The Science and Technology of Rubber

The Science and Technology of Rubber

Fourth Edition

Edited by

Burak Erman

Department of Chemical and Biological Engineering, Koc University, Rumeli Feneri Yolu 34450 Istanbul, Turkey

James E. Mark

Department of Chemistry, University of Cincinnati, Cincinnati, OH 45221-0172, USA

C. Michael Roland

Naval Research Laboratory, Chemistry Division, Code 6120, Washington, DC, USA

Table of Contents

Cover image

Title page

Copyright

Chapter 1. Rubber Elasticity: Basic Concepts and Behavior

1.1 Introduction

1.2 Elasticity of a Single Molecule

1.3 Elasticity of a Three-Dimensional Network of Polymer Molecules

1.4 Comparison with Experiment

1.5 Continuum Theory of Rubber Elasticity

1.6 Second-Order Stresses

1.7 Elastic Behavior Under Small Deformations

1.8 Some Unsolved Problems in Rubber Elasticity

References

Chapter 2. Polymerization: Elastomer Synthesis

2.1 Introduction

2.2 Classification of Polymerization Reactions and Kinetic Considerations

2.3 Polyaddition/Polycondensation

2.4 Chain Polymerization by Free Radical Mechanism

2.5 Emulsion Polymerization

2.6 Copolymerization

2.7 Chain Polymerization by Cationic Mechanism

2.8 Chain Polymerization by Anionic Mechanism

2.9 Stereospecific Chain Polymerization and Copolymerization by Coordination Catalysts

2.10 Graft and Block Copolymerization

References

Chapter 3. Structure Characterization in the Science and Technology of Elastomers

3.1 Introduction

3.2 Chemical Composition

3.3 Sequence Distribution of Repeat Units

3.4 Chain Architecture

3.5 Glass Transition and Secondary Relaxation Processes

3.6 Morphology

References

Chapter 4. The Molecular Basis of Rubberlike Elasticity

4.1 Introduction

4.2 Structure of a Typical Network

4.3 Elementary Molecular Theories

4.4 More Advanced Molecular Theories

4.5 Phenomenological Theories and Molecular Structure

4.6 Swelling of Networks and Responsive Gels

4.7 Enthalpic and Entropic Contributions to Rubber Elasticity: The Force-Temperature Relations

4.8 Direct Determination of Molecular Dimensions

4.9 Single-Molecule Elasticity

References

Chapter 5. The Viscoelastic Behavior of Rubber and Dynamics of Blends

Nomenclature

1 Introduction

)

