505
chain extension 558
emulsions 385
deuterated probes 579–581
EPDM sulfur vulcanisation 386–387
elastic property correlation 575–579
filled rubbers 368–379
elastomers 436–437, 438
carbon-black-filled rubbers 265–266,
369–374, 475, 504 filled 582–584
silica-filled rubbers 378–381 thermoplastic 548–549, 585–588
silica-filled silicone rubbers 374–378 gels 306
multi-quantum spectroscopy 260–262 line shape analysis 527, 561–563,
network structure analysis 355–379 566–575, 580–581
chain entanglements 309–315, 366– mean-field effect 569–571, 575
367, 377–378, 505–506 model systems 561–579
heterogeneity and defects 265–269, molecular dynamics 559–561
360–366 orientation effects 560–561, 569, 571,
molecular mass distribution 364–366 573–575, 579–580
621
PB 436–437, 571–572, 577, 580–581, silica-filled rubbers 378

582, 584, 587–588 thermoplastic elastomers 548
PDMS 263, 527, 528, 561–572, 574– Hydroquinone bis (β-hydroxyethyl) ether
575, 576–579, 584, 585 (HQEE) 189–192
pseudo-solid spin-echoes 561–563 Hydrosilylation
quadrupolar interactions 527, 528, BR 142–144
559, 564–565, 567–571, 588 IR spectra 137, 142–144
rubbery materials 557–596 Hydroxylation, IR spectra 137
semi-crystalline polymers 584–585 2-hydroxyl ethyl methacrylate (HEMA)
silica-filled silicone rubbers 584 188
swollen state 580–581 Hypalon, self-crosslinking blends 97
theory 559–561 Hyphenated techniques see (multi)
time scale of motions 563–566 Hyphenated thermogravimetric analysis
uniaxially deformed model networks (TGA) techniques
566–572, 580, 581
viscoelastic materials 522, 527 I
Homogeneous polarisation field (B0) 247,
250, 257–258, 277, 282, 368 IIR see butyl rubber
Hydroboration, IR spectra 137, 144 Impurities
Hydrocarboxylation, IR spectra 146 (multi) hyphenated TGA 27
Hydroformylation NMR small signals 422, 477
IR spectra 137, 144–145 Infrared spectroscopy (IR) 210–211
polypentenamer 144–145 adhesion 111–113
Hydrogen abstraction 424–425 applications 82–113, 125–165
Hydrogenation blends 92–98, 111–113, 180, 182–183
BR 129–131 bromination 137
CR 134 carbon black 198–199
degree of hydrogenation 128–129 chemically modified rubbers 125–165
ENR 132, 134 chlorination 135–138
IR spectra 126, 127–134, 137, 144– cyclisation 137, 139–142
145 degradation 113
NBR 126, 127–129, 421 halogenation 135–138
NR 131–132 hydroboration 137, 144
SBR 131–132 hydrocarboxylation 146
Hydrogen bonding hydroformylation 137, 144–145
PB 587–588 hydrogenation 126, 127–134, 137,
plasticised PVC 168 144–145
polyureas 192, 193 hydrosilylation 137, 142–144
polyurethanes 100, 190–192 hydroxylation 137
PVC 174 ionomers
rubbery materials 354 blends 151–153, 154–155
622
Index
formation 138, 147–151 374, 376, 380, 474–475

sulfonated 149–151, 198–199 Interfaces, coupled instruments 6, 10
irradiation 153, 156–157, 174–176, Interlaboratory reproducibility 13–14
177 International Confederation for Thermal
isomerisation 139, 212, 214 Analysis (ICTA) recommendations 13
milling 108–109, 111 International Institute of Synthetic
oxidation 137, 146 Rubber Producers (IIRP) 64
phosphonylation 137, 146–147 Interpenetrating networks (IPN) 92
polyamide-12 192–193 Ionic bonding
polyethylenes 93, 168–180, 181 rubbery materials 354
low-density 83, 169, 170–172, 176– viscoelastic materials 383–384
178 Ionomer formation
medium-density 178, 179 blends 151–153, 154–155
polyureas 192, 193 1
polyurethanes 91–92, 98, 100–102, relaxation 383–384
188–194 IR spectra 138, 147–151
PVC 180–188 polyurethanes (PU) 194–195
quantitative analysis 81–82, 88–90 sulfonated 149–151, 198–199
reverse engineering 113 IPN (interpenetrating networks) 92
rubbers 77–124 iPP (isotactic polypropylene) 465–474,
ASTM D3677 83 482–485, 585
peak assignments 83–86
IR see infrared spectroscopy
rubber–filler interaction 102–108
Irradiation see electron beam irradiation;
rubbery materials 167–205
γ-irradiation; radiation
characterisation 125–165
Isomers
sample preparation 80–81, 83
balata 88
silica 198–199
BR 83, 88, 90, 139, 212–213, 403
spectrometers 77
CR 90–91
sulfonation 138, 149–151
cyclisation 140–142
sulfur-vulcanised NR 214
NR 88, 403
theory 77–80
PB 414–415
thermoplastic elastomers 188–199
polyisoprene 88, 323–324, 328, 338–
weathering 138, 153
339, 341, 343, 347, 403
zinc stearate 198–199
see also cis–trans isomerisation; cis-
see also Fourier transform infrared polyisoprene
spectroscopy (FTIR)
Isotactic polypropylene (iPP) 465–474,
Interface properties 482–485, 585
2
H NMR 584
NMR imaging 269–271, 273
rubbery materials 354, 559
J
rubber–filler 102–108, 368, 371, 372, J-resolved spectroscopy 543
623
K M
Kapton film, PA–FTIR spectra 70 Macroradical side reaction suppression
Kinetics 230–232
crystallisation 308–309 Magic-angle spinning (MAS) 522–524
13
gelation 307 C NMR 322, 329–330, 336, 510,
(multi) hyphenated TGA 18 535–536
1
k-space coordinates 252–253, 282 H double-quantum spectroscopy
Kuhn segments 259 540–543
WISE experiment 544, 545–547
L Maleation 92, 421
MAS see magic-angle spinning
Laminates, PA–FTIR spectra 70
Mass spectrometry
Larmour frequencies 248–249, 295, 297,
quadrupole 10
462, 521, 523, 526
and thermogravimetric analysis 2–4,
Latex state
7, 8, 10–12, 15–17, 18–19, 25–29,
high-resolution NMR 446–448
35–36
NR 433 differential scanning calorimetry 2–3,
PB 279–280, 447 11, 15–17, 30, 31
suspensions 306 differential thermal analysis 2–3, 6,
LDPE see low-density polyethylene 11, 15–17, 33–34
Least squares linear regression analysis gas chromatography 34–35
60–65 time-of-flight 10, 28, 36
7
Li NMR 437, 439–440 MDI (4,4′-diphenylene methane
Line shape analysis, 2H NMR 527, 561– diisocyanate) 189–190
563, 566–575, 580–581 Mechanical load, NMR imaging 265
Liquid-like properties 520, 527, 533, Mechanical properties
562–563, 564 EPDM 224–225, 238
see also solution NMR filled rubbers 368
Liquid–lattice theory 491–492 network structure 353, 354, 366–367
Lithium initiators 436, 437, 439 heterogeneity 265–269, 360–366
Living polymers 436–439 oil-extended rubbers 366–367
Loaded polymers 293–294, 315–317 peroxide-curing co-agents 230, 236–
Long-chain molecules 313–315 237
Low-density polyethylene (LDPE) Medium-density polyethylene (MDPE),
ethylene vinyl acetate copolymer 172, IR spectra 178, 179
173 5-methylidene-2-norbonene (MNB),
IR spectra 169, 170–172, 188, 189 EPDM peroxide-curing effects 228
γ-irradiation 176–178 Methyl methacrylate (MMA),
NR blends 188, 189 polyethylene graft copolymers 175–176
sample preparation 83 Microstructure
624
Index
13
C NMR 340 software packages 6
diene rubbers 126 Multinuclear high-resolution NMR 436–
PB 403 443
polyisoprene 403 Multi-quantum NMR 259–264, 538–543
Milling 108–109, 111 Mushroom rubber 431, 432
Mineral aggregates, chain adsorption
293–294, 315–317
Mirror velocity 51, 66–68 N
Mixing, elastomers 264 Nanometer scale probe 291–320
Mobile universal surface explorer (NMR- Natural rubber (NR)
MOUSE) 268–269, 277–284, 388 biosynthesis 402, 403
Molecular dynamics 524–528, 530–538 13
C NMR 324–333
2
H NMR 559–561 crosslinking 226, 326, 329–330
slow motions 526, 529, 548–549, crosslink density 506–508
564–565 deproteinisation 433–435
see also chain dynamics γ-irradiated 339
Molecular properties, rubbery materials 1
H NMR 433, 435
557–558 1
Molecular structure relaxation 267–268, 279
conformation 504 IR spectra 83, 84, 86, 131–132
diene rubbers 126 epoxidation 146
PB 83, 87 hydrogenation 131–132
peroxide-cured elastomers 230 isomers 88, 403, 429
polyisoprene 87 latex state 433
see also chain structure low-density polyethylene blends 188,
Molecular vibration 78–80 189
Monte Carlo method 573 NMR-MOUSE 279–280
Mooney–Rivlin coefficients 578–579 PA–FTIR spectra 56–57
MOUSE see NMR-MOUSE peroxide-curing 226, 338–339
Multidimensional NMR techniques small molecule diffusion 499–500
principle 528–529 structure 88, 433–435
viscoelastic material characterisation sulfur vulcanisation 214, 323–333,
519–555 334
(Multi) hyphenated thermogravimetric accelerated 327–333
analysis (TGA) techniques carbon-black-filled 341–344, 582, 583
applications 14–35 covulcanisation 269–271
calibration 13–14 unaccelerated 327
elastomer characterisation 1–47 swollen 506–508
future prospects 35–36 terminal groups 436
quality control 14, 28, 29 see also epoxidised natural rubber;
polyisoprene
sample requirements 13
625
NBR see nitrile–butadiene rubber topology 353, 361, 378