5.3 The Glass Temperature

5.4 Viscoelastic Behavior Above

5.5 Viscoelastic Behavior of Other Model Elastomers

5.6 Theoretical Interpretation of Viscoelastic Mechanisms and Anomalies

5.7 Component Dynamics of Highly Asymmetric Polymer Blends

References

Chapter 6. Rheological Behavior and Processing of Unvulcanized Rubber

6.1 Rheology

6.2 Linear Viscoelasticity

6.3 Nonlinear Viscoelasticity

6.4 Engineering Analysis

6.5 Practical Processing Considerations

References

Chapter 7. Vulcanization

7.1 Introduction

7.2 Definition of Vulcanization

7.3 Effects of Vulcanization on Vulcanizate Properties

7.4 Characterization of the Vulcanization Process

7.5 Vulcanization by Sulfur without Accelerator

7.6 Accelerated-Sulfur Vulcanization

7.7 Vulcanization by Phenolic Curatives, Benzoquinone Derivatives, or Bismaleimides

7.8 Vulcanization by the Action of Metal Oxides

7.9 Vulcanization by the Action of Organic Peroxides

7.10 Dynamic Vulcanization

References

Chapter 8. Reinforcement of Elastomers by Particulate Fillers

8.1 Introduction

8.2 Preparation of Fillers

8.3 Morphological and Physicochemical Characterization of Fillers

8.4 The Mix: A Nanocomposite of Elastomer and Filler

8.5 Mechanical Properties of Filled Rubbers

References

Chapter 9. The Science of Rubber Compounding

9.1 Introduction

9.2 Polymers

9.3 Filler Systems

9.4 Stabilizer Systems

9.5 Vulcanization System

9.6 Special Compounding Ingredients

9.7 Compound Development

9.8 Compound Preparation

9.9 Environmental Requirements in Compounding

9.10 Summary

References

Chapter 10. Strength of Elastomers

10.1 Introduction

10.2 Initiation of Fracture

10.3 Threshold Strengths and Extensibilities

10.4 Crack Propagation

10.5 Tensile Rupture

10.6 Repeated Stressing: Mechanical Fatigue

10.7 Failure Under Multiaxial Stresses

10.8 surface Cracking by Ozone

10.9 Abrasive Wear

10.10 Computational Approaches to Failure Modeling

Further Reading

References

Chapter 11. The Chemical Modification of Polymers

11.1 Introduction

11.2 Chemical Modification of Polymers Within Backbone and Chain Ends

11.3 Esterification, Etherification, and Hydrolysis of Polymers

11.4 The Hydrogenation of Polymers

11.5 Dehalogenation, Elimination, and Halogenation Reactions in Polymers

11.6 Other Addition Reactions to Double Bonds

11.7 Oxidation Reactions of Polymers

11.8 Functionalization of Polymers

11.9 Miscellaneous Chemical Reactions of Polymers

11.10 Block and Graft Copolymerization

References

Chapter 12. Elastomer Blends

12.1 Introduction

12.2 Thermodynamics and Solubility Parameters

12.3 Preparation

12.4 Miscible Elastomer Blends

12.5 Immiscible Elastomer Blends

12.6 Conclusion

Appendix 1: Acronyms for Common Elastomers

References

Chapter 13. Thermoplastic Elastomers

13.1 Introduction

13.2 Synthesis of Thermoplastic Elastomers

13.3 Morphology of Thermoplastic Elastomers

13.4 Properties and Effect of Structure

13.5 Thermodynamics of Phase Separation

13.6 Thermoplastic Elastomers at Surfaces

13.7 Rheology and Processing

13.8 Applications

References

Chapter 14. Tire Engineering

14.1 Introduction

14.2 Tire Types and Performance

14.3 Basic Tire Design

14.4 Tire Engineering

14.5 Tire Materials

14.6 Tire Testing

14.7 Tire manufacturing

14.8 Summary

References

Chapter 15. Recycling of Rubbers

15.1 Introduction

15.2 Retreading of Tires

15.3 Recycling of Rubber Vulcanizates

15.4 Use of Recycled Rubber

15.5 Pyrolysis and Incineration of Rubber

15.6 Concluding Remarks

References

Index

Copyright

Academic press is an imprint of Elsevier

225 Wyman Street, Waltham, MA 02451, USA

The Boulevard, Langford Lane, Kidlington, Oxford, OX5 1GB, UK

© 2013 Elsevier Inc. All rights reserved.

No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher. Details on how to seek permission, further information about the Publisher’s permissions policies and our arrangements with organizations such as the Copyright Clearance Center and the Copyright Licensing Agency, can be found at our website: www.elsevier.com/permissions.

This book and the individual contributions contained in it are protected under copyright by the Publisher (other than as may be noted herein).

Notices

Knowledge and best practice in this field are constantly changing. As new research and experience broaden our understanding, changes in research methods, professional practices, or medical treatment may become necessary.

Practitioners and researchers must always rely on their own experience and knowledge in evaluating and using any information, methods, compounds, or experiments described herein. In using such information or methods they should be mindful of their own safety and the safety of others, including parties for whom they have a professional responsibility.

To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors, assume any liability for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein.

Library of Congress Cataloging-in-Publication Data

A catalog record for this book is available from the Library of Congress

British Library Cataloging-in-Publication Data

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

ISBN: 978-0-12-394584-6

For information on all Academic Press publications visit our website at http://store.elsevier.com

Printed in the United States of America

12 13 14 9 8 7 6 5 4 3 2 1

Chapter 1

Rubber Elasticity: Basic Concepts and Behavior

A.N. Gent,    The University of Akron, Akron, OH, USA

Acknowledgments

We refer the reader to the classic survey of rubber elasticity by Treloar (1975) and to three recent reviews that give fuller accounts of the molecular theory (Graessley, 2004; Mark and Erman, 1988, 1992). The author thanks Mr. R.A. Paden for drawing several of the figures.

1.1 Introduction

The single most important property of elastomers—that from which their name derives—is their ability to undergo large elastic deformations; that is, to stretch and return to their original shape in a reversible way. Theories to account for this characteristic high elasticity have passed through three distinct phases: (1) the early development of a molecular model relating experimental observations to the known molecular features of rubbery polymers; (2) generalization of this approach by means of symmetry considerations taken from continuum mechanics that are independent of the molecular structure; and (3) a critical reassessment of the basic premises on which these two quantitative theories are founded. In this chapter, the theoretical treatment is briefly outlined and shown to account quite successfully for the observed elastic behavior of rubbery materials. The special case of small elastic deformations is then discussed in some detail because of its technical importance. Finally, attention is drawn to some aspects of rubber elasticity that are still little understood.

1.2 Elasticity of a Single Molecule

The essential requirement for a substance to be rubbery is that it consist of long flexible chainlike molecules. The molecules themselves must therefore have a backbone of many noncollinear single valence bonds, about which rapid rotation is possible as a result of thermal agitation. Some representative molecular subunits of rubbery polymers are shown in Figure 1.1; thousands of these units linked together into a chain constitute a typical molecule of the elastomers listed in Figure 1.1. Such molecules change their shape readily and continuously at normal temperatures by Brownian motion. They take up random conformations in a stress-free state but assume somewhat oriented conformations if tensile forces are applied at their ends (Figure 1.2). One of the first questions to consider, then, is the relationship between the applied tension f and the mean chain end separation r, averaged over time or over a large number of chains at one instant in time.

Figure 1.1 Repeat units for some common elastomer molecules.

Figure 1.2 (a) Random chain and (b) oriented chain (from Gent, 1958).

Chains in isolation take up a wide variety of conformations,¹ governed by three factors: the statistics of random processes; a preference for certain sequences of bond arrangements because of steric and energetic restraints within the molecule; and the exclusion of some hypothetical conformations that would require parts of the chain to occupy the same volume in space. In addition, cooperative conformations are preferred for space-filling reasons in concentrated solutions or in the bulk state.

Flory (1969) has argued that the occupied-volume exclusion (repulsion) for an isolated chain is exactly balanced in the bulk state by the external (repulsive) environment of similar chains, and that the exclusion factor can therefore be ignored in the solid state. Direct observation of single-chain dimensions in the bulk state by inelastic neutron scattering gives values fully consistent with unperturbed chain dimensions obtained for dilute solutions in theta solvents² (Cotton et al., 1972), although intramolecular effects may distort the local randomness of chain conformation.

Flory has again given compelling reasons for concluding that the chain end-to-end distance r in the bulk state will be distributed in accordance with Gaussian statistics for sufficiently long chains, even if the chains are relatively stiff and inflexible over short lengths (Flory, 1969). With this restriction to long chains it follows that the tension-displacement relation becomes a simple linear one,

(1.1)

where f is the tensile force, r is the average distance between the ends of the chain, and A for unstressed chains,

(1.2)

where k is Boltzmann’s constant and T is the absolute temperature.