Network structure uniaxially deformed model networks
chain structure 293–294, 296–307, 566–572
333, 355 see also crosslinking
chemical analysis 354–355 nitrile–butadiene rubber (NBR)
13
C NMR 339, 433–434 acrylonitrile determination 61–62, 88–
defects 360–366 90
disentanglement 361 carbon-black-filled 504
13
end-linked 299, 561, 563, 575, 577–579 C NMR 416, 417–418
entanglements 309–315, 354, 360, crosslink density 506
366–367, 377–378, 505–506, 561 hydrogenation 126, 127–129, 421
EPDM 223–225 IR spectra 83, 85, 88, 89, 111–112,
filled rubbers 377–378 126–129
heterogeneity 264–269, 353, 354, PA–FTIR spectra 61–62
360–366 PVC blends 111–113, 183–186
1
H NMR transverse magnetisation see also carboxylated nitrile rubber
relaxation 355–379, 386 NMR
molten high polymers 292, 309–314, angular dependence 521, 522, 525,
502–504 560–561
ionic viscoelastic materials 383–384 anisotropic effects 322, 519–522, 524,
loaded polymers 293–294 529, 558, 559–560
mechanical property effects 353, 354, branch-points 426–429, 439
360–366 chemical composition 402–422
molecular mass distribution 364–366 accuracy 404–410
molten high polymers 292, 309–314, distribution 410–413
502–504 chemically modified structures 419–
network distribution function 294– 422
295, 299–301 chemical shift 270–271, 277, 325–
NMR 326, 328, 457–458, 469–470, 476–
imaging 271–273 480
models 257–259 chemical structure 324–326, 330–333,
pseudo-solid spin-echoes 298–299, 401–404
310, 561–563 correlation 336, 525–526, 529, 538
oil-extended rubbers 366–367 dipolar-correlation effect 526
PDMS 566–572 heteronuclear correlation experiments
properties relationships 224–225, 543–547
236–237 cross-polarisation 322, 470, 474
randomly crosslinked 302–306, 561 crystallisation 307–309
sulfur vulcanisation 323 dipolar decoupling 322, 522–524,
swollen rubbers and blends 502–510 544, 547
temporary networks 310–312 end-groups 422–426, 436–437, 438
thermo-oxidative degradation 347–348 EPM 307–308
626
Index
field cycling technique 526 NMR imaging

high-resolution 401–456 applications 264–276
latex state 306, 446–448 blends 264, 271, 272, 384
7
Li NMR 437, 439–440 BR 500–501
linewidths 506–510 calibration 256–257
linkage effect 302 chemical shift 270–271, 277, 521
low-resolution 248, 277 contrast 250, 253–264, 270–271, 558
magic-angle spinning 322, 329–330, multi-quantum filtering 259–264
336, 510, 522–524 relaxation-time contrast 255–257,
multidimensional techniques 519–555 271–276
nanometer scale probe 291–320 susceptibility contrast 254–255
orientation dependence 281, 521–522, coordinates 252–253
525, 527, 560–561, 569, 571, 573– CPMG echoes 278–280
575, 579–580 crosslink density 271–273, 279–280,
31
P NMR 442 446
pseudo-solid spin-echoes 298–299, defects 264–269
310, 527, 548, 561–563 elastomers 247–289
quadrupolar interactions 522, 527, Fourier NMR 248–251, 508, 528–
528, 559, 564–565, 567–571, 588 529, 544
quantitative analysis 405–410, 541 Hahn echoes 253–257, 258–259, 262,
residual dipolar coupling 524–526, 265, 266–267, 268, 269–271, 274–
527, 530–531, 533–535, 538–547 275, 277–280, 282, 503, 527
sequence distribution 402, 413–419, heterogeneities 264–269
420, 430, 444, 446 homogeneous polarisation field (B0)
29
Si NMR 440–441 247, 250, 257–258, 277, 282
size exclusion chromatography 404, interface detection 269–271, 273, 559,
410–413 584
small signal assignment 422 Larmour frequencies 248–249, 295,
29
Sn NMR 442–443, 444 297
soft polymeric matter 291–320 low-resolution 248, 277, 384
solid-state methods 321–322 magnetic field gradients 250–254,
273, 280, 282, 369
solution methods 321–322
gradient echo imaging 253–255, 265,
theory 295
271, 462–463, 496–497, 498–500
two-dimensional 519–520 magnetisation 248–249, 252, 255–
see also 13C NMR spectroscopy; 256, 258, 295, 297, 299
dipole–dipole interactions; high- models 257–259
resolution NMR; 1H NMR relaxation; Cohen–Addad–Sotta model 258–259
1
H NMR transverse magnetisation Gotlib–Fedotov–Schneider model
relaxation; 2H NMR; 258, 259
multidimensional NMR techniques;
129 NR 499–500
Xe NMR spectroscopy
polyisoprene 500
627
precession phase 248–250 Orientation effects