If the real molecule is replaced by a hypothetical chain consisting of a large number n of rigid, freely jointed links, each of length l (Figure 1.3), then

(1.3)

Figure 1.3 Model chain of freely jointed links.

is independent of temperature because completely random link arrangements are assumed. The tension f in Eq. (1.1) then arises solely from an entropic mechanism; that is, from the tendency of the chain to adopt conformations of maximum randomness, and not from any energetic preference for one conformation over another. The tension f is then directly proportional to the absolute temperature T.

For real chains, consisting of a large number n of primary valence bonds along the chain backbone, each of length l,

(1.4)

is found to vary from 4 to 10, depending on the chemical structure of the molecule and on temperature, because the energetic barriers to random bond arrangements are more easily overcome at higher temperatures (may thus be regarded as the effective bond length of the real chain, a measure of the stiffness of the molecule.

Equation (1.1) is reasonably accurate only for relatively short distances r, less than about one-third of the fully stretched chain length (Cotton et al., 1972). Unfortunately, no good treatment exists for the tension in real chains at larger end separations. We must therefore revert to the model chain of freely jointed links, for which

(1.5)

denotes the inverse Langevin function. An expansion of this relation in terms of r/nl,

(1.6)

gives a useful indication of where significant departures from Eq. (1.1) may be expected.

Equation . Rubber shows a similar steeply rising relation between tensile stress and elongation at high elongations. Indeed, experimental stress-strain relations closely resemble those calculated using Eq. (1.5) in place of Eq. (1.1) in the network theory of rubber elasticity (outlined in the following section). The deformation at which a small but significant departure is first found between the observed stress and that predicted by small-strain theory, using Eq. (1.1), yields a value for the effective length l of a freely jointed link for the real molecular chain. This provides a direct experimental measure of molecular stiffness. The values obtained are relatively large, of the order of 5–15 main-chain bonds, for the only polymer that has been examined by this method so far, cis-1,4-polyisoprene (Morris, 1964).

Figure 1.4 Tension-displacement relation for a freely jointed chain (Eq. (1.5)), - - -, Gaussian solution (Eq. (1.1)) (from Gent, 1958).

Equation (1.5) has also been used to estimate the force at which a rubber molecule will become detached from a particle of a reinforcing filler (e.g., carbon black) when a filled rubber is deformed (Bueche, 1960, 1961). In this way, a general semiquantitative treatment has been achieved for stress-induced softening (Mullins effect) of filled rubbers (shown in Figure 1.5).

Figure 1.5 Stress-induced softening of a carbon black-filled vulcanizate of a copolymer of styrene and butadiene (25:75); - - -, stress-strain curve of a corresponding unfilled vulcanizate (from Tobolsky and Mark, 1971).

1.3 Elasticity of a Three-Dimensional Network of Polymer Molecules

Some type of permanent structure is necessary to form a coherent solid and to prevent liquidlike flow of elastomer molecules. This requirement is met by incorporating a small number of intermolecular chemical bonds (crosslinks) to make a loose three-dimensional molecular network. Such crosslinks are generally assumed to form in the most probable positions, so that the long sections of molecules between them have the same spectrum of end-to-end lengths as a similar set of uncrosslinked molecules would have. Under Brownian motion each molecular section takes up a wide variety of conformations, as before, but now subject to the condition that its ends lie at the crosslink sites. The elastic properties of such a molecular network are treated later. We consider first another type of interaction between molecules.

High-molecular-weight polymers form entanglements by molecular intertwining, with a spacing (in the bulk state) characteristic of the particular molecular structure (between entanglement sites are given in Table 1.1. Thus, a high-molecular-weight polymeric melt will show transient rubberlike behavior even in the absence of any permanent intermolecular bonds.

Table 1.1

Representative Values of the Average Molecular Weight Between Entanglements for Polymeric Meltsa

aObtained from flow viscosity measurements.

In a crosslinked rubber, many of these entanglements are permanently locked in (Figure 1.6), the more so the higher the degree of crosslinking. If they are regarded as fully equivalent to crosslinks, the effective number N , arising from entanglements and chemical crosslinks, respectively, where

denote the average molecular weights between entanglements and between crosslinks, respectively. The efficiency of entanglements in constraining the participating chains is, however, somewhat uncertain, particularly when the number of chemical crosslinks is relatively small (Langley, 1968; Vilgis, 1992; Graessley, 2004). Moreover, the force-extension relation for an entangled chain will differ from that for a crosslinked chain (Prager and Frisch, 1967), being stiffer initially and nonlinear in form. The effective number N of molecular chains that lie between fixed points (i.e., crosslinks or equivalent sites of molecular entanglement) is therefore a somewhat ill-defined quantity, even when the chemical structure of the network is completely specified.

Figure 1.6 Sketch of a permanent entanglement (from Gent, 1958).

It is convenient to express the elastic behavior of the network in terms of the strain energy density W per unit of unstrained volume. The strain energy w for a single chain is obtained from Eq. (1.1) as

(1.7)

For a random network of N (deformed dimension/undeformed dimension) in the three principal directions (Figure 1.7), W is given by (Treloar, 1975)

(1.8)

denotes the mean square end-to-end distance between chain ends (crosslink points or equivalent junctions) in the undeformed state. The close similarity of .

Figure 1.7 (a) Undeformed and (b) deformed states.

, the corresponding mean square end-to-end distance for unconnected chains of the same molecular length. Because A (Eq. (1.2)), the only molecular parameter that remains in Eq. (1.8) is the number N of elastically effective chains per unit volume. Thus, the elastic behavior of a molecular network under moderate deformations is predicted to depend only on the number of molecular chains and not on their flexibility, provided that they are long enough to obey Gaussian statistics.

for real chains (Eq. , as discussed elsewhere (Flory, 1969; Graessley, 2004).

.

From the general relation for strain energy, Eq. (1.8), the elastic stresses required to maintain any given deformation can be obtained by means of virtual work considerations (Figure 1.7),

. Because of the practical incompressibility of rubbery materials in comparison to their easy deformation in other ways, the original volume is approximately conserved under deformation. The extension ratios then obey the simple relationship

(1.9)

As a result, the stress-strain relations become

where p denotes a possible hydrostatic pressure (which has no effect on an incompressible solid). Thus, only stress differences can be written explicitly

(1.10)

(from Eq. . Hence,

(1.11)

It is customary to express this result in terms of the tensile force f in the unstrained state, where

The corresponding relation is shown in Figure 1.8. It illustrates a general feature of the elastic behavior of rubbery solids: although the constituent chains obey a linear force-extension relationship (Eq. (1.1)), the network does not. This feature arises from the geometry of deformation of randomly oriented chains. Indeed, the degree of nonlinearity depends on the type of deformation imposed. In simple shear, the relationship is predicted to be a linear one with a slope (shear modulus G) given by

(1.12)

is the amount of shear; for example, dx/dy.

Figure 1.8 Force-extension relation for simple extension. - - -, Linear relation obtaining at infinitesimal strains (from Gent, 1958).

Because rubbery materials are virtually incompressible in bulk, the value of Poisson’s ratio is close to 0.5. Young’s modulus E is therefore given by 3G , is linear only for quite small extensions (Figure 1.8), so that Young’s modulus is applicable only for extensions or compressions of a few percent.

All of these stress relations are derived from Eq. (1.8). They are valid therefore only for moderate deformations of the network; that is, for deformations sufficiently small for the chain tensions to be linearly related to their end-to-end distances r (Eq. (1.1)). Unfortunately, no correspondingly simple expression can be formulated for W using Eq. (1.5), the relationship for large strains of the constituent chains, in which the molecular stiffness parameter reappears. Instead, a variety of series approximations must be used, as in Eq. (1.6), to give close approximations to the behavior of rubber networks under large strains (Arruda and Boyce, 1993).

1.4 Comparison with Experiment

Although the treatment of rubber elasticity given in the preceding section is generally rather successful, certain discrepancies are found to occur. The first consists of observed stresses higher than predicted—for example, by Eq. stress" appears to reflect a non-Gaussian characteristic of network chains, which is important only at small values of the chain end-to-end distance rstress and its complex dependence on type and degree of strain, and on degree of swelling, can all be accurately described by a simple additional term in the relation for the strain energy w for a single network chain, Eq. (1.7), which becomes

(1.13)

The second term clearly becomes insignificant at large values of r.

does not appear to be strongly dependent on temperature and therefore does not appear to be associated with the energetics of chain conformations; and it is closely correlated with the tendency of the polymer chains to form molecular entanglements. For example, those polymers with a high density of entanglements in the bulk state (stress component (Graessley, 2004).

discrepancy appears to arise only when the molecular chains are tied into a network.

stress suggest that it is associated with entangled chains in networks (stress.

A second discrepancy between theory and experiment is found when the Gaussian part of the measured stresses is compared with the theoretical result for an ideal network. Numerical differences of up to 50% are obtained between the density of effective chains calculated from the observed stresses and that calculated from the chemistry of crosslinking. This discrepancy may be due to an error in the theoretical treatment as given here. James and Guth (1943, 1949) arrived at stresses only half as large as those given in Eq. (1.10), from a somewhat different theoretical standpoint.

A third and major discrepancy, already referred to, is found at large deformations when the network chains fail to obey Gaussian statistics, even approximately. Considerable success is achieved in this case by using Eq. (1.5) in place of Eq. (1.1) for chain tensions in the network.

Notwithstanding these discrepancies, the simple treatment of rubber elasticity outlined in this chapter has proved to be remarkably successful in accounting for the elastic properties of rubbers under moderate strains, up to about 300% of the unstrained length (depending on the length and flexibility, and hence the extensibility, of the constituent chains). It predicts the general form of the stress-strain relationships correctly under a variety of strains, the approximate numerical magnitudes of the stresses for various chemical structures, and the effects of temperature and of swelling the rubber with an inert mobile liquid on the elastic behavior. It also predicts novel second-order stresses, discussed later, which have no counterpart in classical elasticity theory. In summary, it constitutes a major advance in our understanding of the properties of polymeric materials.

1.5 Continuum Theory of Rubber Elasticity

A general treatment of the stress-strain relations of rubberlike solids was developed by Rivlin (1948, 1956), assuming only that the material is isotropic in elastic behavior in the unstrained state and incompressible in bulk. It is quite surprising to note what far-reaching conclusions follow from these elementary propositions, which make no reference to molecular structure.

Symmetry considerations suggest that appropriate measures of strain are given by three strain invariants, defined as

are the principal stretch ratios (the ratios of stretched to unstretched lengths; , remain. It follows that the strain energy density W is a function of these two variables only:

(1.14)

Furthermore, to yield linear stress-strain relations at small strains, W . Therefore, the simplest possible form for the strain energy function is

(1.15)

equal to the small-strain shear modulus G. Equation .