2
pulsed 250, 253–254, 522, 523, 528, H NMR 560–561, 569, 571, 573–
530, 543–544 575, 579–580
sample deformation 275–276 NMR-MOUSE 281
single-point imaging 267–268, 282, 283 PA–FTIR spectra 71
space encoding 253, 255–256, 282 polyethylene films 169–170
spatial resolution 248, 251–253, 255, soft polymeric matter 291
276–277 viscoelastic materials 527
SPRITE technique 267–268 Outgassing, plasticisers 23–24
swollen rubbers 499–502 Oxidation
theory 248–264 IR spectra 137, 146
vulcanisation process 264, 273–274, photo-oxidation 68, 70
303–304 polyethylene film 68–69
NMR-MOUSE (mobile universal surface see also thermo-oxidative degradation
explorer) 268–269, 277–284, 388 Ozone treated polyethylene film, PA–
NMR relaxation techniques FTIR spectrum 68–69
elastomers 247–248, 271–276
relaxation-time contrast 255–257
P
soft polymeric matter 295
see also 1H NMR relaxation; 1H NMR PA–FTIR see photoacoustic Fourier
transverse magnetisation relaxation transform infrared spectroscopy
NOE (nuclear Overhauser effect) 414, PB see polybutadiene
535–538 PDMS see poly(dimethylsiloxane)
Non-Newtonian liquids 292 PE see polyethylene
NR see natural rubber Percolation 302–304, 315
Nuclear magnetic resonance see NMR Peroxide-curing 210, 225–226, 338–339
Nuclear Overhauser effect (NOE) 414, BR 227, 338–339
535–538 co-agents 229–237, 238
Nylon 6, acrylate rubber blends 194 EPDM 209, 227–237
third monomer effects 227–229, 238
EPM 230–231, 232–234, 235, 236
O NR 226, 338–339
Oil-extended rubbers, network structure PB 227
366–367 polydiene elastomers 225–227
Opacity 51 polyisoprene 339, 506
Optical absorption coefficient 51 scorch retarders 237
Optical spectroscopy Phase-separated systems 547
crosslinking 207–246 Phenol–formaldehyde resins (resols) 209
sulfur vulcanisation 211, 214 Phosphonylation, IR spectra 137, 146–147
see also infrared spectroscopy; NMR; Photoacoustic detectors 50–51, 53
Raman spectroscopy Photoacoustic effect 49–52
628
Photoacoustic Fourier transform infrared carbon-black-filled 582, 584
spectroscopy (PA–FTIR) 81 13
C NMR 414–415, 495–496
background spectrum 53 end-groups 436–437
baseline constructions 64–65 FTIR spectra 88
depth profiling 66–70 high-pressure vulcanisation 339
helium atmosphere 53–54 1
H NMR 314
history 49–50 2
H NMR 436–437, 571–572, 577,
instrumentation 49–51, 52–55, 57 580, 581, 582, 584, 587–588
polarised 71 IR spectra 83–84, 87–88
powders 54–55 isomers 83, 88, 90, 139, 403, 414–
quantitative analysis 60–65 415
rubbers and related materials 49–76 latex state 279–280, 447
samples 54–55 peroxide-curing 227
29
sensitivity 53–54 Sn NMR 442–443, 444
software 55, 57 structure 83, 87
surface analysis 66–70 telechelic 423–424
theory 50–52 temporary network 311
Photoageing Polybutylacrylate 437, 439
polyurethanes 194 Poly-σ-caprolactone (PCL), PVC blends
PVC 187–188 182–183
see also ageing Polychloroprene (PCP)
Photons 77–78 IR spectra 92–94
Photo-oxidation 68, 70 NMR 416, 444, 445
Physical networks see network structure Polydiene rubbers
Physical properties, cured materials 354 crosslinking 210–211, 215–216
Planck’s constant 77 peroxide-curing 225–227
Plasticisers sulfur vulcanisation 211–216
outgassing 23–24 mechanism 214–216
PVC hydrogen bonding 168 Poly(dimethylsiloxane) (PDMS)
Plasticity 167–168 adsorption on mineral aggregates
13 315–316
P NMR 442
crosslink density 505, 561
Polarised photoacoustic Fourier
transform infrared spectroscopy 71 elasticity 305, 567, 576–577
2
Polyamide 6,66,610 copolymer (LPA), IR H NMR 527, 528, 561–572, 574–
spectra 94–95 575, 576–579, 584, 585
Polyamide-12 (PA-12), IR spectra 192– NMR imaging 256–257, 262–263,
193 305
Polyamide–polyether (PA–PEG) block silica-filled silicone rubbers 374–381
copolymers 585–587 uniaxially deformed network 566–572
Polybutadiene (PB) Polyester–polyurethane, chemical
acid-catalysed cyclisation 420–421 degradation 442
629
2
Polyethylene (PE) H NMR 436–437, 438
aluminium laminates 178–179 IR spectra
fibres 459, 460 cyclisation 142
films 68–69, 169–170, 174–178 epoxidation 146
2
H NMR 585 isomers 88, 131
7
IR spectra 93, 168–180, 181 Li NMR 439–440
chain branching 168–169 multi-quantum NMR 260–261
crosslinking 174, 175, 177–178, 179 naturally occurring 429–436
crystallisation 169–170 NMR 416, 500
elongation 169–170 peroxide-curing 339
grafting 92–94, 172, 174, 175–176, structure 87, 88
178 sulfur vulcanisation 324–325
irradiation 176–178
swollen 500
orientation effects 169–170
peroxide-cured 178, 179 telechelic 423–424
129
Xe NMR 485–486
pipe welding 178, 179
single crystals 180 see also natural rubber
surface oxidation 68–69 cis-polyisoprene (IIR), sulfur vulcanisate
129
330–333, 340, 343–344, 347
Xe NMR spectroscopy 459, 460
Polyketones 146
see also high-density polyethylene,
low-density polyethylene, medium- Polymer blends see blends
density polyethylene Polymeric networks see network structure
Poly(ethylene glycol) diacrylate (PEGDA), Polymers
1
H NMR transverse magnetisation aggregated 294
relaxation 358–359, 361–363 amorphous 460, 585
Polyisobutylene (PIB) chemical modification 125
γ-irradiation-induced scission 426 crosslinking 493
1
H NMR 426–427 defects 364
Polyisoprene deformation 167–168
acid-catalysed cyclisation 420–421 γ-irradiation 153
cis–trans isomerisation 323–324, 328, gels 491–492
338–339, 341, 343, 347, 403, 429– living 436–439
435 loaded 293–294, 315–317
13
C NMR spin–lattice relaxation 494– molten 292, 309–314, 502–504, 562,
495 564
crosslink density 506, 543 orientation effects 71, 169–170
end-groups 436–437, 438 quantitative analysis 60–65
γ-irradiation 503–504 semi-crystalline 292–293, 307–309,
1
H NMR 381–382, 481, 584–585
double-quantum spectroscopy 543 solutions 494–496
transverse magnetisation relaxation thermogravimetric analysis samples
503–504 12–13
630
viscoelastic deformation 167–168 hydrogen bonding 174
129
Xe NMR spectroscopy 459–461, IR spectra 180–188
481–482 photoageing 187–188
see also copolymers; soft polymeric plasticised 168
matter substitution reactions 188
Poly(methyl methacrylate) (PMMA) thermo-oxidative degradation 185–
quantitative NMR 404–405 187
129
Xe NMR spectroscopy 460–461 Powders, photoacoustic Fourier
Polypentenamer (PPA), hydroformylation transform infrared spectroscopy 54–55
144–145 PP see polypropylene
Polyprenols 429–433 Principle component analysis (PCA) 36,
Poly-β-propiolactone (PPL), PVC blends 408, 410
182–183 Processing
Polypropylene (PP) elastomers 264–276
blends and block copolymers 465– EPDM 207–208
474, 482–485 quality control 281
isotactic 465–474, 482–485, 585 Product development, (multi) hyphenated
phosphonylation 137, 146–147 TGA 18
Poly(propylene oxide) (PPO) networks, Propylene, determination in EPM 90,
1
H NMR transverse magnetisation 405–410, 414
relaxation 356–357 Propylene–ethylene copolymers see
Polystyrene, PA–FTIR spectra 54, 55 ethylene–propylene copolymers
Polyureas Proton NMR magnetisation relaxation
hydrogen bonding 192, 193 see 1H NMR relaxation; 1H NMR
IR spectra 192, 193 transverse magnetisation relaxation
Polyurethanes (PU) Pseudo-solid spin-echoes 298–299, 310,
degradation 194 527, 548, 561–563
2
H NMR 587 PTA (pulse thermal analysis) 19
hydrogen bonding 100, 190–192 PU see polyurethanes
ionomer formation 194–195 Pulsed field gradient NMR 496–497,
IR spectra 91–92, 98, 100–102, 188– 498–499
129
194 Xe NMR 462–463, 480–482, 483–
phase-separated systems 547 484, 485
31
P NMR 442 Pulse thermal analysis (PTA) 19
Polyurethane–urea 416, 419 PVC see polyvinyl chloride
13
C NMR 416, 419 Pyrolysis methods 35
Polyvinyl chloride (PVC)
blends 180, 182–187
Q
NBR 111–113, 183–186
dechlorination 138, 186 Quadrupolar interactions 522, 527, 528,
grafting 188 559, 564–565, 567–571, 588
631
Quadrupole mass spectrometry (QMS) 10 high-resolution 13C MAS NMR 510