On expanding Eq. (1.15) as a power series in strains e. Thus it necessarily gives good agreement with experiment at small strains, say for values of e up to 10–20%, where higher powers of e are negligibly small. However, considerable confusion has arisen from its application at larger strains, for values of e of 100% or more, when it no longer holds. It is rather unfortunate that experimental stress-strain relations in simple extension appear to be in accord with Eq. (1.15) up to moderately large strains. This fortuitous agreement arises because the particular strain energy function obeyed by rubber, discussed later, depends on strain in such a way that the two stress-strain relations in tension are similar in form. Relations for other types of strain are quite different, even at modest strains (Rivlin and Saunders, 1951).

1.5.1 Stress-Strain Relations

Stresses can be obtained from the derivatives of the strain energy function W:

(1.16)

Rewriting Eq. yields

(1.17)

hereafter.

(Gent and Thomas, 1958):

(1.18)

is a constant. This form of the second term is in reasonably good numerical agreement with the predictions of Thomas’s additional term in the strain energy function for a single chain, Eq. (1.13), and simpler in form.

appears to be rather constant, independent of the degree of crosslinking, and thus it is relatively more important for lightly crosslinked materials. As mentioned earlier, it appears to reflect physical restraints on molecular strands like those represented in the tube model of restricted configurations in the condensed state (Graessley, 2004)—restraints that diminish in importance as the deformation increases or the strands become more widely separated.

(i) Strain-Hardening at Large Strains

Rubber becomes harder to deform at large strains, probably because the long flexible molecular strands that comprise the material cannot be stretched indefinitely. The strain energy functions considered up to now do not possess this feature and therefore fail to describe behavior at large strains. Strain-hardening can be introduced by a simple modification to the first term in Eq. (Gent, 1996):

(1.19)

Equation is small. Thus Eq. (1.19) is probably the simplest possible strain energy function that accounts for the elastic behavior to good approximation over the entire range of strains (Pucci and Saccomandi, 2002). It requires three fitting parameters, two of which are related to the small-strain shear modulus G:

(1.20)

, is proportional to the number N , appears to reflect physical restraints on molecular strands like those represented in the tube model (is inversely proportional to N . Thus the entire range of elastic behavior arises from only two fundamental molecular parameters.

(Ogden, 1984). Although experimental results can be described economically and accurately in this way, the functions employed are empirical and the numerical parameters used as fitting constants do not appear to have any direct physical significance in terms of the molecular structure of the material. On the other hand, the molecular elasticity theory, supplemented by a simple non-Gaussian term whose molecular origin is in principle within reach, seems able to account for the observed behavior at small and moderate strains with comparable success.

to be neglected. Some stress-strain relations are now derived using this approximation to illustrate how such calculations are carried out and to deduce under what conditions the deformations become unstable. Instabilities are interesting from a theoretical point of view because they occur suddenly, at a well-defined deformation, and they are often unexpected on the basis of classical elasticity theory. Moreover, a comparison of the observed onset of instability with the predictions of various strain energy functions W provides, at least in principle, a critical test for the validity of a proposed form for W. From a practical standpoint, unstable states are quite undesirable because the deformation becomes highly nonuniform, leading to premature failure.

(ii) Inflation of a Thin-Walled Tube

in the axial direction, with the wall thickness h because the rubber volume remains constant. The inflation pressure P gives rise to stresses in the circumferential and axial directions:

(1.21)

where r is the tube radius in the unstrained state.

From Eq. , the undefined pressure p is obtained as:

(1.22)

(In a thin-walled tube of large radius the inflating pressure P that it generates, and thus P in determining p.) Inserting this result for p in of the internal volume of the tube:

(1.23)

The relation between inflating pressure P and internal volume of the tube is then obtained as:

(1.24)

This relation is plotted in is infinitely large, the aneurysm is unbounded and failure would then occur immediately on reaching the critical pressure.

Figure 1.9 Pressure-volume relations for a thin-walled tube from Eq. ).

Figure 1.10 Sketch of an aneurysm in an inflated tube.

(iii) Inflation of a Thin-Walled Spherical Balloon

are equal and given by

(1.25)

from Eq. . The inflation pressure P is then given by:

(1.26)

where r and h . In practice, the deformation becomes quite complex (Figure 1.11). The balloon remains roughly spherical in shape but one part is lightly stretched while the remainder is highly stretched. The two states of strain resemble the two deformations that are predicted at a given pressure after the critical point is reached.

Figure 1.11 Inflation of a thin-walled spherical rubber balloon. Solid curve: Eq. (from Gent, 1958).

(iv) Inflation of a Thick-Walled Spherical Shell

The internal pressure P required to inflate a small spherical cavity in the center of a thick block can be obtained by integrating the contributions from concentric shells (thin-walled balloons) given in the preceding section. The result is (Gent and Lindley, 1958)

(1.27)

the pressure P ; that is, about 5G/2, where G is the small-strain shear modulus. For typical rubbery materials where G is about 0.5 MPa, the maximum pressure is thus about 1.2 MPa, or about 12 bar. Any small cavity will expand greatly at this rather modest inflation pressure. Internal fracture is therefore likely to occur in soft rubbery solids at inflation pressures or equivalently, triaxial tensions, of this amount. In practice, all rubbery solids are found to develop internal fractures when supersaturated with gases or liquids at pressures or triaxial tensions about equal to 5G/2 (Gent and Lindley, 1958; Gent and Tompkins, 1969).

Note that the initial radius of the spherical cavity does not appear in Eq. (1.27). Thus, cavities of all sizes are predicted to inflate equally. However, we have neglected surface energy contributions that will tend to stabilize small cavities. When they are taken into account, it appears that only cavities having radii greater than about 100 nm will expand dramatically at the low pressures predicted by Eq. (1.27). Internal fractures suggest that vulcanized rubber must contain many precursor cavities of this effective size or larger.

(v) Surface Instability of Compressed or Bent Blocks

set up in two perpendicular directions in the surface. The critical condition is

(1.28)

and Eq. and the critical compression becomes 33.3%.

along the width is largely unchanged (at unity). Thus, from Eq. is 0.544; that is, when the surface is compressed by about 46%. Experimentally, sharp folds or creases appear suddenly in the inner surface of a bent block at a critical degree of bending (Gent and Cho, 1999); see Figure 1.12. However, the critical compression of the inner surface was considerably smaller than predicted by Biot’s theory, 35% instead of 46%. It is not known why the instability occurred so much sooner than expected. Although rubber follows a more complex strain energy function than the simple form assumed here, it is unlikely that the difference would have such a large effect.

Figure 1.12 Sketch of a bent block showing creases that appear on the inner surface where the compressive strain is greatest.

Rubber articles are often subjected to rather severe bending deformations, for example, in tires. Folds and creases in the interior may pass undetected. Nevertheless, they represent lines of high stress concentration and sites of possible failure. Folds (Schallamach waves) also appear when soft rubber slides over a rigid countersurface (Schallamach, 1971). They appear to be Biot creases caused by frictional compression of the surface.

(vi) Resistance of a Compressed Block to Indentation

When a block is subjected to a sufficiently large equibiaxial compression in the surface plane, it becomes unstable to small indentations. Green and Zerna (1975) expressed the relation between indentation force N and amount of indentation d as:

(1.29)

where G is the shear modulus of the half-space material, R , given by

(1.30)

Values of indentation force N for a given small indentation, from Eq. (1.27), are plotted in Figure 1.13 against the equibiaxial strain e .) The resistance to indentation is seen to decrease sharply as the compressive strain is increased, becoming zero at a compressive strain of 0.333, in agreement with Biot’s result.

Figure 1.13 Force N for a small indentation vs. equibiaxial strain e (from Green and Zerna, 1975).

(vii) Torsional Instability of Stretched Rubber Rods (Gent and Hua, 2004)

Another unstable state is encountered when a stretched rubber rod is subjected to large torsions. A kink suddenly appears at one point along the rod, Figure 1.14, and more kinks form on twisting the rod further.

Figure 1.14 Sketch of a kink that appears on twisting a stretched rubber rod (from Gent and Hua, 2004).

Minimization of the total elastic strain energy suggests that the rod will become unstable at a critical amount of torsion: part of the rod will unwind and form a tight ring while the remainder of the rod will become slightly more stretched. A simple criterion can be derived on this basis for the onset of kinks. For a neo-Hookean material, Eq. (1.8), the condition for forming a kink becomes:

(1.31)

and the rod radius a. Measured values for rods of different radius, stretched to extensions of up to 250%, were found to be in reasonably good agreement with Eq. (1.31), indicating that the sudden formation of kinks in twisted rubber rods is, indeed, a consequence of an elastic instability.

1.6 Second-Order Stresses

(in addition, of course, to simple shear stresses) (.

They are represented schematically in Figures 1.15 and 1.16 for two different choices of the arbitrary hydrostatic pressure pis set equal to zero; this condition would arise near the side surfaces of a sheared block. In each case a compressive is found to be necessary to maintain the simple shear deformation. In its absence the block would tend to increase in thickness on shearing.

Figure 1.15 is set equal to zero (from Gent, 1958).

Figure 1.16 is set equal to zero (from Gent, 1958).

in , which obtain under the imposed deformation state (Rivlin, 1947).

Figure 1.17 in Figures 1.10 and 1.11) over the cross-section of the rod (from Treloar, 1975).

1.7 Elastic Behavior Under Small Deformations

) compared with the shear modulus G ), they may be regarded as relatively incompressible. The elastic behavior under small strains can thus be described by a single elastic constant G. Poisson’s ratio is effectively 1/2, and Young’s modulus E is given by 3G, to good approximation.

A wide range of values for G , but those fillers, which have a particularly pronounced stiffening action, also give rise to stress-softening effects like those shown in Figure 1.5, so that the modulus becomes a somewhat uncertain quantity.

It is customary to characterize the modulus, stiffness, or hardness of rubbers by measuring their elastic indentation by a rigid die of prescribed size and shape under specified loading conditions. Various nonlinear scales are employed to derive a value of hardness from such measurements (Soden, 1952). Corresponding values of shear modulus G for two common hardness scales are given in Figure 1.18.

Figure 1.18 Relations between shear modulus G and indentation hardness: —, Shore A Scale; - - -, International Rubber Hardness Scale (from Tobolsky and Mark, 1971).

Many rubber products are normally subjected to fairly small deformations, rarely exceeding 25% in extension or compression or 75% in simple shear. A good approximation for the corresponding stresses can then be obtained by conventional elastic analysis assuming linear relationships. One particularly important deformation is treated here: the compression or extension of a thin rubber block, bonded on its major surfaces to rigid plates (Figure 1.19). A general treatment of such deformations has been reviewed (Horton, 2002).

Figure 1.19 and shear stress t acting at the bonded surfaces are represented by the upper portions of the diagram (from Tobolsky and Mark, 1971).