Quality control IR spectroscopy 77–124
1
H NMR transverse magnetisation photoacoustic Fourier transform
relaxation 281, 387–388 infrared spectroscopy 49–76
(multi) hyphenated thermogravimetric quantitative analysis 81–82, 88–90
analysis techniques 14, 28, 29 self-diffusion 496–499
NMR-MOUSE 281 swelling 491–493, 502–510
Quantitative analysis vulcanisates 209–210
automated analysis 408 see also natural rubber
copolymers 88–90 Rubbery materials
IR spectroscopy 81–82, 88–90 chemically modified 125–165
NMR 405–410, 541 customised properties 557
PA–FTIR 60–65 deformation 167–168, 557–558, 567–
polymers 60–65 572
1
TGA 18–19 H NMR transverse magnetisation
relaxation 353–400
2
H NMR 557–596
R IR spectroscopy 167–205
characterisation 125–165
Radiation
molecular properties 557–558
IR spectra 153, 156–157, 174–176,
177 swelling 491–518, 580
vulcanisation 339 Rubber–filler interactions 102–108, 368,
371, 372, 374, 376, 380, 474–475, 584
see also electron beam irradiation; γ-
irradiation
Raman spectroscopy 210–211 S
sulfur vulcanisation 211–213, 214,
Samples
238
IR spectroscopy 80–81, 83
see also Fourier transform Raman
spectroscopy NMR imaging 275–276
Reactivity, (multi) hyphenated TGA 18 photoacoustic Fourier transform
infrared spectroscopy 54–55
Real-time 1H NMR transverse
magnetisation relaxation 385–387 thermogravimetric analysis 12–13
Recycling, thermoplastic elastomers 188 SBR see styrene–butadiene rubbers
residual dipolar coupling 524–526, 527, Scorch retarders, peroxide-curing 237
530–531, 533–535, 538–547 Seals see gaskets
Resols (phenol–formaldehyde resins) 209 Self-crosslinking blends, IR spectra 96–
Reverse engineering 113 98, 183, 186
Rouse model 313 Self-diffusion
Rubber blends see blends rubbers 498–499
small molecules 496–500
Rubbers 129
Xe NMR 463, 464, 480–487
crosslinking 209–210, 493
632
diffusion time 485–487 Software
Semi-crystalline materials (multi) hyphenated thermogravimetric
elastomers 381–383 analysis 6
polymers 292–293, 307–309, 381– photoacoustic Fourier transform
382, 481, 584–585 infrared spectroscopy 55, 57
see also crystallinity Solid-like properties 257, 520, 527, 533
Sequence distribution 402, 413–419, 420, Solid state NMR see 13C NMR
424–425, 430, 444, 446, 495–496 Spectroscopy; high-resolution NMR; 1H
Shear modulus 167–168, 530 NMR relaxation
Short-chain molecules 312 Solution NMR 321–322, 519, 522
13
Silica C NMR 494–496, 508–509
2
IR spectra 198–199 H NMR 559
PA–FTIR spectra 56–57 multi-quantum NMR 538
Silica-filled rubbers NOESY–MAS spectroscopy 535
carbon–silica dual phase fillers 108 swollen state 443–446, 500
chain adsorption 315–316, 375–377, see also liquid-like properties
378–379 Solvent effects
13
grafting 379–381 C NMR 444, 445
1
1
H NMR transverse magnetisation H NMR transverse magnetisation
relaxation 378–381 relaxation 361–363, 499
2
PDMS 315–316 H NMR 580–581
cis-polyisoprene 347 see also swollen state
rubber–filler interaction 106, 108, Spatial heterogeneity
109, 110, 584 crosslink density 273
SBR 66–67 network structure 353, 361, 558
sulfur vulcanisation 347 Spatially resolved NMR 276–277, 558
Silica-filled silicon rubbers Spin–lattice relaxation 310, 494–496
1
H NMR transverse magnetisation Spin–spin interactions 295, 297, 298–
relaxation 374–378 299, 300–301, 314, 520–524
2
H NMR 584 manipulation 522–524
Silicone methyl rubber, IR spectra 86 spin diffusion 547
Siloxanes, 29Si NMR 440–441 SPRITE (single-point ramped imaging
Single-point ramped imaging with T1 with T1 enhancement) technique 267–268
enhancement technique (SPRITE) 267–268 Stability see thermal stability
29
Si NMR 440–441 Standardisation
Size exclusion chromatography (SEC) interlaboratory reproducibility 13–14
404, 410–413 see also ASTM standards
29
Sn NMR 442–443, 444 Storage modulus 358–359
Soft polymeric matter Stress/strain effects
NMR 291–320 2
H NMR 573–575, 579–580, 589
NMR imaging 247, 282 NMR imaging 265
633
strain distribution 263–264 vinyl SBR 63–64, 336

Stretching see also carboxylated styrene–
gels 306 butadiene rubber
polyethylene fibres 459, 460 Sulfonation
rubbery materials 567–570, 574 ionomers 149–151, 198–199
see also elasticity; elongation IR spectra 138, 149–151, 198–199
Stretching vibrations 79–80 maleated styrene/ethylene–butylene/
Structure see chain structure; chemical styrene 149–151
structure; microstructure; molecular Sulfur vulcanisation
structure; network structure accelerated 216–217, 327–333, 386–
Styrene/ethylene–butylene/styrene (SEBS), 387
sulfonation 149–151 accelerators 215, 322–323, 328–329,
Styrene–butadiene rubbers (SBR) 330–333, 336, 506–507
carbon-black-filled 24–25, 372, 374, activators 215, 323
582 cis-butadiene rubber 212–213, 333–
13
C NMR 336 338
covulcanisation 269–271 carbon-black-filled rubbers 336–337,
crosslinking 531–533, 541–542, 545 341–347, 582, 583
emulsion SBR 62–63 covulcanisation 269–271
1
H NMR dialkenylsulfides 217–219
double-quantum MAS spectroscopy diene rubbers 323–324
540–542 elastomers 209–210, 264
magnetisation-exchange spectroscopy EPDM 208–209, 216–225, 238, 365,
530–533 386–387, 422
NOESY–MAS spectroscopy 536–538 free radicals 213, 335
1
H NMR transverse magnetisation 1
relaxation 267–268, 279–280 relaxation 273–274
2
H NMR 582 mechanism 322–324, 345–346
hydrogenation 131, 132 model studies 209–210, 216–217
IR spectra 83, 84, 88–89, 127, 131, natural rubber 214, 323–333, 334,
132 582, 583
NMR 408–410, 522, 523 accelerated 327–333
NMR-MOUSE 278–279 unaccelerated 327
PA–FTIR spectra 57–58, 62–63, 66–67 network chain density 333, 334, 344,
silica-filled 66–67 346–347, 367
structure 401 NMR imaging 264, 273–274, 303–
styrene determination 62–63, 88, 89, 304
408–410 optical spectroscopy 211, 214
sulfur vulcanisation 338 pendent groups 323
thermal degradation 24–25 polydiene rubbers 211–216
thermo-oxidative ageing 274–275 mechanism 214–216
634
129
cis-polyisoprene 330–333, 340, 343– Xe NMR spectroscopy 470, 471–
344 473, 476–478, 479
Raman spectroscopy 211–213, 214 Temporary networks 310–312
Fourier transform Raman Tenax absorption cartridges 3–4, 12, 16
spectroscopy 217–223 Terminal groups 424–426, 436–437, 438
SBR 338 see also end-groups
silica-filled rubbers 347 Termonomers, EPDM 90, 91
sulfurating agents 323 Tetramethyl thiuram disulfide (TMTD)
see also vulcanisates 328, 334
Surface analysis 66–70 TGA see (multi) hyphenated
Swollen state Thermogravimetric analysis (TGA)
chain structure 302, 303, 304, 306, techniques; thermogravimetric analysis
344, 346–347 Thermal analysis 1
cyclohexane solvent 266–267, 497– Thermal conductivity 51
498, 501 Thermal decomposition
high-resolution NMR 443–446, 510 carboxynitroso rubber 28
1
H NMR transverse magnetisation EPDM 23–24
relaxation 354, 361–363, 378, 500– pressure behaviour 26–27
506 sulfur vulcanisation 323
2
H NMR 580–581 Thermal degradation
network density 502–510 ABS 21–22, 26
NMR imaging 499–502 Aflas FA-150C rubber 21–22
NR 506–508 carbon-black-filled SBR 24–25
polymer chain motion 494–496 IR spectra 113, 138, 156, 184, 185–
rubbery materials 354, 491–518 187
self-diffusion 496–499 (multi) hyphenated TGA 17, 27–28
small molecule motion 496–500, 510, polyurethanes 194
580, 581 PVC 184, 185–187
Thermal desorption 34–35
Thermal diffusion depth 52, 66–67
T Thermal diffusivity 51–52, 54, 66–67
Tacticity Thermal stability, (multi) hyphenated
EPM rubber 90 TGA 17
IR spectra 90, 111 Thermobalances 3, 6, 13
Telechelic chains 299, 423–424, 439 Thermogravimetric analysis (TGA) 1
Temperature dependence calibration 13–14
chain structure 301–302 high pressure 4
1
H NMR transverse magnetisation multicomponent separation 12–13
relaxation 356–357, 375, 376 quantitative analysis 18–19
log (shear) modulus 167–168 sample requirements 12–13
NMR imaging 276 see also (multi) hyphenated
635
thermogravimetric analysis (TGA) V