It is convenient to assume that the deformation takes place in two stages: a pure homogeneous compression or extension of amount e, and a shear deformation restoring points in the planes of the bonded surfaces to their original positions in these planes. For a cylindrical block of radius a and thickness h, the corresponding shear stress t acting at the bonded surfaces at a radial distance r from the cylinder axis is given by

, given by

(1.32)

These stress distributions are shown schematically in Figure 1.19. Although they must be incorrect right at the edges of the block, because the assumption of a simple shear deformation cannot be valid at these points of singularity, they appear to provide satisfactory approximations over the major part of the bonded surfaces.

over the bonded surface, the total compressive force F is obtained in the form (Gent, 1994)

(1.33)

(given by the right-hand side of Eq. (1.33)) is much larger than the real value E because of the restraints imposed by the bonded surfaces. Indeed, for values of the ratio a/h is now so large that it becomes comparable to the modulus of bulk compression (Gent, 1994) (Figure 1.20).

Figure 1.20 for bonded blocks vs. ratio of radius to thickness a/h (from Tobolsky and Mark, 1971).

A more accurate treatment of the compression of bonded blocks has been given by . However, this term does not yield the correct value of unity for tall blocks; that is, when a/h is small, and it is equivalent to Eq. (1.33) for thin blocks of large radius, when a/h is large. It should therefore be regarded as a better approximation for blocks of intermediate size.

, from Eq. (1.32). Under this outwardly directed tension a small cavity in the central region of the block will expand indefinitely at a critical value of the tension, of about 5E, given approximately by

and at a corresponding critical value of the applied tensile load, obtained by substituting this value of e in Eq. (1.33). To avoid internal fractures of this kind, it is thus necessary to restrict the mean tensile stress applied to thin bonded blocks to less than about E/3.

In compression, on the other hand, quite large stresses can be supported. A stress limit can be calculated by assuming that the maximum shear stress, developed at the bonded edges, should not exceed G; that is, the maximum shear deformation should not exceed about 100%. This yields a value for the allowable overall compressive strain of h/3a, corresponding to a mean compressive stress of the order of E for disks with a/h between about 3 and 10. This calculation assumes that the approximate stress analysis outlined earlier is valid right at the edges of the block, and this is certainly incorrect. Indeed, the local stresses in these regions depend strongly on the detailed shape of the free surface in the neighborhood of the edge.

1.8 Some Unsolved Problems in Rubber Elasticity

We turn now to some features of the elastic response of rubbery materials that are still not fully understood.

As normally prepared, molecular networks comprise chains of a wide distribution of molecular lengths. Numerically, small chain lengths tend to predominate. The effect of this diversity on the elastic behavior of networks, particularly under large deformations, is not known. A related problem concerns the elasticity of short chains. They are inevitably non-Gaussian in character and the analysis of their conformational statistics is likely to be difficult. Nevertheless, it seems necessary to carry out this analysis to be able to treat real networks in an appropriate way.

It is also desirable to treat network topology in greater detail; that is, to incorporate the functionality of crosslinks, their distribution in space, and loop formation. The effect of mutual interaction between chains in the condensed state appears to be accounted for satisfactorily by the tube model for uncrosslinked polymers, but its application to networks seems incomplete. But the problem in greatest need of attention is the response of highly filled elastomers to stress. Filled elastomers are not really elastic; their stress-strain relations are irreversible (see Figure 1.5), and it is therefore inappropriate to describe their response to stress by a strain energy function. Moreover, they appear to become anisotropic on stretching and to some degree after release. At present, the molecular processes that occur on deformation and the mathematical framework suitable for describing them are both unclear.

References

1. Arruda EM, Boyce MC. J Mech Phys Solids. 1993;41:389.

2. Beatty MF. In: New york: American Society of Mechanical Engineers; 1977;Rivlin RS, ed. AMD Finite Elasticity. vol. 27 p. 125.

3. Biot M. Mechanics of Incremental Deformations. New York: Wiley; 1965.

4. Bueche F. J Appl Polym Sci. 1960;4:107.

5. Bueche F. J Appl Polym Sci. 1961;5:271.

6. Cotton JP, Farnoux B, Jannink GJ. J Chem Phys. 1972;57:290.

7. Fetters LJ, Lohse DJ, Graessley WW. J Polym Sci Part B: Polym Phys. 1999;37:1023.

8. Flory PJ. Statistical Mechanics of Chain Molecules. New York: Wiley-Interscience; 1969.

9. Gent AN. J Polym Sci Poly Symp. 1958;28:625.

10. Gent AN. Rubber Chem Technol. 1994;67:549.

11. Gent AN. Rubber Chem Technol. 1996;69:59.

12. Gent AN, Cho IS. Rubber Chem Technol. 1999;72:253.

13. Gent AN, Hua KC. Int J Non-Linear Mech. 2004;39:483.

14. Gent AN, Lindley PB. Proc Roy Soc Lond A. 1958;249:195.

15. Gent AN, Thomas AG. J Polym Sci. 1958;28:625.

16. Gent AN, Tompkins DA. J Appl Phys. 1969;40:2520.

17. Graessley WW. Polymeric Liquids and Networks: Structure and Properties. New York: Taylor and Francis Books; 2004.

18. Green AE, Zerna W. Theoretical Elasticity. 2nd ed. Oxford: Clarendon Press; 1975; p. 135 (Section 4.6).

19. Horton JM, Tupholme GE, Gover MJC. ASME J Appl Mech. 2002;69:836.

20. James HM, Guth E. J Chem Phys. 1943;11:455.

21. James HM, Guth E. J Polym Sci. 1949;4:153.

22. Langley NR. Macromolecules. 1968;1:348.

23. Mark JE, Erman B. Rubberlike Elasticity: A Molecular Primer. New York: John & Wiley Sons; 1988.

24. Mark JE, Erman B, eds. Elastomeric Polymer Networks. Englewood Cliffs, NJ: Prentice-Hall; 1992.

25. Mooney M. J Appl Phys. 1940;11:582.

26. Morris MC. J Appl Polm Sci. 1964;8:545.

27. Ogden RW. Non-Linear Elastic Deformations. Chichester, UK: Ellis Harwood; 1984; (Dover Publications, Mineola, NY, 1997, Chapter 7).