techniques
Thermo-oxidative degradation Vibrational frequencies 78–80, 79, 170–
13 172
C NMR 348
Vinyl acetate 60–61
network structure 347–348
Vinyl silanes, grafting onto polyethylene
NMR imaging 274–275
174, 175, 176
PVC 185–187
Vinyl styrene–butadiene rubber
SBR 274–275 13
C NMR 336
see also ageing; oxidation
PA–FTIR spectrum 63–64
Thermoplastic elastomers (TPE)
2
Viscoelastic materials
H NMR 548–549, 585–588 1
IR spectra 188–199
relaxation 365–366, 383–384, 526–
recycling 188 527
Thermoplastic rubbers 168 2
H NMR 522, 527
Time-of-flight mass spectrometry ionic 383–384
(ToFMS) 10, 28, 36
multidimensional NMR
Topology 353, 361, 378 characterisation 519–555
TPE see thermoplastic elastomers property temperature dependence
Transmittance 82 167–168
Transparency 51 Viscous melts 167
Triallylcyanurate (TAC) 229, 235, 502 Volatiles, residual 26
Trimethylolpropane trimethacrylate Volume-average material properties 360
(TMPTMA) 156–157 Vrentas–Duda equation 497
Two-dimensional NMR techniques 519– Vulcanisates
520, 528–550 13
C NMR 322, 324–333, 333–338,
applications 530–549 338–339
double-quantum MAS spectroscopy high-resolution solid state NMR 321–
540–543 352
exchange spectroscopy 529, 530–535, rubbers 209–210
548–549 oil-extended 366
heteronuclear correlation experiments Vulcanisation see high-pressure
543–547 Vulcanisation; peroxide-curing; sulfur
2
H NMR 548–549 Vulcanisation
multi-quantum spectroscopy 538–543
W
separation spectroscopy 529
Tyres, NMR imaging 267–269 Wavelength 77–78
Weathering, IR spectra 138, 153
U Welding, crosslinked polyethylene pipes
178, 179
Urey–Bradley field theory 126 Wide-line separation experiments (WISE)
636
543–547 diffusion time 485–487
Williams–Landel–Ferry (WLF) equation temperature dependence 470, 471–
494, 497–498, 586 473, 476–478, 479
XNBR see carboxylated nitrile rubber
X XSBR see carboxylated styrene–butadiene
rubber
129
Xe NMR spectroscopy
chemical shift 457–458, 469–470,
476–480 Y
elastomers in blends and composites
Young’s modulus 460
457–489
EPDM 458, 463–464, 474–480, 482–
486 Z
EPM 465–474, 482
exchange model 479–480, 484 Ziegler–Natta catalysis 168, 207
gas phase 458–459 Zinc oxide
PBT/PTMO block copolymer 486–487 blends, IR spectra 151–153, 154–155,
polyisoprene 485–486 195–198
polymers 459–461 sulfur vulcanisation 215, 340
pulsed field gradient 462–463, 480– Zinc stearate
482, 483–484, 485 bloom 66–68
self-diffusion coefficients 461, 463, IR spectra 198–199
464, 480–487 PVC thermal stability 187
637
638