28. Prager S, Frisch HL. J Chem Phys. 1967;46:1475.

29. Pucci E, Saccomandi G. Rubber Chem Technol. 2002;75:839.

30. Rivlin RS. J Appl Phys. 1947;18:444.

31. Rivlin RS. Philos. Trans Roy Soc Lond Ser A. 1948;241:379.

32. Rivlin RS. In: New York: Academic Press; 1956;Eirich FR, ed. Rheology, Theory and Application. vol. 1 (Chapter 10).

33. Rivlin RS, Saunders DW. Philos. Trans Roy Soc Lond Ser A. 1951;243:251.

34. Schallamach A. Wear. 1971;17:301.

35. Soden AL. A Practical Manual of Rubber Hardness Testing. London: Maclaren; 1952.

36. Thomas AG. Trans Faraday Soc. 1955;51:569.

37. Tobolsky AV, Mark HF, eds. Polymer Science and Materials. New York: Wiley; 1971; (Chapter 13).

38. Treloar LRG. The Physics of Rubberlike Elasticity. 3rd ed. Oxford: Clarendon Press; 1975.

39. Vilgis TA. In: Mark JE, Erman B, eds. Elastomeric Polymer Networks. Englewood Cliffs, NJ: Prentice-Hall; 1992; (Chapter 5).

¹Although the terms configuration and conformation are sometimes used interchangeably, the former has acquired a special meaning in organic stereochemistry and designates specific steric structures. Conformation is used here to denote a configuration of the molecule, which is arrived at by rotation of single-valence bonds in the polymer backbone.

²These are (poor) solvents in which repulsion between different segments of the polymer molecule is balanced by repulsion between polymer segments and solvent molecules.

Chapter 2

Polymerization: Elastomer Synthesis

Roderic P. Quirk* and Deanna L. Pickel†,    *The Maurice Morton Institute of Polymer Science, The University of Akron, Akron, OH, USA,    †Center for Nanophase Materials Sciences, Oak Ridge National Laboratory Oak Ridge, TN, USA

2.1 Introduction

The development of synthetic rubber played a special role in the history of polymerization chemistry. This was due primarily to the fact that attempts to synthesize rubber were made long before there was even the faintest idea of the nature of polymerization reactions. Such attempts began very soon after the elegant analytical work of Williams (1859) in 1860, which clearly demonstrated that Hevea rubber was composed of isoprene. Thus, Bouchardat (1879) was actually able to prepare a rubberlike substance from isoprene (which he obtained from rubber pyrolysis), using heat and hydrogen chloride. Tilden (1884) repeated this process in 1884 but used isoprene obtained from pyrolysis of turpentine to demonstrate that it was not necessary to use the mother substance of rubber itself. These explorations were soon followed by the work of Kondakow (1900) with 2,3-dimethylbutadiene that of Thiele (1901) with piperylene, and finally that of Lebedev (1910) on butadiene itself. Mention should also be made of the almost simultaneous, and apparently independent, discoveries in 1910 by Harries (1911) in Germany and Matthews and Strange (1910) in England of the efficient polymerization in isoprene by sodium.

Although all of these attempts had a noble purpose indeed, the means used could hardly be considered a contribution to science, as the transformation of the simple molecules of a diene into the colloidal substance known as rubber was then far beyond the comprehension of chemical science. As a matter of fact, the commercial production of synthetic rubber was already well established, at least in Germany and Russia, before Staudinger laid the basis for his macromolecular hypothesis during the 1920s (Staudinger, 1920). Even such relatively modern synthetic elastomers as polychloroprene and the poly(alkylene sulfides) were already in commercial production by 1930–1931. This was, of course, also before Carothers and coworkers’ pioneering studies on the polymerization of chloroprene (Carothers et al., 1931)!

Hence, it is apparent that it was not the development of an understanding of polymerization that led to the invention of synthetic rubber, but perhaps the reverse. In contrast, it was the new science of organic macromolecules, whose foundations were established by Staudinger, which expanded rapidly during the 1930s and 1940s, and pointed the way to the synthesis of a vast array of new polymeric materials, including synthetic fibers and plastics and even new elastomers. This new science included the classical studies of polycondensation by Carothers and Flory and the establishment of the principles governing free radical chain addition reactions by Schulz, Flory, Mayo, and others (Flory, 1953a; Morawetz, 1985).

Thus it was that the paths of synthetic rubber and macromolecular science finally crossed and became one broad avenue (Morton, 1961). Hence today the design of a new elastomer or the modification of an old one requires the same kind of molecular architecture which applies to any other polymer and is based on an understanding of the principles of polymerization reactions.

2.2 Classification of Polymerization Reactions and Kinetic Considerations

Historically polymers have been divided into two broad classes: condensation polymers and addition polymers (Flory, 1953a; Carothers, 1929,1931). Flory (1953b) has defined these as follows:

Condensation polymers, in which the molecular formula of the structural unit (or units) lacks certain atoms present in the monomer from which it is formed, or to which it may be degraded by chemical means, and addition polymers, in which the molecular formula of the structural unit (or units) is identical with that of the monomer from which the polymer is derived.

Thus, an example of a condensation polymer would be a polyester, formed by