LitvinovVictorM. DePrajnaparamita SpectroscopyofRubbersandRubberyMaterials ISmithersRapraPublishing2011!02!28

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

LitvinovVictorM. DePrajnaparamita SpectroscopyofRubbersandRubberyMaterials ISmithersRapraPublishing2011!02!28

Uploaded by

Copyright:

Available Formats

Spectroscopy

Victor M. Litvinov and Prajna P. De

Rapra Technology Limited

Shawbury, Shrewsbury, Shropshire, SY4 4NR, United Kingdom

Rapra Technology Limited

©2002, Rapra Technology Limited

All rights reserved. Except as permitted under current legislation no part

Typeset by Rapra Technology Limited

2 Photoacoustic Fourier Transform Infrared Spectroscopy of Rubbers

2.3 Theory of Photoacoustic Spectroscopy ................................................ 50

3 Infrared Spectroscopy of Rubbers ................................................................. 77

4 Application of Infrared Spectroscopy to Characterise Chemically

4.3 Hydrogenation ................................................................................... 126

5 Infrared Spectroscopy of Rubbery Materials ............................................... 167

6 Crosslinking of EPDM and Polydiene Rubbers Studied by Optical

6.1.2 EPDM Crosslinking ............................................................... 208

7 NMR Imaging of Elastomers ...................................................................... 247

8 NMR in Soft Polymeric Matter: Nanometer Scale Probe ............................ 291

9 Chemical Characterisation of Vulcanisates by High-Resolution

10 Characterisation of Chemical and Physical Networks in Rubbery

10.5.1 NMR Relaxation of Filled Rubbers ....................................... 368

11 High Resolution NMR of Elastomers ......................................................... 401

11.5 Structural of Naturally Occurring Polyisoprenes ............................... 429

13 Swollen Rubbery Materials: Chemistry and Physical Properties

13.1.3 Reviews of NMR of Crosslinked Rubbers

14 Multidimensional NMR Techniques for the Characterisation

14.3.2 Two-dimensional NMR Techniques and

15 Deuterium NMR in Rubbery Materials ...................................................... 557

Abbreviations .................................................................................................... 597

Contributors ...................................................................................................... 607

Index ................................................................................................................. 611

Chemical structure of rubbery materials: Chapters 1, 2, 3, 4, 5, 6, 9 and 11, describe

Heterogeneity of rubbery materials: Use of NMR imaging, NMR MOUSE (mobile

Functional properties and stability of rubbery materials: Chapters 1, 3, 4, 7, 12 and 13,

Processing of rubbery materials: The effect of milling on the heterogeneity of rubbery

Quality control: Applications of NMR relaxation experiments, NMR imaging, NMR

Victor M. Litvinov and Prajna P. De

Jan C.J. Bart and Charles Raemaekers

Thermal analysis is a group of techniques in which a physical property of a substance is

Figure 1.1 (Multi) hyphenated thermogravimetric analysis techniques. The figures

characterisation. Some thirty different experimental approaches may already be identified,

Table 1.1 Selected references to (multi) hyphenated

1.2 Instrumental Part

Thermogravimetric measurements require a thermobalance. There are many different

Table 1.2 Characteristics of some non-simultaneous TG methods [26]

It is especially important to have simultaneous techniques incorporating the most

• Reduced influence of sample preparation (a single sectioning of the original material).

• Simplified interpretation for perturbation of results arising from thermal history,

• Combined evaluation assuring identical experimental and sampling conditions for

• Correlation of observed effects.

• Weight measurements validating quantitative measurements from DSC experiments.

• Detection of moisture and determination of in-situ dry weight.

Kaisersberger and co-workers [96] discussed hyphenation of a thermo-balance to infrared

Figure 1.2 Simultaneous TG-DSC

Table 1.3 Characteristics of TG-FTIR and TG-MS methods [26]

Simultaneous TG-MS is a very powerful hyphenated technique combining the direct

More complex thermal decomposition processes may require identification instead of

On-line thermal desorption-GC-FTIR-MS can be operated both in parallel and tandem

With a thermal desorption unit capable of accommodating Tenax absorption cartridges

1.2.1.1 Sample Requirements for Polymers

Multicomponent separation of a rubber material performed with TG is typically as follows:

extenders), successive heating up to 600 °C (for decomposition of the rubber component),

For (multi) hyphenated thermogravimetric analysis techniques typically a few mg of